diff --git a/experiments/universality/fold_isolation_diagonal/AUTHOR_ALIGNMENT_ADDENDUM.md b/experiments/universality/fold_isolation_diagonal/AUTHOR_ALIGNMENT_ADDENDUM.md new file mode 100644 index 000000000..91e78340d --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/AUTHOR_ALIGNMENT_ADDENDUM.md @@ -0,0 +1,66 @@ +# Author alignment addendum to the restricted collinear calculation + +16 October 2026 UTC. Author: OpenAI/Codex `/root/core_theorem_audit_20261015`. This is an append-only clarification for handoff. It leaves the two frozen objects below unchanged and makes no additional mathematical or process claim. + +| Frozen object | SHA-256 | +| --- | --- | +| [COALESCING_CONTACT_CALCULATION.md](COALESCING_CONTACT_CALCULATION.md) | `39a63e1b00393b49e7b6d049c0c9ea7be82436352eed64014155a63aa8c7fb36` | +| [NONAUTHOR_COALESCENCE_REVIEW.md](NONAUTHOR_COALESCENCE_REVIEW.md) | `2f766d8aa5541602ace7bbc70fd360e3a342d2cd258c56c44a35503e3e8624d5` | + +The complete review was read. I accept both nonblocking wording corrections. I found no mathematical disagreement with the review's restricted interpretation. The following paragraphs give the precise reading of the original report's basepoint and simplex sentences; they do not rewrite its historical bytes. + +## Basepoint convention + +In each four-site configuration, replace the original coordinates +`z_i=z+x_i u` by + +```text +z' = z+x_1 u = z_1, +x_i' = x_i-x_1, +z_i = z'+x_i' u, x_1'=0. +``` + +Then rename z' as z and x_i' as x_i. Thus §3's limiting gradient jets are at the **first physical site z=z_1**, and the segment is `0=x_1=1 define + +```text +Delta_j = {(t_1,...,t_j): t_i>=0, sum_i t_i<=1}, +t_0=1-sum_i t_i, +volume(Delta_j)=1/j!. +``` + +The convention is + +```text +a[x_0,...,x_j] + = integral_(Delta_j) + a^(j)(t_0 x_0+...+t_j x_j) dt_1...dt_j. +``` + +There is **no additional factor 1/j! in the integrand or outside this integral**. The j=0 convention is point evaluation with mass one. At coincident nodes, this gives `a^(j)(x)/j!`. If probability-simplex averaging is used instead, one multiplies that expectation by 1/j! exactly once; that is only an equivalent alternative description. + +Accordingly, the original §3 limit is read literally as + +```text +(F_u, F_uu, F_uuu/2!, F_uuuu/3!, + F_v, F_uv, F_uuv/2!, F_uuuv/3!)(z_1). +``` + +The phrase in the frozen report about an “average … divided by j!” together with a measure of mass 1/j! must not be read as applying the factorial twice. Its intended corrected interpretation is the unscaled-derivative/simplex-volume formula above. The displayed limiting jets already use that convention. The same convention applies to the Hermite divided differences in the twelve-row discussion, including orders through seven. No covariance determinant power or displayed limiting-jet factorial is changed by this clarification. + +## Review lineage and preserved limits + +The reviewer is OpenAI/Codex `/root/coalescence_review_20261016`, a separate nonauthor agent within the same provider and parent thread. Its frozen review records a complete read of the report, diagnostic program and JSON, pinned source checks, reconstruction of the restricted estimates, diagnostic replay and independent finite algebra. Those are the reviewer's performed checks, not new checks attributed to this author addendum. Organizational-independence credit is **0**; there is no human or blind review. + +Its verdict is **ACCEPT, restricted conventional mathematics only**, with the two wording corrections incorporated by this addendum. That verdict covers the stated collinear bounds (8)/(11), coordinate-aligned ranks and K2 obstruction (16), and the K>=4 exact-contact bound (20). This addendum neither extends the verdict nor converts it into source-owner adoption, formal verification, a graph transition or scientific promotion. + +The fixed-law, planar, positive-full-square spectrum, short embedded chart, four distinct sites, declared pair-separation ratio, canonical zero-gradient regression and fixed positive lower cubic marks remain essential. Constants are not made effective or uniform over vanishing spectral weights, unbounded K or varying L. The twelve-row K2/K3 coordinate strata remain singular. The eight-gradient argument retains its stated all-frame fixed-law scope; the twelve-row obstruction is not a blanket claim about every K2 orientation. + +Actual H0 selectors are consumed only as Borel marks bounded by one. No global pairing or selected-bar counting identity is assumed or proved here. The collinear stratum still has zero full spatial measure. Transverse collision geometry, k_i=O(eta²), uniform finite-radius/contact-limit interchange, other incidence patterns, increasing-domain estimates, global factorial control, Lean interfaces and spatial point-process convergence remain unresolved. Decimal arithmetic remains diagnostic and uncertified; no sampler, seeds or numerical H0/process experiment were added. + +This addendum records the author's accepted interpretation for handoff alongside the frozen report and review. No native write, source claim, PR, merge, source-owner transfer or theorem-status change occurred. diff --git a/experiments/universality/fold_isolation_diagonal/COALESCING_CONTACT_CALCULATION.md b/experiments/universality/fold_isolation_diagonal/COALESCING_CONTACT_CALCULATION.md new file mode 100644 index 000000000..fc6499549 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/COALESCING_CONTACT_CALCULATION.md @@ -0,0 +1,406 @@ +# Coalescing two-contact calculation: a collinear cancellation and a finite-carrier obstruction + +16 October 2026 UTC. Read-only mathematical analysis by OpenAI/Codex `/root/core_theorem_audit_20261015`. No native source write, source claim, PR, graph update, proof-status transition or process theorem. This report derives restricted-geometry estimates on one actual finite Gaussian field. The decimal calculations at the end are diagnostics only. + +## 1. What changes the next action + +There are two concrete results. + +1. **A useful restricted estimate.** For four nearby critical sites lying exactly on one line, the joint eight-gradient density has a Vandermonde singularity `V^-2`. The product of the four physical Hessian determinant weights is bounded by `C V² (1+||F||C5)^8`. These factors cancel in the gradient-pinned four-point Rice kernel. For two disjoint small pairs with center separation eta, requiring a positive lower cubic-gap mark `k_i>=k_->0` forces `||F||C5 >= c k_-/eta²`; consequently the same kernel is at most `C exp(-c k_-²/eta^4)`. Any actual H0 selectors bounded by one can be retained in that upper bound. No global pairing or iid replacement is required. +2. **A real obstruction to the naive height-density route.** On a coordinate-aligned line, the twelve value/gradient observations at four distinct sites have rank 9 for full-square K2 and rank 11 for K3, rather than rank 12. Two limiting six-row fold contacts have the same deficient ranks. Full positivity of the finite square does not fix this. In K2, two distinct aligned exact positive cubic contacts are actually incompatible with the field's support. The infinite-law K7/K5 ranks in the current process source therefore cannot simply be inherited by these finite carriers. + +The next mathematical step should use gradient-pinned Rice plus actual height-gap indicators and a geometry-dependent divided-difference analysis, or strengthen the carrier and stratify the height-observation singularities explicitly. It should not assume a uniform twelve-dimensional density for the K2/K3 catalogue. The collinear estimate still leaves transverse collision geometry and the small-k boundary unestimated; it is not a global factorial or spatial Poisson theorem. + +## 2. Source, law, observable and scale + +Pinned source: Main `85b928b326ff0586715a56ae91f51d3fbd89557a`. + +- [iid_short_bar_process/PROOF.md §6](https://github.com/d6g8k5htny-coder/main/blob/85b928b326ff0586715a56ae91f51d3fbd89557a/experiments/universality/iid_short_bar_process/PROOF.md#L409) proves the separated-center, sufficiently small-radius twofold contribution `O_eta0(t^(4/3))` for the actual **infinite** periodized Gaussian law. Its four-site K7 and two-contact K5 interpolation are explicitly not transferred from finite K2/K3. +- [finite_gaussian_contact/PROOF.md §§2–5](https://github.com/d6g8k5htny-coder/main/blob/85b928b326ff0586715a56ae91f51d3fbd89557a/experiments/universality/finite_gaussian_contact/PROOF.md#L94) supplies the actual finite Fourier law, covariance multiplier identity, literal six observation rows, target and physical Hessian weights. Its joint seven-row covariance is a **single-contact** result. +- [finite_candidate_lifetime/PROOF.md](https://github.com/d6g8k5htny-coder/main/blob/85b928b326ff0586715a56ae91f51d3fbd89557a/experiments/universality/finite_candidate_lifetime/PROOF.md) and [finite_h0_lifetime/PROOF.md](https://github.com/d6g8k5htny-coder/main/blob/85b928b326ff0586715a56ae91f51d3fbd89557a/experiments/universality/finite_h0_lifetime/PROOF.md) keep candidate and selected populations distinct and put the Hessian determinant into the Rice weight once. + +Fix K>=2, L>0 and omega=2pi/L. The field is the source's planar ideal Gaussian Fourier law: + +The four input bodies were independently checked against Main's immutable Git tree: + +| Input | Git blob | Bytes | SHA-256 | +| --- | --- | ---: | --- | +| iid_short_bar_process/PROOF.md | `cd7ef8b8a364cb394be48208f55332658e9fff0b` | 30935 | `320bc988c468bea4938bb777a2d05952c6ab8e0a384f33725b50de847aeec1eb` | +| finite_gaussian_contact/PROOF.md | `5f345201309cfa1781e92d6292f542de4bcb2999` | 25367 | `7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474` | +| finite_candidate_lifetime/PROOF.md | `a27a84150b3e70bfd794672eeeed6915425a1da4` | 29190 | `ed31f562f0ffaaf6d4bd2ceb1ae3d234798e983a55d55c6983b128fce27133f5` | +| finite_h0_lifetime/PROOF.md | `7a3518edd9bd04adf7aac04f3a5a40bdaab78a48` | 27064 | `5ea91111fa5f3d78d7c4188ebe84b3bc2c452b05dfdfc62c74b03123cb2696d1` | + +```text +S_K = {-K,...,K}², +q_n=q_-n>0, sum q_n=1, +F(z)=sqrt(q_0) C_0 + + sum_nonzero_pairs sqrt(2q_n) + [C_n cos(omega n.z)+S_n sin(omega n.z)], +``` + +with independent standard real Gaussian coefficients, one representative per pair. No cutoff, weights, coefficient distribution or source normalization is changed in the analysis. Constants may depend on this fixed law and on L,K. They are not uniform as positive weights approach zero, over unbounded K, or over variable L. + +Choose an orthonormal frame (u,v), a short embedded line segment on the torus and four ordered distinct sites + +```text +z_i = z + x_i u, x_10, +D_i = product_(j!=i)|x_i-x_j|, +G = (F_u(z_1),...,F_u(z_4), F_v(z_1),...,F_v(z_4)), +H_i = Hess F(z_i), +ell_i=F(M_i)-F(S_i), k_i=ell_i/r_i³. +``` + +The exact kernel estimated below is + +```text +R(x_1,...,x_4;B) + = p_G(0) E[ product_i |det H_i| * B(F) | G=0 ], +``` + +where B is any Borel mark in [0,1]. For the compact-positive-gap bound it includes `k_1>=k_-` and `k_2>=k_-`; it may also include `ell_i<=Tt`, maximum/saddle Hessian types and the two actual elder selectors. These marks are all from the **same F**. The proof only bounds selectors by one and never assumes they are a bijection or that contacts are independent. + +This is an integrand on a collinear position stratum. That stratum has zero full spatial measure; a bound there alone cannot be integrated as a bound on all four-site configurations. + +## 3. Uniform gradient-jet rank at K2 already suffices for the collinear density + +Let T_x send four scalar values at x_1,...,x_4 to their successive Newton divided differences of orders 0,1,2,3. Its triangular diagonal gives + +```text +|det T_x|=1/V. +``` + +Apply T_x separately to the four F_u and four F_v evaluations. With `U=diag(T_x,T_x)G`, + +```text +|det(G -> U)|=V^-2, +det Cov(G)=V^4 det Cov(U). (1) +``` + +When the segment diameter tends to zero, the rows of U tend uniformly, including arbitrarily unequal node spacings, to + +```text +J = (F_u, F_uu, F_uuu/2!, F_uuuu/3!, + F_v, F_uv, F_uuv/2!, F_uuuv/3!)(z). (2) +``` + +One elementary justification of uniformity is the integral representation of an order-j divided difference as an average of the jth derivative divided by j!, over the simplex spanned by its nodes. It follows by repeated use of the fundamental theorem of calculus; the positive averaging measure has fixed mass 1/j!. Thus the difference from the coincident-node derivative is at most the derivative's modulus of continuity on the segment. Finite Fourier derivatives have that uniform modulus in mean square, uniformly in frame. + +The covariance of J is positive definite for every frame and K>=2. To see this directly from the source covariance formula, a zero-variance linear combination gives a frequency polynomial that is a linear combination of + +```text +a, a², a³, a^4, c, ac, a²c, a³c, +a=n.u, c=n.v, (3) +``` + +with nonzero omega, i and factorial factors. Its degree in each original frequency coordinate is at most four. Positivity of q makes it vanish on the full grid with at least five coordinates in each direction. The elementary polynomial root argument in each coordinate makes the polynomial identically zero; the invertible frame change makes the eight coefficients vanish. No arbitrary-jet theorem is imported. + +Continuity and compactness of the frame group give a positive covariance floor for J. The uniform row convergence therefore yields constants c,C>0 for all sufficiently short segments and every configuration of four distinct nodes: + +```text +c I <= Cov(U) <= C I. +``` + +Equation (1) gives the exact singular order, uniformly in all node ratios, + +```text +p_G(0) = (2pi)^-4 V^-2 / sqrt(det Cov(U)) = Theta(V^-2). (4) +``` + +For the two-pair scale, + +```text +(1-rho)^4 r_1 r_2 eta^4 <= V + <= (1+rho)^4 r_1 r_2 eta^4. +``` + +Hence the raw gradient density grows like `r_1^-2 r_2^-2 eta^-8`. This is an actual covariance calculation for the finite field. + +## 4. Four pinned gradient zeros force precisely the matching Hessian vanishing + +Set `g(s)=F_u(z+s u)` and `h(s)=F_v(z+s u)`. On the G=0 fiber, both functions vanish at all four x_i. + +For any C4 scalar function a with these four zeros, + +```text +|a'(x_i)| <= D_i * sup |a^(4)| / 4!. (5) +``` + +Proof: for a point s different from the four nodes, subtract from a the degree-four polynomial `a(s) product_j(t-x_j)/product_j(s-x_j)`. The difference has five zeros. Four applications of Rolle's theorem give the Lagrange remainder bound `|a(s)|<=sup|a^(4)| product_j|s-x_j|/4!`. Divide by |s-x_i| and pass to x_i. This establishes (5) without a conditional genericity assumption. + +Apply (5) to g and h: + +```text +|F_uu(z_i)| <= D_i ||F_(u^5)||_infinity /24, +|F_uv(z_i)| <= D_i ||F_(u^4 v)||_infinity /24. (6) +``` + +Take a fixed coordinate C5 norm and let M=1+C||F||C5 dominate all indicated directional derivatives and F_vv. With the segment radius small enough that D_i<=1, + +```text +|det H_i| + = |F_uu F_vv-F_uv²|(z_i) + <= C M² D_i. +``` + +Every pairwise distance appears twice in `product_i D_i`, so + +```text +product_i |det H_i| <= C M^8 product_i D_i + = C M^8 V². (7) +``` + +Under G=0 the standard coefficient vector has canonical law PZ, where P is an orthogonal projection and Z is standard Gaussian. It is centered, even though Cov(G) degenerates as sites approach. In particular `||PZ||<=||Z||`, and the fixed finite expansion gives uniform conditional moments of M of every finite order. Equations (4) and (7) prove: + +**Restricted collinear Rice lemma.** For the fixed finite Gaussian law, all sufficiently short collinear four-site configurations satisfy + +```text +p_G(0) E[product_i |det H_i| * B | G=0] <= C (8) +``` + +for every Borel mark 0<=B<=1. This includes typed or actual-selected four-endpoint marks whenever defined; it does not assert any selector identity. + +In the two-pair regime, (7) has the more informative form + +```text +product_i |det H_i| <= C M^8 r_1² r_2² eta^8. +``` + +The individual fold-pair determinant factors are r_i²; the simultaneous four-pin geometry contributes the extra eta^8 that cancels the additional eta^-8 gradient-density singularity. Dropping that vanishing by bounding determinants merely by field norms would create a spurious divergent bound. + +## 5. Positive cubic marks force a Gaussian large deviation, already at finite radii + +There is a stronger deterministic constraint. The same Rolle remainder gives, for s between the first pair's endpoints, + +```text +|F_u(z+s u)| + <= ||F_(u^5)||_infinity + (s-x_1)(x_2-s)(x_3-s)(x_4-s)/24. +``` + +Each cross-pair factor is at most `(1+rho)eta`. Integrate along the first pair and use `integral_0^r s(r-s) ds=r³/6`. Repeat for the second pair: + +```text +|ell_i| <= ((1+rho)² eta² /144) + ||F_(u^5)||_infinity r_i³, + +|k_i| <= ((1+rho)² eta² /144) + ||F_(u^5)||_infinity. (9) +``` + +This bound does not need the maximum/saddle type or an H0 pairing. It only uses the four longitudinal gradient zeros. Reversing one ordered pair changes the sign of its gap, not the absolute bound. + +If B includes either `k_i>=k_->0`, then (9) forces + +```text +||F||C5 >= c k_- eta^-2. (10) +``` + +The canonical zero-gradient conditional coefficient law PZ has a uniform Gaussian norm tail. Therefore + +```text +E[M^8 1{||F||C5 >= c k_- eta^-2} | G=0] + <= C exp(-c' k_-² eta^-4) +``` + +for eta sufficiently small relative to fixed k_-. Polynomial Gaussian-tail prefactors are absorbed by reducing c'. Combined with the exact cancellation in §4: + +**Compact-positive-mark collinear bound.** Under the same fixed-law assumptions, + +```text +p_G(0) E[product_i |det H_i| + *1{k_1>=k_-,k_2>=k_-}*B | G=0] + <= C exp(-c k_-²/eta^4). (11) +``` + +One can add `0= L_1(v)²/Var(L_1) + >= 20736 k_1² /[m_10(u) eta^4]. (15) +``` + +The reflected formula at the second endpoint gives the same bound with |k_2|. Thus a compact positive cubic mark costs an `eta^-4` Gaussian action, regardless of birth marks. The Gaussian density's exponential is bounded by `exp[-10368 max(k_1²,k_2²)/(m_10(u)eta^4)]`. No conditional density of a pinned coordinate is declared positive. + +For same-oriented positive contacts, summing the two reflected functionals gives the equivalent identity + +```text +F_uuu(0)+F_uuu(eta) + -2[F_uu(eta)-F_uu(0)]/eta + = integral_0^eta s(eta-s)/eta * F_(u^5)(s) ds. +``` + +On the target this is 12(k_1+k_2), and it also yields the deterministic necessary condition `||F_(u^5)||infinity >=72(k_1+k_2)/eta²`. The individual-endpoint formula is safer when directions are reversed. + +These formulas explain more than an unspecified loss of covariance floor: the degenerating Gaussian coordinate itself is identified, its variance is bounded explicitly, and the prescribed positive-contact target lies a distance of order eta^-2 in its standard-deviation units. + +## 7. Why the twelve-dimensional density is unavailable at K2/K3 + +Restrict the actual field to a coordinate line y=y_0. The scalar function F(x,y_0) is a real trigonometric polynomial of degree K, dimension 2K+1. Its value and x-derivative at four sites produce eight longitudinal rows. The transverse derivative F_y(x,y_0) is another degree-K trigonometric polynomial, producing four value rows. + +These two data channels can be varied independently in the full-square coefficient support: use x-only modes for the first and `sin(omega(y-y_0))` times an x trigonometric polynomial for the second. Thus for four distinct sufficiently close sites, the full value/gradient observation rank is + +```text +rank O = min(8,2K+1)+min(4,2K+1) + = 9 (K2), 11 (K3), 12 (K>=4). (16) +``` + +The same count holds for the two limiting contacts: longitudinal jets of orders 0–3 at two centers impose eight Hermite rows, while F_y,F_xy at those centers give four transverse Hermite rows. Standard polynomial/trigonometric root counting proves the ranks. In particular the twelve-dimensional Lebesgue density simply does not exist on this stratum at K2/K3. This is not a numerical small-eigenvalue issue. + +There is an even sharper K2 support obstruction. Suppose two distinct aligned exact contacts had + +```text +F_x=F_xx=0 at both centers, F_xxx=12k_i !=0. +``` + +Then a=F_x restricted to that coordinate circle is a degree-two trigonometric polynomial with zero mean, and has two double zeros. A nonzero degree-two trigonometric polynomial has at most four zeros counting multiplicities. The two double zeros exhaust them, so its sign cannot change on the circle; zero mean then forces a identically zero. Consequently F_xxx is zero, contradicting either positive cubic target. + +Equivalently, in a sufficiently short coordinate-aligned four-site gradient-pinned configuration at K2, the four longitudinal values of F_x are independent observations on the four-dimensional zero-mean derivative space. Their zero target forces F_x identically zero on that line. Hence F_xx=0 at all four pins and none can be a nondegenerate maximum. The typed four-endpoint kernel is zero there. This is compatible with the nonsingular **eight-gradient** covariance proved in §3; full gradient observation rank and conditional maximum admissibility are different questions. + +The K3 height covariance remains singular but its positive two-contact target need not be excluded by this K2 zero-mean argument. A lower-dimensional affine-support treatment would have to be stated and checked; a twelve-dimensional Gaussian density cannot be used. + +## 8. What a stronger finite carrier gives for the height-density route + +For K>=4, the source-style frequency polynomial argument supplies positive covariance for the twelve limiting jet rows + +```text +(F, F_u, ..., F_(u^7), + F_v, F_uv, F_uuv, F_uuuv), (17) +``` + +with factorial factors as appropriate. The total degree is at most seven in each original frequency coordinate, below the nine-point grid available at K4. The framewise covariance floor is positive. This is a sufficient carrier condition for the following collinear normalization; it is not claimed necessary. + +For four values and four longitudinal first derivatives, the confluent Hermite Vandermonde determinant is V^4. The four transverse gradient values have ordinary Vandermonde determinant V. Therefore a twelve-row divided-difference transformation has absolute determinant V^-5: + +```text +det Cov(O)=V^10 det Cov(normalized O), +p_O(target)=V^-5 * bounded_prefactor + * exp(-Gaussian_action/2). (18) +``` + +On a one-scale collision `x_i=epsilon a_i` with fixed distinct a_i this gives raw density prefactor `epsilon^-30`. In the two-pair regime it is `r_1^-5 r_2^-5 eta^-20`. + +For the two limiting contact vectors (12), confluent multiplicities four and four in the longitudinal channel contribute eta^16; multiplicities two and two in the transverse channel contribute eta^4. Hence + +```text +det Cov(U_1,U_2)=Theta(eta^40), +p_(U_1,U_2)(v)=O(eta^-20) exp(-action/2). (19) +``` + +The largest normalized derivative order is seven. The divided-difference transformation has entries bounded by `C eta^-7`, so the normalized target is bounded by `C eta^-7 |v|`. Regression with its uniformly nonsingular normalized covariance bounds conditional coefficient means by `C eta^-7 (1+|v|)`; conditional centered covariance is still a projection. On compact b_i,k_i marks this gives, for A_i=F_vv at the two centers, + +```text +E[A_1² A_2² | U=v] <= C eta^-28. +``` + +The literal single-contact planar weight from the source is + +```text +W_i/r_i² -> 36 k_i² A_i² 1{A_i<0}. +``` + +For any bounded selection factors, the limiting joint compact-positive-mark contact integrand is consequently bounded by + +```text +C eta^-48 exp(-c k_-²/eta^4), (20) +``` + +and is smaller than every power of eta. The exponent 48 is a sufficient crude polynomial bound, not a sharp claim; the exponential suppression is the useful fact. No equality of actual-bar selection or new process law is inferred. + +Equation (20) concerns exact contact limits. Uniform passage from finite radii to those limits as eta shrinks is a separate quantifier problem. The gradient-pinned finite-radius bound (11) avoids that exchange on the restricted collinear geometry. + +## 9. Diagnostic computations, not proof + +Files: + +- `coalescing_diagnostics.py` +- `COALESCING_DIAGNOSTICS.json` + +The script uses Python standard-library Fraction for exact polynomial evaluation-matrix ranks and Decimal at 95 significant digits for covariance determinants and actions. Sin/cos use a truncated Taylor summation without certified intervals. No field is sampled, no seeds are used, and there is no numerical H0 or Poisson test. + +The diagnostic law has uniform positive full-square spectral weights and omega=1 (L=2pi). It is a legitimate fixed finite law of the stated class, but not an identification with the L24 catalogue sampler. The exact rational rank checks give: + +| K | Eight gradient-jet polynomial rank | Twelve height/contact limiting-jet polynomial rank | +| ---: | ---: | ---: | +| 2 | 8 | 9 | +| 3 | 8 | 11 | +| 4 | 8 | 12 | + +For four sites 0,epsilon,2epsilon,3epsilon and K4 contact centers 0,epsilon, the decimal values are: + +| epsilon | det Cov(G_K2)/V^4 | det Cov(O_K4)/V^10 | det Cov(U_K4)/eta^40 | eta^4 * contact action | +| ---: | ---: | ---: | ---: | ---: | +| .2 | 1.27595559 | 62.91004725 | 2591309.75579909 | 1135.96506151 | +| .1 | 1.54859762 | 119.72582867 | 2945034.77606788 | 1147.42709118 | +| .05 | 1.62517191 | 140.51761514 | 3040737.50095657 | 1150.31525609 | +| .025 | 1.64488465 | 146.25015545 | 3065143.34549533 | 1151.03871914 | + +The action target is `k_1=k_2=1, b_1=0, b_2=eta³/5`, the height difference compatible with the local quartic derivative `a(s)=6 eta^-2 s²(s-eta)²`. This choice avoids gratuitously forcing a higher-order inconsistent birth difference. These observations are consistent with (1), (18), (19) and eta^-4 coercivity; they do not certify constants, asymptotic windows, orientation-uniform global bounds or theorem closure. + +The first diagnostic invocation failed because Decimal treats `0**0` as InvalidOperation; monomial evaluation was corrected to form the integer power before Decimal conversion. The successful rerun and its outputs are retained locally. That arithmetic repair supplies no scientific evidence. + +## 10. Scope that remains genuinely open + +The restricted lemma removes a naive covariance-versus-weight divergence in one geometry and demonstrates a support obstruction to another route. It does not settle the physical full-dimensional integral: + +- Four generic sites need not lie on a common straight line. Transverse offsets and independently oriented pair axes change which divided differences remain bounded. A uniform tubular/multiscale extension is needed before the collinear cancellation can control spatial measure. +- The regime k_i=O(eta²) is not suppressed by the compact-positive-mark bound. Its Jacobian, conditional transverse Hessian law, selection and mark integration must be calculated. +- Coalescing patterns where radii and center separation are comparable, endpoint permutations, shared endpoints in candidate factorial counts, and possible higher degeneracies must be partitioned explicitly. Actual distinct H0 bars have four distinct endpoints, but that fact does not furnish all covariance estimates. +- The K2/K3 singular twelve-row strata must be treated with a supported affine incidence formula or avoided by a strengthened sufficient carrier; assigning them a positive Lebesgue density is not an option. +- A local/fixed-law bound does not give uniform increasing-volume constants, mixing, a limiting coefficient as L varies, tightness or a spatial Poisson limit. + +A concrete next calculation is therefore to perturb the collinear configuration transversely with declared ratios to r_i and eta, derive a joint observation normalization with a uniform covariance floor, and determine the smallest transverse scale at which the Vandermonde cancellation survives. That is a finite extension of the demonstrated mechanism. It preserves the existing process owner's unproved coalescing-center and expanding-domain scope. diff --git a/experiments/universality/fold_isolation_diagonal/COUNTERFAMILY_NONAUTHOR_REVIEW.md b/experiments/universality/fold_isolation_diagonal/COUNTERFAMILY_NONAUTHOR_REVIEW.md new file mode 100644 index 000000000..290d64bf8 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/COUNTERFAMILY_NONAUTHOR_REVIEW.md @@ -0,0 +1,152 @@ +# Nonauthor review of the repeated-spectrum candidate counterfamily + +**Verdict: ACCEPT_SCOPED for §§3–5 of the exact frozen obstruction proposal.** I found no blocking mathematical error in the loss of one-contact rank under upper moments alone, the stronger uniformly one-contact-nondegenerate repeated-spectrum family, its macroscopic two-contact degeneracy, the explicit empty-in-probability candidate process, or the separate deterministic diagonal construction preserving a finite positive first moment. The mean-preserving version is an existence construction of smaller positive fills; it is not established here for the explicit choice epsilon_N=N^-12. No source amendment is required within this review's scope. + +This is a new conventional mathematical review by OpenAI/Codex `/root/c261a_local_runtime_20261016`, an actual separate nonauthor agent. The frozen proposal's authors are the core-theorem agent and Root steering. I read their source before reviewing it. This is source-exposed, same-provider OpenAI work with organizational-independence credit0, blind0 and human0. It is not a Lean/formal verification, scientific acceptance, point-process theorem for the project source, actual-global-bar counterexample, source admission, custody review, or authority transfer. The integrating reviewer separately owns §§6–9; I make no verdict on those sections here. + +## 1. Exact object, consumed inputs and execution + +The reviewed object is `JET_NONDEGENERACY_FACTORIZATION_OBSTRUCTION.md`, 34627 UTF-8 bytes, SHA-256 `0dfdddfbcac47b2fd07449f4152b7f5d935f92aed15272fb8333ed9cc60b5766`. I read the whole note to understand the boundary, and substantively checked only §§3–5 (lines39–177 of the frozen local file). The author control script is5818B / `5b1e8fc18a6e32e3a312a253e3e4d1110bb674385f3d49447a45e78ad4df541f`; its frozen JSON is1618B / `1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764`. + +The specific consumed source imports were checked in their retained immutable bytes: + +| Source | Bytes | SHA-256 | Imported scope | +|---|---:|---|---| +| finite_gaussian_contact/PROOF.md | 25367 | 7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474 | actual real-coefficient law, all-frame seven-row rank, small-r augmented rank/regression, determinant-weighted contact limit | +| finite_candidate_lifetime/PROOF.md | 29190 | ed31f562f0ffaaf6d4bd2ceb1ae3d234798e983a55d55c6983b128fce27133f5 | Sard Morse genericity, distinct critical values, all-Borel ordered-pair Rice identity, lifetime Jacobian and coefficient | +| FINITE_FOURIER_CRITICAL_COUNT_DIAGONAL_ROUTE.md | 25699 | 621cee96854aa440498cecf5452a8d1e23f90a4458779c3286c18eba106fe138 | only the accepted unconditional a.s. critical-point cap16K² | +| NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW.md | 21880 | 16152427f7013822ac1bba81142429800334e6bf0f354a6439b23a1997bc4524 | existing scoped acceptance of that finite-law cap; no duplicate review | + +The first two are pinned to Main85b928b326ff0586715a56ae91f51d3fbd89557a. I inspected the relevant local proofs directly, including FG§§1–5 and FC§§4–6,8–10, and opened the [pinned contact source](https://raw.githubusercontent.com/d6g8k5htny-coder/main/85b928b326ff0586715a56ae91f51d3fbd89557a/experiments/universality/finite_gaussian_contact/PROOF.md) and [pinned candidate source](https://raw.githubusercontent.com/d6g8k5htny-coder/main/85b928b326ff0586715a56ae91f51d3fbd89557a/experiments/universality/finite_candidate_lifetime/PROOF.md) as primary-source cross-checks. No new external theorem is required beyond those named imports and elementary finite-dimensional Gaussian, contraction and measure arguments explained below. I did not rerun the earlier accepted norm/count/isolation reviews. + +The actual command records are in `counterfamily-nonauthor-checks-20261016/EXECUTION_RECORD.json`. I ran the frozen author controls and my own independent `counterfamily_checks.py` with `python3 -B`. Both exited0, with empty stderr. The author stdout matches its1618B frozen JSON byte for byte. My187 exact Fraction/integer checks passed. The tests include18 rational oriented/reflected frames, independent moment-based covariance construction, full degree-three grid rank, determinant/trace/cubic-Schur identities, five mixture cases, nonprimitive-resonance residue arithmetic, short lifts across periodic seams, and meaningful negative controls. There are no field draws, approximate roots, numerical eigenvalues, Rice integrals or process simulations. The three author files were hashed before and after both runs and remained unchanged. + +The finite controls support arithmetic. The all-frame proof, a.s. genericity/stability, boundary null sets, diagonal existence and process conclusion are conventional arguments, not consequences inferred from sampled frames or control success. + +## 2. §3: upper moments do not supply a lower variance + +With physical frequencies xi_n=n/N and n in[-2N,2N]², every frequency satisfies |xi_n|²<=8. Thus sum q_n|xi_n|^12<=8^6=262144 for every displayed normalized mixture. Full positive fill gives q_n>0, symmetry and exact total mass1; the dominant DC contribution vanishes under every directional derivative. Therefore Var(F_u)<=8epsilon_N for every unit u. Selecting that coordinate in the seven-row covariance gives the claimed Rayleigh upper bound. Positivity for each fixed positive full square is consistent with collapse of its common lower bound. + +Putting the dominant mass instead on the four physical axial modes of length1/N gives directional gradient variance(1-epsilon_N)/(2N²) plus fill variance<=8epsilon_N. This verifies the second collapse without a dominant DC term. The small remaining DC weight from the positive fill is allowed; no claim that q_0 was removed entirely is necessary. Normalized variance, upper twelfth moments and individual smoothness do not impose any lower derivative variance. + +## 3. §4: actual law, full positivity and every-frame contact floor + +G has25 real independent Gaussian coordinates under the actual sine/cosine convention. Repeating it on the torus of side2piN places its physical frequencies at m, through Fourier indices n=Nm. Its derivatives are consequently the original physical derivatives, not derivatives scaled by1/N. Independent H_N supplies one independent Gaussian sine/cosine coordinate per mode pair plus the DC coordinate on the full square. At overlaps, summing independent Gaussian coefficients produces exactly the variance sum in equation(1); different mode pairs remain independent. The result is the stated centered real ideal Gaussian law with all(4N+1)² positive symmetric mode weights and unit variance. No independence between repeated cells is asserted or used. + +For a seven-row linear functional in any real orthonormal frame, its complex multiplier is a polynomial of total degree<=3 in the two physical frequency coordinates. Vanishing variance with positive weights on all25 grid points forces the polynomial to vanish at each point. Successively interpolating in each coordinate (five points, degree<=3) forces a polynomial identity. The invertible frame change then separates1,a,a²,a³,c,ac,c², forcing all seven real coefficients to vanish. Thus every frame, including reflections, has positive covariance. Entries are continuous in O(2); its compactness gives a common positive floor. This proves the all-frame assertion; an axial LDL calculation alone would not suffice. + +I also derived an independent explicit uniform lower bound for this particular uniform5×5 base. Let u=(c,s), v=(-s,c), c²+s²=1, and q=c²s² in[0,1/4]. For independent uniform X,Y in{-2,-1,0,1,2}, their moments are E X²=2, E X⁴=34/5, E X⁶=26. Odd and even derivative rows decouple by inversion parity. The conditional variance of F_uuu after both gradient rows is + +```text +E(cX+sY)^6 - (|E[X(cX+sY)^3]|²+|E[Y(cX+sY)^3]|²)/2 + = (26+126q) - (578+2106q)/25 + = (72+1044q)/25 >=72/25. +``` + +Changing from(H_xx,H_xy,H_yy) to(H_uu,H_uv,H_vv) has determinant±1. Hence the even block(F,H_uu,H_uv,H_vv) has determinant784/25, using the conditional diagonal-square variance14/5 and mixed variance4. With the two-gradient determinant4, this gives + +```text +det Gamma_G(u)=3136(72+1044q)/625 >=225792/625, +tr Gamma_G(u)=(243+682q)/5 <=827/10. +``` + +Reflections only change signs of rows involving v and do not affect these identities. Since each other eigenvalue is at most the trace, + +```text +m_G >= (225792/625)/(827/10)^6 + =361267200/319913861015774089 >0. +``` + +The polynomial/compactness proof already establishes the required result; this additional conservative floor checks the orientation issue quantitatively. The independent controls verify these identities at18 rational oriented/reflected frames and rank10 for the complete total-degree-three multiplier space. A K=1 negative control has the exact axial annihilator F_u+F_uuu=0, demonstrating why one cannot omit the required support argument. + +The covariance decomposition is a sum of positive semidefinite matrices. Thus Gamma_N>=(1-epsilon_N)Gamma_G and epsilon_N<=1/2 yields the claimed floor m_G/2, uniformly in N and frame. Every positive mixture has M12<=262144. The one-contact cubic variance stays positive; this obstruction does not discard the fold jet. + +## 4. §4: macroscopic coherence and two-contact degeneracy + +For j=1,...,N-1, the integer interval[-2N,2N] has four copies of every residue moduloN and one extra zero residue. Evaluating its character at any nontrivial Nth root gives sum1, whether or not the root is primitive. The transverse sum contributes4N+1. Therefore the normalized fill covariance at h=(2pij,0) is exactly1/(4N+1). The repeated G and all its physical derivatives agree at those translations, so R_N(h)=1-epsilon_N+epsilon_N/(4N+1). + +The unit two-value difference vector has variance1-R_N(h)=epsilon_N(1-1/(4N+1))=4, so this is a genuinely far resonance. Value correlation tends to1; far covariance decay/factorization cannot be deduced from the one-contact floor and upper moments. + +This spectral obstruction alone would not disprove every Poisson conclusion: an empty limit is Poisson with zero intensity, and failure of one sufficient covariance estimate is not failure of a limit theorem. The actual process argument in§5 is therefore necessary and was reviewed separately below. + +## 5. §5: the global C4 estimate and uniform repeated-cell stability + +There are15 coordinate multiindices alpha with |alpha|<=4. At each x, the physical coefficient row for partial^alpha H_N has squared norm equal to its derivative variance, at most64²: normalized mass1 and |xi|<=2sqrt2 give row norm<=(2sqrt2)^4=64, with the same bound for lower orders. Cauchy–Schwarz against the real coefficient vector Z_N gives the global supremum bound for each row, including mixed derivatives. Summing gives the stated coordinate sum-of-suprema bound||H_N||C4<=960||Z_N||. There are exactly(4N+1)² real coordinates, not twice that number. Thus the noise squared-norm expectation is at most960²epsilon_N(4N+1)²=O(N^-10) when epsilon_N=N^-12. + +The amplitude adjustment of G is also harmless: 1-sqrt(1-epsilon)<=epsilon and G has a finite C4 second moment in25 coordinates. Its repeated physical C4 norm equals its base norm. Combining the two terms, with a factor2 when squaring their sum, proves the required global perturbation convergence. The constant960 is for the stated coordinate C4 sum; ordinary directional/operator conventions differ by fixed dimension constants and do not change convergence or the deterministic radius choice. No local-norm theorem from the previous review is imported. + +The fixed-base source establishes that G is Morse almost surely before Rice, and uses the all-Borel pair formula to prove distinct critical values. Its critical set is finite by compactness and isolation. Take disjoint lifted balls around its critical points, smaller than the base injectivity scale. At each p, write H_p=Hess G(p), and arrange||H_p^-1(Hess G(x)-H_p)||op<1/4 throughout the ball. Choose the perturbation radius to make the analogous additional Hessian term<1/4. Then + +```text +T_p(x)=x-H_p^-1 grad f(x) +``` + +has Lipschitz constant<1/2. Existence, as well as uniqueness, follows by also choosing||H_p^-1 grad f(p)||<=R_p/2: T_p maps the closed ball of radiusR_p into itself and is a contraction. This supplies a unique critical point there. The sufficiently small Hessian error preserves its Morse index. A positive base gradient floor on the complement excludes every other root. Repeating those finite balls and the same gradient floor proves all assertions with a radius independent of the number of cells. Large perturbation frequency cannot create additional roots when this global C2 bound holds. + +A C1-only statement would be false: for a local base derivative x, a perturbation sin(m²x)/m is uniformly small but gives signs+,-,+ at pi/m²,3pi/(2m²),2pi/m² for m>=8, hence two additional roots besides0. Its Hessian error is of orderm, so it violates precisely the C2 hypothesis used above. This is retained as a negative control against silently weakening stability. + +Let Delta(G) be the minimum absolute height gap between a base maximum and a base saddle. It is positive almost surely. Shrink the stability radius so each perturbed critical height is withinDelta/4 of its corresponding base height. Distinct repetitions of the same base critical point have the same Morse index and cannot provide a maximum–saddle pair with base gap0. Thus every positive typed candidate gap in the large torus is at leastDelta/2. Height ties between repetitions of one point are allowed; asserting distinct global critical values for G_repeated would be false. The endpoint role restriction is essential here. + +For completeness the random-radius argument needs both the radius and the gap. Given a small error tolerance, choose deterministic d,D>0 with P(delta(G)<=d or Delta(G)<=D) small. For largeN, Tt_N=d tends to0. On the remaining event there is no candidate of gap<=Tt_N. This uses an unconditional union bound, not Gaussian conditioning at a data-selected critical point. A measurable positive radius can be chosen from countable coefficient neighborhoods of the finite base Morse branches; no probabilistic independence of that radius and the perturbation is required. + +It follows that every bounded-lifetime compact candidate mark window has zero mass with probability tending to1. Rescaling locations byL_N does not change that statement. For a nonnegative continuous compactly supported test, the difference of its Laplace functional from1 is bounded by the probability that the supporting window contains a candidate. Therefore the process tends to the empty process on this sector. This excludes a universal **nonzero** Poisson limit under U plus the one-contact floor; it remains compatible with zero intensity. It does not prove a statement about an elder/global H0 selector. + +## 6. §5: fixed-base restricted coefficient is finite and strictly positive + +The additional compact window b in[-1,1], k in[1,2], A_M in[-2,-1] is an allowed coefficient/position Borel mark in the literal ordered-pair Rice identity. For sufficiently small r, the seven-row augmented covariance is uniformly positive over all frames. Consequently A_M conditioned on the six exact pins has positive scalar variance; at contact its Gaussian density is positive on every point of[-2,-1]. It need not be centered. The six-pin density at v_0=(b,0,0,12k,0,0) is likewise strictly positive. Thus z_0^W>0 pointwise, and compact continuity/domination gives a finite strictly positive integral. No coefficient positivity is postulated or inferred from an axial numerical test. + +The indicator on A has zero limiting Gaussian boundary mass. Conditional Gaussian coupling plus uniform moments therefore transfers the determinant-weighted contact limit with that window. On this compact negative-curvature interval, type limits have no additional singular A=0 boundary. The contact determinant product36k²A² and the six-pin density occur exactly once each. + +I checked the constants independently from the literal ledger: + +```text +(r dr d sigma)(r^3 db dk)(12r^-5 pin_density)(r^2 V_r) + =12r dr db dk d sigma (pin_density V_r), +r dr/dell = (1/3)ell^-1/3 k^-2/3. +``` + +Thus the lifetime density has coefficient4 times the displayed b,k,u integral. Integration over[t/2,t] multiplies by(3/2)(1-2^-2/3)t^(2/3), giving coefficient6V_0, exactly as in equation(8). V_0=(2pi)² is the base torus volume. The arclength measure has total2pi; it is not normalized. Endpoint roles are ordered with u from maximum to saddle and v=Ju, so there is no extra1/2 or reflected-frame multiplicity. This verifies c_W and its positive finite limiting mean without computing its numerical value. + +## 7. §5: exact short lifts and null boundaries for diagonal continuity + +Choose N sufficiently large that N^-1=1 imply r<=N^-1. Hence r0 is an inactive cutoff: its boundary cannot occur in the window. The endpoints lie below the base injectivity scale. Projecting a short pair to the base torus is unique; for each lifted maximum there is one short lift of its specified base saddle. This remains true across a base-cell seam. There are exactlyN² translations, notN^4 arbitrary combinations of endpoint lifts. My independent finite cyclic-lift enumeration checks both ordinary and seam-crossing examples and rejects unrestricted multiplicity as a substitute. + +For each fixed sufficiently largeN, the base field has finitely many critical branches and is away from type degeneracy almost surely. Relevant births/lifetimes lie off fixed boundaries by the source's absolutely continuous two-height Rice disintegration. Boundaries k=1 or2 are null by the positive height-to-k Jacobian r³ at each nonzero separation. Boundaries A_M=-2 or-1 are null using the small-r positive conditional scalar variance and the same Borel Rice identity. The determinant weight is integrable for this fixed finite field; multiplying a zero-probability scalar boundary event by it does not produce mass. Torus cut loci are absent in the relevant short-radius window. Orientation has no additional restricted boundary. + +The copied field's same-type height ties do not invalidate this argument: every relevant typed short pair projects to two different base critical points, where the null-boundary proof is applied. Base max–saddle gaps are not0; copied points cannot change index under the sufficiently small perturbation. There are only finitely many branches/pairs at each fixedN. Away from the finitely many mark-boundary events the count is therefore locally constant in C4 (indeed C2 suffices for these branch/curvature marks). Since F_epsilon,N tends almost surely to G_repeated in C4 as epsilon decreases at fixedN, bounded convergence of the disagreement indicator proves P(C_epsilon,N!=C_dom,N)->0. + +The proof only needs sufficiently largeN; arbitrary initial finitely many fills can be chosen positive. It does not need null curvature-boundary claims for a non-small-r band where the augmented-rank source supplied no such assertion. + +## 8. §5: deterministic diagonal, count cap and first-moment transfer + +For each sufficiently large fixedN, choose a deterministic dyadic epsilon_N>0 with epsilon_N<=N^-12 and disagreement probability<=N^-10. Existence follows from the fixed-N limit just proved. For example take the first dyadic satisfying both inequalities; the eligible set is nonempty. The probabilities concern the unconditioned fixed law of(G,H_N), so this is a choice of a spectral coefficient sequence before the field draw, not an adaptive sample-dependent choice or conditioning. No computable rate, evaluated probability or numerical epsilon sequence is claimed. + +For every positive epsilon_N, the full coefficient law on K=2N is absolutely continuous. The accepted fixed-K critical bound can be applied separately at eachN, retaining its explicit K dependence: Ncrit<=16(2N)²=64N² a.s. Thus the total endpoint-candidate cap is4096N^4. No new covariance/Rice weight is inserted in this direct count bound. The sparse repeated dominant field is not assigned full-carrier absolute continuity; its short count is bounded instead by the exactN² lifts of the fixed-base count<=4096, giving4096N². This distinction avoids an invalid application of the generic bound to the degenerate sparse coefficient law. + +On the disagreement event, the absolute difference is at most the sum of those caps. Therefore + +```text +E|C_epsilon_N,N-C_dom,N| + <= (4096N^4+4096N²)N^-10 + <=8192N^-6 ->0. +``` + +A merely N^-4 disagreement estimate would not establish this transfer: its first cap contribution remains4096. That failure is checked independently. All constants needed from the earlier count lemma are tracked here; no uniform tail/cutoff constants from that earlier fixed-law review are being imported acrossN. + +The exact lift identity and fixed-base asymptotic give + +```text +E C_dom,N(N^-3) + =N² E C_G(N^-3) + -> c_W(1-2^-2/3), with 0=k_->0`, and transverse curvature at each maximum is at most `-a<0`. Candidate identity means the ordered endpoint pair `(M,S)`, not a frame presentation. Distinct pairs sharing a maximum are included and controlled below; the identical-pair diagonal is excluded. Any additional Borel selection factors bounded by one may be retained. No four-site Rice density or conditional covariance inverse is used. + +For gaps `ell<=Tt`, compact positive `k` forces endpoint radii `r<=(Tt/k_-)^(1/3)`. With `eta(t)=c0/log(1/t)` and `c0=c_*/64`, sufficiently small `t` therefore gives + +```text +E D(t,eta(t)) <= B_K^4 2^(N/2) t^4, +t^(-4/3) E D(t,eta(t)) = O(t^(8/3)) -> 0. (2) +``` + +This is an integrated near-diagonal estimate on a **fixed** finite carrier and fixed positive mark window. The far-center region with a moving cutoff, small transverse/cubic marks, increasing-domain uniformity, global bar pairing and a full point-process limit remain separate. + +The isolation inputs come from the unchanged frozen [QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md](QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md), SHA-256 `827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30`. Its controls and output retain SHA-256 `13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5` and `ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf`. This report does not amend those bytes or silently broaden their stated actual-center ratio hypothesis; Section 6 proves the required broader cutoff statement afresh. + +## 2. Pinned field, source convention and classical imports + +Keep Main `85b928b326ff0586715a56ae91f51d3fbd89557a`, [finite_gaussian_contact/PROOF.md](https://github.com/d6g8k5htny-coder/main/blob/85b928b326ff0586715a56ae91f51d3fbd89557a/experiments/universality/finite_gaussian_contact/PROOF.md), Git blob `5f345201309cfa1781e92d6292f542de4bcb2999`, 25,367 bytes, SHA-256 `7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474`. Its exact local body and tree identity were checked. Lines 41–57 supply the finite law, 68–76 the literal endpoint-distance convention, 145–186 the divided differences and factor 12, and 253–307 the physical Hessian types and weights. + +Fix `K>=2`, `L>0`, `omega=2pi/L`, `S_K={-K,...,K}^2`, positive symmetric weights `q_n=q_-n`, and `sum q_n=1`. With one representative of every nonzero pair and `N=(2K+1)^2` independent standard real Gaussian coefficients `Z`, + +```text +F(x,y)=sqrt(q_0)C_0 + + sum_pairs sqrt(2q_n)[C_n cos(omega n.(x,y)) + +S_n sin(omega n.(x,y))]. (3) +``` + +All constants in this report are for this fixed law. Positivity of all weights makes its real trigonometric coefficient distribution nondegenerate on the whole real Fourier parameter space. Its complex Fourier coefficients `a_n` satisfy `a_-n=conjugate(a_n)`. Neither rotational isotropy nor independent fields are substituted. + +Classical imports, with primary sources inspected during this analysis: + +* **Planar Bézout inequality:** coprime polynomials over a field have at most the product of their degrees many common affine zeros. This is Larry Guth's [MIT polynomial-method lecture 13, Theorem 0.1, page 1](https://math.mit.edu/~lguth/PolyMethod/lect13.pdf). Taking the field to be `C` permits reducible coprime polynomials. MIT [Algebraic Geometry, §7.8.1 and the following note, printed page 179](https://math.mit.edu/classes/18.721/ag-may31-2021.pdf) also explicitly allows relatively prime homogeneous forms without irreducibility, using zero divisors with multiplicities. We need only the affine inequality, not an exact multiplicity calculation. +* **Projective closed images:** projective varieties are proper; a morphism from a proper variety into a separated variety has closed image. The precise imports are Stacks [Tag 01WC, Lemma 29.44.5](https://stacks.math.columbia.edu/tag/01WC) and [Tag 01W6, Lemma 29.42.7](https://stacks.math.columbia.edu/tag/01W6). Section 4 constructs the actual multiplication maps to which these imports apply. +* **Measurable projections:** projections of Borel sets in Polish product spaces are analytic; analytic sets are measurable for the completion of a Borel probability measure. The universal-measurability conclusion is also stated in Larson–Neeman–Shelah, [*Universally measurable sets in generic extensions*, §1, printed page 175](https://shelah.logic.at/files/94751/947.pdf). This is a named classical measure-theoretic import, not a new formal proof. + +Saved primary-source copies and SHA-256 identities are in [critical-count-primary-sources/SOURCE_IDENTITIES.json](critical-count-primary-sources/SOURCE_IDENTITIES.json): Guth PDF 83,539 bytes, SHA-256 `92ed0fc0f940179871a25f3a244eb181aaa0176ee5738ac1f8f68a8bacd072a9`; Stacks 01WC HTML 15,419 bytes, SHA-256 `08a39f68a8159e34e436df9c4b8f5e8886f58b3c0d5163c15610e12aef302737`; Stacks 01W6 HTML 14,844 bytes, SHA-256 `060b8c179f815d59977f8a31057b5f5f9e1f2104351d46c33f1fb6d9ed95a202`. The MIT algebraic-geometry and Larson–Neeman–Shelah statements were inspected through browsing; no locally saved byte identity for those two PDFs is claimed. These are classical imports, not new results attributed to this proposal. + +## 3. Polynomial reduction and full-degree witness + +Put `z=exp(i omega x)`, `w=exp(i omega y)` and view `F=sum a_(m,n) z^m w^n` as a Laurent polynomial. Define + +```text +P(z,w)=z^K w^K sum_(m,n) m a_(m,n) z^m w^n, +Q(z,w)=z^K w^K sum_(m,n) n a_(m,n) z^m w^n. (4) +``` + +The omitted common gradient factor is `i omega`, which is nonzero. Exponents in each variable lie between 0 and `2K`, so both total degrees are at most `D=4K`. On `(C*)^2`, multiplication by `z^K w^K` creates no zeros. The real torus embeds injectively in the unit-circle product, so every distinct torus critical point gives a distinct common affine zero of `P,Q`. Zeros on coordinate axes, or extra complex zeros, only enlarge the count used for the upper bound. + +Take the real, permitted trigonometric coefficient witness + +```text +F_*(x,y)=cos(K omega x)+cos(K omega y) + +c cos(K omega(x+y)), c real, c!=0. +``` + +All three modes lie in the actual full square; the strictly positive mode variances allow this deterministic coefficient vector in its real support. Multiplying (4) by the harmless common factor `2/K`, and writing `Z0=z^K`, `W0=w^K` to distinguish these monomials from the Gaussian coefficient vector, gives + +```text +P_*=W0(Z0^2-1)+c(Z0^2 W0^2-1), +Q_*=Z0(W0^2-1)+c(Z0^2 W0^2-1), +P_*-Q_*=(Z0-W0)(Z0 W0+1). (5) +``` + +Both have exact total degree `4K`, with nonzero leading monomial `c z^(2K)w^(2K)`. To prove coprimeness **after lifting back to z,w**, work in the UFD `C[z,w]`. A common irreducible factor must divide the difference in (5). Its possible factors are therefore + +```text +z-zeta w, zeta^K=1, +zw-rho0, rho0^K=-1. (6) +``` + +The first family consists of linear irreducibles. For the second, `zw-rho0` is irreducible: its quotient ring is `C[w,w^-1]` under `z=rho0/w`, an integral domain; equivalently the polynomial is primitive and degree one in `z` over `C(w)`. There are no coordinate factors in the factored difference. + +On each line `z=zeta w`, we have `Z0=W0`; the restriction is + +```text +P_*|line = W0(W0^2-1)+c(W0^4-1), (7) +``` + +which is not identically zero, since its fourth-degree coefficient in `W0` is `c!=0`. On each hyperbola `zw=rho0`, we have `Z0 W0=-1`; in its Laurent quotient ring, + +```text +P_*|hyperbola = W0^(-1)-W0, (8) +``` + +also nonzero. Thus none of (6) divides `P_*`, and `gcd(P_*,Q_*)=1` in `C[z,w]`. These restrictions prove the all-K statement; checking a few numerical values of `K,c` is not its proof. + +Now homogenize both polynomials to the **fixed degree** `D=4K` using a third variable `T0`. Since both affine polynomials have degree exactly `D`, neither homogenization is divisible by `T0`. Any common homogeneous irreducible factor other than `T0` would dehomogenize to a nonconstant common affine factor: a homogeneous polynomial whose dehomogenization is a nonzero constant is a scalar power of `T0`. Hence the homogeneous witness forms are coprime as well. Their restrictions to `T0=0` equal the same nonzero monomial `c z^(2K)w^(2K)`. They may intersect at points at infinity, but they have **no common line or curve there**. + +The tempting lower-degree witness `cos(K omega x)+cos(K omega y)` has affine coprime gradients, but their degrees are only `3K`. Homogenizing them to `4K` supplies a common factor `T0^K`. That witness therefore does **not** prove that the fixed-degree projective bad locus below is proper on the real parameter family. The additional diagonal mode fixes exactly this problem; the failed witness is retained as a negative control. + +## 4. The actual common-factor bad set is algebraic and proper + +Let `V_d` be the complex vector space of homogeneous degree-d forms in `z,w,T0`, and let `D=4K`. For each integer `1<=d<=D`, consider the morphism + +```text +mu_d : P(V_d) x P(V_(D-d) direct_sum V_(D-d)) + -> P(V_D direct_sum V_D), +([H],[A,B]) |-> [HA,HB]. (9) +``` + +Its coordinates are bihomogeneous of degree `(1,1)`. They do not vanish simultaneously: `H!=0`, at least one of `A,B` is nonzero, and a polynomial ring over `C` has no zero divisors. Thus (9) is everywhere a morphism, rather than a rational map with unresolved base points. The domain is a product of projective spaces and hence projective, for example by the Segre embedding. The closed-image imports in Section 2 show that `image(mu_d)` is Zariski closed. Their finite union is precisely the projective locus of nonzero form pairs with a positive-degree common factor, including the case that exactly one form is zero. + +Take the affine cone over this closed projective locus and include the zero pair. Call it `Bad_D`. It is a complex algebraic closed subset of `V_D direct_sum V_D`. Homogenizing (4) to fixed degree `D` is a linear map `Lambda` from the real Fourier coefficient space into this complex vector space. Its coefficients can be complex constants; its dependence on the real sine/cosine coefficients is still real-linear. Define + +```text +Exceptional = {real coefficients theta : Lambda(theta) in Bad_D}. (10) +``` + +The witness in Section 3 belongs to this **same real** coefficient space and lies outside `Bad_D`. Consequently there exists a polynomial equation of `Bad_D` whose pullback is nonzero at that real witness. At least one of the real or imaginary parts of this pullback is a nonzero real polynomial in the actual real parameters. Thus `Exceptional` is contained in a proper real polynomial zero set. This is the step that excludes a merely complex-generic argument with no real witness. + +A nonzero real polynomial has Lebesgue-null zero set: induct on the number of variables, split off the null set where every coefficient in the last variable vanishes, and use that a nonzero one-variable polynomial has finitely many roots on every remaining fiber. Fubini completes the induction. Since (3) has a nondegenerate Gaussian density in its real coefficient coordinates, `P(Exceptional)=0`. + +Outside this single null set the two fixed-degree homogeneous forms have no common factor. Their dehomogenizations in particular have no common affine factor. This bad set is a sufficient exceptional set; it deliberately also includes some lower-degree pairs whose affine gradients are coprime, because fixed-degree homogenization can introduce `T0`. No claim that it is the smallest possible exceptional set is needed. + +By planar Bézout over `C`, + +```text +Ncrit(F) <= # common affine zeros of P,Q + <= (deg P)(deg Q) <= (4K)^2 = 16K^2 (11) +``` + +almost surely. Distinct real critical points are counted once, without multiplicity. Degenerate isolated critical points are included. This proof does not establish Morse regularity or distinct critical values, and neither is needed for (11). It is an almost-sure deterministic bound for fixed K, rather than a bound for every coefficient vector: for example a field depending only on x can have entire critical circles and belongs to the exceptional family. + +## 5. Exact count and orientation conventions + +Fix `a,k_->0`, optionally upper/height compact mark bounds, `T>0`, and `00, u=(S-M)/r, v=J u, +ell=F(M)-F(S)>0, k=ell/r^3>=k_-, +F_vv(M)<=-a, ell<=Tt. (12) +``` + +Here `J` is the fixed positive 90-degree rotation. The reflected choice of v gives the same curvature but is **not** counted again. If a different parameterization counts both reflected frames, its fixed presentation multiplicity must be explicitly divided out or bounded separately; `B_K^4` below concerns endpoint-pair identity. Its center is the geodesic midpoint `z_p`. No persistence or elder-pairing requirement is imposed. Such a selector can only reduce the count if later supplied as a Borel factor in `[0,1]`. + +Let `C_t` be this candidate population. For `eta>0`, define the ordered factorial near-diagonal count + +```text +D(t,eta)=sum_(p,q in C_t, p!=q) + 1{d(z_p,z_q)<=eta} + 1{r_p<=rho eta, r_q<=rho eta}. (13) +``` + +The equality `p=q` means both endpoint roles agree. Distinct pairs may share their maximum or saddle. This is the product count on two **different candidates**, not the full square count that includes `p=q`. An unordered convention divides (13) by two. No further factor `1/2`, spatial volume, angular normalization or Hessian determinant is included in (13). + +On the probability-one set (11), there are at most `Ncrit^2` endpoint candidates, hence at most `Ncrit^4<=B_K^4` ordered pairs of candidates. Thus + +```text +0<=D(t,eta)<=B_K^4 1{D(t,eta)>0} (14) +``` + +almost surely. One may set the count to zero on the exceptional null set. This is the finite combinatorial bound that turns an occurrence estimate into an expected **integrated** count. + +For measurability, work on the completed Gaussian probability space. Each event `{D>=m}` can be expressed by the existence of m distinct qualifying ordered candidate-pair configurations, a Borel incidence condition in coefficient and finite position variables. Requiring distinctness and all the derivatives, strict types and short midpoint conditions is Borel. Its projection is analytic, and hence measurable in the completed Gaussian law by the classical import in Section 2. There are countably many integers m. This supplies a measurable count without introducing a joint Rice formula. The same reasoning covers additional specified Borel selector marks, or one can simply dominate their weighted sum by the measurable unselected count (13). + +## 6. Fresh isolation lemma for a bounding cutoff, not an actual-center ratio + +The frozen isolation proposal's two-pair statement assumes `r_i<=rho s` where `s=|z_2-z_1|` is the **actual** center separation. Small lifetimes imply small radii, but `r_i<=C t^(1/3)` and `s<=eta(t)` do not imply this condition if `s` is much smaller than `t^(1/3)`. For example `eta=1/10`, `rho=1/4`, `s=1/10000`, `r=1/1000` satisfies `r<=rho eta` but violates `r<=rho s`. We do not import the frozen tail statement with this missing premise. + +Instead fix `R0<=L/16` and `0=c_*/eta. (17) +``` + +To prove this, suppose `W^2=1` is used in bounding the `C` term by `beta_4 W^2`. They give its oriented soft eigenframe cubic at least `12k_1-6k_->=tau`. The anchored chart radius satisfies + +```text +R>=kappa/W^2>kappa eta/c_*>=(1+rho)eta. (18) +``` + +There are now two incidence cases. + +* **Distinct maxima:** `|M_2-M_1|<=|z_2-z_1|+(r_1+r_2)/2<=(1+rho)eta=sqrt(c_*/eta)-1>=sqrt(c_*/eta)/2`. Since `Z` has N independent standard real coordinates, + +```text +P(D(t,eta)>0) + <=P(||Z||^2>=c_*/(4eta)) + <=E exp(||Z||^2/4) exp[-c_*/(16eta)] + =2^(N/2) exp[-c_*/(16eta)]. (19) +``` + +Multiplying by (14) proves (1). This avoids conditional jet-density singularities by using a global a.s. finite bound on the number of candidate configurations. It does not assert a pointwise bound on a four-site Rice density or justify conditional covariance formulas on singular strata. + +## 7. Logarithmic cutoff and literal normalization + +Let `A=(T/k_-)^(1/3)`. Every candidate (12) has `r<=A t^(1/3)`. Choose + +```text +c0=c_*/64, +n=log(1/t), eta(t)=c0/n, 00`, so eventually every such candidate radius satisfies the cutoff condition `r<=rho eta(t)`. This is a comparison to the **outer cutoff**, not to an individual pair's actual center distance. Thus eventually (13) counts every two distinct compact-fold candidates with center distance at most `eta(t)`. + +One explicit sufficient small-t threshold is `t<=exp(-n0)`, where + +```text +n0=max(1, c0/eta0, 4c0/c_*, + 6 log_+(6A/(rho c0))), +log_+(x)=max(0,log x). (21) +``` + +Indeed `n exp(-n/3)<=6exp(-n/6)` follows from `n<=6exp(n/6)`. The last term in (21) therefore ensures `A n exp(-n/3)<=rho c0`, while the other terms ensure `eta(t)<=eta0`, `eta(t)<=c_*/4`, and `n>=1`. Equality causes no problem in the non-strict geometric assumptions. + +Equations (1) and (20) now give (2), with constants for this fixed law, a, k_-, rho, T, R0 and eta0. More generally any `04/3` and a vanishing `t^(-4/3)` normalization. The chosen `c0=c_*/64` supplies a simple margin. On the torus of area `V=L^2`, normalizing the factorial count per ordered pair of spatial volumes replaces the prefactor by `B_K^4 2^(N/2)/V^2`. No such factor is present in (13) itself. A per-one-volume convention would instead divide by V once; the exponent is unchanged because V is fixed. + +This calculation uses cumulative candidate counts in a lifetime band `0 this agent's independent factor attack, proof reconstruction, new cutoff/incidence argument and finite exact controls -> pending separate nonauthor review. All current authors belong to OpenAI/Codex in the same parent thread, so provider/organizational-independence credit remains 0. The previous three isolation files are unchanged. New mathematical proposals and saved primary-source copies exist only locally; no production or remote write and no scientific-status transition occurred. diff --git a/experiments/universality/fold_isolation_diagonal/FOLD_ISOLATION_CONTROLS.json b/experiments/universality/fold_isolation_diagonal/FOLD_ISOLATION_CONTROLS.json new file mode 100644 index 000000000..7f51cec43 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/FOLD_ISOLATION_CONTROLS.json @@ -0,0 +1,23 @@ +{ + "classification": "EXACT_FINITE_ALGEBRA_CONTROLS_ONLY", + "arithmetic": "Python Fraction; no floating arithmetic", + "sampling": "NONE", + "process_claim": "NONE", + "scientific_effect": "NONE", + "checks": { + "all_mixed_reduced_third_derivative_identity": "PASS", + "endpoint_distance_and_factor12": "PASS", + "quartic_positive_mark_zero_endpoint_cubic_counterexample": "PASS", + "mixed_cubic_Bsquared_over_a_counterexample": "PASS", + "false_full_height_barrier_counterexample": "PASS", + "two_maxima_derivative_scale_control": "PASS", + "chart_margin_constants": "PASS" + }, + "exact_values": { + "cubic_k": "1/12", + "mixed_cubic_sign_loss_x": "-2/49", + "quartic_gap": "40/9317", + "horizontal_escape_height": "7/60000", + "two_maxima_separation": "1/17" + } +} diff --git a/experiments/universality/fold_isolation_diagonal/JET_FACTORIZATION_OBSTRUCTION_CONTROLS.json b/experiments/universality/fold_isolation_diagonal/JET_FACTORIZATION_OBSTRUCTION_CONTROLS.json new file mode 100644 index 000000000..ce40a1f8c --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/JET_FACTORIZATION_OBSTRUCTION_CONTROLS.json @@ -0,0 +1,49 @@ +{ + "classification": "EXACT_FINITE_ALGEBRA_CONTROLS_ONLY", + "arithmetic": "Python Fraction and integers; no floating arithmetic", + "Gaussian_sampling": "NONE", + "process_simulation": "NONE", + "scientific_effect": "NONE", + "checks": { + "fixed_base_axial_seven_jet_exact_LDL_positive": "PASS", + "positive_full_square_normalized_uniform_physical_M12": "PASS", + "single_jet_mixture_identity_and_resonant_two_value_collapse": "PASS", + "resonant_uniform_fill_geometric_sum": "PASS", + "DC_single_jet_failure_bookkeeping": "PASS", + "small_fill_C4_dimension_ledger_and_expectation_transfer_cap": "PASS", + "ordered_fold_Jacobian_and_middle_far_kernel_exponents": "PASS", + "consumed_frozen_files_unchanged": "PASS" + }, + "base_axial_LDL_pivots": [ + "1", + "2", + "14/5", + "72/25", + "2", + "4", + "14/5" + ], + "finite_family_cases": [ + { + "N": 4, + "M12": "837542830378721313/17592186044416", + "two_value_small_eigenvalue": "1/17825792" + }, + { + "N": 5, + "M12": "1251428171060188861397852/26285648345947265625", + "two_value_small_eigenvalue": "4/1025390625" + }, + { + "N": 8, + "M12": "695556060602940566395768283/14609821306377964879872", + "two_value_small_eigenvalue": "1/70866960384" + }, + { + "N": 12, + "M12": "3863598447035119394611266277217/81153031520128153386221568", + "two_value_small_eigenvalue": "1/9101852540928" + } + ], + "scope": "Finite checks support arithmetic; all-frame positivity, actual process obstruction, and covariance-kernel estimates require the conventional report proofs." +} diff --git a/experiments/universality/fold_isolation_diagonal/JET_NONDEGENERACY_FACTORIZATION_OBSTRUCTION.md b/experiments/universality/fold_isolation_diagonal/JET_NONDEGENERACY_FACTORIZATION_OBSTRUCTION.md new file mode 100644 index 000000000..c3358bfda --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/JET_NONDEGENERACY_FACTORIZATION_OBSTRUCTION.md @@ -0,0 +1,405 @@ +# Uniform smooth moments do not supply spatial jet rank or factorization + +16 October 2026 UTC. Local mathematical proposal by OpenAI/Codex `/root/core_theorem_audit_20261015`, following root's targeted obstruction request and independently constructing the examples and estimates below. Earlier isolation/norm sketches are joint root/agent work; this document is a new obstruction/factorization analysis, not a first review of them. Organizational-independence credit **0**. Pending separate nonauthor review. No production/native/remote edit, source claim, owner transfer, Lean execution or scientific-status transition. + +## 1. Main findings + +Normalized spectral mass and a uniform twelfth physical moment supply upper smoothness bounds. They supply neither lower jet variance nor decorrelation. There are two exact positive full-square counterfamilies: + +1. A dominant DC or long-wavelength component makes the single-contact covariance floor tend to zero, despite normalized mass and uniformly bounded physical moments. +2. More sharply, a repeated fixed finite Gaussian spectrum plus an arbitrarily small positive full-square fill has a **uniformly positive seven-row one-contact covariance floor**, while its two-contact covariance floor tends to zero and correlations at macroscopic torus separations tend to one. + +The second example also gives an actual candidate-process obstruction, not just a failed covariance estimate. On t_N=N^-3 and L_N=2pi N, the compact-fold candidate process tends to the empty process in probability. A deterministic sufficiently small positive fill can at the same time preserve a finite nonzero expected compact-fold count, using the fixed-base source's literal marked kernel and finite-N continuity. Thus a universal nonzero spatial Poisson conclusion cannot follow from uniform smoothness plus one-contact rank, even with this nonzero first-moment scale. No assertion about global persistence bars is transferred from this endpoint-candidate example. + +A possible replacement is quantitative physical covariance control: a uniform single-contact floor, an explicit two-contact floor on the moving middle band, and decay of cross-contact covariance on the far band. Sections 6–8 prove what these hypotheses buy for the actual determinant-weighted candidate kernel, with a precise middle/far ledger. They are finite covariance/derivative estimates to establish for a source, rather than certificates that assume a point-process conclusion. Section 9 identifies a concrete Gaussian truncation family that could satisfy them; it retains the estimates and review work still needed before source adoption or process closure. + +## 2. Authoritative consumed source and exact observable + +The repository pin is Main `85b928b326ff0586715a56ae91f51d3fbd89557a`. The complete retained source identities were rechecked against its immutable tree: + +| Source | Git blob | Bytes | SHA-256 | +| --- | --- | ---: | --- | +| [finite_gaussian_contact/PROOF.md](https://github.com/d6g8k5htny-coder/main/blob/85b928b326ff0586715a56ae91f51d3fbd89557a/experiments/universality/finite_gaussian_contact/PROOF.md) | `5f345201309cfa1781e92d6292f542de4bcb2999` | 25367 | `7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474` | +| [finite_candidate_lifetime/PROOF.md](https://github.com/d6g8k5htny-coder/main/blob/85b928b326ff0586715a56ae91f51d3fbd89557a/experiments/universality/finite_candidate_lifetime/PROOF.md) | `a27a84150b3e70bfd794672eeeed6915425a1da4` | 29190 | `ed31f562f0ffaaf6d4bd2ceb1ae3d234798e983a55d55c6983b128fce27133f5` | +| [iid_short_bar_process/PROOF.md](https://github.com/d6g8k5htny-coder/main/blob/85b928b326ff0586715a56ae91f51d3fbd89557a/experiments/universality/iid_short_bar_process/PROOF.md) | `cd7ef8b8a364cb394be48208f55332658e9fff0b` | 30935 | `320bc988c468bea4938bb777a2d05952c6ab8e0a384f33725b50de847aeec1eb` | + +The finite source proves contact rank for each fixed positive square spectrum. Its constants are expressly not uniform over vanishing weights, unbounded K or varying L. The process source's §§1–5 prove an iid-copy limit on one fixed torus for its actual **infinite** periodized Gaussian law. Its §6 gives a separated-center sufficiently-small-radius twofold O_eta0(t^(4/3)) bound at fixed L and eta0, using the actual K7/K5 interpolation blocks. It explicitly does not prove a spatial increasing-domain limit or a moving-separation uniform bound. None of those source qualifications is removed here. + +The candidate endpoint convention remains + +```text +M=x-r u/2, S=x+r u/2, grad F(M)=grad F(S)=0, +ell=F(M)-F(S)=k r^3>0, r=|S-M|, +H_M<0, det H_S<0, v=J u, +A_M=F_vv(M). +``` + +r is the endpoint distance. u points from maximum to saddle and ranges once over S^1; reflected v is not another endpoint candidate. The six conditioning rows U_r and physical determinant weights are the finite source's literal ones. This analysis counts ordered endpoint candidates, without an elder/global selector. The short-bar density ell^(-1/3), cumulative t^(2/3) scale and twofold t^(4/3) scale are distinguished throughout. + +## 3. Upper moments do not imply even one-contact rank uniformly + +Write the ideal law with physical frequencies xi_n=(2pi/L)n and normalized symmetric positive weights q_n. Uniform assumption U is + +```text +sum q_n=1, q_n=q_-n>0, L>=Lmin, +sum q_n |xi_n|^12 <= M12 uniformly. (U) +``` + +Take L_N=2pi N, K_N=2N, N>=4, and let the fill have uniform weights on {-2N,...,2N}^2. Put mass 1-eps_N at n=0 and eps_N times this fill, with eps_N->0. Every square weight is positive, total mass is one, and |xi_n|<=2sqrt(2), so M12<=262144. Nevertheless + +```text +Var(F_u)<=eps_N (2sqrt(2))^2=8eps_N +``` + +for every unit direction u. The seven-contact covariance has a diagonal Rayleigh quotient at most 8eps_N. It is positive for each N by the source, but has no common positive floor. Removing the dominant DC mode alone does not fix the issue: dominant mass on the four axial modes (+/-1,0),(0,+/-1) has physical frequency 1/N, so gradient variance is O(N^-2+eps_N) and again collapses. These are lower-variance failures within U, not failures of individual smoothness. + +## 4. One-contact rank can stay uniform while distant contacts cohere + +Fix the base field G on T_(2pi)^2 with **uniform positive weights qbar_m=1/25 on {-2,...,2}^2** and the actual real Gaussian coefficients. Let H_N be the normalized uniform full-square ideal field on T_(2pi N)^2 with K_N=2N, independent of G. Define + +```text +F_N(x)=sqrt(1-eps_N) G(x mod 2pi)+sqrt(eps_N) H_N(x), +q^(N)_n=(1-eps_N) sum_m qbar_m 1{n=N m} + +eps_N/(4N+1)^2. (1) +``` + +For every 00`. The physical frequencies of the repeated G are exactly m, not m/N. Therefore + +```text +Gamma_N(u)=(1-eps_N)Gamma_G(u)+eps_N Gamma_fill,N(u), +lambda_min Gamma_N(u)>=(1-eps_N)m_G. (3) +``` + +For eps_N<=1/2, all single-contact floors are at least m_G/2, uniformly in N and frame. This stronger counterexample does not obtain its failure by losing the one-fold cubic variance. + +Now take a displacement h_N=(2pi j_N,0), with 1<=j_N0` in probability, in the physical norm on the whole growing torus. + +Almost surely G is Morse and has distinct critical values by the fixed-base finite source. It has a finite critical set and a positive minimum absolute height difference Delta(G) between a maximum and a saddle; if one of these types were absent, the no-candidate conclusion would be immediate. There is also a positive C4 stability radius delta(G), **uniform over all repetitions**: choose the finite critical neighborhoods and a gradient floor on one base torus, then repeat them. More explicitly, choose convex lifted balls about each base point p so that `||H_p^(-1)(Hess G(x)-H_p)||<1/4` there. Choose the perturbation C2 radius so that the corresponding bound for `Hess f-Hess G` is below 1/4. Then `H_p^(-1)grad f` differs from the identity linear map by a map of Lipschitz constant below 1/2, and hence is injective on that ball. The implicit branch supplies the root there and its type persists; no additional oscillatory roots can appear in the same ball. A gradient floor on the remaining compact region excludes every other root. The minimum of these finitely many base choices works in every cell for every N, however rapidly the perturbation oscillates. Shrink delta(G) further to make every perturbed critical height differ from its base height by less than Delta(G)/4. + +Consequently, if the global perturbation norm is below delta(G), every positive maximum–saddle candidate gap on T_(2pi N)^2 is at least Delta(G)/2. Distinct copies of one base critical point have the same type; they cannot create a maximum–saddle pair with base gap zero. The repeated field has cross-cell height ties, but these do not invalidate the candidate argument or its Morse stability. We do not call it a distinct-critical-value global bar field. + +For every fixed T>0, (6), delta(G)>0 almost surely and t_N->0 yield + +```text +P(any candidate with 0 0. (7) +``` + +One can justify the random radius explicitly: first choose a deterministic delta for which P(delta(G)<=delta) is small, then use convergence in probability of the perturbation below that delta, and finally use Delta(G)>0. This is not conditioning Gaussian noise at a selected critical point. It is a deterministic stability implication on the same global field. + +For a marked process put birth locations at M/L_N in the unit torus, lifetime at ell/t_N, and retain b,k,u and, if desired, A_M. On every compact bounded-lifetime mark window its mass tends to zero in probability by (7). Its Laplace functional against any nonnegative continuous test supported in such a window tends to one. Hence no nonzero Poisson random measure on that sector can be its limit. This is an actual candidate-process counterexample to a universal **nonzero** conclusion from U plus the one-contact floor. It is compatible with the zero-intensity Poisson process. No statement about actual global H0 bars is inferred. + +### Optional strengthening: nonzero mean with the same empty limit + +This paragraph uses a proved fixed-base contact kernel, not assumed coefficient positivity. On the base torus count candidates with r0 at every such target, pi_0>0, and c_W is finite and strictly positive. Compact domination and zero Gaussian mass on A-window boundaries give this restricted limit by the same source DCT. This is a specialization derived here from its exact proved kernel, not a claim that the source already stated this particular window. The coefficient 6 comes from integrating the literal density coefficient 4 in lifetime; u has unnormalized arclength 2pi and no extra 1/2. + +For t_N=N^-3 and N large, k>=1 makes every radius at most N^-1c_W(1-2^(-2/3))>0, +C_dom,N(t_N)->0 in probability. (9) +``` + +The first equality is N^2 for **short endpoint pairs**. An unrestricted count of all spatial endpoint combinations would have different cross-cell multiplicity and is not substituted here. + +For each fixed N and t_N, the count is continuous under sufficiently small C4 perturbations away from critical/mark-boundary events. Those events have probability zero under G: lifetime/birth boundaries use the source's continuous height densities, k boundaries use their positive r change, and A boundaries on the relevant small-r band use the seven-row covariance and its conditional scalar density. The finite critical-branch description makes this pointwise count continuity precise. Global copied height ties within one type cause no max–saddle boundary at ell=0 because base max/saddle heights are distinct. + +It follows, for fixed N as eps decreases to zero in the coupling (1), that `P(C_eps,N(t_N)!=C_dom,N(t_N))->0`. Choose a deterministic positive eps_N<=N^-12 so small that this probability is at most N^-10. Such a sequence exists by the just-proved fixed-N continuity; it may, for example, be chosen from dyadic eps values after their eventual probability bound. No explicit numerical eps values or rate of this continuity are claimed. + +The accepted a.s. count bound for K=2N gives at most `(16(2N)^2)^2=4096N^4` candidates. The dominant short count is at most 4096N^2. Therefore + +```text +E|C_eps_N,N-C_dom,N| + <=(4096N^4+4096N^2)N^-10 <=8192N^-6 ->0. (10) +``` + +This positive full-square counterfamily satisfies U and (3), has an empty-in-probability compact candidate process, and retains the finite nonzero mean in (9). It exhibits failure of uniform integrability, so the first moment does not identify a nonzero Poisson law. This strengthened version is a deterministic existence construction of a coefficient sequence, not the unproved assertion that the explicit eps_N=N^-12 already transfers the mean. Equations (6)–(7) hold for that explicit sequence regardless. + +## 6. A useful quantitative Gaussian factorization lemma + +Here is a replacement estimate that does follow from stated covariance information. Suppose the two seven-contact vectors have covariance + +```text +Sigma=[[Gamma_1,E],[E^T,Gamma_2]], +m0 I<=Gamma_i<=M0 I, +rho_cross=||Gamma_1^(-1/2) E Gamma_2^(-1/2)||op<=1/2. (11) +``` + +Let D=diag(Gamma_1,Gamma_2), B=Sigma-D, and Sigma_s=D+sB. Relative quadratic-form comparison gives `(1-rho_cross)D<=Sigma_s<=(1+rho_cross)D`. The centered Gaussian density satisfies + +```text +d/ds p_(Sigma_s)(y) + = (1/2)p_(Sigma_s)(y) + [y^T Sigma_s^(-1) B Sigma_s^(-1)y + -tr(Sigma_s^(-1)B)]. +``` + +All dimensions here are fixed at fourteen. Since ||B||<=M0 rho_cross, integrating this identity gives constants depending only on m0,M0 such that + +```text +|p_Sigma(y_1,y_2)-p_Gamma1(y_1)p_Gamma2(y_2)| + <= C rho_cross (1+|y|^2) exp(-c|y|^2). (12) +``` + +This is a proved density interpolation estimate, not a claim that small scalar covariance alone controls an arbitrary nonlinear/global event. + +To apply it, let U be the six source contact rows and A=F_vv, with J=(U,A). At a compact target v_0=(b,0,0,12k,0,0), the normalized contact determinant weight is `V0=36k^2 A^2 1{A<0}`. Let w_i be bounded local marks (for example a Lipschitz function of b,k,u,A, supported in fixed compact b,k and negative-A windows). Define the one-fold contact amplitude + +```text +K_i^0=12*36 k_i^2 integral_(A_i<0) + A_i^2 w_i p_(J_i)(v_i,A_i) dA_i. +``` + +The corresponding twofold amplitude is + +```text +K_12^0=12^2*36^2 k_1^2 k_2^2 + integral_(A_1,A_2<0) A_1^2 A_2^2 w_1 w_2 + p_(J_1,J_2)(v_1,A_1,v_2,A_2) dA_1 dA_2. +``` + +Integrating (12) against the displayed polynomial weights yields, uniformly on the compact b,k marks, + +```text +|K_12^0-K_1^0 K_2^0|<=C rho_cross. (13) +``` + +The type and determinant factors occur exactly once. This remains true for bounded Borel **local** marks in the contact variables; the finite-r convergence rate below uses Lipschitz marks. A full-field elder selector is not a local jet mark and cannot be factorized by (12). It needs a separate weighted localization/selection estimate. + +For a stationary physical covariance R(h), every entry of E is a coordinate derivative of R of total order at most six, expanded in the two frames. If + +```text +d(Rsep)=sup_(dist(h,0)>=Rsep, |alpha|<=6) + |partial^alpha R(h)|, +``` + +the coefficient l1 bounds in two dimensions give `||E||op<=56 d(Rsep)`: each entry has at most a factor `(sqrt(2))^6=8`, and the 7-by-7 operator norm is at most seven times its maximum entry. Therefore `rho_cross<=56d(Rsep)/m0`. A quantitative physical covariance derivative bound is a concrete sufficient far input. Counterexample (4) violates even its order-zero requirement. + +## 7. Finite-r kernels: the actual row coupling and a conservative bound + +One must also align the contact amplitude with the finite-radius observable. For two center/frame configurations, write U_r=(U_r1,U_r2), dimension twelve, with the source's literal divided differences and targets. Assume on the middle band that the **joint seven-row contact covariance** has least eigenvalue at least 0=m I/2 if r_1+r_2<=c m. + +The physical normalized endpoint Hessians need more than the seven conditioned rows. Use the following **unconditioned regular row extensions**, with the first pair written from M to S: + +```text +alpha_M#=[F_uu(M)-(F_u(S)-F_u(M))/r]/r, +alpha_S#=[F_uu(S)-(F_u(S)-F_u(M))/r]/r, +beta_M# =[F_uv(M)-(F_v(S)-F_v(M))/r]/r, +beta_S# =[F_uv(S)-(F_v(S)-F_v(M))/r]/r, +A_M=F_vv(M), A_S=F_vv(S). +``` + +For example `alpha_M#=-integral_0^1(1-s)F_uuu(M+rsu)ds`, and alpha_S# uses +s; beta uses F_uuv. These are uniformly bounded L2 rows and have O(r) L2 errors from their contact limits. Under the exact gradient pins they equal the physical alpha=F_uu/r and beta=F_uv/r. This step avoids asserting bounded unconditional covariance of the singular rows F_uu/r or inventing a full augmented-jet rank. + +Canonical Gaussian conditioning at each r can be coupled using the same independent standard coefficient vector Z': + +```text +c_r=L_r^T (L_rL_r^T)^(-1) v_r, +P_r=I-L_r^T(L_rL_r^T)^(-1)L_r, +conditional row vector = L_E,r c_r + L_E,r P_r Z'. +``` + +The compact target has v_r-v_0=O(r), where r=max(r_i). Inverse perturbation and (14) give `||c_r-c_0||<=C r m^-2`, `||P_r-P_0||op<=C r m^-2`. Because the E row dimension is fixed and its Hilbert–Schmidt norm is bounded, the conditional E-row difference has every fixed Lp norm bounded by C_p r m^-2. Its moments are bounded by C_p m^-1. These statements use finite Gaussian vectors only and do not require the conditional E covariance to be invertible. + +The congruence-scaled physical Hessian is + +```text +Htilde_i=[[alpha_i,sqrt(r) beta_i], + [sqrt(r) beta_i,A_i]], +det Htilde_i=det H_i/r. +``` + +Under the coupling it differs in Lp from diag(-6k,A_0) at M and diag(6k,A_0) at S by at most C_p sqrt(r)m^-2. The maximum weight is the product of negative eigenvalue parts, and the saddle weight is (-det)_+. These are continuous, locally Lipschitz functions with quadratic growth. Their product at two endpoints has degree four; the product for two folds has degree eight and local Lipschitz growth degree seven. For bounded Lipschitz marks in A, the mark difference separately contributes O(r m^-2) times a degree-eight moment. Consequently, on the specified compact targets, + +```text +E|V_r1 V_r2 w_r1 w_r2 - V_01 V_02 w_01 w_02| + <=C[sqrt(r)m^-9+r m^-10] <=C sqrt(r)m^-10. +``` + +The twelve-dimensional pin density is at most C m^-6. Its perturbation at compact targets is at most C r m^-8 by the same differentiated-density identity as Section 6. Combining these bounds gives conservative, unoptimized estimates + +```text +|K_12^r|<=C m^-14, +|K_12^r-K_12^0|<=C sqrt(r)m^-16, (15) +K_12^r=12^2 p_(U_r1,U_r2)(v_r1,v_r2) + E[V_r1 V_r2 w_r1 w_r2|U_r=v_r]. +``` + +For (15), moments of the degree-eight weight are bounded by C m^-8. The density error contributes r m^-16; the unmarked weight error contributes sqrt(r)m^-15; the additional Lipschitz mark error contributes r m^-16. Each is bounded by C sqrt(r)m^-16 when r,m<=1. No inverse physical Hessian, inverse endpoint normalizer or covariance of every Hessian entry was used. Hard additional A-window indicators require a boundary anti-concentration argument instead of a Lipschitz rate; continuous compact test marks avoid that extra step. Their upper bound still follows by dropping the mark. + +## 8. What quantitative middle/far hypotheses actually suffice + +For a varying source law let V=L^2, r_t=(Tt/k_-)^(1/3), and choose cutoffs + +```text +r_t << eta_t << R_t << L, +V t^(2/3)=O(1). +``` + +The last condition is a stipulated total rare-intensity scale, not a one-point theorem. In addition to U and fixed positive a,k_-, the following are explicit covariance estimates to prove: + +* A uniform one-contact bound m0 I<=Gamma_0(u)<=M0 I, and convergence of these seven-row covariances to a specified limiting covariance if an intensity is to be identified. +* On eta_t<=dist(x_1,x_2)<=R_t, a joint contact floor `m_t>0`, with `r_t<=c m_t` and `V R_t^2 t^(4/3) m_t^-14->0`. +* On dist(x_1,x_2)>=R_t, cross derivative decay d_t(R_t)->0 through order six, small enough eventually that 56d_t/m0<=1/2. + +For the middle domain, finite-r rank in (14) implies rank of the original twelve value/gradient observations by the actual invertible per-pair transforms. The source's coefficient-space area formula extends in the same finite dimension to four distinct critical endpoints: choose an invertible eight-gradient minor, differentiate to get the four physical Hessian blocks, and partition the position domain. This gives the genuine four-point Borel identity with the determinant product once. Endpoints cannot overlap on the middle/far domains when r_t<0. The near diagonal can be handled by the separate local-norm/count proposal if its hypotheses and review permit reuse; it is not supplied by (16)–(17). + +For illustration, if `m_t>=c eta_t^nu` for some fixed finite nu, `eta_t` is of logarithmic order, `R_t^2=O(log(1/t))`, and V t^(2/3)=O(1), then (16) is O(t^(2/3) times a fixed power of log(1/t)) and vanishes. Also r_t/m_t->0. A finite exponent is enough; one need not assume a separation-independent floor at eta_t->0. + +These are sufficient twofold **candidate** estimates on continuous compact local marks. They do not prove convergence of all factorial moments or a Laplace functional. In particular the iid-source weighted-cluster theorem is not a spatial mixing theorem. Actual global H0 selectors need uniform weighted local-selection/remote-exclusion error estimates before the local kernels can replace them; finite-dimensional jet decorrelation cannot factorize a global selector by itself. + +## 9. A concrete source-specific Gaussian route, with residual work visible + +The iid source's actual covariance is the infinite periodization of exp(-|h|^2/2), with physical Fourier weights proportional to exp(-|xi_n|^2/2). A natural finite family that respects this identity is the exact normalized truncation + +```text +q_(K,L),n = exp(-|(2pi/L)n|^2/2) + /sum_(|m|infinity<=K) exp(-|(2pi/L)m|^2/2), +|n|infinity<=K, +Lambda=(2pi/L)K. +``` + +This is a specific candidate source family, not the arbitrary positive spectra of U. For L large and Lambda>=4, elementary Gaussian grid-sum comparison gives a uniform physical moment bound and a derivative tail through order six, + +```text +delta_6(K,L) + =sum_(|n|infinity>K) q_infinite,L,n(1+|xi_n|^6) + <=C exp(-Lambda^2/64). (18) +``` + +A route to (18) is explicit: the infinite normalizer is bounded below by c omega^-2 using frequency cells in a fixed physical square; compare each tail grid point to its omega-cell, omega<=1, use `|xi_n|^2>=|x|^2/2-1`, and integrate outside radius Lambda-1. A polynomial Gaussian tail is bounded by C exp(-Lambda^2/64). Normalizing the truncation changes every covariance derivative of order <=6 by at most C delta_6, once the omitted mass is <=1/2. The same comparison at moment twelve supplies U. Constants in this paragraph are existence bounds from elementary Gaussian integrals, not evaluated numerical enclosures. + +For the infinite periodization, differentiating the Gaussian images gives + +```text +sup_(|alpha|<=6) |partial^alpha R_infinite,L(h)| + <=C exp[-dist_T(h,0)^2/8], +``` + +uniformly for L bounded below away from zero (and in particular for L large). Absorb the finite Hermite polynomials into exp(-|x|^2/4), factor out exp(-dist_T^2/8), and bound the remaining separated lattice Gaussian sum. Thus the truncated far input has the concrete candidate bound + +```text +d_t(R)<=C[exp(-R^2/8)+delta_6(K,L)]. (19) +``` + +For a middle lower bound, the Euclidean Gaussian covariance has strictly positive two-contact covariance at distinct centers and every frame. Its positive spectral density makes a vanishing variance imply that an exponential-polynomial multiplier vanishes identically. Its inverse Fourier distribution is a sum of finite jets supported at the two different centers; their disjoint supports force each local jet functional to vanish, and the seven contact rows at one point are independent. This is a finite named-jet argument, not an arbitrary rank claim for a finite support. + +There is a possible qualitative-to-power improvement for this one fixed analytic kernel. Parameterize the separation as s e, s real near zero, and the frames by angle charts. The determinant of its fourteen-row covariance is real analytic. It is positive for s!=0 and zero at s=0, because fourteen third-order rows at one point lie in a ten-dimensional full third jet. A compact real-analytic distance inequality then gives `det>=c s^nu` for some finite nu, uniformly in e and frames on a fixed small s band. Dividing by a uniform bound on the other thirteen eigenvalues gives a least-eigenvalue lower bound of the same power. Compactness handles the remaining bounded separations; Gaussian derivative decay handles arbitrarily large separations. + +The named classical import for that last implication is the compact real-analytic distance inequality, inspected as Theorem 3.1 in Teissier's [*Some resonances of Lojasiewicz inequalities*](https://webusers.imj-prg.fr/~bernard.teissier/documents/Lojasiewicz_2012.pdf), which explicitly attributes it to Lojasiewicz. The consumed read is the theorem statement, not the complete paper or a fresh read/proof of the cited original 1959 work. A saved copy and exact identity are in [jet-factorization-primary-sources/SOURCE_IDENTITY.json](jet-factorization-primary-sources/SOURCE_IDENTITY.json). On the compact angular charts the analytic zero set is s=0, so its distance is comparable to |s|. The import supplies some finite exponent, not an evaluated exponent or effective constant. + +To transfer that floor to a finite truncation at eta_t, one needs + +```text +delta_6(K,L)+C exp(-c L^2)=o(eta_t^nu). (20) +``` + +On a fixed bounded separation band the periodization differs from the Euclidean covariance by the displayed image error; on the remaining large-separation band (19) gives a constant floor after choosing a fixed sufficiently large distance. Weyl's inequality then transfers the middle lower bound. Choosing Lambda^2 comparable to log(1/t), eta_t of logarithmic order, and L of order t^(-1/3) makes (18), (20), the r_t/m_t requirement and (16)–(17) compatible. K is then of order L sqrt(log(1/t)), so the logarithmic near cutoff also remains compatible with r_t<=4 exact-contact upper bound (20). Two nonblocking wording corrections are +listed below. Acceptance here does not extend the estimates off the collinear +stratum, make constants effective, identify a global selected-bar measure, or +establish a point-process limit. None of those conclusions follows from this +review. + +## Exact inputs and source exposure + +| Input | SHA-256 | +| --- | --- | +| `COALESCING_CONTACT_CALCULATION.md` | `39a63e1b00393b49e7b6d049c0c9ea7be82436352eed64014155a63aa8c7fb36` | +| `coalescing_diagnostics.py` | `6fe14ca7f518d3c42236421ddd5b8f7470accfec855f8889d17e8aaa62b89e2b` | +| `COALESCING_DIAGNOSTICS.json` | `946a962bf47589a76affae03977f3d03376a463e26b7943ea675cceb4ad2799a` | + +I independently compared all four report-named local source copies with their +Git objects at Main `85b928b326ff0586715a56ae91f51d3fbd89557a`. Their bytes, Git +blob identities, sizes and SHA-256 values match the report's table. The substantive +source inputs consumed were the literal finite Fourier covariance multiplier, +six-row pin/target and physical Hessian weight in +`finite_gaussian_contact/PROOF.md` §§2–5, and the actual infinite-law/separated +center boundary in `iid_short_bar_process/PROOF.md` §6. I did not rerun the +underlying global H0 selection/counting proofs or import their conclusions into +this local estimate. The candidate/H0 source copies were checked for source +identity and their distinction is preserved. + +The current [owner mandate](../gaussian-jet-coefficient-20261016-root01a11cbf/checkout/governance/OWNER_MANDATE_20261015.md) +and [standing delegation](../gaussian-jet-coefficient-20261016-root01a11cbf/checkout/governance/OP-AUTONOMY-20260923-v2.1.md) +were read. They authorize this work and do not supply mathematical acceptance. +For the elementary interpolation remainder I also checked the primary +[NIST DLMF interpolation reference, equation 3.3.5](https://dlmf.nist.gov/3.3.E5). +The proof below is reconstructed directly from Rolle/Taylor arguments rather +than treating that reference as an acceptance authority. + +## Reconstruction of the eight-gradient density and determinant cancellation + +For a fixed positive full-square Gaussian spectrum, write the field as a linear +map of a fixed finite standard Gaussian coefficient vector. A vanishing variance +of a linear combination of the eight limiting gradient jets forces its complex +frequency polynomial to vanish at every point of the full square. Its degree +in each original frequency coordinate is at most four. A grid with at least +five points in each coordinate therefore forces polynomial identity. After the +invertible orthogonal change to frequency coordinates `(a,c)`, the monomials +`a,a²,a³,a⁴,c,ac,a²c,a³c` are distinct, so every coefficient vanishes. The factors +of `i`, omega and factorials are nonzero row scalars and do not alter this +argument. It works for arbitrary positive symmetric weights; isotropy or D4 +symmetry is unnecessary. + +For each fixed law the limiting covariance depends continuously on the frame. +Compactness of O(2) gives a positive minimum eigenvalue. Finite Fourier rows have +uniform derivative bounds, and the divided-difference integral representation +gives convergence in mean square with an error controlled by the segment +diameter, independently of unequal node spacings. This last qualification is +essential: a Taylor expansion with denominator-dependent errors would not +suffice. The proposed integral argument supplies the needed uniformity. + +The Newton map for four values has determinant `1/V`. Applying it separately to +the two gradient components gives determinant `V^-2`, hence +`det Cov(G)=V^4 det Cov(U)` and `p_G(0)=Theta(V^-2)`. The density normalization +`(2pi)^-4` is correct for eight real observations. Stationarity handles the +basepoint, and all constants remain fixed-law constants. + +For a C4 scalar function vanishing at all four nodes, subtracting its value at +a fifth point times the monic nodal polynomial and applying Rolle four times +gives `|a(s)| <= ||a^(4)|| infinity product_j |s-x_j| / 24`. Dividing by +`|s-x_i|` and taking the endpoint limit proves +`|a'(x_i)| <= D_i ||a^(4)|| infinity / 24`. The argument holds also for the +two interior nodes; no common intermediate Rolle point for several functions +is asserted or needed. + +With `a=F_u` and then `a=F_v`, the required derivatives are exactly `F_(u^5)` +and `F_(u^4 v)`. Bounding `F_vv` by a fixed coordinate C5 norm gives +`|det H_i| <= C M² D_i` after shrinking the physical segment so `D_i<=1`. +The squared mixed-Hessian term is included. Since `product_i D_i=V²`, the +product of all four physical determinant weights is at most `C M^8 V²`. + +Conditioning on the zero gradient target has canonical coefficient law `PZ`, +where P is an orthogonal projection. It has zero mean even near covariance +degeneracy. Coupling it to the same standard Z gives `||PZ||<=||Z||`, so all +needed conditional norm moments are bounded independently of the node ratios. +This proves the bounded Rice integrand (8). It is not a four-dimensional +position integration or a new Kac-Rice counting identity. + +## Positive cubic marks and the action identity + +Under `r_i<=rho eta`, `0=4 bound + +On a coordinate line the longitudinal restriction has dimension `2K+1`; its +value/first-derivative observations provide eight Hermite rows. The transverse +restriction has four rows and can be prescribed independently by using a sine +factor in the other coordinate. The exact ranks are therefore 9,11,12 for +K=2,3,>=4 on sufficiently short distinct configurations. Injectivity when the +number of constraints exceeds dimension follows by trigonometric zero counting; +surjectivity of the smaller Hermite lists follows equivalently by the ordinary +polynomial Hermite matrix after multiplying the Laurent polynomial by its +lowest-frequency monomial. For the two exact contact centers the multiplicity +lists are four/four longitudinal and two/two transverse, giving the same ranks. + +At K2, `F_x` on that circle is a zero-mean degree-two trigonometric polynomial. +Two double zeros with nonzero second derivative exhaust its possible four zeros +counting multiplicity and prevent a sign change. Zero mean then forces the +polynomial to vanish identically, contradicting a nonzero contact cubic target. +The separate four-pin statement uses the local nonsingularity of four value +evaluations on the four-dimensional nonzero-frequency derivative space. Hence +the longitudinal derivative vanishes on the whole coordinate line, and +`F_xx=0` at the pins. A nondegenerate maximum is impossible there, although +saddles and nonsingular eight-gradient observation covariance remain possible. +This proves the claimed typed-kernel zero and avoids conflating observation +rank with maximum admissibility. + +These obstructions are coordinate-stratum statements. My exact rank checks give +12 for the twelve coincident normalized jet polynomials at the rational frames +`u=(3/5,4/5)` and `u=(5/13,12/13)` even at K2. They do not supply a uniform +floor as those frames approach a singular orientation. This supports the +report's call for geometry stratification rather than a blanket K2 rank claim. + +For K>=4 the twelve limiting monomials have degree at most seven in each +original coordinate and are independent on the nine-point square. Hermite +normalization is uniform also for arbitrary small node ratios. Four sites +with multiplicity two in the longitudinal channel give determinant `V^4`, +and four transverse values give `V`, proving the combined density factor +`V^-5`. At two contact centers the respective determinant powers are eta^16 +and eta^4. With literal, unfactored derivative rows, the longitudinal constant +is 144; squaring the combined determinant gives `Theta(eta^40)` for the +covariance determinant. This confirms (18)-(19), including the one-scale +epsilon^-30 and the two-pair `r1^-5 r2^-5 eta^-20` prefactors. + +The normalized inverse covariance is bounded; transformation entries are +`O(eta^-7)`. Consequently the conditional finite coefficient mean and each +transverse curvature mean are `O(eta^-7)` on compact contact targets. The +centered covariance stays bounded by identity, giving the stated crude +`E[A1² A2² | U=v] <= C eta^-28`. Multiplying by eta^-20 and the contact +action bound gives (20). The single-contact physical weight has exactly the +source factor `36 k_i² A_i² 1{A_i<0}`. No extra Hessian determinant or independent +contact assumption was inserted. Uniform interchange of finite-radius and +coalescing-center limits is deliberately not supplied. + +## Replay and independent finite algebra + +`NONAUTHOR_COALESCENCE_CHECKS.py` and its JSON retain my local checks. The author +program replay produced JSON equal to the supplied JSON, including all number +strings, using bundled Python 3.12.14. Independent Fraction checks verify the +ordinary and multiplicity-two Hermite determinants, the two-center constants +144 and 1, 39 polynomial instances of L1, and nine rational-frame rank cases. +Those finite cases support, rather than replace, the all-frame argument above. + +Independent exact covariance algebra at the coincident normalized jets gives +these limit constants for the diagnostic law and target: + +| Normalized diagnostic | Exact limit | +| --- | --- | +| `det Cov(G_K2)/V^4` | `129024/78125 = 1.6515072` | +| `det Cov(O_K4)/V^10` | `1550348800000/10460353203` | +| `det Cov(U_K4)/eta^40` | `396889292800000/129140163` | +| `eta^4 * contact action` | `28782/25 = 1151.28` | + +For the last identity, the degree-five Hermite interpolant is +`F(s)=2s³-3eta^-1 s^4+(6/5)eta^-2 s^5`. In the repeated-node Newton coordinates, +eta² times the target tends to `(6/5)e5`; evaluating its quadratic form against +the exact limiting covariance gives the displayed rational action. The +reported decimals approach all four constants consistently. + +Repeating the decimal computations at 130 rather than 95 significant digits +changes the supplied normalized values by relative differences between roughly +10^-90 and 10^-67. This is a useful precision crosscheck, not certified interval +arithmetic. The original Taylor sin/cos routine remains uncertified. No seed, +field sample, numerical H0 test, asymptotic window or error constant was created. + +## Nonblocking wording corrections and remaining boundary + +1. In §3 explicitly rebase `z=z1, x1=0`, or put the limiting jets at `z+x1 u`. + Small diameter alone does not force the four numerical x_i to approach zero. + Stationarity and uniformity already make this harmless to every bound. +2. State the divided-difference simplex convention once: either integrate the + derivative against simplex volume of total mass `1/j!`, or average under a + probability measure and multiply by `1/j!`. Do not apply both normalizations. + The displayed limiting jets already have the correct factorials. + +The assumptions essential to this review are a fixed finite planar Gaussian +law with strictly positive full-square weights, an embedded short torus chart, +four distinct ordered sites, the declared disjoint-pair ratio when eta is used, +canonical zero-target regression, and fixed positive lower cubic marks for +exponential suppression. Constants are not uniform over vanishing spectral +weights, increasing K, varying L or singular orientation families for the +twelve-row K2/K3 route. H0 selectors occur solely as Borel marks bounded by one. +The transverse collision tube, the small-k boundary, endpoint incidence +partition, increasing-domain estimates, Lean interfaces and global selected-bar +or process laws remain work to perform. A collinear set has zero full position +measure, so these accepted local estimates alone cannot resolve any of them. diff --git a/experiments/universality/fold_isolation_diagonal/NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW.md b/experiments/universality/fold_isolation_diagonal/NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW.md new file mode 100644 index 000000000..1b1313813 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW.md @@ -0,0 +1,130 @@ +# Nonauthor review of the fixed finite Fourier count and diagonal route + +**Verdict: ACCEPT_SCOPED.** I found no blocking mathematical error in the frozen proposal's almost-sure bound `Ncrit<=16K^2`, fresh outer-cutoff isolation lemma, integrated ordered factorial-count bound, or logarithmic-cutoff normalization. No amendment is required for equations (1)–(21) with the stated fixed-law, endpoint-identity, type, mark, cutoff and embedding hypotheses. This is a local conventional mathematical review, not source adoption or scientific promotion. + +Actual reviewer: OpenAI/Codex `/root/c261a_local_runtime_20261016`, 16 October 2026 UTC. Root and `/root/core_theorem_audit_20261015` are the authors; I did not author their sketch, the frozen extension or its controls. I read all three new files in full, checked the cited primary imports, and consumed the exact previously accepted isolation result with its hypotheses unchanged. I did not rerun the old isolation controls or claim another first review of the unchanged chart. Source exposure is explicit; **organizational-independence credit is 0**. An attempted separate subagent dispatch returned `agent thread limit reached`; no additional agent participated in this finite-count review. + +## 1. Frozen inputs, primary imports and actual checks + +| New frozen file | Bytes | SHA-256 | +|---|---:|---| +| `FINITE_FOURIER_CRITICAL_COUNT_DIAGONAL_ROUTE.md` | 25,699 | `621cee96854aa440498cecf5452a8d1e23f90a4458779c3286c18eba106fe138` | +| `finite_fourier_count_controls.py` | 5,153 | `fd6f29739280a55017945526cb75fb46ebc80d86c012ff4a1ff65baa404c31de` | +| `FINITE_FOURIER_COUNT_CONTROLS.json` | 849 | `bbe147fdf409a3e37b450d4cbcd5d2a927140f95bd3ce0f36cafe16861123bfd` | + +The consumed isolation proposal/control/output retain SHA-256 `827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30`, `13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5`, and `ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf`. My earlier `NONAUTHOR_FOLD_ISOLATION_REVIEW.md` remains 18,409 bytes, SHA-256 `f06c61e3f3e60c795a1a14e824640f1689d4143dba4f6698cada7a41929eec90`. + +The pinned finite source remains Main `85b928b326ff0586715a56ae91f51d3fbd89557a`, `experiments/universality/finite_gaussian_contact/PROOF.md`, 25,367 bytes, SHA-256 `7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474`, Git blob `5f345201309cfa1781e92d6292f542de4bcb2999`. Its actual real finite coefficient law, rather than a complex-generic unrelated law, is used below. + +I verified the saved primary `SOURCE_IDENTITIES.json` at 796 bytes / SHA-256 `2c42d92a4732f09a58aebd9b9ef8672b51b672ab99f3bbdad61681a0dcf4f504` and all three listed source byte identities. I directly read the complete six-page Guth lecture and the full two named Stacks lemmas/proofs. Guth's Theorem 0.1 is the coprime affine-polynomial inequality over a field, sufficient here over C; it does not require irreducibility. [Guth, MIT lecture 13](https://math.mit.edu/~lguth/PolyMethod/lect13.pdf) + +The projective maps below use that a projective source is proper, and a map from a proper source to a separated target has closed image. The cited tags supply precisely these implications. [Stacks 01WC](https://stacks.math.columbia.edu/tag/01WC), [Stacks 01W6](https://stacks.math.columbia.edu/tag/01W6) + +I also directly read the MIT text's Section 7.8.1 and following reducible-form note/proof on PDF pages 178–179, and Larson–Neeman–Shelah Section 1 through the analytic-set universal-measurability conclusion on printed pages 174–175. I do not claim to have read the entire 216-page text or the entire measure-theory paper, or to have saved byte identities for those two live-browsed PDFs. [MIT algebraic geometry text](https://math.mit.edu/classes/18.721/ag-may31-2021.pdf), [Larson–Neeman–Shelah](https://shelah.logic.at/files/94751/947.pdf) + +At 2026-10-16T01:59:41Z I ran the new author controls and my separate `nonauthor-finite-count-checks/review_checks.py`, using `/Library/Developer/CommandLineTools/usr/bin/python3 -B` (Python 3.9.6). Both exited 0 with empty stderr. The author stdout matches the frozen 849-byte output exactly. My 226 exact-rational finite cases passed. Full commands, timestamps, wall times, stdout/stderr and digests are in `nonauthor-finite-count-checks/CHECK_RUN_SUMMARY.json` and its two process-record sets. Twelve consumed input identities were verified before and after, unchanged. The exact finite tests do not mechanically prove genericity, Bézout, all-K coprimeness or the analytic cutoff lemma; the following arguments are my conventional review. + +## 2. The full-degree witness belongs to the actual real family + +The passage to Laurent variables is correctly normalized. With `z=e^(i omega x)` and `w=e^(i omega y)`, the two real gradient coordinates vanish exactly when the two Laurent sums with multipliers m,n vanish: their omitted factor `i omega` is nonzero. Clearing by `z^K w^K` changes no zero on `(C*)^2`. Each resulting polynomial has exponents between 0 and 2K in each variable, so total degree is at most 4K. The torus map to the unit-circle product is injective modulo the original period L; there is no K-fold covering factor in the critical-point count. + +The witness consists of the axial K-modes and diagonal `(K,K)` mode, together with their negative-frequency conjugates, all within the full square. For any fixed strictly positive q, the real sine/cosine Gaussian coordinates map invertibly onto all real trigonometric coefficients. Set the three relevant sine coordinates to zero and choose the cosine amplitudes to give `1,1,c`; the other coordinates can be zero. This gives a permitted real parameter point with `a_-n=conjugate(a_n)`. It need not have positive probability as one exact draw: its role is to certify a nonzero polynomial on that same real parameter space. + +Computing the actual Laurent derivatives gives the displayed normalized pair + +`P_*=W0(Z0^2-1)+c(Z0^2W0^2-1)`, + +`Q_*=Z0(W0^2-1)+c(Z0^2W0^2-1)`, + +and `P_*-Q_*=(Z0-W0)(Z0W0+1)`. The normalization is `2/K`, so both top terms really have coefficient c and total degree 4K when `c!=0`. + +Checking coprimeness only in `C[Z0,W0]` would not be the stated proof after substitution. The proposal instead checks the actual factors in `C[z,w]`, which is adequate. Every common irreducible divisor divides the difference and hence is one of `z-zeta w` with `zeta^K=1`, or `zw-rho0` with `rho0^K=-1`, up to a unit. The latter are irreducible because their quotient identifies z with `rho0/w` and is a Laurent polynomial domain. There is no hidden coordinate-axis factor. + +On every line factor the restriction is the same nonzero polynomial `W0^3-W0+c(W0^4-1)`, whose top coefficient c is nonzero. On every hyperbola factor the restriction is `W0^(-1)-W0`, nonzero in its Laurent quotient. Thus no member of either family divides `P_*`, for every K and every nonzero real c. This is a valid all-K factor argument, not an inference from the finite K tests. + +Both affine forms have exact degree D=4K. Their degree-D homogenizations therefore have no factor T0. Any other common homogeneous irreducible would dehomogenize to a nonconstant common affine factor; a homogeneous polynomial whose dehomogenization is a nonzero constant can only be a power of T0. The witness homogenizations are consequently coprime. Their equal restrictions at infinity give intersection points there, not a shared curve or line at infinity. + +The failed `c=0` witness is retained correctly. It has affine coprime derivatives but degree 3K, so homogenizing both to the fixed degree 4K inserts a common `T0^K`. It cannot certify properness of the selected fixed-degree bad cone. The diagonal mode supplies exactly the missing degree. No change to this part is required. + +## 3. Closedness, real restriction and the null exceptional set are justified + +For each `1<=d<=D`, the map from `P(V_d) x P(V_(D-d) direct_sum V_(D-d))` sends `[H],[A,B]` to `[HA,HB]`. Joint projectivization of A,B is the right parameter space: it includes pairs with exactly one zero form and allows their relative scale. The coordinates transform by the same product of the two source scalars. They cannot all vanish, since H is nonzero, at least one of A,B is nonzero, and the polynomial ring has no zero divisors. The map is everywhere a morphism defined by sections of `O(1,1)`; there is no unresolved rational-map base locus. + +Its source is a product of projective spaces, hence projective/proper over C. Its target is projective, hence separated. The named closed-image import therefore applies to each actual map. The finite union over d captures exactly the nonzero homogeneous pairs having a positive-degree common factor: the common factor can be chosen homogeneous and the quotients have degree D-d. The case of one zero form is included, for example by choosing the other whole form when d=D. The affine cone, with the zero pair included, is an algebraic closed set. + +The witness shows this cone is proper **on the image of the actual real parameter family**, not merely proper in the ambient complex form space. In particular some defining complex polynomial is nonzero on the witness image. Pulling it back through the real-linear homogenization map gives a complex-coefficient polynomial in the real Gaussian coordinates. At least one real/imaginary part is a nonzero real polynomial, since its value at the real witness is nonzero. The exceptional preimage is contained in that real polynomial zero set. + +The induction/Fubini proof that a nonzero real polynomial has a Lebesgue-null zero set is sound. The fixed positive weights make the real coefficient distribution nondegenerate; multiplying independent standard Gaussian coordinates by fixed nonzero scales preserves a full real density. The constant coefficient lies in the gradient map's kernel, but does not obstruct the argument: the pullback remains a nonzero polynomial in the full parameter space, possibly independent of that coordinate. Hence the exceptional set has probability zero under the actual Gaussian law. + +It is harmless that the chosen exceptional set also removes some lower-degree affine-coprime pairs. Only a sufficient null exceptional set is needed. No unsupported conclusion about the smallest discriminant or global Morse regularity is being used. + +## 4. Bézout counts distinct real torus critical points without a Morse premise + +Off that one null set, the affine P,Q are coprime over C. The affine Bézout inequality bounds all common complex affine zeros by `(deg P)(deg Q)<=16K^2`. Every real torus critical point maps to a distinct such zero with z,w nonzero. Extra complex zeros and coordinate-axis zeros only enlarge this upper bound. Degenerate isolated roots are still distinct zeros and each consumes at least one intersection in the inequality; transversality and a Morse assumption are unnecessary. + +This establishes the almost-sure deterministic bound on all distinct critical points for the fixed K law. It is not claimed for every coefficient vector. A one-variable field may have critical circles and lie in the exceptional family. Setting the later count to zero on that null exceptional set is an allowed measurable modification under the completed Gaussian measure. + +The proof's full-real-density step is unconditional. This almost-sure argument cannot automatically be transported to an arbitrary zero-target conditioned coefficient subspace, which might lie in the algebraic exceptional set. The previous isolation review's centered conditional norm tail alone does not supply a conditional version of this global critical-point theorem. No such conditioned count claim occurs here, so this is a retained boundary rather than a blocker. + +## 5. Candidate identity, finite count and measurability are adequate + +Each candidate is an ordered pair of actual endpoints `(M,S)` with negative-definite Hessian at M and negative determinant at S. Its displacement is the unique short geodesic displacement; r,u, midpoint, gap, k and the fixed `v=Ju` follow from the endpoints and field. A reflected v is not a new candidate. There is no continuously free frame label being counted. Reverse endpoint order fails the maximum/saddle roles and positive gap. This preserves the endpoint-pair identity required for `B_K^4`. + +On the full-probability finite set, the number of endpoint candidates is no more than `Ncrit^2`, and the number of ordered pairs of different candidates is no more than `Ncrit^4`. The sharper bounds `Ncrit(Ncrit-1)` and `C(C-1)` are unnecessary. Counting once by endpoints, rather than by frame presentations, makes the displayed coarse bound valid. An unordered convention would divide by two; no such division is hidden in D. + +The identical candidate `p=q` is explicitly excluded. Distinct candidates can share a maximum or saddle. Cross-role equality is impossible because the physical Hessian at the same point cannot be negative definite and have negative determinant. This is a pointwise type fact in the same field, not an independence statement. + +The count is measurable on the completed Gaussian space. For each integer m, `D>=m` can be expressed by m distinct ordered endpoint-pair configurations satisfying Borel equalities/inequalities in the finite coefficient vector and position variables. Short displacement/midpoint conventions are Borel on their specified domain. Projection gives an analytic event, which is measurable for the completed probability law. The exceptional set is Borel and null, so modifying there is legitimate. A nonnegative Borel weighted count with weights at most one is pointwise dominated by D. Its measurability can also be obtained by the same finite-configuration projection argument, using existence of finitely many distinct configurations whose weights sum above a rational threshold. No joint Rice density is needed for either step. + +## 6. The fresh outer-cutoff lemma closes the actual-separation gap + +The proposal correctly identifies that `r_i<=rho eta` is weaker than `r_i<=rho s` when the realized center distance s can be far smaller than eta. Its explicit old-ratio counterexample is valid. The new Section 6 is a new argument; it does not claim that the old hypothesis follows from short lifetimes. + +The common-lift geometry is sufficient. With `eta<=eta0<=R0/4` and `R0<=L/16`, choose the shortest center displacement and the two short endpoint displacements. The endpoints lie within distance at most `(1+rho/2)eta` of the first center, and their displacements from the first maximum are bounded by `(1+rho)eta`. These scales are far below torus injectivity. The rotated chart itself embeds since its square lies in a ball of radius `sqrt(2)R0=kappa/W^2>kappa eta/c_*>=(1+rho)eta`. + +If the maxima differ, the triangle inequality places the second maximum inside that chart, contradicting unique-maximum isolation. This includes the shared-saddle case. If they share their maximum, distinct endpoint identity forces two different saddles, each within `rho eta=c_*/eta` under the new bounding-cutoff hypothesis with the same displayed constants. + +## 7. The expected count bound is a valid unconditioned integrated estimate + +For `eta<=min(eta0,c_*/4)`, any nonzero D implies `||Z||^2>=c_*/(4eta)`. The finite standard Gaussian exponential moment is `2^(N/2)` at parameter 1/4, so the displayed norm-tail estimate follows. This is a global deterministic inclusion for the random set's existence event, not Gaussian conditioning at the field-selected sites. No adaptive-conditioning assumption, covariance inverse, density normalization or Hessian/Rice weight is hidden in it. + +Almost surely, + +`D<=B_K^4 1{D>0}`. + +Taking expectations therefore gives exactly + +`E D(t,eta)<= (16K^2)^4 2^(N/2) exp[-c_*/(16eta)]`. + +This is the new finite combinatorial input absent from an occurrence estimate by itself. It works because K and the carrier are fixed and the endpoint candidate population has the almost-sure uniform finite cap. It bounds an integrated ordered factorial count, not a pointwise four-site Rice kernel. The fixed finite-support height-covariance obstructions remain untouched. Borel selectors in `[0,1]` can reduce this count, but no actual selected-bar identity is supplied by that domination. + +## 8. Lifetime cutoff, explicit threshold and normalization are correct + +For a candidate with `ell<=Tt` and `k>=k_-`, the literal identity `ell=kr^3` gives `r<=A t^(1/3)`, `A=(T/k_-)^(1/3)`. The comparison needed is to the outer cutoff `rho eta(t)`, not to s. For `n=log(1/t)` and `eta(t)=c0/n`, that comparison is `A n exp(-n/3)<=rho c0`. + +The explicit threshold (21) is sufficient. The elementary bound `n<=6 exp(n/6)` gives `n exp(-n/3)<=6 exp(-n/6)`. If `6A/(rho c0)>1`, the last term of n0 makes its right side times A at most `rho c0`; if that ratio is at most one, the same conclusion follows already from `6A<=rho c0`. The other threshold terms ensure `n>=1`, `eta<=eta0`, and `eta<=c_*/4`. Equality in the non-strict geometric bounds is allowed; the contradiction radius remains strict. Thus for all `00`. More generally `0=tau>0`, `a,tau,R0>0`, and `C>=0`. With `BR<=a/32`, the actual hard-curvature estimate is even stronger than the displayed relaxation: `F_yy<=-a+2BR<=-15a/16<-a/2`. The bound `|F_y(x,0)|<=Bx^2/2` uses both `F_y(0)=0` and `F_xy(0)=0`. It supplies the strict opposing face margins `31aR/64`. Thus for every closed-interval x, including `x=+/-R`, the unique vertical zero lies strictly between the y-faces. It cannot escape through a corner or a horizontal face. + +The implicit-function theorem applies locally at each point of the ridge, including its two endpoints, because the domain is an open neighborhood and `F_yy` is nonzero. The roots agree on overlaps by uniqueness. C3 regularity and the local endpoint continuations are sufficient for all displayed derivatives. Integrating the negative y-derivative gives `|h(x)|<=Bx^2/a<=R/32`, and `h'=-F_xy/F_yy` gives `|h'|<=4BR/a<=1/8` without assuming the ridge remains on `y=0`. + +I independently differentiated the reduced function. Differentiating `f''=F_xx+F_xy h'` first gives + +`f'''=F_xxx+2F_xxy h'+F_xyy(h')^2+F_xy h''`. + +The second differentiated ridge equation is + +`F_yy h''=-(F_xxy+2F_xyy h'+F_yyy(h')^2)`. + +Using `F_xy=-F_yy h'` produces exactly the proposal's coefficients `1,3,3,1`. Thus no unbounded `h''` term was silently discarded. The operator fourth-derivative bound controls the variation of `F_xxx` by `C(|x|+|h|)<=2CR`. For `p<=1/8`, `3p+3p^2+p^3<=4p`; the remaining contribution is at most `16B^2R/a`. Finally `2C+16B^2/a<=16(C+B^2/a)` and the third radius term give the fixed-sign margin `tau/2`. This verifies the chain-rule and radius constants in (3)–(5). + +Put `g=f'`. Strict convexity or concavity gives at most two zeros even on the closed x-interval: an interior third zero between two others contradicts the strict chord inequality. Nondegeneracy of every root does not follow from strict convexity alone. Here the anchor supplies the missing information: `g(0)=0` and `g'(0)<0`. If `g''>0`, a second zero must be to the right; Rolle gives a point with derivative zero strictly between the roots and strict increase makes the second derivative value `g'(x_2)>0`. If `g''<0`, the second zero must be to the left and the same argument with decreasing `g'` again gives `g'(x_2)>0`. This works when the second zero is on an x-face. A double second root cannot coexist with this anchored maximum. + +At every critical point, `det Hess F=F_yy f''`. With `F_yy<0`, the second simple root has negative determinant and is an ambient index-one saddle. The negative-hard Schur sign is correct; there is no sign reversal turning it into a second maximum. The anchor has two negative eigenvalues. The language expressly excludes constrained extrema of the rectangle, so a boundary-constrained maximum is not misclassified as an ambient one. + +My independent boundary controls take `a=tau=B=1`, `C=0`, `R0=1`, giving the proposed `R=1/32`, and `mu=1/64`. The cubic fold has its second critical point exactly at `x=R`, with positive reduced second derivative; changing the cubic sign places it exactly at `x=-R`. Both are saddles. This attacks the closed-square case directly, beyond the author's existing finite checks. + +## 2. The height and mixed-cubic obstructions are preserved correctly + +Integrating `F_yy<=-a/2` twice from the ridge gives the vertical estimate `F(x,y)<=f(x)-a(y-h(x))^2/4`. Since `|h|<=R/32`, the y-face drop is at least `aR^2/8`. It is relative to `f(x)`, not to `F(0)`. No enclosing superlevel barrier or path-trapping conclusion follows. + +The false full-height-barrier example remains valid at the actual radius, not merely at an arbitrary test point. With the radius `R=1/32` above and `mu=1/1000`, the positive cubic example gives + +`F(R,0)-F(0)=113/24576000>0`. + +The proposal retains this obstruction and does not infer an elder relation from the vertical barrier. + +The `D4F=0` mixed-cubic example also works exactly: substituting `h=bx^2/(2a)` produces `b^2x^4/(8a)`, hence `f'''=tau+3b^2x/a`. The reduced cubic can change sign despite a zero ambient fourth derivative. This explains the `B^2/a` term; it is not presented as two maxima inside the valid isolation square. My separate check confirms the sign change at `-2/49` for the author's rational parameters. + +The two-maxima polynomial has cubic magnitudes `tau` at its maxima and fourth derivative magnitude `2tau/delta`. It is a valid obstruction to an isolation radius independent of derivative bounds. These deterministic polynomial controls are not claimed to occur with positive probability as exact samples of the finite Fourier law. All three negative boundaries must remain attached to reuse of this review. + +## 3. Literal pair marks and arbitrary orientation are handled without losing factors + +The source has `M=z-ru/2`, `S=z+ru/2`: r is the full endpoint distance. For a univariate restriction `p(s)=F(M+su)`, direct integration by parts gives + +`integral_0^r [6s(r-s)/r^3] p'''(s) ds = (6/r^3){r[p'(0)+p'(r)]-2[p(r)-p(0)]}`. + +At critical endpoints this is `12k`. The positive kernel has unit mass and mean `r/2`, proving the endpoint error `Cr/2`. My separate monomial-integration controls verify this identity before and after imposing endpoint criticality for degrees 3 through 9 and three rational radii. This avoids relying only on the normal-form example. The example's `k=1/12` is correct. The quartic with positive gap but zero endpoint cubic correctly saturates the `Cr/2` error and prevents dropping that error hypothesis. + +The vector Taylor formula is also correctly normalized: dividing the gradient remainder by r gives `||Hu||<=Br/2`. At a nondegenerate maximum, the hard eigenvalue has magnitude at least a by the transverse Rayleigh quotient. The soft eigenvalue magnitude is at most `||Hu||<=a/2` under `rB<=a`; the eigenspaces are distinct. Projecting onto the hard eigenvector bounds its u-component by `Br/(2a)`. + +With the soft unit vector oriented toward u, `||u-e||^2=2(1-sqrt(1-t^2))<=2t^2`, where `t=|u.v_h|`. The multilinear telescoping difference of the two cubic tensors is at most `3B||u-e||`, giving the sharper constant `(3sqrt(2)/2)B^2r/a`; the proposal's `3B^2r/a` is conservative. Therefore (12)–(13) really imply the anchored eigenframe cubic `F_eee(M)>=6k_-`. No pair-frame diagonalization, covariance isotropy, upper cubic mark, or saddle-type inference is smuggled into this implication. + +## 4. Fixed-law norm radius and two-pair dichotomy check out + +For the literal finite law, the two sine/cosine coefficient squares for each frequency pair add to `2q_n` times the squared product of directional frequency projections. Thus for arbitrary unit derivative slots the entire coefficient-row squared norm is + +`sum_(n in S_K) q_n product_j (omega n.v_j)^2 <= beta_j^2`. + +Cauchy–Schwarz is simultaneous in location and all slots, so taking their supremum proves the full multilinear operator bound. No extra coordinate factor or rotation-invariant spectrum is needed. The constant Gaussian coefficient does not affect derivatives but its inclusion in `||Z||` is harmlessly conservative. Positive nonzero modes in the pinned K>=2 law ensure `beta_3,beta_4>0`. + +For `W>=1`, the first two radius terms dominate their respective constants divided by `W^2`, and `beta_4 W+beta_3^2W^2/a <=(beta_4+beta_3^2/a)W^2` handles the third. Hence `R>=kappa/W^2`. A second maximum at Euclidean distance less than R lies inside the first chart's square regardless of its own eigenframe. The embedded-scale assumptions avoid torus-lift ambiguity. + +For two pairs let s be their actual center distance. The full-distance convention gives `|M_2-M_1|<=s+(r_1+r_2)/2<=(1+rho)s`. Under `W^2(1+rho)s`. The other maximum is strictly inside the chart, a contradiction. This verifies all factors and the strict inequality used to prove `W^2>=c_*/s`. + +For the cutoff event `s<=eta`, the radius hypothesis must remain `r_i<=rho s` for each realized configuration. It cannot be weakened silently to `r_i<=rho eta`. The proposal's paragraph after (20) uses the correct pointwise interpretation. Its positive constants depend on a, `k_-`, rho, `eta0` and the fixed spectrum through `beta_j`, kappa and N. No uniformity as a or `k_-` vanishes, or as K,L,spectra vary, is established. The bound can exceed one at the outer cutoff; it remains a valid bound and may simply be capped at one. + +A nonblocking redundancy: with `rho<1`, `s>0` and `r_i<=rho s`, one already has `|M_2-M_1|>=(1-rho)s>0`; a shared maximum is geometrically impossible within this regime. Explicitly requiring distinct maxima is safe. Likewise the second pair's derivative/mark conditions are stronger than the contradiction needs; it only uses its other maximum. Neither redundancy invalidates the statement or warrants changing the frozen source. + +## 5. Unconditional and zero-target conditioned Gaussian tails + +The deterministic obstruction places the entire occurrence event inside one global coefficient-norm tail. For `t<=c/4`, where c is kappa or `c_*`, `W^2>=c/t` implies `||Z||>=sqrt(c/t)-1>=sqrt(c/t)/2`. Since `E exp(||Z||^2/4)=2^(N/2)`, Markov gives exactly + +`P(E_t)<=2^(N/2) exp[-c/(16t)]`. + +There is no union bound over locations or angles: the derivative and norm implications already hold simultaneously in all of them. The outer-probability formulation avoids an unstated measurability premise and is sufficient for the stated estimate. The finite-dimensional configuration event can also be exhausted by compact sets after bounding coefficients and separating all strict degeneracies, but no count or selected-bar measurability is being inferred here. + +The requested zero-target extension is valid with the following precise law. Let A be any deterministic linear observation map, possibly depending on fixed sites, frames and scale, of rank q. Under the canonical Gaussian disintegration on the support of AZ, set `P=I-A^dagger A`, the orthogonal projection onto `ker A`. Then + +`Z | AZ=0` has the law of `P G`, with `G` standard Gaussian in `R^N`, + +`E[exp(||Z||^2/4) | AZ=0]=2^((N-q)/2)`. + +This statement includes redundant or rank-deficient observations; the target zero always belongs to the support. It requires no observation covariance lower bound. The same pointwise deterministic obstruction therefore gives + +`P_A^0(E_t)<=2^((N-q)/2) exp[-c/(16t)]`, for `t<=c/4`. + +Keeping the original `2^(N/2)` prefactor is uniform over all deterministic A, even if its rank changes with geometry. Zero-gradient pins at fixed sites are an example. This is a canonical Gaussian regression statement, not conditioning on an equality event as if it had positive probability, and not an arbitrary regular-conditional version assigned at a null target. This extension is recorded under this reviewer's derivation and the Gaussian cross-check; it is not retroactively attributed to the frozen proposal. + +For a supported nonzero target y the mean is `m=A^dagger y`, orthogonal to `ker A`, and the exact MGF gains `exp(||m||^2/4)`. Thus a valid estimate is + +`P_A^y(E_t)<=2^((N-q)/2) exp[||m||^2/4-c/(16t)]`. + +Uniform use needs a uniform bound on m. A bounded target alone is insufficient: `epsilon Z_1=1` forces `Z_1=1/epsilon`. The pinned source supplies mean control for its particular stable `U_r=v_r` observations on compact birth/cubic windows; it supplies no automatic arbitrary four-site-height extension. Adaptive field-selected locations, Palm/Rice weighting and nonsingular observation densities need separate arguments. None of these missing extensions is asserted by the present proposal, so they are boundaries rather than blockers to this verdict. + +## 6. Precise disposition and limits + +There are **no blocking findings** against equations (1)–(20) in their stated domain. The following are retained conditions or separate obligations, not waived defects: + +- C4 bounds are needed on the actual chart and pair segment; operator norms and embedded lifts cannot be replaced by unverified coordinate or torus-distance conventions. +- The anchor is a nondegenerate maximum, the hard pivot is bounded away from zero, and the cubic lower bound is fixed. The second root classification uses the anchor slope, not strict convexity alone. +- Pair marks require the displayed `rB` and cubic-error controls or the full norm dichotomy. Positive k alone does not give a nonzero endpoint cubic. +- Occurrence probability is distinct from a factorial kernel or expectation. Rank-deficient conditional norm tails do not create ordinary pin densities, weighted determinant bounds, normalizer bounds or a finite-support counting theorem. +- The vertical barrier supplies no enclosing superlevel region, actual elder selection or global persistence pairing. The D4-zero mixed-cubic and full-height-barrier obstructions remain attached. +- Small a or cubic marks, pairs not short relative to their actual center distance, other collision geometries, global H0 decomposition, essential classes, higher homology, complement estimates, same-field expanding-domain control, process convergence and literal lifetime-density asymptotics remain outside this acceptance. + +My independent finite cases comprise 96 mixed jet cancellations, 21 general kernel/endpoint identities, 2 closed-face saddle examples, 18 negative-hard Schur classifications, 2 retained obstruction checks, 16 exact anisotropic spectral row bounds, 96 inverse-W-squared radius inequalities and 4 singular centered-regression projector/MGF checks. They use Fraction arithmetic and no field sampling. No production or frozen-source edit, remote action, Lean build, kernel audit, formal theorem, human/provider-distinct approval, global bar result, point-process result or scientific acceptance occurred. + +The Gaussian cross-check note is `nonauthor-fold-isolation-checks/gaussian-crosscheck/GAUSSIAN_OPERATOR_TAIL_CROSSCHECK_20261016T014803Z.md`, 13,955 bytes, SHA-256 `c06161e9a8b697b2cb6fd0c6ff597c2dcdff91f653d074c3e47564cada6f2f67`. The final integrity receipt and manifest bind this report, the check source, actual process records and that note, while keeping the three frozen author files unchanged. diff --git a/experiments/universality/fold_isolation_diagonal/NONAUTHOR_FOLD_ISOLATION_REVIEW_MANIFEST.json b/experiments/universality/fold_isolation_diagonal/NONAUTHOR_FOLD_ISOLATION_REVIEW_MANIFEST.json new file mode 100644 index 000000000..803621c60 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/NONAUTHOR_FOLD_ISOLATION_REVIEW_MANIFEST.json @@ -0,0 +1,79 @@ +{ + "state": "FROZEN_NONAUTHOR_CONVENTIONAL_REVIEW_ORG0_NO_SCIENTIFIC_ACCEPTANCE", + "created_at_utc": "2026-10-16T01:51:07.107581+00:00", + "review_verdict": "ACCEPT_SCOPED", + "reviewed_author_inputs": [ + { + "path": "QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md", + "bytes": 19138, + "sha256": "827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30" + }, + { + "path": "fold_isolation_controls.py", + "bytes": 6308, + "sha256": "13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5" + }, + { + "path": "FOLD_ISOLATION_CONTROLS.json", + "bytes": 795, + "sha256": "ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf" + } + ], + "files": [ + { + "path": "NONAUTHOR_FOLD_ISOLATION_REVIEW.md", + "bytes": 18409, + "sha256": "f06c61e3f3e60c795a1a14e824640f1689d4143dba4f6698cada7a41929eec90" + }, + { + "path": "nonauthor-fold-isolation-checks/CHECK_RUN_SUMMARY.json", + "bytes": 3567, + "sha256": "638503d7560f87424e55cb3f118ae63b3bbc355e01350a86e40076b155e7429d" + }, + { + "path": "nonauthor-fold-isolation-checks/author-controls.meta.json", + "bytes": 702, + "sha256": "932a196891241aba910293fbdf326b116e8bb341dfe4f3d0f37158aae2a9d05f" + }, + { + "path": "nonauthor-fold-isolation-checks/author-controls.stderr", + "bytes": 0, + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "path": "nonauthor-fold-isolation-checks/author-controls.stdout", + "bytes": 795, + "sha256": "ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf" + }, + { + "path": "nonauthor-fold-isolation-checks/gaussian-crosscheck/GAUSSIAN_OPERATOR_TAIL_CROSSCHECK_20261016T014803Z.md", + "bytes": 13955, + "sha256": "c06161e9a8b697b2cb6fd0c6ff597c2dcdff91f653d074c3e47564cada6f2f67" + }, + { + "path": "nonauthor-fold-isolation-checks/independent-controls.meta.json", + "bytes": 734, + "sha256": "3606ca471c372e5f86b936c842a40eccd5b7ee9c17471f062872742bc6b156cc" + }, + { + "path": "nonauthor-fold-isolation-checks/independent-controls.stderr", + "bytes": 0, + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "path": "nonauthor-fold-isolation-checks/independent-controls.stdout", + "bytes": 777, + "sha256": "4b1eb60b2f44d9e67caf3469dbba06d90fa29d7d32a74cb6fe46488b988bd8ec" + }, + { + "path": "nonauthor-fold-isolation-checks/review_checks.py", + "bytes": 6772, + "sha256": "0c9155f38b4aa7b9baae0835b8f9c6681f5cfe1de890d84178605bafbdea6eec" + }, + { + "path": "nonauthor-fold-isolation-checks/run_checks.py", + "bytes": 2831, + "sha256": "eabae610abb5c0a658d2705144e30dab6ed9cbd14972bfc220540ca1d18a9ba0" + } + ] +} diff --git a/experiments/universality/fold_isolation_diagonal/NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW.json b/experiments/universality/fold_isolation_diagonal/NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW.json new file mode 100644 index 000000000..61eeb54b5 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW.json @@ -0,0 +1,132 @@ +{ + "accepted_claims": [ + "Upper moments alone do not give a contact floor", + "Positive full-square repeated spectra retain a uniform one-contact floor but lose two-contact floor and far mixing", + "Explicit epsilon=N^-12 has an empty compact candidate process in probability at t=N^-3", + "Separate deterministic smaller positive fills preserve a finite positive compact first moment", + "Fourteen-dimensional Gaussian interpolation controls bounded local contact weights under the stated whitened cross bound", + "Canonical finite-r coupling and conservative m^-14 / sqrt(r)m^-16 bounds under joint floor and small-r premises", + "Distinct-endpoint four-point Borel identity and stated moving middle/far candidate bounds in dimension two", + "Qualitative normalized Gaussian truncation grid/image bounds, analytic power floor and compatible logarithmic regime" + ], + "actual_nonauthor": true, + "author_frozen_json_reproduced": true, + "authors": "OpenAI/Codex core-theorem agent with Root steering", + "blind_credit": 0, + "blocking_findings": [], + "completed_utc": "2026-10-16T02:38:13.687091+00:00", + "counterfamily_independent_exact_checks": 187, + "counterfamily_manifest": { + "bytes": 4458, + "path": "COUNTERFAMILY_NONAUTHOR_REVIEW_MANIFEST.json", + "sha256": "614f84154b75505c0004f70ca2e81d974c9dec49a9f5c64c422b63a787d82a56" + }, + "counterfamily_report": { + "bytes": 23286, + "path": "COUNTERFAMILY_NONAUTHOR_REVIEW.md", + "sha256": "26b1b1a9a1e978edd036ee2d1e244e8b8b2089d3c6a48edb94d78d37577cc2c5" + }, + "counterfamily_subreview_scope": "Sections 3-5 only; read and incorporated at exact identity. No duplicate first check by integrating reviewer.", + "evidence_scope": "Finite algebra controls plus conventional mathematical proof review; no stochastic draws, numerical eigenvalues/Rice integrals or process simulations.", + "formal_verification": false, + "frozen_author": { + "JET_FACTORIZATION_OBSTRUCTION_CONTROLS.json": { + "bytes": 1618, + "path": "JET_FACTORIZATION_OBSTRUCTION_CONTROLS.json", + "sha256": "1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764" + }, + "JET_NONDEGENERACY_FACTORIZATION_OBSTRUCTION.md": { + "bytes": 34627, + "path": "JET_NONDEGENERACY_FACTORIZATION_OBSTRUCTION.md", + "sha256": "0dfdddfbcac47b2fd07449f4152b7f5d935f92aed15272fb8333ed9cc60b5766" + }, + "jet_factorization_obstruction_controls.py": { + "bytes": 5818, + "path": "jet_factorization_obstruction_controls.py", + "sha256": "5b1e8fc18a6e32e3a312a253e3e4d1110bb674385f3d49447a45e78ad4df541f" + } + }, + "frozen_input_count": 41, + "human_credit": 0, + "kernel_controls_checks": { + "canonical_conditioning_projector_with_singular_extension": "PASS", + "density_weight_radial_volume_exponent_ledger": "PASS", + "fourteen_dimensional_gaussian_interpolation_derivative": "PASS", + "frozen_author_source_identities_unchanged": "PASS", + "literal_six_row_transform_and_twelve_row_block_jacobian": "PASS", + "physical_hessian_congruence_type_and_contact_normalization": "PASS", + "regular_endpoint_integral_rows_degrees_3_through_8": "PASS" + }, + "kernel_negative_controls_rejected": { + "missing_12_in_each_fold": true, + "omitted_nonzero_observation_pin_detected": true, + "wrong_r_instead_of_sqrt_r_offdiagonal": true, + "zero_value_crosscovariance_does_not_factor_local_A_weight": true + }, + "new_hosted_execution": false, + "not_established": [ + "Mean preservation for explicit epsilon=N^-12", + "Effective analytic exponent or numerical constants/window", + "Admission or implementation of an actual changing project source sequence", + "One-point intensity/remainder/tight marks for that sequence joined to the near theorem", + "Full candidate point-process convergence or all higher configurations", + "Nonlocal elder selector factorization, actual-global-bar universality or higher homology" + ], + "organizational_independence_credit": 0, + "primary_import": { + "bytes": 586327, + "consumed_scope": "Theorem 1 and Corollary 4 only, together giving a local finite positive distance exponent for an analytic determinant. No full paper or original Lojasiewicz proof claimed.", + "sha256": "b11ceb8df55335260ed8eaef4a71406cfa592565869ced020f757354124913c1", + "teissier_saved_pdf": { + "bytes": 325166, + "consumed_scope": "Theorem 3.1 on PDF page 7, rendered and visually inspected. Printed exponent >=0; the standard positive exponent in the nonempty-zero-set application is made explicit by Feehan.", + "relative_path": "../jet-factorization-primary-sources/teissier-lojasiewicz-2012.pdf", + "sha256": "cb88d0410e5c0a629f85592864809b987612663593a04dd8ea2928502cac6873" + }, + "url": "https://arxiv.org/html/1708.09775v6", + "version": "arXiv:1708.09775v6, 11 June 2018" + }, + "prior_review_mutations": 0, + "process_runs": [ + { + "command": [ + "python3", + "-B", + "jet_factorization_obstruction_controls.py" + ], + "label": "author", + "reproduces_frozen_json_byte_for_byte": true, + "returncode": 0, + "started_utc": "2026-10-16T02:35:18.553011+00:00", + "stderr_bytes": 0, + "stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "stdout_bytes": 1618, + "stdout_sha256": "1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764", + "wall_seconds": 0.49629570869728923 + }, + { + "command": [ + "python3", + "-B", + "nonauthor-jet-factorization-review/independent_kernel_controls.py" + ], + "fixture_refinement": "Nonzero exact cross-quadratic term 19/45 exercises the Gaussian derivative sign.", + "label": "independent", + "returncode": 0, + "started_utc": "2026-10-16T02:36:01.939405+00:00", + "stderr_bytes": 0, + "stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "stdout_bytes": 4385, + "stdout_sha256": "1e94c994a8111bd1b741d83e8404a242109a69ac5331cd07eddf8b0930a46330", + "wall_seconds": 0.3060152088291943 + } + ], + "remote_mutations": 0, + "reviewer": "OpenAI/Codex /root/c261a_successor_preparation_review_20261016", + "same_provider": true, + "scientific_effect": "NONE", + "source_admission": false, + "source_exposed": true, + "source_mutations": 0, + "verdict": "ACCEPT_SCOPED" +} diff --git a/experiments/universality/fold_isolation_diagonal/NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW.md b/experiments/universality/fold_isolation_diagonal/NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW.md new file mode 100644 index 000000000..289037fab --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW.md @@ -0,0 +1,132 @@ +# Nonauthor review of the jet obstruction and local factorization route + +**Verdict: ACCEPT_SCOPED.** I found no blocking mathematical error in the frozen note's candidate-process obstruction, separate existence-only positive-mean diagonal, local Gaussian factorization lemma, finite-radius conditional coupling, conservative middle/far kernel bounds, or qualitative Gaussian truncation route. No source amendment is required for those claims with the premises and scope recorded below. The actual obstruction is stronger than a failed covariance estimate: the displayed positive full-square family has an empty compact candidate process in probability despite a uniform single-contact floor; a separate deterministic diagonal retains a positive limiting mean. The latter mean transfer has not been established for the explicit fill `epsilon_N=N^-12`. + +This is an actual nonauthor mathematical review by OpenAI/Codex `/root/c261a_successor_preparation_review_20261016`, completed 16 October 2026 UTC. The authors are the core-theorem agent with Root steering. I read the frozen note and controls before reviewing them. The separate agent `/root/c261a_local_runtime_20261016` reviewed only §§3–5; its sealed report is incorporated at its exact identity, without rewriting it. I substantively reviewed §§6–9, the source-row/area-formula joins, and the integration boundaries. Both reviews are source-exposed and from the same OpenAI provider. **Organizational-independence credit 0; blind credit 0; human credit 0.** This is conventional mathematical review, not Lean verification, a hosted test, a scientific transition, a source-admission decision, or a review of the earlier hosted packet. + +## Exact object and evidence + +| Frozen input | Bytes | SHA-256 | +|---|---:|---| +| `JET_NONDEGENERACY_FACTORIZATION_OBSTRUCTION.md` | 34627 | `0dfdddfbcac47b2fd07449f4152b7f5d935f92aed15272fb8333ed9cc60b5766` | +| `jet_factorization_obstruction_controls.py` | 5818 | `5b1e8fc18a6e32e3a312a253e3e4d1110bb674385f3d49447a45e78ad4df541f` | +| `JET_FACTORIZATION_OBSTRUCTION_CONTROLS.json` | 1618 | `1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764` | +| `COUNTERFAMILY_NONAUTHOR_REVIEW.md` | 23286 | `26b1b1a9a1e978edd036ee2d1e244e8b8b2089d3c6a48edb94d78d37577cc2c5` | +| `COUNTERFAMILY_NONAUTHOR_REVIEW_MANIFEST.json` | 4458 | `614f84154b75505c0004f70ca2e81d974c9dec49a9f5c64c422b63a787d82a56` | + +The retained source inputs are Main `85b928b326ff0586715a56ae91f51d3fbd89557a`: `finite_gaussian_contact/PROOF.md`, 25367 B / `7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474`; `finite_candidate_lifetime/PROOF.md`, 29190 B / `ed31f562f0ffaaf6d4bd2ceb1ae3d234798e983a55d55c6983b128fce27133f5`; and `iid_short_bar_process/PROOF.md`, 30935 B / `320bc988c468bea4938bb777a2d05952c6ab8e0a384f33725b50de847aeec1eb`. I consumed their actual six-row divided differences, Gaussian regression/contact formulas, coefficient-space Borel area formula, and normalized periodized covariance. The previously accepted local-norm, critical-count and isolation results remain inputs at their frozen identities. I did not repeat their first reviews or rerun their controls. + +The author controls exited 0 with empty stderr and reproduced all 1618 frozen JSON bytes exactly. My separate `nonauthor-jet-factorization-review/independent_kernel_controls.py` also exited 0 with empty stderr. It checks the actual six-row observation transform and twelve-row block determinant; regular endpoint integral identities through polynomial degree eight; a critical maximum–saddle polynomial with nonzero mixed Hessians; the fourteen-dimensional density derivative with a nonzero cross-quadratic term; and twelve-pin canonical projections in coefficient dimensions 16, 20 and 28, including singular conditional extension covariance. Four negative controls reject a missing fold factor 12, an incorrect off-diagonal scaling, an omitted nonzero observation pin, and a scalar-only factorization inference. These are independent exact Fraction/integer controls, not field samples or numerical proofs of the analytic estimates. + +The controls, stdout/stderr, process receipt, primary import receipt, machine verdict and source identities are retained separately. The new manifest excludes itself and hashes this review and its evidence without a self-hash cycle. The report and manifest from the separate counterfamily reviewer have been read and their complete 11-input/7-output manifest closure verified. All three new author files and consumed frozen inputs remain unchanged. + +## The counterfamily is an actual candidate-process obstruction + +I incorporate the separate §§3–5 review at the exact identity above. Its 187 independent finite checks support, but do not replace, its conventional arguments. The report establishes every-frame seven-row rank by degree-three polynomial interpolation on the full 5-by-5 frequency grid and compactness of `O(2)`, rather than extrapolating an axial LDL calculation. It additionally derives the conservative explicit base floor + +`m_G >= 361267200/319913861015774089 > 0`. + +The covariance mixture therefore preserves a uniform one-contact floor while the macroscopic resonant two-value difference has variance less than `epsilon_N`. The physical twelfth moment remains at most `8^6`. This disproves the inference from upper moments plus a one-contact floor to a uniform two-contact floor or far covariance decay. Upper moments alone fail even the one-contact lower-bound inference. + +The process conclusion uses additional work. The full physical perturbation has global `C4` second moment `O(N^-10)` for the explicit fill `N^-12`. The real coefficient dimension is `(4N+1)^2`; the coordinate derivative estimate uses 15 multiindices and row norm at most 64, giving the stated constant 960. The fixed base is almost surely Morse with distinct base critical values. A uniform repeated-cell `C2` stability radius follows from contraction on each critical ball, including existence of a root, together with a gradient floor off those balls. It preserves the critical roles and excludes extra roots. Repeated copies of the same base critical point have the same role, so their height ties do not provide a maximum–saddle gap of zero. The random positive base role-gap and stability radius yield the empty compact candidate process at `t_N=N^-3` by an unconditional probability bound. + +The separate positive-mean argument is correctly limited to a smaller deterministic sequence `00`. The latter follows from the stated `R_t<=c eta_t^nu`, `eta_t` is an inverse fixed power of `log(1/t)`, and `R_t^2=O(log(1/t))`, the middle bound is `O(t^(2/3)` times a fixed logarithmic power). Also `r_t/m_t->0`. These statements hold for any fixed finite exponent; no separation-independent floor at shrinking eta is needed. All dimensions and determinant powers here are planar. They are not a theorem for arbitrary spatial dimension. + +The near estimate may only be joined at its existing accepted hypotheses, including the lifetime-radius inclusion condition and fixed positive curvature/cubic lower marks. Neither middle/far estimates nor the iid-source weighted-cluster theorem supply that join by themselves. These calculations establish the stated twofold candidate estimates; they do not establish all factorial moments or a Laplace functional. + +## The Gaussian truncation route is qualitatively coherent + +The retained iid source's covariance is the **normalized** image sum `sum_j exp(-|h+Lj|^2/2)/sum_j exp(-|Lj|^2/2)`. Its physical Fourier weights are precisely proportional to `exp(-|xi_n|^2/2)`. The finite normalized square truncation in §9 is a specific changing source family consistent with that identity, rather than arbitrary weights permitted by U. The denominator of the image sum is bounded above and below for `L>=Lmin`; for large L it differs from 1 by `O(exp(-cL^2))`. The estimates below include this normalization. + +The grid comparison is sufficient as an existence bound. For `omega<=1` the central frequency cells give normalizer at least `c omega^-2`. On each cell, `|xi|^2>=|x|^2/2-1` and `|xi|<=|x|+1` give an integrable polynomial-Gaussian majorant. A frequency beyond the square cutoff has cell points outside Euclidean radius `Lambda-1`. For `Lambda>=4`, absorbing a fixed polynomial into a weaker Gaussian proves the generous `C exp(-Lambda^2/64)` bound in (18). The same argument with degree twelve proves uniform physical twelfth moments for the normalized truncations, since their central-square normalizers have the same lower bound. Omitting tail mass changes the normalized covariance derivatives through order six by at most `C delta_6` once that mass is at most 1/2. + +For the image sum, each derivative of order at most six is a fixed Hermite polynomial times a Gaussian. Absorb it into `C exp(-|h+Lj|^2/4)`, factor out `exp(-dist_T(h,0)^2/8)`, and bound the remaining separated lattice Gaussian sum uniformly for positive Lmin. This proves (19), with the normalization just noted. On a fixed bounded separation band, all noncentral images and the normalization error are `O(exp(-cL^2))` as L grows. + +The Euclidean Gaussian spectral density is strictly positive. If a linear combination of the fourteen named contact rows at distinct centers has zero variance, its exponential-polynomial spectral multiplier vanishes identically. Inverting it as a distribution gives two sums of derivatives of point masses with disjoint supports. Each local polynomial must vanish. The seven rows at one center are independent under an invertible frame change. This proves positive joint covariance at every nonzero separation and all frames; it is not an assertion about arbitrary finite-support augmented jets. + +For the power floor, parameterize the separation as `s e(theta)` and each frame by an angle; reflected frames only change the signs of appropriate rows and preserve the determinant. The Euclidean determinant is real analytic in these four real parameters. It is zero exactly at `s=0` in a small signed-separation band: at zero fourteen rows lie in the ten-dimensional full third jet, while nonzero separation has the positivity just proved. On compact angular charts the distance to that zero set is comparable to `|s|`. A finite cover of the charts and the analytic distance inequality give `det>=c|s|^nu` for some finite positive exponent. The other thirteen eigenvalues have a uniform upper bound, so the same power lower-bounds the smallest eigenvalue. Compactness handles the remaining bounded separations and Gaussian cross-decay handles all sufficiently large separations. + +I inspected the retained Teissier PDF's Theorem 3.1 on PDF page 7, including a rendered visual read. Its printed statement permits a nonnegative reciprocal exponent. To make explicit the positive finite exponent actually used in the nonempty-zero-set application, I additionally read the statements of Theorem 1 and Corollary 4 of [Feehan, *Resolution of Singularities and Geometric Proofs of the Lojasiewicz Inequalities*, arXiv:1708.09775v6](https://arxiv.org/html/1708.09775v6). Together they give the local distance power for a real analytic function; compact covering gives the form consumed here. This bounded import corroborates the standard application of [Teissier's Theorem 3.1](https://webusers.imj-prg.fr/~bernard.teissier/documents/Lojasiewicz_2012.pdf). I did not review either entire paper, resolution of singularities, or the original historical proof. The saved primary identities and consumed scope are in `nonauthor-jet-factorization-review/PRIMARY_IMPORT_RECEIPT.json`. + +Covariance entry errors through order six control the fourteen-row matrix error uniformly in the angles. Weyl therefore transfers the Euclidean floor at separation at least eta when `delta_6+C exp(-cL^2)=o(eta^nu)`. For separations beyond a fixed sufficiently large constant, (19) and the one-contact floor give a constant joint floor directly. Thus the transfer covers the entire growing middle band, not just a fixed compact set of separations. The one-contact covariances converge uniformly to their Euclidean values by the same estimate. + +The proposed choices are mutually compatible: take `L` comparable to `t^-1/3`, `Lambda^2` a positive fixed multiple of `log(1/t)`, `R_t^2` a positive fixed multiple of that logarithm, and eta an inverse logarithmic power compatible with the earlier near cutoff. Then `delta_6` decays by a positive power of t, image errors are faster, and both are smaller than every fixed inverse logarithmic power. The floor, `r_t/m_t`, middle and far conditions follow. The integer cutoff is `K=ceil(L Lambda/(2pi))=O(L sqrt(log(1/t)))`; its logarithm and the covering logarithm have the size required by the accepted near-diagonal existence regime. These are eventual qualitative bounds. No evaluated exponent, effective constants, numerical window or tested sequence is supplied. + +## Acceptance boundaries and remaining work + +There are no blocking findings requiring an author-file change within the frozen scope. The accepted claims and the remaining boundaries should travel together: + +| Claim | Disposition | +|---|---| +| U alone need not give a one-contact floor | ACCEPT_SCOPED: explicit ideal-Gaussian counterfamilies | +| U plus a one-contact floor need not give far mixing or a nonzero candidate Poisson law | ACCEPT_SCOPED: actual repeated-spectrum empty-process obstruction | +| Positive limiting compact mean despite empty probability limit | ACCEPT_SCOPED for the deterministic smaller-fill existence diagonal; explicit `N^-12` mean transfer remains unproved | +| Local fourteen-row density interpolation and contact factorization | ACCEPT_SCOPED under (11), bounded local marks and fixed compact targets | +| Finite-radius `m^-14` upper and `sqrt(r)m^-16` error | ACCEPT_SCOPED under the joint floor and `r<=c m`; rate requires Lipschitz local marks | +| Four-point Borel identity and moving middle/far estimates | ACCEPT_SCOPED for distinct planar candidate endpoints and the stated covariance/scale premises | +| Normalized Gaussian truncation, derivative decay and qualitative power floor | ACCEPT_SCOPED as an eventual source-specific route with the bounded classical import above | +| Admission of a project source sequence or full candidate point-process theorem | NOT ESTABLISHED by this review | +| Nonlocal elder-selector factorization, actual-global-bar universality or higher homology | NOT ESTABLISHED | + +A source adoption would still have to declare the actual changing weights, K,L,t and marks, join the accepted near theorem and one-point intensity/remainder/tight-mark results for that same sequence, and supply higher-configuration or Laplace/dependency control. Actual persistence bars additionally need source-bound weighted localization/remote-exclusion estimates for the global selector and the remaining topology/complement work. The existing finite K2/K3 enlarged-height singularities remain negative evidence; they cannot be relabeled as the new changing Gaussian carrier. The frozen note is not amended by this appended review, and no old source status is silently changed. + +All work was local and bounded. No author source, prior review, hosted packet, Lean file, native record, remote status, claim, Ready state or scientific register was edited. No field draws, approximate critical roots, floating eigenvalues, Poisson simulations, numerical Rice integrals, Lean execution, new hosted run or provider-distinct approval occurred. The exact controls and conventional mathematical arguments support the scoped acceptance stated here, with scientific effect NONE and organizational-independence credit 0. diff --git a/experiments/universality/fold_isolation_diagonal/NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW_MANIFEST.json b/experiments/universality/fold_isolation_diagonal/NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW_MANIFEST.json new file mode 100644 index 000000000..c8a311649 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW_MANIFEST.json @@ -0,0 +1,288 @@ +{ + "helper_report_preserved_verbatim": true, + "inputs": [ + { + "bytes": 23286, + "path": "COUNTERFAMILY_NONAUTHOR_REVIEW.md", + "sha256": "26b1b1a9a1e978edd036ee2d1e244e8b8b2089d3c6a48edb94d78d37577cc2c5" + }, + { + "bytes": 4458, + "path": "COUNTERFAMILY_NONAUTHOR_REVIEW_MANIFEST.json", + "sha256": "614f84154b75505c0004f70ca2e81d974c9dec49a9f5c64c422b63a787d82a56" + }, + { + "bytes": 849, + "path": "FINITE_FOURIER_COUNT_CONTROLS.json", + "sha256": "bbe147fdf409a3e37b450d4cbcd5d2a927140f95bd3ce0f36cafe16861123bfd" + }, + { + "bytes": 25699, + "path": "FINITE_FOURIER_CRITICAL_COUNT_DIAGONAL_ROUTE.md", + "sha256": "621cee96854aa440498cecf5452a8d1e23f90a4458779c3286c18eba106fe138" + }, + { + "bytes": 795, + "path": "FOLD_ISOLATION_CONTROLS.json", + "sha256": "ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf" + }, + { + "bytes": 1618, + "path": "JET_FACTORIZATION_OBSTRUCTION_CONTROLS.json", + "sha256": "1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764" + }, + { + "bytes": 34627, + "path": "JET_NONDEGENERACY_FACTORIZATION_OBSTRUCTION.md", + "sha256": "0dfdddfbcac47b2fd07449f4152b7f5d935f92aed15272fb8333ed9cc60b5766" + }, + { + "bytes": 21880, + "path": "NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW.md", + "sha256": "16152427f7013822ac1bba81142429800334e6bf0f354a6439b23a1997bc4524" + }, + { + "bytes": 18409, + "path": "NONAUTHOR_FOLD_ISOLATION_REVIEW.md", + "sha256": "f06c61e3f3e60c795a1a14e824640f1689d4143dba4f6698cada7a41929eec90" + }, + { + "bytes": 23887, + "path": "NONAUTHOR_UNIFORM_LOCAL_NORM_REVIEW.md", + "sha256": "ff552dfacadede53405530a219302c1163dca8ba97d518f02d407970d3de3c1c" + }, + { + "bytes": 4763, + "path": "NONAUTHOR_UNIFORM_LOCAL_NORM_REVIEW_MANIFEST.json", + "sha256": "1ccd38374adcf4931b87abeacdefaa64990080a9b3b98835a7c140bde2f505a7" + }, + { + "bytes": 19138, + "path": "QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md", + "sha256": "827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30" + }, + { + "bytes": 1790, + "path": "UNIFORM_LOCAL_NORM_CONTROLS.json", + "sha256": "5d6cdb2d4925c6ba0b1b545d7a5aca1977655c1d18aa1ea29a82273f30845ece" + }, + { + "bytes": 26687, + "path": "UNIFORM_LOCAL_NORM_DIAGONAL_PROPOSAL.md", + "sha256": "8e1eefca9888d43824a9fd9caca34c44d2de4bf8e30c668b474fa2097491c76b" + }, + { + "bytes": 2560, + "path": "counterfamily-nonauthor-checks-20261016/EXECUTION_RECORD.json", + "sha256": "8ae808eed6b7f8fec67435a3b97c8422a08212e7f18ababbced3b367380e04fb" + }, + { + "bytes": 0, + "path": "counterfamily-nonauthor-checks-20261016/author_controls.stderr", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 1618, + "path": "counterfamily-nonauthor-checks-20261016/author_controls.stdout.json", + "sha256": "1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764" + }, + { + "bytes": 8080, + "path": "counterfamily-nonauthor-checks-20261016/counterfamily_checks.py", + "sha256": "0718a7f0e415f75110b5b08b5850fdc5e76c601e515ee64bb0770ceaa21834e9" + }, + { + "bytes": 0, + "path": "counterfamily-nonauthor-checks-20261016/independent_controls.stderr", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 12359, + "path": "counterfamily-nonauthor-checks-20261016/independent_controls.stdout.json", + "sha256": "3c53ed54756c911292180c5686750de9a47cf4a54c56cc0d2616155c3f5af35b" + }, + { + "bytes": 29190, + "path": "experiments/universality/finite_candidate_lifetime/PROOF.md", + "sha256": "ed31f562f0ffaaf6d4bd2ceb1ae3d234798e983a55d55c6983b128fce27133f5" + }, + { + "bytes": 25367, + "path": "experiments/universality/finite_gaussian_contact/PROOF.md", + "sha256": "7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474" + }, + { + "bytes": 30935, + "path": "experiments/universality/iid_short_bar_process/PROOF.md", + "sha256": "320bc988c468bea4938bb777a2d05952c6ab8e0a384f33725b50de847aeec1eb" + }, + { + "bytes": 5153, + "path": "finite_fourier_count_controls.py", + "sha256": "fd6f29739280a55017945526cb75fb46ebc80d86c012ff4a1ff65baa404c31de" + }, + { + "bytes": 6308, + "path": "fold_isolation_controls.py", + "sha256": "13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5" + }, + { + "bytes": 427, + "path": "jet-factorization-primary-sources/SOURCE_IDENTITY.json", + "sha256": "683fb8d29a62c9e73f01a93bf3b288c1e7189f050da4b77eb831bb063ac88492" + }, + { + "bytes": 325166, + "path": "jet-factorization-primary-sources/teissier-lojasiewicz-2012.pdf", + "sha256": "cb88d0410e5c0a629f85592864809b987612663593a04dd8ea2928502cac6873" + }, + { + "bytes": 5818, + "path": "jet_factorization_obstruction_controls.py", + "sha256": "5b1e8fc18a6e32e3a312a253e3e4d1110bb674385f3d49447a45e78ad4df541f" + }, + { + "bytes": 7279, + "path": "nonauthor-uniform-local-norm-checks/CHECK_RUN_SUMMARY.json", + "sha256": "480e1e4786933f2201a6dadb93ee720dc21b632c5a749a91ea9323f1fc506364" + }, + { + "bytes": 1419, + "path": "nonauthor-uniform-local-norm-checks/PRIMARY_IMPORT_AUDIT.json", + "sha256": "98c521f7c773e1338b2ed6bac74ebba0c3f146cdc0b989abc7dd67390b07239b" + }, + { + "bytes": 715, + "path": "nonauthor-uniform-local-norm-checks/author-controls.meta.json", + "sha256": "0e8fd7097e06de071dd233b8f6a8ff7623d370e77d78421d7b11a88b9deffac8" + }, + { + "bytes": 0, + "path": "nonauthor-uniform-local-norm-checks/author-controls.stderr", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 1790, + "path": "nonauthor-uniform-local-norm-checks/author-controls.stdout", + "sha256": "5d6cdb2d4925c6ba0b1b545d7a5aca1977655c1d18aa1ea29a82273f30845ece" + }, + { + "bytes": 738, + "path": "nonauthor-uniform-local-norm-checks/independent-controls.meta.json", + "sha256": "c88b17ff52d1eb833679da97a4cb93a2f6068345c785b8e24c8a989ac2c55a34" + }, + { + "bytes": 0, + "path": "nonauthor-uniform-local-norm-checks/independent-controls.stderr", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 1439, + "path": "nonauthor-uniform-local-norm-checks/independent-controls.stdout", + "sha256": "bd553b02c66ad4cf761700f53043ac4b03f7371b5e15aa03ad08b4d4feef2fce" + }, + { + "bytes": 8199, + "path": "nonauthor-uniform-local-norm-checks/review_checks.py", + "sha256": "d764331f77125a27f7f6a6809a813e5b4a2cacc65f66d0dc889300347074f8c5" + }, + { + "bytes": 4309, + "path": "nonauthor-uniform-local-norm-checks/run_checks.py", + "sha256": "98f86202f1ac2a15a18058a7b8cd9a49bbea180f9ab416536b224d6967f639af" + }, + { + "bytes": 443, + "path": "uniform-local-norm-primary-sources/SOURCE_IDENTITY.json", + "sha256": "92bf5297b309e947f42ee1dc837487775e7341bc7c1051e493899963ff85891a" + }, + { + "bytes": 940975, + "path": "uniform-local-norm-primary-sources/simon-lecs-on-pde.pdf", + "sha256": "e1f1b2f51d2558aa467a8c064fe82e7b9e157ad28c9c75f01e38596cde656870" + }, + { + "bytes": 6486, + "path": "uniform_local_norm_controls.py", + "sha256": "4b84755b771598ed34193c075718a41da447719f90a2ff5c6a17d48d0abc3dc4" + } + ], + "manifest_excludes_self": true, + "no_remote_or_source_mutations": true, + "nonauthor": true, + "organizational_independence_credit": 0, + "outputs": [ + { + "bytes": 25600, + "path": "NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW.md", + "sha256": "531b9d9a16831383f2f217e61bad7dea84aca941a934e4bcaa771bedcc4e73af" + }, + { + "bytes": 6389, + "path": "NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW.json", + "sha256": "13395e24282bbf7d3005b7a236c7de52f703f2d0e0878fe350dfcc6df8df91a5" + }, + { + "bytes": 8076, + "path": "nonauthor-jet-factorization-review/FROZEN_INPUT_INTEGRITY.json", + "sha256": "7c724d8a3227b41de638a6289d103361fe8bc7c3db5eb3d45ac702159c449989" + }, + { + "bytes": 4385, + "path": "nonauthor-jet-factorization-review/INDEPENDENT_KERNEL_CONTROLS.json", + "sha256": "1e94c994a8111bd1b741d83e8404a242109a69ac5331cd07eddf8b0930a46330" + }, + { + "bytes": 821, + "path": "nonauthor-jet-factorization-review/PRIMARY_IMPORT_RECEIPT.json", + "sha256": "3a52a5cd5a098b551bacb0c2dbad7c48030eaa613d36fae110e29daa4410265b" + }, + { + "bytes": 1265, + "path": "nonauthor-jet-factorization-review/PROCESS_RUN_RECEIPT.json", + "sha256": "0bfc877da8ead5d680745aeba4889baabd131d5eddea0b41efa8b8700f42b049" + }, + { + "bytes": 0, + "path": "nonauthor-jet-factorization-review/author_stderr.txt", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 1618, + "path": "nonauthor-jet-factorization-review/author_stdout.json", + "sha256": "1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764" + }, + { + "bytes": 586327, + "path": "nonauthor-jet-factorization-review/feehan-1708.09775v6.html", + "sha256": "b11ceb8df55335260ed8eaef4a71406cfa592565869ced020f757354124913c1" + }, + { + "bytes": 11734, + "path": "nonauthor-jet-factorization-review/independent_kernel_controls.py", + "sha256": "132f8fa628df68082aaf3c42e727584540e00f293ce6b0223f6cf74dae5a3924" + }, + { + "bytes": 0, + "path": "nonauthor-jet-factorization-review/independent_stderr.txt", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 4385, + "path": "nonauthor-jet-factorization-review/independent_stdout.json", + "sha256": "1e94c994a8111bd1b741d83e8404a242109a69ac5331cd07eddf8b0930a46330" + }, + { + "bytes": 230248, + "path": "nonauthor-jet-factorization-review/tmp/pdfs/teissier-page7.png", + "sha256": "e95528c20ec62ce8ea3a14f5d6672ea52d891fca7bc2e75fbd6cc44076851b29" + } + ], + "reviewer": "OpenAI/Codex /root/c261a_successor_preparation_review_20261016", + "scientific_effect": "NONE", + "sealed_utc": "2026-10-16T02:38:13.687091+00:00", + "source_exposed": true, + "source_frozen_unchanged": true, + "state": "FROZEN_LOCAL_NONAUTHOR_MATHEMATICAL_REVIEW_ORG0", + "verdict": "ACCEPT_SCOPED" +} diff --git a/experiments/universality/fold_isolation_diagonal/NONAUTHOR_UNIFORM_LOCAL_NORM_REVIEW.md b/experiments/universality/fold_isolation_diagonal/NONAUTHOR_UNIFORM_LOCAL_NORM_REVIEW.md new file mode 100644 index 000000000..2137f18c3 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/NONAUTHOR_UNIFORM_LOCAL_NORM_REVIEW.md @@ -0,0 +1,172 @@ +# Nonauthor review of the uniform local-norm diagonal proposal + +**Verdict: ACCEPT_SCOPED.** I found no blocking mathematical error in the frozen proposal's physical spectral-moment implication, explicit cutoff/trace calculation, finite PSD Gaussian exponential moment, quantitative H6-to-C4 operator estimate, deterministic covering argument, or growing-K/L diagonal cancellation. No changes are required for equations (1)–(24) under the stated hypotheses. This verdict concerns the new local mathematical argument. It is not source-law admission, a formal proof, or point-process/scientific promotion. + +Actual reviewer: OpenAI/Codex `/root/c261a_local_runtime_20261016`, 16 October 2026 UTC. Root and `/root/core_theorem_audit_20261015` are the authors. I did not author the frozen proposal or its control files. I read all new files fully and checked the bounded primary Fourier import directly. I consumed the two earlier lemmas at their exact accepted identities; I did not duplicate their first reviews or rerun their controls. This is a separate actual nonauthor, source-exposed review in the same OpenAI lineage. **Organizational-independence credit: 0.** No other reviewer is implied by this report. + +## 1. Exact reviewed object and evidence + +| New frozen input | Bytes | SHA-256 | +|---|---:|---| +| `UNIFORM_LOCAL_NORM_DIAGONAL_PROPOSAL.md` | 26,687 | `8e1eefca9888d43824a9fd9caca34c44d2de4bf8e30c668b474fa2097491c76b` | +| `uniform_local_norm_controls.py` | 6,486 | `4b84755b771598ed34193c075718a41da447719f90a2ff5c6a17d48d0abc3dc4` | +| `UNIFORM_LOCAL_NORM_CONTROLS.json` | 1,790 | `5d6cdb2d4925c6ba0b1b545d7a5aca1977655c1d18aa1ea29a82273f30845ece` | +| `uniform-local-norm-primary-sources/SOURCE_IDENTITY.json` | 443 | `92bf5297b309e947f42ee1dc837487775e7341bc7c1051e493899963ff85891a` | + +The saved Simon PDF is 940,975 bytes, SHA-256 `e1f1b2f51d2558aa467a8c064fe82e7b9e157ad28c9c75f01e38596cde656870`; I verified those bytes. I directly read Lecture 5's Fourier properties (a)–(h), including the unitary derivative, inversion and Parseval/Plancherel formulas, and its displayed Cauchy–Schwarz Sobolev proof. I do not claim to have read the entire 118-page PDF. The proposal's explicit derivative constants are calculated below using its own stated norm, rather than borrowed from an unspecified embedding constant. [Leon Simon, Lectures on PDE, Lecture 5](https://math.stanford.edu/~lms/lecs-on-pde.pdf) + +The inherited isolation proposal/review remain SHA-256 `827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30` / `f06c61e3f3e60c795a1a14e824640f1689d4143dba4f6698cada7a41929eec90`. The finite-count proposal/review remain `621cee96854aa440498cecf5452a8d1e23f90a4458779c3286c18eba106fe138` / `16152427f7013822ac1bba81142429800334e6bf0f354a6439b23a1997bc4524`. The exact source law is still Main `85b928b326ff0586715a56ae91f51d3fbd89557a`, `finite_gaussian_contact/PROOF.md`, 25,367 bytes, SHA-256 `7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474`. The six preceding author files and both preceding reviews were verified unchanged. + +At 2026-10-16T02:16:05Z I ran the new author controls and my separate `nonauthor-uniform-local-norm-checks/review_checks.py` using `/Library/Developer/CommandLineTools/usr/bin/python3 -B`, Python 3.9.6. Both exited 0 with empty stderr. Author stdout reproduces all 1,790 frozen bytes exactly. My 2,132 exact finite cases passed, including 2,001 unequal/singular PSD spectra and a separately reindexed trace-constant calculation. Commands, timestamps, wall times, stdout/stderr and hashes are retained in `nonauthor-uniform-local-norm-checks/CHECK_RUN_SUMMARY.json` and its process-record sets. Fourteen consumed input identities were verified before and after, unchanged. No previous control program was rerun. + +These finite checks supplement the mathematical review; they do not mechanically establish continuous derivative bounds, Fourier theorems, Gaussian integrals, isolation, genericity or asymptotic limits. + +## 2. The uniform family hypothesis is sufficient and is not inferred from smoothness + +The actual law is centered finite Gaussian with deterministic spectral parameters, independent standard real sine/cosine coefficients, positive symmetric full-square weights, and period `L>=Lmin>0`. The weights may vary between family members; they are not selected from a realized field. For a coordinate derivative of order j<=6 its exact variance is the sum of the corresponding squared physical-frequency monomials times q. The inequality + +`r^(2j)<=1+r^12`, `0<=j<=6`, + +therefore gives the required bound `1+M12` under normalized total mass. It covers the zeroth derivative as well as all orders through six. With total mass merely bounded by a fixed M0, the same calculation gives `M0+M12` throughout. This replacement propagates to the trace T and the eventual exponential constant c. A fixed M0 and a uniform physical M12 are sufficient for this variant. + +M12 alone does not bound the full localized H6 norm. A growing independent DC Gaussian changes the zeroth variance and, after multiplication by the cutoff, all its cutoff-derivative contributions, while leaving every positive physical spectral moment unchanged. The proposed `F_A=A Z0+G` example makes the trace diverge as `A^2||chi_b||_H6^2`. This is a correct falsifier of the stated full-norm trace inference without mass control. Its purpose concerns this full-norm method; it does not claim that adding a constant changes critical locations or gap marks. + +Likewise, finite smoothness at each K does not bound physical derivative variances across a family. At fixed L, a fixed positive normalized mass on the axial K-mode pair makes the sixth derivative variance grow as `(omega K)^12`; the inner-cutoff integral gives the stated trace lower bound. Fourth derivative variance already grows as K^8. These examples are outside (U), rather than counterexamples to the theorem with (U). + +For normalized weights and `K/L<=Lambda`, every physical frequency is at most `2pi sqrt(2)Lambda`, so the displayed uniform M12 is valid. No positive lower bound on individual weights is needed for the moment/trace argument. Strict positivity remains a separate requirement for the accepted almost-sure real-parameter critical-count theorem. The proposal correctly refrains from certifying (U) for arbitrary project spectra or moving families without a source-bound check. + +All asserted constants can depend on fixed `Lmin`, M12 (or M0,M12), a, `k_-`, rho and the fixed lifetime-band cap. They are uniform in K,L and coefficient dimension within those bounds. Allowing Lmin to vanish or the lower marks to vanish is not covered. + +## 3. The explicit smooth cutoff and trace calculation check out + +With `b=min(1,Lmin/16)`, one has `b>0`, `b<=1` and `6b=1, exactly the range required for transition derivatives. The reciprocal-denominator bound is at most `exp(2)<8`; differentiating `D times D^-1=1` gives the stated En recurrence. The Leibniz formula then gives Gn. For positive derivative orders theta's support has total length two, giving `d_j=2G_j^2`; order zero instead uses support length six and the actual bound `|theta|<=1`. The different d0 choice is correct. + +Scaling the two one-dimensional derivative integrals gives + +`||partial^gamma chi_b||_2^2 <= b^(2-2|gamma|) d_gamma1 d_gamma2`. + +Since `|gamma|<=6` and b<=1, the exponent is at least -10 and its factor is at most `b^-10`. There is no missing power of b or extra dimension factor. + +The H6 norm is explicitly the unitary Fourier norm with weight `(1+|xi|^2)^6`. Expanding `(1+xi1^2+xi2^2)^6` and applying Plancherel gives the exact multinomial weights w_alpha in (9), not an untracked norm equivalence. + +For a localized derivative the pointwise weighted Cauchy inequality is + +`|sum_gamma c_gamma X_gamma|^2 <= (sum_gamma c_gamma) sum_gamma c_gamma |X_gamma|^2`, + +with `c_gamma=binom(alpha,gamma)` and sum `2^|alpha|`. Using the variance bound for each field derivative and integrating proves (10). Correlation between the terms is allowed. Omitting that weight sum by pretending the derivatives were independent would be an error; the proposal does not do so. My separate maximally correlated summand checks retain this negative control. + +I independently recomputed A_theta by a different aggregation. Summing first over `alpha-gamma`, the remaining multinomial contribution is `(1+2+2)^(6-|gamma|)=5^(6-|gamma|)`. Thus the author's nested sum is exactly + +`sum_(|gamma|<=6) [6!/(gamma1! gamma2! (6-|gamma|)!)] 2^|gamma| 5^(6-|gamma|) d_gamma1 d_gamma2`. + +This gives the same integer `430299201784277045424464881586844996`. My cutoff coefficients were obtained from the closed alternating factorial/binomial formula, and the reciprocal constants from a truncated exponential generating-series inverse; they match all Aj,Ej,Gj,dj values. These are algebraic cross-checks of the displayed constants, not sampled derivative maxima. + +It follows that `T=(1+M12)b^-10 A_theta` uniformly bounds `E||f_g||_H6^2` for every deterministic center g. Its very large magnitude is honestly retained. Nothing in the proof establishes a useful finite numerical window. + +## 4. The finite PSD Gaussian MGF is genuinely dimension-free + +For a fixed deterministic g, the localized field is a finite real linear combination of deterministic compact smooth functions. Its H6 Gram matrix Qg is real symmetric PSD. Under the unconditioned standard coefficient vector, + +`S_g^2=Z^T Qg Z`, `tr Qg=E S_g^2<=T`. + +A finite orthogonal diagonalization preserves the independent standard Gaussian law. Each eigenvalue is nonnegative and at most T, so the one-dimensional Gaussian integrals at exponent `1/(4T)` are finite and give exactly the displayed product. Setting `a_j=lambda_j/(2T)`, all are nonnegative and their sum is at most 1/2. Induction yields + +`product_j(1-a_j)>=1-sum_j a_j>=1/2`. + +Therefore `E exp(S_g^2/(4T))<=sqrt(2)`, with no rank or dimension factor. Zero eigenvalues are harmless. A rank-one matrix with eigenvalue T attains the product bound, so the constant is valid at the extreme trace allocation, not merely for many equal small eigenvalues. My 2,001 rational spectra include unequal allocations, zeros, changing dimension and the rank-one extreme. + +PSD is essential to this deduction: a hypothetical matrix with eigenvalues `3T,-2T` has trace T but its positive direction makes the proposed exponential moment diverge. The actual Qg cannot have this defect because it is a Gram matrix. The tail (14) now follows directly by Markov for every nonnegative v. No infinite-dimensional Gaussian result, covariance-rank floor or unrelated jet regression is imported. + +The centered, unconditioned law matters. The document does not assert that the whole count theorem transfers to arbitrary regressions, and it correctly retains conditional mean/algebraic-exception boundaries. + +## 5. The quantitative H6-to-C4 operator constants are correct + +With the displayed unitary transform in dimension two, inversion contributes `(2pi)^-1`. For any collection of unit derivative slots, its Fourier multiplier has magnitude at most `|xi|^j`. Cauchy–Schwarz against the exact H6 weight therefore controls the full multilinear operator norm, simultaneously in all locations and directions. This does not require the original field's covariance to be rotationally invariant. + +Changing variables `u=r^2`, then `v=u/(1+u)`, gives + +`I_j/pi=integral_0^1 v^j(1-v)^(4-j) dv=j!(4-j)!/5!`, for j<=4. + +In particular I3=pi/20 and I4=pi/5. Combining with the unitary inversion factor gives squared constants `1/(80pi)` and `1/(20pi)`. The rational majorants `p3=1/15`, `p4=2/15` already follow from pi>3. My separate polynomial integrations of the beta integral produce all five displayed values. Absolute convergence also supplies the asserted continuity; for these compact smooth functions classical differentiation is already available. + +Where the cutoff equals one, the derivatives of f_g are exactly the physical derivatives of F. The operator estimate controls arbitrary rotated eigenframes directly; no hidden coordinate-partial inflation or extra K/L factor appears. Orders through six in the trace are sufficient for this C4 estimate, so no unmentioned uniform higher spectral moment is needed. + +## 6. The deterministic physical cover supplies the required domains + +For each L the grid uses `n=ceil(L/b)`, giving spacing at most b and nearest-center coordinate distance at most b/2 in short torus lifts. Since b<=1, + +`n<=(L+b)/b<=(L+1)/b`. + +This proves the stated `n^2<=(1+L)^2/b^2` bound. The grid depends on deterministic family parameters, not on the field draw. + +A rotated radius-b square about its assigned point M lies within coordinate distance `b/2+sqrt(2)b<2b` of g. Thus the entire closed chart, not just its anchor or pair endpoints, lies in the plateau of chi_b. Its support side 6b is strictly less than L, so the translated support embeds even across a canonical torus seam. The chart square also lies below torus injectivity. The first pair segment, of length at most `rho eta<=rho b/4`, stays well inside the plateau. The center and endpoint short lifts can be chosen consistently at these scales. + +My separate controls test rational rotations and a seam where subtracting canonical `[0,L)` coordinates would fail. The underlying all-orientation margin is the analytic bound `|cos|+|sin|<=sqrt(2)<3/2`, not inference from a sampled set of angles. + +For an eta-cluster, the displacement between its maxima is at most `(1+rho)eta<=b/2`. If maxima coincide, both distinct saddle endpoints remain within rho eta of that anchor. These are precisely the displacements needed by the two incidence cases. No bound relative to the actual center separation is introduced. + +## 7. Localization preserves the inherited isolation hypotheses + +For the covering center of the first maximum, set `S=||chi_b F(g+.)||_H6`, `W=1+S`. The local operator bounds `B=p3 W`, `C=p4 W` hold on the whole radius-b chart and on the pair segment by the preceding two sections. This domain check is the substantive new bridge to the accepted deterministic chart. + +Supposing `W^2=kappa/W^2>(1+rho)eta`. Different maxima contradict maximum uniqueness; different candidates sharing the same maximum give three distinct critical points, contradicting the chart count. A shared saddle is included in the different-maxima case. Cross-role equality is incompatible with the physical Hessian types. The identical candidate is excluded. This proves (19) and preserves the older height-barrier/mixed-cubic falsifiers: no enclosing superlevel region or elder selection is inferred. + +The first maximum and a suitable center can depend on the realized field. That does not justify taking an unconditioned norm law at a random maximum; the proposal does not do so. It only establishes that the event implies a large norm at at least one member of the predetermined finite grid. Ties in choosing a nearest center need no measurable tie-break for this existence-to-finite-union inclusion. + +## 8. The unconditional probability and integrated-count estimates follow + +For `eta<=min(eta0,gamma/4)`, the large-W condition gives `S_g^2>=gamma/(4eta)` at some deterministic grid center. Applying (14) separately to every center and taking a union bound gives + +`P(cluster)<=sqrt(2)b^-2(1+L)^2 exp[-gamma/(16T eta)]`. + +No independence of cubes is assumed. Every probability is under the same unconditioned field law, so no field-selected covariance, conditional mean or Palm/Rice weight is hidden in the argument. + +The accepted fixed-degree critical-count result applies almost surely to each individual positive full-square law. No uniform lower spectral weight is required to say a proper polynomial zero set has measure zero under that law. The exceptional set may depend on the family member; an expectation inequality for each member with common deterministic constants does not require a common probability-one event over an uncountable family. This is sufficient for the asserted limit of expectations and does not provide a joint increasing-domain coupling. + +Endpoint identity and physical types are preserved: candidates have no reflected-frame duplicates, and the ordered factorial sum excludes `p=q` while including shared endpoints. On the source-specific full-probability set, `D<=(16K^2)^4 1{D>0}`. The completed-space finite-configuration measurability argument is consumed at its accepted scope. Borel selectors bounded by one only reduce the count. + +Consequently the explicit bound is + +`E D(t,eta)<=Ccov 16^4 K^8 (1+L)^2 exp(-c/eta)`, + +with `Ccov=sqrt(2)b^-2` and `c=gamma/(16T)>0`. These constants are independent of K,L,N under the fixed family/mark assumptions. The remaining K^8 is openly retained and handled by the new cutoff. There is no second Hessian determinant weight or spatial-area normalization inserted into this integrated endpoint count. + +## 9. The K/L/t cancellation and radius condition are correctly separated + +For `H(t)=8log K+2log(1+L)+4log(1/t)` and `eta=c/H`, the equality + +`exp(-c/eta)=K^-8 (1+L)^-2 t^4` + +is exact. Multiplying by the coarse count and cover powers leaves `Ccov 16^4 t^4`. The requested normalization therefore gives `O(t^(8/3))`. + +H is positive since K>=2, L>0 and 0=4log(1/t)` forces eta below the fixed eta_cap eventually. With fixed `Tgap,k_-`, every lifetime-band radius is at most `A t^(1/3)`, `A=(Tgap/k_-)^(1/3)`. The additional condition `t^(1/3)H(t)->0` then ensures `A t^(1/3)<=rho c/H(t)` eventually. Only at this point may the radius-restricted D be said to include every near-center candidate pair in the stated lifetime band. Without that condition, the bound for D with its explicit radius restriction remains true, but full-band inclusion is not established. + +Polynomial growth of K,L implies `H=O(log(1/t))`, so it satisfies the condition. In the example `L=t^-1/3`, `K=ceil L`, one has K>=2 and L>=Lmin eventually, K/L<=2, and `H=(22/3)log(1/t)+O(1)`. Thus (U) holds for arbitrary positive symmetric normalized weights and the lifetime-radius ratio tends to zero. This is a valid example of the new diagonal estimate, without establishing a constant-total-intensity regime. + +The generalized coefficients p>=8, q>=2 and u>4/3 also work with their corresponding H: the residual K and `(1+L)` powers are nonpositive, and the normalized t exponent is positive. The corresponding radius condition must still be imposed. + +The fast-growth falsifier is correctly distinguished from moment failure. At fixed L, normalized positive weights proportional to `exp(-|n|^2)` on each finite square give a uniform physical M12, because their normalizer is bounded below by the constant-mode contribution and the full lattice moment sum converges. Hence (U) can coexist with `log K` of order `t^-1/2`. The selected eta is then at most of order `t^1/2`, and the generic lifetime radius upper bound `t^1/3` is not o(eta). This does not disprove the radius-restricted estimate or assert that every actual pair attains the upper bound; it prevents claiming automatic inclusion from that bound. + +If Tgap grows, the condition must contain `Tgap(t)^(1/3)`. The example `Tgap=1/t` makes the radius upper bound constant, so it cannot be automatically included in a cutoff tending to zero. Moving upper birth/cubic marks do not harm this proof because only fixed positive lower marks and the lifetime cap are used. Vanishing lower curvature/cubic marks remove the fixed positive gamma and are outside the claim. + +Optional division by L^2 or L^4 is bounded above by an Lmin-dependent constant and would change the prefactor, not the t exponent. The displayed count is total and contains no such division. Its cumulative normalization proves neither the one-point `t^(2/3)` law nor density `ell^(-1/3)`. + +## 10. Disposition and retained limits + +There are **no blocking findings** within the frozen scope. The following boundaries remain necessary: + +- Uniform total spectral mass and physical M12, with fixed positive Lmin, are real family assumptions. Finite smoothness, a cutoff degree alone, or the word Gaussian does not verify them for a source sequence. +- Gaussianity, centeredness and deterministic spectral parameters underlie the Gram MGF. The proof does not cover non-Gaussian coefficients or adaptive conditioning of the whole count theorem. +- Full positive square support is used by the inherited almost-sure count theorem; it is not needed by the local trace estimate alone. No infinite-carrier polynomial count cap is supplied. +- The cover and operator estimates control the actual chart and segment. They do not license a pointwise derivative bound without a surrounding norm domain, or a Gaussian law conditioned at a random maximum. +- Ordered endpoint identity, short displacements, positive a and `k_-`, the distinct-candidate exclusion and the radius cutoff remain part of D. A full square count including the identical-pair diagonal has a different scale. +- Growing K,L require the explicit radius condition to claim the whole lifetime sector. Growing Tgap changes that condition. The very loose constants do not certify a useful experimental window. +- A uniform integrated diagonal estimate does not supply moving far-center bounds, one-point intensity/remainders, global H0 bar correspondence, essential classes, higher homology, mixing, increasing-domain tightness, an infinite-field statement, or a limiting point process. Existing K2/K3 enlarged height-covariance singularities remain unchanged. + +My 2,132 exact finite cases comprise eight independent cutoff/reindexed-trace checks, 28 H6 multinomial coefficients, seven beta-integral/operator checks, 2,001 unequal/rank-deficient PSD product cases, 28 maximally correlated Leibniz cases, 42 lower-moment inequalities, 12 cover/rotation parameter cases, one periodic seam and five growth/normalization identities. Negative controls retain the failures from dropping PSD, dropping the correlation-safe weight sum, dropping mass or uniform M12, excessive log-degree growth, and a growing lifetime cap. No fields, numerical Sobolev integrals, approximate critical roots, or process experiments were sampled. + +All new author files, previous author/review files, pinned source and saved primary source retain their identities. The new manifest and integrity receipt bind this standalone review and its own artifacts. No production/native/remote claim, source adoption, Lean execution, kernel audit, human/provider-distinct approval or scientific promotion occurred. Actual separate OpenAI/Codex nonauthor, source-exposed, organizational-independence credit 0. diff --git a/experiments/universality/fold_isolation_diagonal/NONAUTHOR_UNIFORM_LOCAL_NORM_REVIEW_MANIFEST.json b/experiments/universality/fold_isolation_diagonal/NONAUTHOR_UNIFORM_LOCAL_NORM_REVIEW_MANIFEST.json new file mode 100644 index 000000000..dd2bf28f6 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/NONAUTHOR_UNIFORM_LOCAL_NORM_REVIEW_MANIFEST.json @@ -0,0 +1,134 @@ +{ + "state": "FROZEN_LOCAL_NONAUTHOR_REVIEW_ORG0_NO_SCIENCE_PROMOTION", + "created_at_utc": "2026-10-16T02:18:44.729937+00:00", + "verdict": "ACCEPT_SCOPED", + "reviewed_inputs_unchanged": [ + { + "path": "UNIFORM_LOCAL_NORM_DIAGONAL_PROPOSAL.md", + "bytes": 26687, + "sha256": "8e1eefca9888d43824a9fd9caca34c44d2de4bf8e30c668b474fa2097491c76b" + }, + { + "path": "uniform_local_norm_controls.py", + "bytes": 6486, + "sha256": "4b84755b771598ed34193c075718a41da447719f90a2ff5c6a17d48d0abc3dc4" + }, + { + "path": "UNIFORM_LOCAL_NORM_CONTROLS.json", + "bytes": 1790, + "sha256": "5d6cdb2d4925c6ba0b1b545d7a5aca1977655c1d18aa1ea29a82273f30845ece" + }, + { + "path": "QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md", + "bytes": 19138, + "sha256": "827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30" + }, + { + "path": "fold_isolation_controls.py", + "bytes": 6308, + "sha256": "13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5" + }, + { + "path": "FOLD_ISOLATION_CONTROLS.json", + "bytes": 795, + "sha256": "ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf" + }, + { + "path": "FINITE_FOURIER_CRITICAL_COUNT_DIAGONAL_ROUTE.md", + "bytes": 25699, + "sha256": "621cee96854aa440498cecf5452a8d1e23f90a4458779c3286c18eba106fe138" + }, + { + "path": "finite_fourier_count_controls.py", + "bytes": 5153, + "sha256": "fd6f29739280a55017945526cb75fb46ebc80d86c012ff4a1ff65baa404c31de" + }, + { + "path": "FINITE_FOURIER_COUNT_CONTROLS.json", + "bytes": 849, + "sha256": "bbe147fdf409a3e37b450d4cbcd5d2a927140f95bd3ce0f36cafe16861123bfd" + }, + { + "path": "NONAUTHOR_FOLD_ISOLATION_REVIEW.md", + "bytes": 18409, + "sha256": "f06c61e3f3e60c795a1a14e824640f1689d4143dba4f6698cada7a41929eec90" + }, + { + "path": "NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW.md", + "bytes": 21880, + "sha256": "16152427f7013822ac1bba81142429800334e6bf0f354a6439b23a1997bc4524" + }, + { + "path": "uniform-local-norm-primary-sources/SOURCE_IDENTITY.json", + "bytes": 443, + "sha256": "92bf5297b309e947f42ee1dc837487775e7341bc7c1051e493899963ff85891a" + }, + { + "path": "uniform-local-norm-primary-sources/simon-lecs-on-pde.pdf", + "bytes": 940975, + "sha256": "e1f1b2f51d2558aa467a8c064fe82e7b9e157ad28c9c75f01e38596cde656870" + }, + { + "path": "experiments/universality/finite_gaussian_contact/PROOF.md", + "bytes": 25367, + "sha256": "7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474" + } + ], + "files": [ + { + "path": "NONAUTHOR_UNIFORM_LOCAL_NORM_REVIEW.md", + "bytes": 23887, + "sha256": "ff552dfacadede53405530a219302c1163dca8ba97d518f02d407970d3de3c1c" + }, + { + "path": "nonauthor-uniform-local-norm-checks/CHECK_RUN_SUMMARY.json", + "bytes": 7279, + "sha256": "480e1e4786933f2201a6dadb93ee720dc21b632c5a749a91ea9323f1fc506364" + }, + { + "path": "nonauthor-uniform-local-norm-checks/PRIMARY_IMPORT_AUDIT.json", + "bytes": 1419, + "sha256": "98c521f7c773e1338b2ed6bac74ebba0c3f146cdc0b989abc7dd67390b07239b" + }, + { + "path": "nonauthor-uniform-local-norm-checks/author-controls.meta.json", + "bytes": 715, + "sha256": "0e8fd7097e06de071dd233b8f6a8ff7623d370e77d78421d7b11a88b9deffac8" + }, + { + "path": "nonauthor-uniform-local-norm-checks/author-controls.stderr", + "bytes": 0, + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "path": "nonauthor-uniform-local-norm-checks/author-controls.stdout", + "bytes": 1790, + "sha256": "5d6cdb2d4925c6ba0b1b545d7a5aca1977655c1d18aa1ea29a82273f30845ece" + }, + { + "path": "nonauthor-uniform-local-norm-checks/independent-controls.meta.json", + "bytes": 738, + "sha256": "c88b17ff52d1eb833679da97a4cb93a2f6068345c785b8e24c8a989ac2c55a34" + }, + { + "path": "nonauthor-uniform-local-norm-checks/independent-controls.stderr", + "bytes": 0, + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "path": "nonauthor-uniform-local-norm-checks/independent-controls.stdout", + "bytes": 1439, + "sha256": "bd553b02c66ad4cf761700f53043ac4b03f7371b5e15aa03ad08b4d4feef2fce" + }, + { + "path": "nonauthor-uniform-local-norm-checks/review_checks.py", + "bytes": 8199, + "sha256": "d764331f77125a27f7f6a6809a813e5b4a2cacc65f66d0dc889300347074f8c5" + }, + { + "path": "nonauthor-uniform-local-norm-checks/run_checks.py", + "bytes": 4309, + "sha256": "98f86202f1ac2a15a18058a7b8cd9a49bbea180f9ab416536b224d6967f639af" + } + ] +} diff --git a/experiments/universality/fold_isolation_diagonal/PUBLICATION_MANIFEST.json b/experiments/universality/fold_isolation_diagonal/PUBLICATION_MANIFEST.json new file mode 100644 index 000000000..6bcfdfee1 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/PUBLICATION_MANIFEST.json @@ -0,0 +1,813 @@ +{ + "actual_publisher": "OpenAI/Codex Root01a11cbf", + "canonical_pickup": "https://github.com/d6g8k5htny-coder/main/issues/419#issuecomment-6206833950", + "external_reference_only": [ + { + "bytes": 83539, + "consumed_scope": "B\u00e9zout isolated projective intersection bound, with the full-degree witness and exceptional-set argument supplied in the proof note.", + "disposition": "EXTERNAL_REFERENCE_ONLY_NOT_REPUBLISHED", + "path": "critical-count-primary-sources/guth-lect13.pdf", + "review_manifest_reference_preserved": [ + "NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW_MANIFEST.json:reviewed_inputs_unchanged" + ], + "sha256": "92ed0fc0f940179871a25f3a244eb181aaa0176ee5738ac1f8f68a8bacd072a9", + "source_url": "https://math.mit.edu/~lguth/PolyMethod/lect13.pdf" + }, + { + "bytes": 14844, + "consumed_scope": "Proper morphisms have the named closed-image/composition properties used in the exceptional-locus argument.", + "disposition": "EXTERNAL_REFERENCE_ONLY_NOT_REPUBLISHED", + "path": "critical-count-primary-sources/stacks-01W6.html", + "review_manifest_reference_preserved": [ + "NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW_MANIFEST.json:reviewed_inputs_unchanged" + ], + "sha256": "060b8c179f815d59977f8a31057b5f5f9e1f2104351d46c33f1fb6d9ed95a202", + "source_url": "https://stacks.math.columbia.edu/tag/01W6" + }, + { + "bytes": 15419, + "consumed_scope": "Projective morphisms are proper; used only for the projective multiplication image.", + "disposition": "EXTERNAL_REFERENCE_ONLY_NOT_REPUBLISHED", + "path": "critical-count-primary-sources/stacks-01WC.html", + "review_manifest_reference_preserved": [ + "NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW_MANIFEST.json:reviewed_inputs_unchanged" + ], + "sha256": "08a39f68a8159e34e436df9c4b8f5e8886f58b3c0d5163c15610e12aef302737", + "source_url": "https://stacks.math.columbia.edu/tag/01WC" + }, + { + "bytes": 325166, + "consumed_scope": "Theorem3.1 on PDF7; distance-power import for a real analytic determinant on compact angular charts.", + "disposition": "EXTERNAL_REFERENCE_ONLY_NOT_REPUBLISHED", + "path": "jet-factorization-primary-sources/teissier-lojasiewicz-2012.pdf", + "review_manifest_reference_preserved": [ + "NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW_MANIFEST.json:inputs" + ], + "sha256": "cb88d0410e5c0a629f85592864809b987612663593a04dd8ea2928502cac6873", + "source_url": "https://webusers.imj-prg.fr/~bernard.teissier/documents/Lojasiewicz_2012.pdf" + }, + { + "bytes": 586327, + "consumed_scope": "Theorem1 and Corollary4 only, making the finite positive analytic distance exponent explicit.", + "disposition": "EXTERNAL_REFERENCE_ONLY_NOT_REPUBLISHED", + "path": "nonauthor-jet-factorization-review/feehan-1708.09775v6.html", + "review_manifest_reference_preserved": [ + "NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW_MANIFEST.json:outputs" + ], + "sha256": "b11ceb8df55335260ed8eaef4a71406cfa592565869ced020f757354124913c1", + "source_url": "https://arxiv.org/html/1708.09775v6" + }, + { + "bytes": 230248, + "consumed_scope": "Rendered visual inspection of Teissier Theorem3.1; derivative of the externally referenced PDF.", + "disposition": "EXTERNAL_REFERENCE_ONLY_NOT_REPUBLISHED", + "path": "nonauthor-jet-factorization-review/tmp/pdfs/teissier-page7.png", + "review_manifest_reference_preserved": [ + "NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW_MANIFEST.json:outputs" + ], + "sha256": "e95528c20ec62ce8ea3a14f5d6672ea52d891fca7bc2e75fbd6cc44076851b29", + "source_url": "https://webusers.imj-prg.fr/~bernard.teissier/documents/Lojasiewicz_2012.pdf" + }, + { + "bytes": 940975, + "consumed_scope": "Lecture5 Fourier inversion, differentiation, Parseval/Plancherel on PDF26 and displayed Sobolev Cauchy-Schwarz argument on PDF27.", + "disposition": "EXTERNAL_REFERENCE_ONLY_NOT_REPUBLISHED", + "path": "uniform-local-norm-primary-sources/simon-lecs-on-pde.pdf", + "review_manifest_reference_preserved": [ + "NONAUTHOR_UNIFORM_LOCAL_NORM_REVIEW_MANIFEST.json:reviewed_inputs_unchanged", + "NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW_MANIFEST.json:inputs" + ], + "sha256": "e1f1b2f51d2558aa467a8c064fe82e7b9e157ad28c9c75f01e38596cde656870", + "source_url": "https://math.stanford.edu/~lms/lecs-on-pde.pdf" + } + ], + "files": [ + { + "bytes": 5592, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "AUTHOR_ALIGNMENT_ADDENDUM.md", + "role": "exact_historical_context_no_acceptance_extension", + "sha256": "73e3117e739a072ce0238b45e44f01389ad18c4127e1a1b6f0c818ae5cde4a30" + }, + { + "bytes": 25314, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "COALESCING_CONTACT_CALCULATION.md", + "role": "exact_historical_context_no_acceptance_extension", + "sha256": "39a63e1b00393b49e7b6d049c0c9ea7be82436352eed64014155a63aa8c7fb36" + }, + { + "bytes": 23286, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "COUNTERFAMILY_NONAUTHOR_REVIEW.md", + "role": "frozen_review_or_evidence", + "sha256": "26b1b1a9a1e978edd036ee2d1e244e8b8b2089d3c6a48edb94d78d37577cc2c5" + }, + { + "bytes": 4458, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "COUNTERFAMILY_NONAUTHOR_REVIEW_MANIFEST.json", + "role": "frozen_review_or_evidence", + "sha256": "614f84154b75505c0004f70ca2e81d974c9dec49a9f5c64c422b63a787d82a56" + }, + { + "bytes": 849, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "FINITE_FOURIER_COUNT_CONTROLS.json", + "role": "frozen_author_control_output", + "sha256": "bbe147fdf409a3e37b450d4cbcd5d2a927140f95bd3ce0f36cafe16861123bfd" + }, + { + "bytes": 25699, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "FINITE_FOURIER_CRITICAL_COUNT_DIAGONAL_ROUTE.md", + "role": "reviewed_proof", + "sha256": "621cee96854aa440498cecf5452a8d1e23f90a4458779c3286c18eba106fe138" + }, + { + "bytes": 795, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "FOLD_ISOLATION_CONTROLS.json", + "role": "frozen_author_control_output", + "sha256": "ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf" + }, + { + "bytes": 1618, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "JET_FACTORIZATION_OBSTRUCTION_CONTROLS.json", + "role": "frozen_author_control_output", + "sha256": "1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764" + }, + { + "bytes": 34627, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "JET_NONDEGENERACY_FACTORIZATION_OBSTRUCTION.md", + "role": "reviewed_proof", + "sha256": "0dfdddfbcac47b2fd07449f4152b7f5d935f92aed15272fb8333ed9cc60b5766" + }, + { + "bytes": 14415, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "NONAUTHOR_COALESCENCE_REVIEW.md", + "role": "exact_historical_context_no_acceptance_extension", + "sha256": "2f766d8aa5541602ace7bbc70fd360e3a342d2cd258c56c44a35503e3e8624d5" + }, + { + "bytes": 21880, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW.md", + "role": "frozen_review_or_evidence", + "sha256": "16152427f7013822ac1bba81142429800334e6bf0f354a6439b23a1997bc4524" + }, + { + "bytes": 4392, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW_MANIFEST.json", + "role": "frozen_review_or_evidence", + "sha256": "b4cd749e3d552f9d943d6116ac77288d971c616cbd022f881049bb42ffa5e7d0" + }, + { + "bytes": 18409, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "NONAUTHOR_FOLD_ISOLATION_REVIEW.md", + "role": "frozen_review_or_evidence", + "sha256": "f06c61e3f3e60c795a1a14e824640f1689d4143dba4f6698cada7a41929eec90" + }, + { + "bytes": 2834, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "NONAUTHOR_FOLD_ISOLATION_REVIEW_MANIFEST.json", + "role": "frozen_review_or_evidence", + "sha256": "3beb504fbd6bb216cfc8ba50c420a20e0782b5e77f11a1ca62ac7eea77c8fe05" + }, + { + "bytes": 6389, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW.json", + "role": "frozen_review_or_evidence", + "sha256": "13395e24282bbf7d3005b7a236c7de52f703f2d0e0878fe350dfcc6df8df91a5" + }, + { + "bytes": 25600, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW.md", + "role": "frozen_review_or_evidence", + "sha256": "531b9d9a16831383f2f217e61bad7dea84aca941a934e4bcaa771bedcc4e73af" + }, + { + "bytes": 10638, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW_MANIFEST.json", + "role": "frozen_review_or_evidence", + "sha256": "4f455ebb8d1aea06c967fac1653a499dc7c3446580e7f23939f07da5022431e2" + }, + { + "bytes": 23887, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "NONAUTHOR_UNIFORM_LOCAL_NORM_REVIEW.md", + "role": "frozen_review_or_evidence", + "sha256": "ff552dfacadede53405530a219302c1163dca8ba97d518f02d407970d3de3c1c" + }, + { + "bytes": 4763, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "NONAUTHOR_UNIFORM_LOCAL_NORM_REVIEW_MANIFEST.json", + "role": "frozen_review_or_evidence", + "sha256": "1ccd38374adcf4931b87abeacdefaa64990080a9b3b98835a7c140bde2f505a7" + }, + { + "bytes": 19138, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md", + "role": "reviewed_proof", + "sha256": "827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30" + }, + { + "bytes": 7200, + "disposition": "NEW_PUBLICATION_GLUE", + "path": "README.md", + "role": "new_publication_scope_and_review_crosswalk", + "sha256": "d0f2dc0f7d023a56bf0b47db514f6a8e29b3564854574857beaba24e1d3bd55d" + }, + { + "bytes": 8213, + "disposition": "NEW_PUBLICATION_GLUE", + "path": "SOURCE_BOUNDARY.md", + "role": "new_explicit_publication_boundary", + "sha256": "e11d1ae4960770ba316c7e12dc5a2764a66a7dd4f8f8446dc1fdb7d9a0a3ba78" + }, + { + "bytes": 1790, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "UNIFORM_LOCAL_NORM_CONTROLS.json", + "role": "frozen_author_control_output", + "sha256": "5d6cdb2d4925c6ba0b1b545d7a5aca1977655c1d18aa1ea29a82273f30845ece" + }, + { + "bytes": 26687, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "UNIFORM_LOCAL_NORM_DIAGONAL_PROPOSAL.md", + "role": "reviewed_proof", + "sha256": "8e1eefca9888d43824a9fd9caca34c44d2de4bf8e30c668b474fa2097491c76b" + }, + { + "bytes": 0, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "counterfamily-nonauthor-checks-20261016/author_controls.stderr", + "role": "frozen_review_or_evidence", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 1618, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "counterfamily-nonauthor-checks-20261016/author_controls.stdout.json", + "role": "frozen_review_or_evidence", + "sha256": "1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764" + }, + { + "bytes": 8080, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "counterfamily-nonauthor-checks-20261016/counterfamily_checks.py", + "role": "frozen_review_or_evidence", + "sha256": "0718a7f0e415f75110b5b08b5850fdc5e76c601e515ee64bb0770ceaa21834e9" + }, + { + "bytes": 0, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "counterfamily-nonauthor-checks-20261016/independent_controls.stderr", + "role": "frozen_review_or_evidence", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 12359, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "counterfamily-nonauthor-checks-20261016/independent_controls.stdout.json", + "role": "frozen_review_or_evidence", + "sha256": "3c53ed54756c911292180c5686750de9a47cf4a54c56cc0d2616155c3f5af35b" + }, + { + "bytes": 796, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "critical-count-primary-sources/SOURCE_IDENTITIES.json", + "role": "external_import_identity_receipt", + "sha256": "2c42d92a4732f09a58aebd9b9ef8672b51b672ab99f3bbdad61681a0dcf4f504" + }, + { + "bytes": 29190, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "experiments/universality/finite_candidate_lifetime/PROOF.md", + "role": "pinned_source_mirror_no_new_acceptance", + "sha256": "ed31f562f0ffaaf6d4bd2ceb1ae3d234798e983a55d55c6983b128fce27133f5" + }, + { + "bytes": 25367, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "experiments/universality/finite_gaussian_contact/PROOF.md", + "role": "pinned_source_mirror_no_new_acceptance", + "sha256": "7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474" + }, + { + "bytes": 30935, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "experiments/universality/iid_short_bar_process/PROOF.md", + "role": "pinned_source_mirror_no_new_acceptance", + "sha256": "320bc988c468bea4938bb777a2d05952c6ab8e0a384f33725b50de847aeec1eb" + }, + { + "bytes": 5153, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "finite_fourier_count_controls.py", + "role": "frozen_author_finite_control", + "sha256": "fd6f29739280a55017945526cb75fb46ebc80d86c012ff4a1ff65baa404c31de" + }, + { + "bytes": 6308, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "fold_isolation_controls.py", + "role": "frozen_author_finite_control", + "sha256": "13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5" + }, + { + "bytes": 427, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "jet-factorization-primary-sources/SOURCE_IDENTITY.json", + "role": "external_import_identity_receipt", + "sha256": "683fb8d29a62c9e73f01a93bf3b288c1e7189f050da4b77eb831bb063ac88492" + }, + { + "bytes": 5818, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "jet_factorization_obstruction_controls.py", + "role": "frozen_author_finite_control", + "sha256": "5b1e8fc18a6e32e3a312a253e3e4d1110bb674385f3d49447a45e78ad4df541f" + }, + { + "bytes": 815, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-finite-count-checks/FINAL_REVIEW_INTEGRITY.json", + "role": "frozen_review_or_evidence", + "sha256": "ac591f6c84f086c1e4717e553a84ed11cbb0e64bafe42783ba0d180878b7db92" + }, + { + "bytes": 2345, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-finite-count-checks/PRIMARY_IMPORT_AUDIT.json", + "role": "frozen_review_or_evidence", + "sha256": "a1faa2d1e04cae111b146f5158d423121434ed9216ffeccad6455f0d7fdb1d4f" + }, + { + "bytes": 0, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-finite-count-checks/author-controls.stderr", + "role": "frozen_review_or_evidence", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 849, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-finite-count-checks/author-controls.stdout", + "role": "frozen_review_or_evidence", + "sha256": "bbe147fdf409a3e37b450d4cbcd5d2a927140f95bd3ce0f36cafe16861123bfd" + }, + { + "bytes": 0, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-finite-count-checks/independent-controls.stderr", + "role": "frozen_review_or_evidence", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 978, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-finite-count-checks/independent-controls.stdout", + "role": "frozen_review_or_evidence", + "sha256": "c2fd4842cc1bc9ba9e7a947d8630a256386064bd63e94b724ffa699258ae7046" + }, + { + "bytes": 6984, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-finite-count-checks/review_checks.py", + "role": "frozen_review_or_evidence", + "sha256": "e79983d895b90c4d9f4c9a89bc64471e62f7bd79acded7102b53674337e53302" + }, + { + "bytes": 3905, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-finite-count-checks/run_checks.py", + "role": "historical_capture_wrapper_not_turnkey_without_external_downloads", + "sha256": "6014f124e595a81f726650b7bd3fa55f1eba28ca8af41fc5071ef7f8fa4836b6" + }, + { + "bytes": 724, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-fold-isolation-checks/FINAL_REVIEW_INTEGRITY.json", + "role": "frozen_review_or_evidence", + "sha256": "3a26736035434fb95a0c8fea31d640e6cc43ee56fa5daeb03274da7b75532cd3" + }, + { + "bytes": 0, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-fold-isolation-checks/author-controls.stderr", + "role": "frozen_review_or_evidence", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 795, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-fold-isolation-checks/author-controls.stdout", + "role": "frozen_review_or_evidence", + "sha256": "ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf" + }, + { + "bytes": 13955, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-fold-isolation-checks/gaussian-crosscheck/GAUSSIAN_OPERATOR_TAIL_CROSSCHECK_20261016T014803Z.md", + "role": "frozen_review_or_evidence", + "sha256": "c06161e9a8b697b2cb6fd0c6ff597c2dcdff91f653d074c3e47564cada6f2f67" + }, + { + "bytes": 0, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-fold-isolation-checks/independent-controls.stderr", + "role": "frozen_review_or_evidence", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 777, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-fold-isolation-checks/independent-controls.stdout", + "role": "frozen_review_or_evidence", + "sha256": "4b1eb60b2f44d9e67caf3469dbba06d90fa29d7d32a74cb6fe46488b988bd8ec" + }, + { + "bytes": 6772, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-fold-isolation-checks/review_checks.py", + "role": "frozen_review_or_evidence", + "sha256": "0c9155f38b4aa7b9baae0835b8f9c6681f5cfe1de890d84178605bafbdea6eec" + }, + { + "bytes": 2831, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-fold-isolation-checks/run_checks.py", + "role": "historical_capture_wrapper_not_turnkey_without_external_downloads", + "sha256": "eabae610abb5c0a658d2705144e30dab6ed9cbd14972bfc220540ca1d18a9ba0" + }, + { + "bytes": 8076, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-jet-factorization-review/FROZEN_INPUT_INTEGRITY.json", + "role": "frozen_review_or_evidence", + "sha256": "7c724d8a3227b41de638a6289d103361fe8bc7c3db5eb3d45ac702159c449989" + }, + { + "bytes": 4385, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-jet-factorization-review/INDEPENDENT_KERNEL_CONTROLS.json", + "role": "frozen_review_or_evidence", + "sha256": "1e94c994a8111bd1b741d83e8404a242109a69ac5331cd07eddf8b0930a46330" + }, + { + "bytes": 821, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-jet-factorization-review/PRIMARY_IMPORT_RECEIPT.json", + "role": "external_import_identity_receipt", + "sha256": "3a52a5cd5a098b551bacb0c2dbad7c48030eaa613d36fae110e29daa4410265b" + }, + { + "bytes": 1265, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-jet-factorization-review/PROCESS_RUN_RECEIPT.json", + "role": "frozen_review_or_evidence", + "sha256": "0bfc877da8ead5d680745aeba4889baabd131d5eddea0b41efa8b8700f42b049" + }, + { + "bytes": 0, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-jet-factorization-review/author_stderr.txt", + "role": "frozen_review_or_evidence", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 1618, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-jet-factorization-review/author_stdout.json", + "role": "frozen_review_or_evidence", + "sha256": "1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764" + }, + { + "bytes": 11734, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-jet-factorization-review/independent_kernel_controls.py", + "role": "frozen_review_or_evidence", + "sha256": "132f8fa628df68082aaf3c42e727584540e00f293ce6b0223f6cf74dae5a3924" + }, + { + "bytes": 0, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-jet-factorization-review/independent_stderr.txt", + "role": "frozen_review_or_evidence", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 4385, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-jet-factorization-review/independent_stdout.json", + "role": "frozen_review_or_evidence", + "sha256": "1e94c994a8111bd1b741d83e8404a242109a69ac5331cd07eddf8b0930a46330" + }, + { + "bytes": 929, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-uniform-local-norm-checks/FINAL_REVIEW_INTEGRITY.json", + "role": "frozen_review_or_evidence", + "sha256": "7ed09294cdaa061fda89253bbb1ae1cec6d1c5d8eb5fde75a024b7afb12232d3" + }, + { + "bytes": 1419, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-uniform-local-norm-checks/PRIMARY_IMPORT_AUDIT.json", + "role": "frozen_review_or_evidence", + "sha256": "98c521f7c773e1338b2ed6bac74ebba0c3f146cdc0b989abc7dd67390b07239b" + }, + { + "bytes": 0, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-uniform-local-norm-checks/author-controls.stderr", + "role": "frozen_review_or_evidence", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 1790, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-uniform-local-norm-checks/author-controls.stdout", + "role": "frozen_review_or_evidence", + "sha256": "5d6cdb2d4925c6ba0b1b545d7a5aca1977655c1d18aa1ea29a82273f30845ece" + }, + { + "bytes": 0, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-uniform-local-norm-checks/independent-controls.stderr", + "role": "frozen_review_or_evidence", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 1439, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-uniform-local-norm-checks/independent-controls.stdout", + "role": "frozen_review_or_evidence", + "sha256": "bd553b02c66ad4cf761700f53043ac4b03f7371b5e15aa03ad08b4d4feef2fce" + }, + { + "bytes": 8199, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-uniform-local-norm-checks/review_checks.py", + "role": "frozen_review_or_evidence", + "sha256": "d764331f77125a27f7f6a6809a813e5b4a2cacc65f66d0dc889300347074f8c5" + }, + { + "bytes": 4309, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "nonauthor-uniform-local-norm-checks/run_checks.py", + "role": "historical_capture_wrapper_not_turnkey_without_external_downloads", + "sha256": "98f86202f1ac2a15a18058a7b8cd9a49bbea180f9ab416536b224d6967f639af" + }, + { + "bytes": 3129, + "disposition": "DERIVATIVE_NOT_ORIGINAL", + "original_relative_path": "counterfamily-nonauthor-checks-20261016/EXECUTION_RECORD.json", + "original_sha256": "8ae808eed6b7f8fec67435a3b97c8422a08212e7f18ababbced3b367380e04fb", + "path": "publication-projections/counterfamily-nonauthor-checks-20261016/EXECUTION_RECORD.json", + "role": "labeled_publication_projection", + "sha256": "51332efafef10ac12574475ef20844cd0cc4dc2654002f14f89d20c7e2952cb1" + }, + { + "bytes": 7179, + "disposition": "DERIVATIVE_NOT_ORIGINAL", + "original_relative_path": "nonauthor-finite-count-checks/CHECK_RUN_SUMMARY.json", + "original_sha256": "5faeb21e56f042394de6772e54081fd16eded69b42abd89109773d5193898132", + "path": "publication-projections/nonauthor-finite-count-checks/CHECK_RUN_SUMMARY.json", + "role": "labeled_publication_projection", + "sha256": "ae4f4a6fdb7da00737d17b9bc97f5ab085306d1beb5a2544f5165ad969023312" + }, + { + "bytes": 1211, + "disposition": "DERIVATIVE_NOT_ORIGINAL", + "original_relative_path": "nonauthor-finite-count-checks/author-controls.meta.json", + "original_sha256": "07fa0d98d0f407f912444426ffdacaf57f5b0f51928f52be197937ef00bd1463", + "path": "publication-projections/nonauthor-finite-count-checks/author-controls.meta.json", + "role": "labeled_publication_projection", + "sha256": "0107aeb802b10a33ab7349102dd7d26911dcd89877e6e3e6033a744ffcf28f4f" + }, + { + "bytes": 1240, + "disposition": "DERIVATIVE_NOT_ORIGINAL", + "original_relative_path": "nonauthor-finite-count-checks/independent-controls.meta.json", + "original_sha256": "49e1a99d70a759a0e2e108b1a0e40303fbd05d6f3c6afb8ad45384abfebd4edf", + "path": "publication-projections/nonauthor-finite-count-checks/independent-controls.meta.json", + "role": "labeled_publication_projection", + "sha256": "2c9e34a1b31cb0eb5ee88c7641d5073b4ffdce0c8eb5c8dfef5ff66fdb5e13a4" + }, + { + "bytes": 3785, + "disposition": "DERIVATIVE_NOT_ORIGINAL", + "original_relative_path": "nonauthor-fold-isolation-checks/CHECK_RUN_SUMMARY.json", + "original_sha256": "638503d7560f87424e55cb3f118ae63b3bbc355e01350a86e40076b155e7429d", + "path": "publication-projections/nonauthor-fold-isolation-checks/CHECK_RUN_SUMMARY.json", + "role": "labeled_publication_projection", + "sha256": "756624416102842360ec4e896e6a6918a6bdd212901b760f4c37aa3e23cf80b1" + }, + { + "bytes": 1198, + "disposition": "DERIVATIVE_NOT_ORIGINAL", + "original_relative_path": "nonauthor-fold-isolation-checks/author-controls.meta.json", + "original_sha256": "932a196891241aba910293fbdf326b116e8bb341dfe4f3d0f37158aae2a9d05f", + "path": "publication-projections/nonauthor-fold-isolation-checks/author-controls.meta.json", + "role": "labeled_publication_projection", + "sha256": "fc11d07b212c7959641fffa7799a4c770d56de0884abd9cc31e7e599e00f086e" + }, + { + "bytes": 1235, + "disposition": "DERIVATIVE_NOT_ORIGINAL", + "original_relative_path": "nonauthor-fold-isolation-checks/independent-controls.meta.json", + "original_sha256": "3606ca471c372e5f86b936c842a40eccd5b7ee9c17471f062872742bc6b156cc", + "path": "publication-projections/nonauthor-fold-isolation-checks/independent-controls.meta.json", + "role": "labeled_publication_projection", + "sha256": "b6d406a4bb27db4abf19183ad1b627fbb082a67cd9792cb419d47b5831e8a164" + }, + { + "bytes": 8029, + "disposition": "DERIVATIVE_NOT_ORIGINAL", + "original_relative_path": "nonauthor-uniform-local-norm-checks/CHECK_RUN_SUMMARY.json", + "original_sha256": "480e1e4786933f2201a6dadb93ee720dc21b632c5a749a91ea9323f1fc506364", + "path": "publication-projections/nonauthor-uniform-local-norm-checks/CHECK_RUN_SUMMARY.json", + "role": "labeled_publication_projection", + "sha256": "9b55ad001c8547323785b5a42d2d75813a9f6071ea024cf687f24f5a3f611e3b" + }, + { + "bytes": 1215, + "disposition": "DERIVATIVE_NOT_ORIGINAL", + "original_relative_path": "nonauthor-uniform-local-norm-checks/author-controls.meta.json", + "original_sha256": "0e8fd7097e06de071dd233b8f6a8ff7623d370e77d78421d7b11a88b9deffac8", + "path": "publication-projections/nonauthor-uniform-local-norm-checks/author-controls.meta.json", + "role": "labeled_publication_projection", + "sha256": "e47e4e5dabb38b7d7a33497785acaa3d00bd4abdf09d540c5cd5211368835186" + }, + { + "bytes": 1243, + "disposition": "DERIVATIVE_NOT_ORIGINAL", + "original_relative_path": "nonauthor-uniform-local-norm-checks/independent-controls.meta.json", + "original_sha256": "c88b17ff52d1eb833679da97a4cb93a2f6068345c785b8e24c8a989ac2c55a34", + "path": "publication-projections/nonauthor-uniform-local-norm-checks/independent-controls.meta.json", + "role": "labeled_publication_projection", + "sha256": "3142170824e4181d0ba1501bc2db1f984dba69bac033c56ddfaec13d0e0fce29" + }, + { + "bytes": 443, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "uniform-local-norm-primary-sources/SOURCE_IDENTITY.json", + "role": "external_import_identity_receipt", + "sha256": "92bf5297b309e947f42ee1dc837487775e7341bc7c1051e493899963ff85891a" + }, + { + "bytes": 6486, + "disposition": "EXACT_ORIGINAL_BYTES", + "path": "uniform_local_norm_controls.py", + "role": "frozen_author_finite_control", + "sha256": "4b84755b771598ed34193c075718a41da447719f90a2ff5c6a17d48d0abc3dc4" + } + ], + "full_research_goal": "ACTIVE_UNMET", + "payload_entry_count": 82, + "schema": "fold-isolation-publication-v1", + "scientific_effect": "NONE", + "scope": "Exact frozen reviewed conventional proofs and finite evidence; new publication glue/projections. No new accepted-control execution, source-specific full PP theorem, Lean or scientific acceptance.", + "selected_publication_base": "7ec8e92f4c7f18153f022d07a754eb247b235b2b", + "self_hash_rule": "This manifest excludes itself. The final Git tree binds all83files. Original review manifests are unchanged.", + "source_exposed_same_provider_org0": true, + "source_mirrors": [ + { + "bytes": 29190, + "commit": "85b928b326ff0586715a56ae91f51d3fbd89557a", + "current_base_bytes_equal": true, + "git_blob": "a27a84150b3e70bfd794672eeeed6915425a1da4", + "git_repository": "d6g8k5htny-coder/main", + "path": "experiments/universality/finite_candidate_lifetime/PROOF.md", + "publication_base": "7ec8e92f4c7f18153f022d07a754eb247b235b2b", + "sha256": "ed31f562f0ffaaf6d4bd2ceb1ae3d234798e983a55d55c6983b128fce27133f5" + }, + { + "bytes": 25367, + "commit": "85b928b326ff0586715a56ae91f51d3fbd89557a", + "current_base_bytes_equal": true, + "git_blob": "5f345201309cfa1781e92d6292f542de4bcb2999", + "git_repository": "d6g8k5htny-coder/main", + "path": "experiments/universality/finite_gaussian_contact/PROOF.md", + "publication_base": "7ec8e92f4c7f18153f022d07a754eb247b235b2b", + "sha256": "7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474" + }, + { + "bytes": 30935, + "commit": "85b928b326ff0586715a56ae91f51d3fbd89557a", + "current_base_bytes_equal": true, + "git_blob": "cd7ef8b8a364cb394be48208f55332658e9fff0b", + "git_repository": "d6g8k5htny-coder/main", + "path": "experiments/universality/iid_short_bar_process/PROOF.md", + "publication_base": "7ec8e92f4c7f18153f022d07a754eb247b235b2b", + "sha256": "320bc988c468bea4938bb777a2d05952c6ab8e0a384f33725b50de847aeec1eb" + } + ], + "source_repository": "d6g8k5htny-coder/main", + "withheld_originals": [ + { + "disposition": "HASH_IDENTIFIED_WITHHELD_ORIGINAL_PRIVATE_CAPTURE_PATHS", + "original_bytes": 2560, + "original_relative_path": "counterfamily-nonauthor-checks-20261016/EXECUTION_RECORD.json", + "original_sha256": "8ae808eed6b7f8fec67435a3b97c8422a08212e7f18ababbced3b367380e04fb", + "projection_bytes": 3129, + "projection_path": "publication-projections/counterfamily-nonauthor-checks-20261016/EXECUTION_RECORD.json", + "projection_sha256": "51332efafef10ac12574475ef20844cd0cc4dc2654002f14f89d20c7e2952cb1" + }, + { + "disposition": "HASH_IDENTIFIED_WITHHELD_ORIGINAL_PRIVATE_CAPTURE_PATHS", + "original_bytes": 6479, + "original_relative_path": "nonauthor-finite-count-checks/CHECK_RUN_SUMMARY.json", + "original_sha256": "5faeb21e56f042394de6772e54081fd16eded69b42abd89109773d5193898132", + "projection_bytes": 7179, + "projection_path": "publication-projections/nonauthor-finite-count-checks/CHECK_RUN_SUMMARY.json", + "projection_sha256": "ae4f4a6fdb7da00737d17b9bc97f5ab085306d1beb5a2544f5165ad969023312" + }, + { + "disposition": "HASH_IDENTIFIED_WITHHELD_ORIGINAL_PRIVATE_CAPTURE_PATHS", + "original_bytes": 717, + "original_relative_path": "nonauthor-finite-count-checks/author-controls.meta.json", + "original_sha256": "07fa0d98d0f407f912444426ffdacaf57f5b0f51928f52be197937ef00bd1463", + "projection_bytes": 1211, + "projection_path": "publication-projections/nonauthor-finite-count-checks/author-controls.meta.json", + "projection_sha256": "0107aeb802b10a33ab7349102dd7d26911dcd89877e6e3e6033a744ffcf28f4f" + }, + { + "disposition": "HASH_IDENTIFIED_WITHHELD_ORIGINAL_PRIVATE_CAPTURE_PATHS", + "original_bytes": 741, + "original_relative_path": "nonauthor-finite-count-checks/independent-controls.meta.json", + "original_sha256": "49e1a99d70a759a0e2e108b1a0e40303fbd05d6f3c6afb8ad45384abfebd4edf", + "projection_bytes": 1240, + "projection_path": "publication-projections/nonauthor-finite-count-checks/independent-controls.meta.json", + "projection_sha256": "2c9e34a1b31cb0eb5ee88c7641d5073b4ffdce0c8eb5c8dfef5ff66fdb5e13a4" + }, + { + "disposition": "HASH_IDENTIFIED_WITHHELD_ORIGINAL_PRIVATE_CAPTURE_PATHS", + "original_bytes": 3567, + "original_relative_path": "nonauthor-fold-isolation-checks/CHECK_RUN_SUMMARY.json", + "original_sha256": "638503d7560f87424e55cb3f118ae63b3bbc355e01350a86e40076b155e7429d", + "projection_bytes": 3785, + "projection_path": "publication-projections/nonauthor-fold-isolation-checks/CHECK_RUN_SUMMARY.json", + "projection_sha256": "756624416102842360ec4e896e6a6918a6bdd212901b760f4c37aa3e23cf80b1" + }, + { + "disposition": "HASH_IDENTIFIED_WITHHELD_ORIGINAL_PRIVATE_CAPTURE_PATHS", + "original_bytes": 702, + "original_relative_path": "nonauthor-fold-isolation-checks/author-controls.meta.json", + "original_sha256": "932a196891241aba910293fbdf326b116e8bb341dfe4f3d0f37158aae2a9d05f", + "projection_bytes": 1198, + "projection_path": "publication-projections/nonauthor-fold-isolation-checks/author-controls.meta.json", + "projection_sha256": "fc11d07b212c7959641fffa7799a4c770d56de0884abd9cc31e7e599e00f086e" + }, + { + "disposition": "HASH_IDENTIFIED_WITHHELD_ORIGINAL_PRIVATE_CAPTURE_PATHS", + "original_bytes": 734, + "original_relative_path": "nonauthor-fold-isolation-checks/independent-controls.meta.json", + "original_sha256": "3606ca471c372e5f86b936c842a40eccd5b7ee9c17471f062872742bc6b156cc", + "projection_bytes": 1235, + "projection_path": "publication-projections/nonauthor-fold-isolation-checks/independent-controls.meta.json", + "projection_sha256": "b6d406a4bb27db4abf19183ad1b627fbb082a67cd9792cb419d47b5831e8a164" + }, + { + "disposition": "HASH_IDENTIFIED_WITHHELD_ORIGINAL_PRIVATE_CAPTURE_PATHS", + "original_bytes": 7279, + "original_relative_path": "nonauthor-uniform-local-norm-checks/CHECK_RUN_SUMMARY.json", + "original_sha256": "480e1e4786933f2201a6dadb93ee720dc21b632c5a749a91ea9323f1fc506364", + "projection_bytes": 8029, + "projection_path": "publication-projections/nonauthor-uniform-local-norm-checks/CHECK_RUN_SUMMARY.json", + "projection_sha256": "9b55ad001c8547323785b5a42d2d75813a9f6071ea024cf687f24f5a3f611e3b" + }, + { + "disposition": "HASH_IDENTIFIED_WITHHELD_ORIGINAL_PRIVATE_CAPTURE_PATHS", + "original_bytes": 715, + "original_relative_path": "nonauthor-uniform-local-norm-checks/author-controls.meta.json", + "original_sha256": "0e8fd7097e06de071dd233b8f6a8ff7623d370e77d78421d7b11a88b9deffac8", + "projection_bytes": 1215, + "projection_path": "publication-projections/nonauthor-uniform-local-norm-checks/author-controls.meta.json", + "projection_sha256": "e47e4e5dabb38b7d7a33497785acaa3d00bd4abdf09d540c5cd5211368835186" + }, + { + "disposition": "HASH_IDENTIFIED_WITHHELD_ORIGINAL_PRIVATE_CAPTURE_PATHS", + "original_bytes": 738, + "original_relative_path": "nonauthor-uniform-local-norm-checks/independent-controls.meta.json", + "original_sha256": "c88b17ff52d1eb833679da97a4cb93a2f6068345c785b8e24c8a989ac2c55a34", + "projection_bytes": 1243, + "projection_path": "publication-projections/nonauthor-uniform-local-norm-checks/independent-controls.meta.json", + "projection_sha256": "3142170824e4181d0ba1501bc2db1f984dba69bac033c56ddfaec13d0e0fce29" + } + ] +} diff --git a/experiments/universality/fold_isolation_diagonal/QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md b/experiments/universality/fold_isolation_diagonal/QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md new file mode 100644 index 000000000..86c9715b9 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md @@ -0,0 +1,293 @@ +# Quantitative fold isolation in arbitrary pair directions + +16 October 2026 UTC. Local conventional-mathematics proposal by OpenAI/Codex `/root/core_theorem_audit_20261015`, independently reconstructing and attacking the sketch supplied by OpenAI/Codex `/root`. Organizational-independence credit: **0**. This document has not yet received a nonauthor review. It changes no production source, native record, claim, graph node or scientific status. In particular, it is not a Lean theorem, a factorial Rice bound, or a spatial point-process limit. + +## 1. Concrete advance and source binding + +An explicit Lyapunov–Schmidt chart isolates **at most two ambient critical points**, including at most one maximum, around a maximum with a nonvanishing soft cubic and a uniformly negative hard eigenvalue. Its radius is bounded below by a constant times the inverse square of a finite Gaussian coefficient norm. A literal positive cubic-gap mark can imply the required eigenframe cubic, but only with the derivative errors displayed below. Combining these statements gives an arbitrary-angle **occurrence-probability** estimate `C exp(-c/eta)` for two nearby short pairs on fixed positive transverse/cubic mark windows. It does not bound their factorial counting kernel. + +The source consumed for the pair convention and finite Gaussian law is Main `85b928b326ff0586715a56ae91f51d3fbd89557a`, [experiments/universality/finite_gaussian_contact/PROOF.md](https://github.com/d6g8k5htny-coder/main/blob/85b928b326ff0586715a56ae91f51d3fbd89557a/experiments/universality/finite_gaussian_contact/PROOF.md). Git blob `5f345201309cfa1781e92d6292f542de4bcb2999`, 25,367 bytes, SHA-256 `7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474`. Its complete local body was checked against this tree identity. Its lines 41–57 define the actual finite source, 68–76 define the pins and radius, 145–186 give the divided differences, and 253–307 specify physical Hessian weights. The source is pinned, rather than asserted to be the latest moving head. + +The previous frozen [COALESCING_CONTACT_CALCULATION.md](COALESCING_CONTACT_CALCULATION.md), [NONAUTHOR_COALESCENCE_REVIEW.md](NONAUTHOR_COALESCENCE_REVIEW.md), and [AUTHOR_ALIGNMENT_ADDENDUM.md](AUTHOR_ALIGNMENT_ADDENDUM.md) were read. This proposal is a separate arbitrary-angle calculation. Their review verdict does not extend to this document. Their collinear `exp(-c/eta^4)` Rice-kernel estimate is **not** the estimate derived here. Their four-site height-covariance rank obstructions remain in force. + +## 2. Deterministic chart, including domain and constants + +Use Euclidean coordinates in a Hessian eigenframe at an anchored critical point, moved to the origin. Let `F` be `C^4` on an open neighborhood of the closed square `Q_R0=[-R0,R0]^2`, with `R0>0`. Suppose + +```text +grad F(0)=0, F_xy(0)=0, +F_xx(0)<0, F_yy(0)<=-a<0, +|F_xxx(0)|>=tau>0. +``` + +Let `B>=sup_Q_R0 ||D^3 F||op` and `C>=sup_Q_R0 ||D^4 F||op`. These are multilinear operator norms with unit vectors in each slot, so the bounds survive frame changes. In particular, each coordinate partial is bounded by the corresponding constant. Necessarily `B>0`. If only the maximum absolute coordinate partials `M3,M4` in an initial orthonormal frame are available, the conservative substitutions `B=2 sqrt(2) M3`, `C=4 M4` suffice by `||v||_1<=sqrt(2)||v||_2` in each slot. + +Set + +```text +D = C+B^2/a > 0, +R = min(R0, a/(32B), tau/(32D)). (1) +``` + +The following statements hold on the **closed** square `Q_R`. Classifications refer to ambient critical points of `F`, not constrained extrema of the square. + +For each `|x|<=R`, there is a unique `h(x)` in `(-R,R)` with `F_y(x,h(x))=0`. The ridge has `C^3` regularity (locally extending at the two endpoints), and + +```text +h(0)=h'(0)=0, +|h(x)|<=B x^2/a<=R/32, +|h'(x)|<=4BR/a<=1/8. (2) +``` + +For `f(x)=F(x,h(x))`, + +```text +|f'''(x)-F_xxx(0)|<=2CR+16B^2 R/a<=tau/2. (3) +``` + +Thus `f'''` has the fixed sign of `F_xxx(0)` and magnitude at least `tau/2`. There are at most two critical points of `F` in `Q_R`. The anchored maximum is its only local maximum. If a second critical point exists, including on `x=+/-R`, it is nondegenerate and has Morse index one. The faces `y=+/-R` contain no ambient critical points. This is an isolation result containing a possible maximum–saddle pair, not an ordinary quadratic Morse chart about the maximum. + +### Proof of the chart and its vertical barrier + +On the square, integration along coordinate segments gives + +```text +|F_yy(x,y)-F_yy(0)|<=B(|x|+|y|)<=2BR, +F_yy(x,y)<=-a/2. +``` + +Since `F_y(0)=F_xy(0)=0`, the one-dimensional Taylor integral on `y=0` gives `|F_y(x,0)|<=B x^2/2`. Hence + +```text +F_y(x,+R)<=B R^2/2-aR/2<=-31 aR/64, +F_y(x,-R)>=-B R^2/2+aR/2>=+31 aR/64. (4) +``` + +The intermediate value theorem and strict decrease in `y` give the unique root. The implicit function theorem gives its stated regularity; the strict face margins permit local continuation at the endpoints. Subtracting `F_y(x,0)` from `F_y(x,h)=0` bounds `|h|<=B x^2/a`. Also + +```text +h'=-F_xy/F_yy, +|F_xy(x,h)|<=B(|x|+|h|)<=2BR, +|h'|<=4BR/a. +``` + +All derivatives in the next display are evaluated at `(x,h(x))`. Exact differentiation, using the first and second derivatives of `F_y(x,h(x))=0`, yields + +```text +f'=F_x, +f''=F_xx+F_xy h'=F_xx-F_xy^2/F_yy, +f'''=F_xxx+3F_xxy h'+3F_xyy(h')^2+F_yyy(h')^3. (5) +``` + +For example, differentiating `f''=F_xx+F_xy h'` initially produces `F_xy h''`; the second derivative of `F_y=0` and `F_xy=-F_yy h'` convert this into the remaining mixed terms in (5). Omitting that cancellation would be an error. + +The first term differs from `F_xxx(0)` by at most `2CR`. For `p=|h'|<=1/8`, `3p+3p^2+p^3<=4p`. The remaining terms have total absolute value at most `4Bp<=16B^2R/a`, proving (3). Here `2C+16B^2/a<=16(C+B^2/a)` and (1) give the final `tau/2` bound. + +Put `g=f'`. Fixed-sign `g''=f'''` makes `g` strictly convex or strictly concave, so it has at most two zeros. Every critical point of `F` in the square lies on `h`, and its `x` coordinate is a zero of `g`. At the anchored zero, `g'(0)=f''(0)=F_xx(0)<0`. If a second zero exists, strict monotonicity of `g'`, or Rolle's theorem between the two roots, implies its derivative is strictly positive. Thus the second root is simple and `f''>0`. At a critical point, + +```text +det Hess F = F_yy f''. +``` + +The second point has a negative determinant and is an index-one saddle. This also proves that a second degenerate root cannot coexist with the anchored nondegenerate maximum. Without that anchor hypothesis, a double root `g=g'=0` is possible; the two-root bound by itself would not prove nondegeneracy. No critical point outside the square or constrained boundary extremum is covered. + +Finally, integrating `F_yy<=-a/2` twice from the ridge gives the useful **vertical** height barrier + +```text +F(x,y)<=f(x)-(a/4)(y-h(x))^2, +F(x,+/-R)<=f(x)-aR^2/8. (6) +``` + +The last bound follows from `|h|<=R/32` and `(31/32)^2/4>=1/8`. It is relative to `f(x)`, rather than to the height at the anchored maximum. In particular, a distinct maximum at Euclidean distance strictly less than `R` from the origin is excluded, regardless of its own eigenframe or pair direction. + +## 3. Falsification attempts that must remain attached + +**A height drop on the entire boundary relative to `F(0)` is false.** For + +```text +F(x,y)=-mu x^2/2+tau x^3/6-a y^2/2, mu,a,tau>0, +``` + +the origin is a nondegenerate maximum with the required transverse curvature and cubic. But `F(R,0)>F(0)` whenever `mu0`, let + +```text +F(x,y)=-A x^2(x-delta)^2-a y^2/2, A=tau/(12delta). +``` + +There are nondegenerate maxima at `x=0,delta`, both with `|F_xxx|=tau`; but `|F_xxxx|=2tau/delta`. Their distance can vanish as the fourth-derivative bound diverges. They share a local saddle; they are not being asserted to represent two actual finite persistence bars. + +These polynomial examples are deterministic local controls. They do not assert positive-probability occurrences within a particular finite Fourier law. + +## 4. Literal endpoint distance and conversion of marks to the eigenframe + +The pinned source uses `M=z-r u/2`, `S=z+r u/2`, so **`r=|S-M|` is the endpoint distance**, and `ell=F(M)-F(S)=k r^3`. It is not the half-distance. Assume the whole endpoint segment is in the domain on which the operator bounds `B,C` hold, and both endpoints are critical. The source's fourth divided-difference row becomes exactly + +```text +12k = integral_0^r [6s(r-s)/r^3] F_uuu(M+s u) ds. (8) +``` + +The nonnegative kernel has mass one and mean `s=r/2`. Consequently, + +```text +|F_uuu(M)-12k|<=C r/2. (9) +``` + +This is an endpoint estimate. It must not be replaced by a midpoint estimate without moving the evaluation point. + +Let `H=Hess F(M)`. Vector Taylor expansion of `grad F(S)-grad F(M)=0` gives + +```text +H u = -(1/r) integral_0^r (r-s) D^3F(M+s u)[u,u,.] ds, +||H u||<=B r/2. (10) +``` + +Suppose `M` is a nondegenerate maximum and its most negative eigenvalue is at most `-a`. For example the source transverse condition `F_vv(M)<=-a` for `v` perpendicular to the **pair** direction implies this by the Rayleigh quotient; it does not assert that the pair frame diagonalizes `H`. Assume `rB<=a`. The soft eigenvalue has absolute value at most `Br/2<=a/2`, since the smallest singular value of `H` is at most `||Hu||`. Thus the hard and soft eigenspaces are distinct. Let `v_h` be a unit hard eigenvector and `e` a unit soft eigenvector oriented by `e.u>=0`. Projecting (10) onto the hard eigenspace gives + +```text +|u.v_h|<=Br/(2a)<=1/2, +||u-e||<=sqrt(2)|u.v_h|, +|F_eee(M)-F_uuu(M)|<=3B||e-u||<=3B^2 r/a. (11) +``` + +The last constant is deliberately conservative. Combining (9) and (11), + +```text +F_eee(M)>=12k-r(C/2+3B^2/a). (12) +``` + +Therefore `k>=k_->0`, `rB<=a`, and + +```text +r(C/2+3B^2/a)<=6k_- (13) +``` + +imply `F_eee(M)>=tau:=6k_-`. Choose the isolation chart's `x` direction to be this `e` and its `y` direction the hard eigenvector. The original pair direction can have any angle; only the estimate (11) is used. An upper mark bound is not needed for this implication. The derivative conditions are retained explicitly. + +The normalization check requested by the sketch author is exact: + +```text +F=x^3/6-delta^2 x/2-a y^2/2, +M=-delta, S=+delta, r=2delta, +ell=2delta^3/3, k=1/12, F_xxx=1=12k. (14) +``` + +Conversely, positive `k` alone does **not** give a nonzero cubic at the maximum. The polynomial + +```text +F=(k/r)x^4-2kr x^2-a y^2/2 +``` + +has a nondegenerate maximum at `M=0`, a saddle at `S=r`, and gap `k r^3`, but `F_xxx(M)=0` and `F_xxxx=24k/r`. The error `Cr/2=12k` in (9) is exactly saturated. This falsifies a uniform mark-to-endpoint-cubic inference without controlling `rC`. It asserts no global elder relation. + +## 5. Consequence for the actual fixed finite Gaussian source + +Keep exactly the pinned source law: fixed `L>0`, `omega=2pi/L`, square support `S_K={-K,...,K}^2`, fixed positive symmetric weights `q_n` summing to one, and its `N=(2K+1)^2` independent standard real coefficients `Z`. No covariance isotropy or independent-copy replacement is made. For `j=3,4`, put + +```text +beta_j = [sum_(n in S_K) q_n (omega|n|)^(2j)]^(1/2)>0, +W=1+||Z||_2. +``` + +For arbitrary unit vectors in the `j` derivative slots, the squared Euclidean norm of the coefficient row for that directional derivative is at most `beta_j^2`: within each sine/cosine pair the squares add to `2q_n` times the squared product of the frequency projections. Cauchy–Schwarz then gives, uniformly in location and frame, + +```text +sup_X ||D^jF||op <= beta_j ||Z||_2 <= beta_j W. (15) +``` + +Fix an embedded square scale `R0>0` small enough for all translated/rotated charts being used, for example `R0<=L/16`, and use global derivative bounds `B=beta_3 W`, `C=beta_4 W`. For a maximum satisfying the direct transverse/cubic hypotheses with fixed `a,tau>0`, define + +```text +kappa=min(R0, a/(32beta_3), + tau/[32(beta_4+beta_3^2/a)])>0. +``` + +Because `W>=1`, each member of the minimum in (1) is at least the corresponding constant divided by `W^2`. Thus its isolation radius obeys + +```text +R>=kappa/W^2. (16) +``` + +If another distinct maximum lies at Euclidean chart distance `delta`, (16) implies `W^2>=kappa/delta`. In particular, for any measurable occurrence event `E_delta` of two such maxima at positive distance at most `delta`, and `delta<=kappa/4`, + +```text +P(E_delta)<=2^(N/2) exp[-kappa/(16delta)]. (17) +``` + +Indeed `W>=sqrt(kappa/delta)` forces `||Z||>=sqrt(kappa/delta)-1>=sqrt(kappa/delta)/2`, and `E exp(||Z||^2/4)=2^(N/2)`. Markov's inequality yields (17). The same statement applies to outer probability if an occurrence set's measurability has not been separately supplied. It does not require a first/second-jet covariance inverse. + +### Pair marks without silently assuming their errors small + +Here is a precise norm dichotomy that incorporates the errors (13). Fix `a,k_->0`, `00`. Suppose two candidate pairs `(M_i,S_i)`, with distinct maxima `M_1!=M_2`, satisfy + +```text +grad F(M_i)=grad F(S_i)=0, +M_i is a nondegenerate local maximum, +u_i=(S_i-M_i)/r_i, r_i=|S_i-M_i|>0, +F(M_i)-F(S_i)=k_i r_i^3, k_i>=k_-, +F_(v_i v_i)(M_i)<=-a, v_i perpendicular to u_i, +00. (18) +``` + +Then these two pairs force + +```text +W^2>=c_*/eta. (19) +``` + +For otherwise `W^2=kappa/W^2>kappa eta/c_*>=(1+rho)eta>=delta`, placing the other maximum strictly inside the isolation square, a contradiction. Global derivative bounds justify using the same `B,C` along the first pair segment and throughout this chart. + +For an occurrence event `E_eta` of such configurations with center distance at most `eta<=min(eta0,c_*/4)`, (19) gives + +```text +P(E_eta)<=2^(N/2) exp[-c_*/(16eta)]. (20) +``` + +For varying center distances `s<=eta`, apply (19) with `s` and obtain the stronger lower bound `W^2>=c_*/s>=c_*/eta`. Thus this is a bound on the whole close-pair occurrence event, not just one fixed configuration. It is fixed-law and fixed-mark-window; no uniformity in vanishing `a` or `k_-`, unbounded `K`, varying `L`, or varying spectra is claimed. + +## 6. What this does and does not change + +The next finite mathematical task can use (19) as an arbitrary-angle collision obstruction on positive transverse/cubic windows. This removes a collinearity restriction from this **occurrence** bound and avoids demanding a nonexistent uniform twelve-row height density on the K2/K3 coordinate stratum. It does not repair that density, calculate a conditional four-site covariance determinant, or supply conditional determinant/selection-weight estimates. In particular, an exponentially small probability of some collision does not by itself bound the expectation of a possibly large factorial count. A separate weighted counting argument, with its actual conditioning list and finite-support rank control, remains necessary. + +Small transverse curvature, small cubic marks, pairs whose radii are not bounded by a fixed fraction of center separation, shared-maximum incidences, and other collision geometries are outside (18)–(20). The height barrier (6) proves no actual local selection probability. Global H0 max–saddle pairing, interval decomposition, essential-class handling, higher homology, leading complement estimates, same-field expanding-domain control, process convergence, and the literal lifetime density `ell^(-1/3)` rather than merely cumulative `t^(2/3)` remain separate obligations. No theorem closure is inferred from this local normal form. + +## 7. Finite algebra controls and review custody + +[fold_isolation_controls.py](fold_isolation_controls.py) is a standard-library exact-rational polynomial checker. [FOLD_ISOLATION_CONTROLS.json](FOLD_ISOLATION_CONTROLS.json) is its output. It checks the full mixed reduced-third-derivative identity on a prescribed nontrivial stationary ridge, the endpoint-distance factor 12, the positive-mark/zero-endpoint-cubic counterexample, the mixed-cubic sign-loss example with `D^4F=0`, the false full-height-barrier example, the two-maxima derivative-scale example, and the displayed rational chart margins. Every check passed using Python `Fraction`; no floating arithmetic, field draws, sampling, persistence experiment or process test occurred. + +These controls catch finite algebra and normalization mistakes. They do not mechanically verify the analytic proof, geometric selection, measure-theoretic counting, or a Lean statement. The current review lineage is root sketch -> this agent's independent conventional derivation plus falsification controls -> pending separate nonauthor review. All are OpenAI/Codex in the same parent thread; there is no provider independence or human review. The report and its control source/output are to be frozen by SHA-256 for that next review. Any correction after freezing should be an append-only addendum. diff --git a/experiments/universality/fold_isolation_diagonal/README.md b/experiments/universality/fold_isolation_diagonal/README.md new file mode 100644 index 000000000..e5f8b5de5 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/README.md @@ -0,0 +1,34 @@ +# Fold isolation, near-diagonal counts and covariance obstructions + +This packet supplies four reviewed conventional mathematical arguments for planar finite Fourier Gaussian fields. It proves a quantitative bound for nearby distinct short maximum–saddle candidates, and shows why uniform spectral moments and a one-contact jet floor do not suffice for a nonzero Poisson limit. It also gives conditional local and middle/far estimates for a specified Gaussian truncation route. It does not establish a complete spatial point-process theorem or identify these candidates with actual global persistence bars. + +The source publisher is OpenAI/Codex Root01a11cbf. The proof notes identify their actual authors; each was subsequently reviewed by separate, source-exposed OpenAI agents. All reviews have organizational-independence, blind and human credit **0**. Publication preserves the exact reviewed proof bytes, including their historical statements that review was pending. The completed reports listed below supersede those historical pending statements for their exact scopes. Scientific effect **NONE**. + +| Argument | Exact proof and completed review | Accepted scope | +| --- | --- | --- | +| Arbitrary-angle isolation | [Proof](QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md), [review](NONAUTHOR_FOLD_ISOLATION_REVIEW.md) | A quantitative ridge chart contains at most two critical points and at most one maximum under the displayed derivative and fold-jet bounds. Its height barrier is vertical; an enclosing barrier remains false. | +| Finite Fourier count | [Proof](FINITE_FOURIER_CRITICAL_COUNT_DIAGONAL_ROUTE.md), [review](NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW.md) | Positive full-square weights give the almost-sure cap `Ncrit <= 16K^2`; the count argument includes shared endpoint incidences and the outer radius cutoff. | +| Growing-family diagonal | [Proof](UNIFORM_LOCAL_NORM_DIAGONAL_PROPOSAL.md), [review](NONAUTHOR_UNIFORM_LOCAL_NORM_REVIEW.md) | A localized Gaussian H6 trace bound and deterministic cover yield a uniform near-diagonal counting estimate under normalized variance and a uniform physical twelfth moment. | +| Covariance obstruction and replacement estimates | [Proof](JET_NONDEGENERACY_FACTORIZATION_OBSTRUCTION.md), [integrated review](NONAUTHOR_JET_FACTORIZATION_OBSTRUCTION_REVIEW.md), [separate counterfamily review](COUNTERFAMILY_NONAUTHOR_REVIEW.md) | An actual empty-candidate counterfamily, a separate existence-only positive-mean diagonal, local contact factorization, finite-radius regression bounds, planar middle/far estimates and an eventual qualitative Gaussian truncation route. | + +For the diagonal result, let `D(t,eta)` count ordered distinct endpoint candidates in the stated compact lifetime/cubic/transverse sectors whose centers are within `eta` and whose endpoint radii are at most `rho eta`. Duplicate frame presentations of the same pair are excluded; distinct pairs sharing a maximum or saddle are included. Assume `L >= Lmin > 0`, positive symmetric full-square weights with total mass one, a uniform physical twelfth spectral moment, fixed positive transverse and cubic lower marks and a fixed bounded lifetime sector. The reviewed argument gives constants independent of `K,L` and coefficient dimension such that + +```text +E D(t,eta) <= C 16^4 K^8 (1+L)^2 exp(-c/eta). +H(t) = 8 log K(t) + 2 log(1+L(t)) + 4 log(1/t), +eta(t) = c/H(t). +``` + +If `t^(1/3) H(t) -> 0`, the endpoint radii fit this cutoff eventually, `E D = O(t^4)` and `t^(-4/3) E D -> 0`. These are candidate counting statements, with explicit family hypotheses. A uniform moment upper bound does not imply a positive jet variance, a uniform two-contact floor or mixing. + +The obstruction uses positive full-square weights on `L_N=2pi N`, `K_N=2N`, consisting of a repeated fixed Gaussian field and a small independent full-square fill. It retains a uniform one-contact floor in every frame and a physical twelfth moment at most `2^18`, but has macroscopic coherent resonances. With `epsilon_N=N^-12`, the compact candidate process tends to the empty process at `t_N=N^-3`. A **separate deterministic, potentially smaller** sequence `0=Lmin>0, +sum_n q_n=1, q_n=q_-n>0, +sup_family sum_n q_n |(2pi/L)n|^12 <= M12 < infinity. (U) +``` + +The last condition is a **uniform physical spectral moment**, in addition to the individual smoothness of each finite field. The normalization `sum q_n=1` is an essential separate uniform variance bound, including the DC mode. Without normalization one must also assume `sum q_n<=M00 independent of K,L and coefficient dimension give + +```text +P(two distinct qualifying candidates form an eta-cluster) + <= C (1+L)^2 exp(-c/eta), +E D(t,eta) <= C (16K^2)^4 (1+L)^2 exp(-c/eta). (1) +``` + +The cluster requires center distance at most eta and both endpoint radii at most rho eta. Distinct candidates sharing their maximum or saddle are included; the identical-candidate diagonal is excluded. The count is by ordered endpoint identity, not by continuous frame labels. All constants and small-scale restrictions are supplied below. + +With K=K(t), L=L(t), define + +```text +H(t)=8log K(t)+2log(1+L(t))+4log(1/t), +eta(t)=c/H(t), 00`, the compact-mark lifetime radius bound lies below rho eta(t) eventually, and (1) gives + +```text +E D(t,eta(t)) <= C 16^4 t^4, +t^(-4/3) E D(t,eta(t)) = O(t^(8/3)) -> 0. (3) +``` + +In particular polynomial growth of K and L in 1/t satisfies the radius condition. A family with K/L uniformly bounded automatically satisfies (U), regardless of its positive normalized weights. Thus K and L comparable to `t^(-1/3)` are allowed by this diagonal estimate. No one-point intensity asymptotic, far-field mixing, global selected-bar identity, infinite-carrier result or spatial point-process limit is inferred. + +## 2. Exact inherited source and review identities + +The finite source convention is Main `85b928b326ff0586715a56ae91f51d3fbd89557a`, [experiments/universality/finite_gaussian_contact/PROOF.md](https://github.com/d6g8k5htny-coder/main/blob/85b928b326ff0586715a56ae91f51d3fbd89557a/experiments/universality/finite_gaussian_contact/PROOF.md), Git blob `5f345201309cfa1781e92d6292f542de4bcb2999`, 25,367 bytes, SHA-256 `7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474`. Its actual law and literal endpoint-distance normalization are preserved. Its original fixed-law arguments do not themselves establish family hypothesis (U). + +The following local inputs remain unchanged: + +| Input | SHA-256 | +| --- | --- | +| [QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md](QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md) | `827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30` | +| [NONAUTHOR_FOLD_ISOLATION_REVIEW.md](NONAUTHOR_FOLD_ISOLATION_REVIEW.md) | `f06c61e3f3e60c795a1a14e824640f1689d4143dba4f6698cada7a41929eec90` | +| [FINITE_FOURIER_CRITICAL_COUNT_DIAGONAL_ROUTE.md](FINITE_FOURIER_CRITICAL_COUNT_DIAGONAL_ROUTE.md) | `621cee96854aa440498cecf5452a8d1e23f90a4458779c3286c18eba106fe138` | +| [NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW.md](NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW.md) | `16152427f7013822ac1bba81142429800334e6bf0f354a6439b23a1997bc4524` | + +The finite-count review was read in full. Its verdict is ACCEPT_SCOPED for that exact fixed-law argument. It explicitly does not supply expanding-domain uniformity. The accepted local chart and a.s. critical count are reused with their premises; their verdicts do not extend automatically to this new norm/cover argument. All reviews/authors are OpenAI/Codex, source-exposed, organizational-independence credit 0. The previous six author files and both review files have their frozen hashes checked by the new control program. + +The classical analytic import is Fourier inversion, differentiation and Parseval/Plancherel on Schwartz functions. These were inspected in Leon Simon, [*Lectures on PDE*, Lecture 5, Fourier properties (a)–(h), PDF page 26, with its Sobolev proof on page 27](https://math.stanford.edu/~lms/lecs-on-pde.pdf). Simon uses the unitary convention adopted below. The explicit H6-to-C4 operator constants here are derived by a displayed Cauchy–Schwarz calculation, rather than imported as an unnamed uniform embedding constant. Only the relevant portions of Lecture 5 were read, not the entire text. A saved primary PDF and byte identity are in [uniform-local-norm-primary-sources/SOURCE_IDENTITY.json](uniform-local-norm-primary-sources/SOURCE_IDENTITY.json). + +## 3. Uniform physical derivative variances + +For each K>=2 and L>=Lmin, keep the actual ideal law + +```text +omega=2pi/L, S_K={-K,...,K}^2, +F(x)=sqrt(q_0) C_0 + + sum_nonzero_pairs sqrt(2q_n) + [C_n cos(omega n.x)+S_n sin(omega n.x)], (4) +``` + +with N=(2K+1)^2 independent standard real Gaussian coordinates Z. Weights may vary with K,L but obey (U). The law is centered, stationary and variance one. It need not be isotropic. All reasoning remains on the same F; there are no independent-copy replacements. + +For a coordinate multi-index beta of order j<=6, the sine/cosine pair variances add exactly to give + +```text +Var(partial^beta F(x)) + = sum_n q_n product_l ((omega n_l)^(2beta_l)) + <= sum_n q_n |omega n|^(2j) + <= 1+M12. (5) +``` + +The last inequality uses `r^(2j)<=1+r^12` for r>=0 and j<=6; it includes j=0. This bound is uniform in x,K,L. In covariance notation, `sum q_n|omega n|^12=(-Delta)^6 Cov(F(0),F(h))|_(h=0)`. Condition (U) controls this physical twelfth-order quantity. A cutoff K or period L by itself is not that condition. + +A simple nonempty family with automatic (U) is + +```text +K/L<=Lambda0, R0=b, eta0=b/4. +``` + +Define the flat function and smooth step + +```text +psi(t)=0 for t<=0, exp(-1/t) for t>0, +J(t)=psi(t)/(psi(t)+psi(1-t)), +theta(s)=1-J(|s|-2), +chi_b(x1,x2)=theta(x1/b) theta(x2/b). (6) +``` + +The denominator is positive. The step is 0 on t<=0 and 1 on t>=1. Exponential flatness at zero makes it C-infinity. Although absolute value appears in theta, theta is identically one near s=0, so there is no nonsmooth point there. Thus chi_b is a fixed C-infinity cutoff, equals one on `[-2b,2b]^2`, vanishes outside `[-3b,3b]^2`, and has values in [0,1]. Since 6b=1. Since `sup_(u>=1) u^m exp(-u)<=max(1,m^m)`, its absolute value is at most A_j. On 0<=t<=1 the denominator in J is at least exp(-2), because at least one of t,1-t is at least 1/2. Its jth derivative is bounded by 2A_j. Differentiating the identity between that denominator and its reciprocal gives the recurrence E_n, since exp(2)<8. Then the Leibniz formula gives `|J^(n)|<=G_n`. For j>=1, theta's derivative is supported on the two intervals of total length 2; consequently + +```text +integral_R |theta^(j)(s)|^2 ds <= d_j, +||partial^gamma chi_b||_L2(R^2)^2 + <= b^(2-2|gamma|) d_(gamma1)d_(gamma2). (8) +``` + +For j=0, `0<=theta<=1` and support length 6 give d_0=6. The elementary exponential bound used above follows, for example, by bounding the tail of the e series after 1/4!: `e<=65/24+1/100<11/4`, hence e^2<8. These are finite explicit constants, not evaluated derivative suprema from sampled points. + +Use the unitary Fourier transform on R^2, + +```text +hat f(xi)=(2pi)^(-1) integral exp(-i x.xi) f(x) dx, +||f||_H6^2=integral (1+|xi|^2)^6 |hat f(xi)|^2 dxi. +``` + +By the derivative rule and Plancherel, + +```text +||f||_H6^2=sum_(|alpha|<=6) w_alpha ||partial^alpha f||_2^2, +w_alpha=6!/[(6-|alpha|)! alpha1! alpha2!]. (9) +``` + +For any deterministic torus basepoint g, let `f_g(x)=chi_b(x)F(g+x)` using the periodic lift. The derivative Leibniz sum and weighted Cauchy–Schwarz give + +```text +E ||partial^alpha f_g||_2^2 + <= (1+M12) 2^|alpha| + sum_(gamma<=alpha) binom(alpha,gamma) + ||partial^gamma chi_b||_2^2. (10) +``` + +For clarity, the weight sum is `sum_(gamma<=alpha) binom(alpha,gamma)=2^|alpha|`. The random derivative in each summand has variance bounded by (5); no independence between summands is assumed. Since b<=1 and |gamma|<=6, `b^(2-2|gamma|)<=b^(-10)`. Define the fixed integer + +```text +A_theta=sum_(|alpha|<=6) w_alpha 2^|alpha| + sum_(gamma<=alpha) binom(alpha,gamma) + d_(gamma1)d_(gamma2) + =430299201784277045424464881586844996, +T=(1+M12)b^(-10) A_theta > 0. (11) +``` + +Equations (8)–(10) prove `E||f_g||_H6^2<=T`, uniformly in K,L,g and weights satisfying (U). The constant is intentionally very loose and is not a usable numerical asymptotic-window claim. A smaller trace bound using the actual cutoff derivative integrals would also work, but is not needed to establish uniformity. + +## 5. Dimension-free Gaussian exponential moment + +For each individual finite field, write `f_g=sum_(j=1)^N Z_j phi_(g,j)` in H6, where every deterministic phi is a compactly supported smooth function. Let Q_g be its real H6 Gram matrix. It is positive semidefinite, and + +```text +S_g^2:=||f_g||_H6^2=Z^T Q_g Z, +tr Q_g=E S_g^2<=T. (12) +``` + +An orthogonal diagonalization in this finite real coefficient space gives nonnegative eigenvalues lambda_j with sum at most T. Gaussian rotational invariance and the one-dimensional Gaussian integral yield + +```text +E exp(S_g^2/(4T)) + = product_j (1-lambda_j/(2T))^(-1/2) + <= sqrt(2). (13) +``` + +Indeed, for a_j=lambda_j/(2T)>=0 with sum at most 1/2, induction gives `product(1-a_j)>=1-sum a_j>=1/2`. The estimate includes zero eigenvalues and arbitrary finite rank. It uses neither an infinite-dimensional Gaussian theorem nor a rank-dependent prefactor. In particular, + +```text +P(S_g^2>=v)<=sqrt(2) exp[-v/(4T)], v>=0. (14) +``` + +This is the unconditioned Gaussian law (4). We do not transport the almost-sure Bézout count or this trace bound to an arbitrary regression subspace; a conditional mean or an algebraically exceptional subspace would need its own analysis. + +## 6. Direct Fourier derivative constants, including rotation + +For any unit vectors v_1,...,v_j and j<=4, Fourier inversion and Cauchy–Schwarz give + +```text +|D^j f(x)[v_1,...,v_j]| + <= (2pi)^(-1) [integral |xi|^(2j)/(1+|xi|^2)^6 dxi]^(1/2) + ||f||_H6. +``` + +Polar integration and u=r^2, followed by v=u/(1+u), give exactly + +```text +I_j=integral_R2 |xi|^(2j)/(1+|xi|^2)^6 dxi + =pi integral_0^infinity u^j/(1+u)^6 du + =pi j!(4-j)!/5!, 0<=j<=4. (15) +``` + +Thus the full multilinear operator norms, without a coordinate-frame inflation, satisfy + +```text +sup ||D^3 f||op <= ||f||_H6/sqrt(80pi) <= p3 ||f||_H6, +sup ||D^4 f||op <= ||f||_H6/sqrt(20pi) <= p4 ||f||_H6, +p3=1/15, p4=2/15. (16) +``` + +The rational majorants follow already from pi>3. Absolute convergence of these derivative Fourier integrals also gives continuity. For the present compact smooth f, it is ordinary classical differentiation; the display supplies the quantitative H6-to-C4 control. Arbitrary directions enter through `product |xi.v_l|<=|xi|^j`, so no isotropy of F or frame-wise coordinate norm assumption is used. Where chi_b=1, all these derivatives are the physical derivatives of F. + +## 7. Deterministic cover and the local cluster obstruction + +For each L choose `n=ceil(L/b)` evenly spaced points in each torus coordinate, giving n^2 deterministic grid centers. Their spacing is L/n<=b. Every torus point M has a grid center g for which its short lift satisfies + +```text +|M-g|_infinity<=b/2, +n^2<=(1+L)^2/b^2. (17) +``` + +Any rotated closed square of radius R0=b about M is contained in the region where chi_b(x-g)=1, since its coordinates relative to g are bounded by `b/2+sqrt(2)b<2b`. The cutoff support has side 6b=k_-, and F_vv(M)<=-a. Let D(t,eta) be its ordered sum over **different** candidates with gaps at most Tgap t, center distance at most eta and both radii at most rho eta. Here Tgap>0 is fixed; it is unrelated to the norm-trace constant T in (11). Bounded Borel additional selectors may reduce this count. We use the completed unconditioned Gaussian probability space and the finite-configuration measurability argument of the accepted report. + +For one grid center write `S=S_g`, `W=1+S`, and use the local derivative bounds `B=p3 W`, `C=p4 W`. Define constants depending only on b,a,k_-,rho: + +```text +tau=6k_-, +kappa=min(b, a/(32p3), tau/[32(p4+p3^2/a)]), +D0=p4/2+3p3^2/a, +gamma=min(a^2/(rho^2 p3^2 eta0), + 6k_-/(rho D0), kappa/(1+rho))>0. (18) +``` + +Suppose two distinct candidates form the stated eta-cluster, and choose a covering grid center g for its first maximum M_1. Then + +```text +(1+S_g)^2 >= gamma/eta. (19) +``` + +Here is the direct proof, keeping the localization premise visible. If W^2=12k_1-r_1(C/2+3B^2/a). +``` + +The conditions needed for that conversion hold: + +```text +r_1 B<=rho p3 sqrt(gamma eta)<=a, +r_1(C/2+3B^2/a)<=rho eta D0 W^2=kappa/W^2>(1+rho)eta. (20) +``` + +If the maxima are distinct, the second is at distance at most (1+rho)eta from the first, contradicting maximum uniqueness in the chart. If they share their maximum, distinct candidate identity forces two different saddles, each at distance at most rho eta=gamma/(4eta)` for at least one deterministic covering center. Consequently + +```text +{D(t,eta)>0} subset union_(deterministic g) + {S_g^2>=gamma/(4eta)}, +P(D(t,eta)>0) + <=sqrt(2) n^2 exp[-gamma/(16T eta)] + <=sqrt(2)b^(-2)(1+L)^2 exp(-c/eta), +c=gamma/(16T)>0. (21) +``` + +Although the witnessing maximum and its covering center depend on F, the union is over a predetermined finite cover. Each probability is unconditioned. We do not condition a local norm at a chosen maximum or assume the different cubes are independent. The same trace bound T applies to each deterministic center. + +The accepted algebraic argument gives `Ncrit<=16K^2` almost surely for **each** positive full-square finite law, including every member of a varying family. It needs no uniform positive lower weight, no uniform covariance inverse and no Morse premise. The candidate population contains at most Ncrit^2 endpoint pairs. Thus, almost surely, + +```text +D(t,eta)<=(16K^2)^4 1{D(t,eta)>0}. +``` + +There is no need for a common probability-one event over an uncountable family of laws: this expectation estimate is proved for each member with the same displayed deterministic constants. Combining it with (21) proves (1), explicitly + +```text +E D(t,eta) + <= Ccov 16^4 K^8 (1+L)^2 exp(-c/eta), +Ccov=sqrt(2)b^(-2). (22) +``` + +This is an integrated candidate factorial count, not a pointwise Rice-density bound. Setting the count to zero on the source-specific exceptional null set is harmless. No four-site observation covariance or Hessian determinant weight is inserted a second time. + +For the full lifetime band 0=k_-, let `A=(Tgap/k_-)^(1/3)`. Then every radius is at most A t^(1/3). Choose H(t) and eta(t) from (2). The exact algebra is + +```text +exp(-c/eta(t))=K(t)^(-8)(1+L(t))^(-2)t^4. +``` + +Since H(t)>=4log(1/t), eta(t)<=eta_cap eventually. If additionally + +```text +t^(1/3)H(t)->0, (23) +``` + +then `A t^(1/3)<=rho eta(t)` eventually. Thus (13) of the accepted report, with this eta, counts every two distinct candidates in this compact-fold lifetime sector whose centers have distance at most eta(t). Equation (22) proves (3). A sufficient general growth condition is + +```text +log K(t)+log(1+L(t))=o(t^(-1/3)). (24) +``` + +Polynomial K,L in 1/t satisfy it. More generally one may replace the logarithmic coefficients 8,2,4 by p>=8, q>=2, u>4/3 to obtain a vanishing t^(-4/3) normalization, with the corresponding radius condition. These are exact cancellations of the coarse count/cover powers, not an asymptotic independence argument. + +For example take L=t^(-1/3) and K=ceil(L), with arbitrary positive symmetric normalized weights. For t sufficiently small, K/L<=2, so (U) holds with a fixed M12. The numerator/denominator scales in (2) give `H(t)=(22/3)log(1/t)+O(1)` and eta of logarithmic order; A t^(1/3)/eta tends to zero. The bound is therefore compatible with this increasing-domain regime. Calling this a constant-total-intensity regime additionally requires a **separate** one-point intensity asymptotic; none is proved or assumed here. + +The displayed D is a total ordered endpoint count. Dividing it by L^2 or L^4, if a later explicitly stated convention requires per-volume or per-volume-pair normalization, can only reduce the bound uniformly up to an Lmin-dependent constant. No such division is hidden in (22). The requested t^(-4/3) factor is a twofold cumulative-lifetime scale. It is not a proof of lifetime density ell^(-1/3) or of cumulative mass t^(2/3). + +## 9. Falsifiers and limits of reuse + +**M12 alone does not control variance.** If normalization is dropped, take `F_A=A Z_0+G`, with Z_0 an independent standard real Gaussian and G a fixed finite full-square field with all nonconstant mode weights positive. Its twelfth physical spectral moment is independent of A because the DC mode contributes zero. But `E||chi_b F_A||_H6^2=A^2||chi_b||_H6^2+E||chi_b G||_H6^2` diverges. This explicitly falsifies a uniform local trace from M12 alone. It is outside (U), whose normalized total mass controls DC and every lower moment. A separate uniform M0 bound is the alternative stated above. + +**Individual smoothness does not imply a uniform local trace.** Fix a physical period, let K grow, put a fixed positive fraction of the normalized spectral mass on the two axial frequencies +/-K, and distribute the remaining mass positively over the full square. Then M12 grows at least as K^12. On the inner cutoff square chi_b=1, so + +```text +E ||chi_b F||_H6^2 + >= integral_inner E |partial_x^6 F|^2 + >= (positive fixed area and mass) times (omega K)^12. +``` + +A dimension-free bound with a fixed T would imply a uniform second moment and is impossible for this family. Even the fourth derivative at one fixed physical point has variance of order K^8. This is a literal obstruction to removing (U) merely because each field is a finite C-infinity Fourier sum. It is not a counterexample under (U). + +**Very fast growth can defeat the radius comparison despite (U).** If `log K(t)` grows as t^(-1/2), then the cutoff (2) is at most of order t^(1/2), whereas the lifetime radius upper bound is of order t^(1/3). The latter is not o(eta). Uniform (U) can still be arranged by assigning sufficiently small positive weight to high physical frequencies, so this failure is distinct from the moment failure. It shows that the count/cover logarithmic route cannot cover the whole stated lifetime sector at that growth rate by this bound alone. It does not disprove the occurrence bound when its radius cutoff is actually imposed, or any point-process conjecture. + +**Moving mark restrictions must be identified by what they affect.** Growing upper height or upper cubic-mark thresholds do not harm (21)–(22) when the fixed lower a,k_- and fixed lifetime cap Tgap remain: this proof uses no such upper bounds. Increasing lower transverse/cubic thresholds also only restricts the counted population, so fixed positive lower bounds still suffice. In contrast, allowing the lower transverse curvature or cubic mark to tend to zero removes the uniform gamma in (18). Allowing Tgap=Tgap(t) to grow changes (23) to `Tgap(t)^(1/3)t^(1/3)H(t)->0`; for example Tgap=1/t makes the radius bound constant, so it cannot be automatically below a cutoff tending to zero. Those are scope failures, not conclusions against the estimate with its actual premises. + +The full-square assumption is used for the a.s. critical-count theorem, not for the local Gram trace. Infinite Fourier carriers have no finite polynomial Bézout cap supplied here. The same is true of arbitrary conditioned coefficient laws that may lie inside the algebraic bad set. Non-Gaussian coefficients need a different exponential-moment argument. The argument uses fixed positive geometric lower marks and a physical Lmin. Neither small-mark integration nor a source-law admission to (U) is filled by the word smooth. + +Even under (U), the moving far-center complement, uniform one-point remainders/intensity, actual global H0 bar correspondence, essential classes, higher homology, mixing, increasing-domain process tightness and identification of a limiting point process remain separate. No independence follows from the deterministic union. The accepted K2/K3 enlarged height-covariance singularities are unchanged; this integrated route avoids that density for this sector rather than repairing it. + +## 10. Controls and review custody + +[uniform_local_norm_controls.py](uniform_local_norm_controls.py) and [UNIFORM_LOCAL_NORM_CONTROLS.json](UNIFORM_LOCAL_NORM_CONTROLS.json) use exact integers and Fraction arithmetic, with no floating arithmetic or field samples. They verify the beta-integral constants and rational operator majorants, the explicit cutoff recurrence and trace integer, multinomial/Leibniz sums, 256 finite product margins across varying dimensions, deterministic cover/rotation margins, logarithmic cancellation, the growing-lifetime-cap radius failure, and unchanged identities for the six prior author files and two accepted reviews. All checks passed. + +The controls do not mechanically establish the continuous cutoff derivative estimates, Plancherel, all-parameter trace bound, Gaussian integral, local analytic chart, algebraic genericity, or asymptotic theorem. Those are the conventional derivations/imports stated in this report. The large explicit constants are not certified useful experimental windows. No Gaussian draws, approximate critical roots, PRNG seeds, Rice integrals, field-level experiments or process simulations occurred. + +Current lineage is root sketch -> this agent's independent analytic reconstruction and falsification controls -> pending separate nonauthor review of these new bytes. The two previous ACCEPT_SCOPED reviews are consumed only within their frozen scope; this proposal needs its own review. All authors/reviewers so far are OpenAI/Codex, with organizational-independence credit 0. Previous files and reviews remain unchanged. These new local files and a primary-source copy have no production/native/remote or scientific-status effect. diff --git a/experiments/universality/fold_isolation_diagonal/counterfamily-nonauthor-checks-20261016/author_controls.stderr b/experiments/universality/fold_isolation_diagonal/counterfamily-nonauthor-checks-20261016/author_controls.stderr new file mode 100644 index 000000000..e69de29bb diff --git a/experiments/universality/fold_isolation_diagonal/counterfamily-nonauthor-checks-20261016/author_controls.stdout.json b/experiments/universality/fold_isolation_diagonal/counterfamily-nonauthor-checks-20261016/author_controls.stdout.json new file mode 100644 index 000000000..ce40a1f8c --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/counterfamily-nonauthor-checks-20261016/author_controls.stdout.json @@ -0,0 +1,49 @@ +{ + "classification": "EXACT_FINITE_ALGEBRA_CONTROLS_ONLY", + "arithmetic": "Python Fraction and integers; no floating arithmetic", + "Gaussian_sampling": "NONE", + "process_simulation": "NONE", + "scientific_effect": "NONE", + "checks": { + "fixed_base_axial_seven_jet_exact_LDL_positive": "PASS", + "positive_full_square_normalized_uniform_physical_M12": "PASS", + "single_jet_mixture_identity_and_resonant_two_value_collapse": "PASS", + "resonant_uniform_fill_geometric_sum": "PASS", + "DC_single_jet_failure_bookkeeping": "PASS", + "small_fill_C4_dimension_ledger_and_expectation_transfer_cap": "PASS", + "ordered_fold_Jacobian_and_middle_far_kernel_exponents": "PASS", + "consumed_frozen_files_unchanged": "PASS" + }, + "base_axial_LDL_pivots": [ + "1", + "2", + "14/5", + "72/25", + "2", + "4", + "14/5" + ], + "finite_family_cases": [ + { + "N": 4, + "M12": "837542830378721313/17592186044416", + "two_value_small_eigenvalue": "1/17825792" + }, + { + "N": 5, + "M12": "1251428171060188861397852/26285648345947265625", + "two_value_small_eigenvalue": "4/1025390625" + }, + { + "N": 8, + "M12": "695556060602940566395768283/14609821306377964879872", + "two_value_small_eigenvalue": "1/70866960384" + }, + { + "N": 12, + "M12": "3863598447035119394611266277217/81153031520128153386221568", + "two_value_small_eigenvalue": "1/9101852540928" + } + ], + "scope": "Finite checks support arithmetic; all-frame positivity, actual process obstruction, and covariance-kernel estimates require the conventional report proofs." +} diff --git a/experiments/universality/fold_isolation_diagonal/counterfamily-nonauthor-checks-20261016/counterfamily_checks.py b/experiments/universality/fold_isolation_diagonal/counterfamily-nonauthor-checks-20261016/counterfamily_checks.py new file mode 100644 index 000000000..731874125 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/counterfamily-nonauthor-checks-20261016/counterfamily_checks.py @@ -0,0 +1,150 @@ +"""Independent exact controls for obstruction §§3–5; not a proof or simulation.""" +from fractions import Fraction as Q +from math import comb +from itertools import product +import json + +ROWS = [(0,0),(1,0),(2,0),(3,0),(0,1),(1,1),(0,2)] +checks = [] +def check(name, condition): + if not condition: raise AssertionError(name) + checks.append(name) + +def uniform_moments(k,scale=1): + return [sum(Q(i,scale)**p for i in range(-k,k+1))/Q(2*k+1) for p in range(7)] + +def directional_moment(p,q,c,s,moment,reflection=1): + return reflection**q * sum(Q(comb(p,i)*comb(q,j))*c**i*s**(p-i)*(-s)**j*c**(q-j)*moment(i+j,p+q-i-j) for i in range(p+1) for j in range(q+1)) + +def gram(c,s,moment,reflection=1): + out=[] + for p,q in ROWS: + row=[] + for r,t in ROWS: + phase=((p+q)-(r+t))%4 + sign={0:1,1:0,2:-1,3:0}[phase] + row.append(sign*directional_moment(p+r,q+t,c,s,moment,reflection)) + out.append(row) + return out + +def ldl(a): + n=len(a); low=[[Q(i==j) for j in range(n)] for i in range(n)]; d=[] + for i in range(n): + pivot=a[i][i]-sum(low[i][j]**2*d[j] for j in range(i)) + if pivot<=0: raise ValueError('nonpositive pivot') + d.append(pivot) + for k in range(i+1,n): + low[k][i]=(a[k][i]-sum(low[k][j]*low[i][j]*d[j] for j in range(i)))/pivot + return d + +def determinant(a): + a=[row[:] for row in a]; det=Q(1); n=len(a) + for i in range(n): + pivot=next((j for j in range(i,n) if a[j][i]),None) + if pivot is None:return Q(0) + if pivot!=i: a[i],a[pivot]=a[pivot],a[i];det=-det + v=a[i][i];det*=v + for j in range(i+1,n): + fac=a[j][i]/v + for k in range(i+1,n):a[j][k]-=fac*a[i][k] + return det + +def rank(a): + a=[list(map(Q,r)) for r in a]; lead=0 + for col in range(len(a[0])): + p=next((j for j in range(lead,len(a)) if a[j][col]),None) + if p is None:continue + a[lead],a[p]=a[p],a[lead] + v=a[lead][col];a[lead]=[x/v for x in a[lead]] + for j in range(lead+1,len(a)): + v=a[j][col] + a[j]=[x-v*y for x,y in zip(a[j],a[lead])] + lead+=1 + if lead==len(a):break + return lead + +def cyclic_distance(x,y,n): + v=abs(x-y) + return min(v,n-v) + +def run(): + base=uniform_moments(2) + check('base_even_moments',[base[i] for i in (0,2,4,6)]==[Q(1),Q(2),Q(34,5),Q(26)]) + mb=lambda i,j:base[i]*base[j] + monomials=[(i,d-i) for d in range(4) for i in range(d+1)] + vander=[[Q(x)**i*Q(y)**j for i,j in monomials] for x,y in product(range(-2,3),repeat=2)] + check('full_total_degree_three_grid_rank_ten',rank(vander)==10) + frames=[] + lower=Q(225792,625)/Q(827,10)**6 + for z in [Q(0),Q(1,3),Q(-1,3),Q(1,2),Q(-1,2),Q(1),Q(-1),Q(2),Q(3)]: + c=(1-z*z)/(1+z*z);s=2*z/(1+z*z) + for reflection in (-1,1): + label=f'{z}:{reflection}' + g=gram(c,s,mb,reflection) + piv=ldl(g); q=c*c*s*s + schur=g[3][3]-(g[3][1]**2+g[3][4]**2)/2 + check('all_frame_cubic_schur_'+label,schur==(72+1044*q)/25) + check('all_frame_determinant_'+label,determinant(g)==Q(3136,625)*(72+1044*q)) + trace=sum(g[i][i] for i in range(7)) + check('all_frame_trace_'+label,trace==(243+682*q)/5 and trace<=Q(827,10)) + check('all_frame_LDL_positive_'+label,all(p>0 for p in piv)) + check('det_trace_uniform_lower_'+label,determinant(g)/trace**6>=lower) + frames.append({'t':str(z),'reflection':reflection,'c':str(c),'s':str(s),'conditional_cubic_variance':str(schur),'determinant':str(determinant(g))}) + # An insufficient support genuinely loses the axial cubic row. + thin=uniform_moments(1);thin_gram=gram(Q(1),Q(0),lambda i,j:thin[i]*thin[j]) + annih=[Q(0)]*7;annih[1]=annih[3]=Q(1) + check('negative_K1_Fu_plus_Fuuu_annihilator',sum(annih[i]*thin_gram[i][j]*annih[j] for i in range(7) for j in range(7))==0) + family=[] + for n in [4,5,7,11,17]: + eps=Q(1,n**12);fill=uniform_moments(2*n,n) + mm=lambda i,j:(1-eps)*mb(i,j)+eps*fill[i]*fill[j] + # Twelfth moment requires moments through12, independent of the Gram helper. + fm=[sum(Q(i,n)**p for i in range(-2*n,2*n+1))/Q(4*n+1) for p in range(13)] + bm=[sum(Q(i)**p for i in range(-2,3))/5 for p in range(13)] + m12=sum(Q(comb(6,j))*((1-eps)*bm[2*j]*bm[12-2*j]+eps*fm[2*j]*fm[12-2*j]) for j in range(7)) + check('positive_normalized_mass_'+str(n),sum(eps/Q((4*n+1)**2)+(1-eps)/25*int(i%n==0 and j%n==0) for i,j in product(range(-2*n,2*n+1),repeat=2))==1) + check('uniform_M12_'+str(n),m12<=8**6==262144) + check('strict_all_mode_fill_'+str(n),eps/Q((4*n+1)**2)>0 and eps<1) + for z in [Q(0),Q(1,2),Q(2)]: + c=(1-z*z)/(1+z*z);s=2*z/(1+z*z) + g=gram(c,s,mm); gb=gram(c,s,mb);gf=gram(c,s,lambda i,j:fill[i]*fill[j]) + check('mixture_covariance_'+str(n)+':'+str(z),all(g[i][j]==(1-eps)*gb[i][j]+eps*gf[i][j] for i,j in product(range(7),repeat=2))) + ldl(g) + residues=[sum(i%n==r for i in range(-2*n,2*n+1)) for r in range(n)] + check('geometric_cycles_not_primitive_only_'+str(n),residues==[5]+[4]*(n-1)) + # The cycles evaluate to zero at every nontrivial Nth root, primitive or not. + corr=1-eps+eps/Q(4*n+1) + check('unit_two_value_Rayleigh_'+str(n),1-corr==eps*(1-Q(1,4*n+1))=Q(1,4)) + check('coefficient_C4_dimension_bound_'+str(n),eps*(4*n+1)**2<=Q(25,n**10)) + check('proper_N_cubed_lifetime_scale_'+str(n),Q(n*n)*Q(1,n**2)==1) + err=(4096*n**4+4096*n*n)*Q(1,n**10) + check('mean_transfer_count_cap_'+str(n),err<=Q(8192,n**6)) + check('negative_weak_probability_rate_not_enough_'+str(n),4096*n**4*Q(1,n**4)==4096) + check('negative_wrong_frequency_scaling_'+str(n),Q(2,n*n)=2, L>0 and omega=2 pi/L. Let + + S_K={-K,...,K}^2, + q_n>0, q_n=q_-n, sum_{n in S_K} q_n=1. + +Choose one member of each nonzero pair {n,-n}. With independent standard real +Gaussians C_0,C_n,S_n, put + + F(x)=sqrt(q_0) C_0 + + sum_pairs sqrt(2q_n) + [C_n cos(omega n.x)+S_n sin(omega n.x)]. (1) + +The torus is X=R^2/(L Z^2), with area V=L^2. The real coefficient vector xi +has dimension N=(2K+1)^2 and standard Gaussian Lebesgue density rho_N. +The field is centered, variance one, stationary and C-infinity. It need not +be rotationally isotropic. Every fixed C^j norm is bounded by C_j ||xi||. + +For I a Borel subset of (0,infinity), define + + N_cand(I)=sum_{p!=q, gradF(p)=gradF(q)=0} + 1{H_p<0} 1{det H_q<0} 1{F(p)-F(q) in I}. (2) + +Here H is the physical two-dimensional Hessian. The first role is a local +maximum; the second is a nondegenerate index-one saddle. Both roles are +ordered. This population includes every such positive-gap pair, without +requiring a persistence pairing or excluding a maximum because it supports +the essential class. Set mu_cand(I)=E N_cand(I)/V. + +**Candidate theorem.** For every fixed law (1), mu_cand has a finite continuous +density nu_cand(ell) on ell>0, and + + nu_cand(ell) ~ C_loc ell^(-1/3), ell decreasing to zero, + 0=2,L,q above. They are the named +inputs consumed here, with short mechanisms recorded to make their scope +auditable: + +**R1 — seven contact rows (FG Sections 2–3).** +(F,F_u,F_uu,F_uuu,F_v,F_uv,F_vv) has positive unconditional covariance in every +orthonormal frame. Its Fourier multiplier is a polynomial of degree at most +three in each original frequency coordinate. If its variance is zero, positivity +of every q_n makes that polynomial zero on the full square grid. At least five +points in each coordinate force polynomial identity, and its distinct +homogeneous terms force every coefficient to vanish. Frame compactness gives +a contact covariance floor. Literal divided-difference coefficient matrices +extend continuously to contact, so that floor persists for every sufficiently +small r, uniformly over basepoints and frames. + +**R2 — original two-site first jets (FG Section 7).** +For every distinct torus pair, + + O_t=(F(p),gradF(p),F(q),gradF(q)) + +has positive covariance. Choose a coordinate in which the torus sites differ, +and let its relative phase be t!=1. Holding the other frequency fixed, a +zero-variance combination has form (a+bj)+(c+dj)t^j. Four consecutive j values +give the invertible confluent Vandermonde determinant t(t-1)^4. Varying the +other frequency then kills the remaining value and transverse-gradient +coefficients. K>=2 suffices. On any fixed compact off-diagonal domain this +six-row covariance and its four-gradient submatrix have uniform floors. +The two endpoint heights conditional on the four gradients have positive +two-dimensional covariance by the Schur complement. + +**R3 — actual conditional law and moments (FG Section 4).** +For any full-row-rank observation matrix R, conditioning R xi=z gives mean +R^T(RR^T)^(-1)z and centered covariance +P=I-R^T(RR^T)^(-1)R. P is an orthogonal projection. Uniform observation +covariance floors give bounded compact-target means and polynomial growth +in arbitrary targets. Finite coefficient expansion gives every fixed-order +field moment. Further scalar regression supplies the independent transverse +residual used in FG; no arbitrary residual jet rank is asserted. + +**R4 — literal contact normalizer (FG Section 5 with E1).** +For M=x-r u/2,S=x+r u/2, exact pins + + F(M)=b, F(S)=b-k r^3, gradF(M)=gradF(S)=0 + +give U_r target v_r=(b-k r^3/2,-k r^2,0,12k,0,0). The original six-pin +observation transform has absolute determinant 12 r^-5. With + + W_r=|det H_M det H_S| 1{H_M<0} 1{det H_S<0}, + Z_r=E_Q W_r, + +one has, uniformly on compact b windows and 0 z_0=36 k^2 E[A_0^2 1{A_0<0}]>0, (6) + +where A_0 is F_vv conditioned on U_0=(b,0,0,12k,0,0). +Endpoint gradient averages give |F_uu(i)|/r,|F_uv(i)|/r<=M_3/2, while the +height pins give F_uu(M)/r -> -6k and F_uu(S)/r -> 6k. The correct congruence +diag(r^-1/2,1) has off-diagonal sqrt(r) beta. The transverse Schur variance, +Gaussian density bound and uniform moments give type convergence and uniform +integrability. This is compact-mark convergence, not a uniform inverse-Z +bound over all marks. + +**R5 — all-mark unnormalized envelope (FG Section 6).** +On one small band 00, original six-pin density at +(b,0,b-ell,0), 00 and s_u^2>0. The centered scalar law A_u|V_u=0 has +E[A_u^2 1{A_u<0}|V_u=0]=s_u^2/2. These factors are continuous in the frame. + +No part of this proof consumes positive covariance of both full two-site +second jets. That stronger fact holds at K>=3 and fails in an axial K2 +example. Neither it nor an arbitrary finite-jet premise is needed here. + +## 4. Sard genericity before any Rice argument + +This step avoids a Morse-to-Rice-to-Morse cycle. At a physical point p, write + + gradF(p)=B(p) xi. + +B(p) has row rank two: positive coordinate-mode weights give positive +gradient variance in every nonzero direction. Cover the torus by coordinate +patches on which some two coefficient columns B_C(p) are invertible. A +finite or countable such cover exists by continuity and the finite set of +possible minors. In each patch split xi=(zeta,eta), with zeta in R^2. +The critical-point equation is equivalent to + + zeta=a(p,eta)=-B_C(p)^(-1) B_D(p) eta. + +The smooth map + + Psi(p,eta)=(a(p,eta),eta) + +has N-dimensional domain and target. Differentiating the constraint at +xi=Psi(p,eta), while taking the physical derivative of gradF at fixed xi, +gives + + D_p a=-B_C(p)^(-1) H_F(p), + |det D Psi|=|det H_F(p)|/|det B_C(p)|. (8) + +If a field has a degenerate critical point in this patch, its coefficient +vector is a critical value of Psi. Sard makes those vectors Lebesgue-null. +A countable patch union is still null, and the coefficient law has Lebesgue +density rho_N. Thus F is Morse almost surely, before any counting formula. + +The critical set is closed in the compact torus. Nondegenerate critical +points are isolated; an infinite critical set would accumulate at a critical +point and contradict local uniqueness. Hence the set is finite. + +The Morse coefficient set is open. Near a Morse coefficient vector, finitely +many implicit-function critical branches persist; outside their chosen +neighborhoods the gradient stays bounded away from zero. Countable local +coefficient neighborhoods therefore make the resulting finite root counts, +with arbitrary Borel location/coefficient marks, Borel. On the null non-Morse +set define a count as zero. No claim about every fixed conditional pin law +has entered this argument. + +## 5. Direct coefficient-space area formula for all Borel pair marks + +Let T=(XxX) minus the torus diagonal and t=(p,q). Define + + G_t=(gradF(p),gradF(q))=R(t) xi. + +R2 implies row rank four on T. On a coordinate patch U in T choose an +invertible four-column minor R_C(t). Split xi=(zeta,eta), zeta in R^4. +Then G_t=0 is equivalent to + + zeta=a(t,eta)=-R_C(t)^(-1) R_D(t) eta, + Phi(t,eta)=(a(t,eta),eta). + +Phi is smooth from UxR^(N-4) to R^N. Its preimages of a fixed coefficient +vector correspond exactly to that field's pair-gradient zeros in U. +At constraints, + + D_t a=-R_C(t)^(-1) D_tG, + D_tG=diag(H_p,H_q), + |det D Phi|=Delta_t/|det R_C(t)|, + Delta_t=|det H_p det H_q|. (9) + +Apply (AF) with rho_N(Phi(t,eta)) times any nonnegative Borel mark w(t,xi) +and a Borel restriction to a chosen subset E of U. Initially allow both sides +to be infinite. It gives + + E sum_{t in E:G_t=0} w(t,xi) + = integral_E integral_R^(N-4) + rho_N(a(t,eta),eta) w(t,(a(t,eta),eta)) + Delta_t / |det R_C(t)| d eta dt. (10) + +Localize in t and eta to apply the locally Lipschitz theorem, then use +nonnegative monotone convergence. A countable Borel partition of T subordinate +to the minor patches avoids counting a pair twice. + +For fixed t, the linear transformation (zeta,eta) -> (G_t,eta) has absolute +Jacobian |det R_C(t)|. Thus + + integral rho_N(a(t,eta),eta) f(a(t,eta),eta) + / |det R_C(t)| d eta + =p_Gt(0) E[f(xi)|G_t=0], (11) + +using the canonical Gaussian regression law. Substituting f=Delta_t w and +combining the patches yields the all-Borel pair Rice identity + + E sum_{t in D:G_t=0} w(t,xi) + =integral_D p_Gt(0) E[Delta_t w(t,xi)|G_t=0] dt. (12) + +The root-count formula was not assumed to establish genericity or finiteness. +On a compact D separated from the diagonal, the four-gradient covariance +floor bounds p_Gt(0), and conditional centered coefficients have covariance +<=I and mean zero. Delta_t is bounded by C||xi||^4. Consequently (12) with +w=1 is finite on D. This derives the local finiteness needed by a separate +E2 two-finite-measure proof; it is not a hidden premise of (10). + +Sard applied also to Phi shows regular pair roots almost surely directly. +The single-point construction in Section 4 already establishes that result +via a Morse field, so neither direction relies on a Rice/genericity cycle. + +The typed mark is + + tau_t=1{H_p<0} 1{det H_q<0}, + W_t=Delta_t tau_t. + +It is Borel, and its type sets are open. In fact W_t, unlike the bare +indicator, is continuous in both Hessians: changing type at a boundary +requires a zero determinant, which kills the jump. It has degree-four +polynomial growth. Equation (12) contains Delta exactly once, not an +additional determinant inserted into the mark. + +## 6. Distinct critical values and the independently checked AAL/E2 route + +R2 makes the two endpoint heights conditional on G_t=0 nondegenerate Gaussian. +Thus the conditional probability of F(p)=F(q) is zero for each distinct t. +Since Delta_t is finite and integrable under that conditional law, (12) with +w=1{F(p)=F(q)} gives expected tie count zero on every compact off-diagonal +domain. Countably exhaust T, for example by torus separation at least 1/j +and finitely many coordinate domains. No two distinct critical points have +equal heights almost surely. This proof uses two-site first jets only. + +For comparison, actual AAL hypotheses hold: G is C-infinity and Gaussian; +Cov(G_t) is positive definite; its density has a compact-domain bound; and + + law(xi|G_t=z)=law(R(t)^T(R(t)R(t)^T)^(-1)z+P_t eta), + P_t=I-R(t)^T(R(t)R(t)^T)^(-1)R(t) + +is continuous in (t,z) through the whole C^j field. These verify the Gaussian +base and conditional requirements. The continuous weighted base can use xi +as a constant extra Gaussian field. Its bounded continuous cylinders generate +Borel(DxR^N); equality of the two finite counting/kernel measures then extends +to the typed and height marks by E2's monotone-class mechanism. +This is an independently checked alternative to Section 5, not an assertion +that AAL's lower-semicontinuity condition covers every arbitrary Borel mark. +E2's global elder event is not used. + +## 7. Height disintegration and an actual density version + +For t=(p,q), let Q^t_(b,s) be continuous Gaussian regression on the six +observations O_t=(F(p),gradF(p),F(q),gradF(q)) at (b,0,s,0). Define + + K_t(b,s)=p_Ot(b,0,s,0) E_Q^t_(b,s) W_t. (13) + +The order of observations may be permuted, and physical gradients rotated +orthogonally, with absolute Jacobian one. The six-row rank makes the Gaussian +height disintegration legitimate. Equation (12), (11) and nonnegative +Tonelli give the expected typed-pair height measure with kernel K_t(b,s). +Push it forward under ell=b-s; ds has absolute Jacobian one. Therefore + + nu_cand(ell)=(1/V) integral_T integral_R + K_(p,q)(b,b-ell) db dp dq, ell>0, (14) + +and mu_cand(I)=integral_I nu_cand(ell) d ell for every positive-lifetime +Borel I. Initially this is a nonnegative density version, possibly infinite; +the next sections prove finiteness, continuity and its asymptotic directly. +No differentiation of a cumulative asymptotic is used. + +For distinct sites, canonical conditional coefficient means and covariance +square roots depend continuously on locations and heights. The continuous +weighted type W and degree-four growth, with finite Gaussian moments, imply +continuity of its conditional expectation. This argument allows singular +conditional Hessian covariance; it does not need full two-site second-jet rank. + +## 8. Near-diagonal ledger and dominated convergence + +Choose the R1/R5 band 00. The height change has absolute Jacobian r^3. +The literal P contact transform and R4 give original-pin density +12 r^-5 pi_r(v_r), v_r=(b-k r^3/2,-k r^2,0,12k,0,0). +The determinant expectation is Z_r. The factors are thus + + (r dr d sigma) (r^3 db dk) (12 r^-5 pi_r) Z_r + =r A_r dr db dk d sigma, + A_r=12 pi_r Z_r/r^2. (15) + +Endpoint roles are ordered. u ranges once over the full circle; no factor +1/2 or normalized-angle convention is introduced. + +At fixed k the pushforward ell=k r^3 has +dr/dell=1/(3k r^2). Formula (15), or equivalently the fixed-ell change +r=(ell/k)^(1/3), gives + + ell^(1/3) nu_near(ell) + =integral_Rx(0,infinity)xS^1 + 1{k>=ell/r_0^3} A_(ell/k)^(1/3)(b,k,u) + /(3k^(2/3)) db dk d sigma. (16) + +For each fixed b,k>0,u, R4 gives A_r -> 12 pi_0(v_0) z_0, and the indicator +in (16) tends to one. R5 supplies the integrable majorant H/(3k^(2/3)). +Dominated convergence gives + + ell^(1/3) nu_near(ell) -> + 4 integral k^(-2/3) pi_0(v_0) z_0 db dk d sigma + =C_loc. (17) + +This conclusion handles all birth heights and all positive k; its proof uses +the unnormalized product pi_r Z_r. There is no global inverse-Z estimate, +no division by k in a conditional probability, and no selection probability. + +Equation (16) bounds nu_near(ell) by C ell^-1/3 for every ell>0. It also proves +continuity for each positive ell: for ell_j -> ell>0, the integrand converges +at every fixed b,k,u except k=ell/r_0^3, a measure-zero boundary. Conditional +kernel continuity from Section 7 and the same H majorant permit dominated +convergence. This yields a particular continuous Radon–Nikodym density. + +## 9. Far contribution, finiteness and the full theorem + +On dist(p,q)>=r_0 the six-pin covariance has a compact-domain floor. +For 00. + +There is also no hidden infinite-total-count premise. For the near population, +integrating (15) over all positive gaps and all b,k is bounded by +integral_0^r0 r dr times integral H db dk d sigma, which is finite. +For the far population, the target (b,s) has Gaussian decay in both heights, +and its conditional determinant moment has degree-four polynomial growth; +the db ds integral is finite. Hence mu_cand((0,infinity)) 0. + +Each C_loc,j is positive; taking a finite minimum also gives uniform relative +leading asymptotics across these twenty cases. There is no asserted uniform +window or constant over arbitrary q approaching the boundary of the positive +weight simplex, over unbounded K, or over variable L. For each fixed positive +q, integer K>=2 and L>0, the theorem holds with its own constants. + +The K1 contact dependence and the K2 second-jet singularity in FG are retained +negative boundaries. The seventh-order axial annihilator at K3 blocks reuse +of P's arbitrary finite-jet premise but does not obstruct this finite list +of ranks. No counterexample to (3) under its stated assumptions was found. + +## 12. Exact remaining scope + +This preparation supplies a complete conventional candidate-count derivation +from the named FG inputs and the stated classical Sard/area imports. There +is no additional unproved model-specific analytic premise in that route. +Independent exact-source review is recorded separately in [REVIEW.md](REVIEW.md); +the earlier review of FG and pre-file route discussions do not substitute +for an incorporated-source review. No formal Lean implementation or kernel +execution is claimed. + +Actual persistence bars require further source-bound elder selection, +global once-counting and the appropriate complement proof. Conditional +all-point genericity for enlarged cap/collision/factorial-moment lists has +not been proved here and is not silently imported. Higher homology and a +process limit remain separate. This theorem also does not certify numerical +sampling/coupling, a usable asymptotic window, blind-field validation, +publication acceptance or external uptake. + +In particular the actual finite-word sampler is not the Gaussian coefficient +law used by Sard/null-set and Gaussian density arguments. The non-Gaussian +coefficient laws are not covered. Exact mathematical weights, actual +conditioning, density versus cumulative normalization, field-dependent +coefficient and per-area versus whole-torus counts must be preserved in any +consumer. Existing catalogue admissibility labels, applicability arrays, +graphs and scientific-status records remain unchanged. diff --git a/experiments/universality/fold_isolation_diagonal/experiments/universality/finite_gaussian_contact/PROOF.md b/experiments/universality/fold_isolation_diagonal/experiments/universality/finite_gaussian_contact/PROOF.md new file mode 100644 index 000000000..5f3452013 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/experiments/universality/finite_gaussian_contact/PROOF.md @@ -0,0 +1,554 @@ +# Positive finite Gaussian spectra: the contact and regression interfaces + +Actual author: OpenAI/Codex /root, task +01a11cbf-068d-7102-861e-75814e715c98, 8 October 2026. +This proof discharges named analytic interfaces for the ideal finite Gaussian +fields in the [fifty-law catalogue](../models.json). Its review is recorded in +[REVIEW.md](REVIEW.md). It is a source proposal, not a scientific-status +register, a Lean proof, or an admission to the full persistence theorem. + +## 1. Source, field and exact scope + +The consumed parent P is +[Math-@13af1089fd7991105a3d8828539bdd9a41ca87c3, +UNIFORM_MATRIX_CAP_AND_LIFETIME.md](https://github.com/d6g8k5htny-coder/Math-/blob/13af1089fd7991105a3d8828539bdd9a41ca87c3/imports/lifetime_parent_20260925/UNIFORM_MATRIX_CAP_AND_LIFETIME.md): +Git blob dfed3b8d318a3ab1950957f393307733a4bef3f2, 40,261 UTF-8 bytes, +SHA256 9350ad6eaba6626b93c3dedeef9e2ff816e5cdf1c8318e85fb27499141c84bc7. +Root fetched its complete content and independently reproduced that byte count +and SHA256. The proof uses P's literal rows (3.1), target (3.3), transverse +block at M, endpoint identities (5.1)-(5.3), and the unnormalized intensity +convention (13.4). The finite support argument below replaces the named uses +of P §2; it does not inherit §2's arbitrary finite-jet assertion. + +Read P with its [D1 reconciliation](https://github.com/d6g8k5htny-coder/Math-/blob/13af1089fd7991105a3d8828539bdd9a41ca87c3/reviews/d1_chain_reconciliation_20260928/RECONCILIATION.md) +and its complete reading-rule identity set. The consumed congruence correction +is [E1](https://github.com/d6g8k5htny-coder/Math-/blob/13af1089fd7991105a3d8828539bdd9a41ca87c3/imports/lifetime_parent_20260925/ERRATUM_CONGRUENCE.md), +blob 213594d6ca6a86fb938110f4d166d9ce275a02d0, 1,782 bytes, +SHA256 bad7ef609c4ad8c41ad6af562c1b6807921e19a9d556ed793ad1a0db6e202028; +root read and independently checked its complete bytes. D1 also names +[CAP](https://github.com/d6g8k5htny-coder/Math-/blob/13af1089fd7991105a3d8828539bdd9a41ca87c3/imports/lifetime_parent_20260925/MARKED_CYLINDER_CAP_PROOF.md), +blob 0633aca3c2a2882b0de4399da0a75d64c2e6b2e1, SHA256 +0bf922b9203c29088b12388807aa0e2ecd020485eb0f6e919679841b5b2636fc, +and [E2](https://github.com/d6g8k5htny-coder/Math-/blob/13af1089fd7991105a3d8828539bdd9a41ca87c3/reviews/d1_section9_borel_repair_20260925/REPAIR.md), +blob fe9b9ce4999908bb3814b500ee2d0ceb0c6f704a, SHA256 +845abf9f9c99d672c2a10a887b5a2e7206a3d2de3d876f35f75ff6e2dc13e62f, +plus its own W1/embedding correction. CAP and E2 are named dependencies for +the remaining bar-selection/counting obligations, not results transferred to +the finite model by this proof. Their identities here are read from D1. +Our explicit finite-coefficient coupling below supplies its stated convergence +in probability; it does not silently reuse the parent's unamended wording. + +Fix L>0, omega=2 pi/L, an integer K>=2, and the square support + + S_K = {-K,...,K}^2. + +Take fixed real weights q_n>0 with q_n=q_-n and sum q_n=1. +Let P_K contain one member of each nonzero pair {n,-n}. For independent +standard real Gaussians C_0, C_n, S_n, define the ideal field on the square +torus X=R^2/(L Z^2) by + + F(x) = sqrt(q_0) C_0 + + sum_{n in P_K} sqrt(2 q_n) + [C_n cos(omega n.x) + S_n sin(omega n.x)]. (1) + +There are (2K+1)^2 real Gaussian coordinates. F is centered, stationary and +variance one. Translation rotates each independent sine/cosine pair +orthogonally; this proves stationarity. General spatial rotations need not +preserve its covariance. + +At L=24 and K=2 or 3, each of the ten catalogue spectral formulas is strictly +positive and symmetric on this support. Thus the results apply to the ten +ideal Gaussian coefficient laws, including gaussian__gaussian. The constants +can be chosen uniformly across this finite set of twenty fixed profiles. +They are not uniform over all possible positive spectra, K, L or dimension. +No coefficient draw or floating PRNG realization is certified by (1). +The rounded weights retained by models.py and its continuum-admissibility +label remain unchanged. + +Let (u,v) be any orthonormal frame, including reflected frames, and x any +basepoint. For 00. + +The last line denotes continuous Gaussian regression on these observations; +it is not an event of positive probability. All conclusions below specify +the exact conditioning coordinates. + +We prove: + +* Positive contact covariance of the seven named rows for every frame. +* Uniformly nonsingular small-r divided-difference observations, including + the transverse Hessian at M. +* Uniform compact-mark Gaussian regression, independent residual and endpoint + normalizer bounds. +* An integrable unnormalized all-mark intensity envelope. +* Positive original value-and-gradient covariance at every two distinct sites, + and its uniform bound on a fixed off-diagonal domain. +* Finiteness and positivity of the local coefficient functional. + +None of these conclusions supplies the actual-bar selection probability, +a global once-counted persistence identity, higher homology, a process limit, +or a usable numerical asymptotic window. + +## 2. The named contact covariance + +In frame coordinates, define the UNCONDITIONED observation vector + + J_0 = (F, F_u, F_uu, F_uuu, F_v, F_uv, F_vv)(x). + +For real lambda in R^7, let a=n.u and c=n.v. The Fourier multiplier of the +corresponding derivative functional is + + P_lambda(n) = + lambda_0 + i omega lambda_1 a - omega^2 lambda_2 a^2 + - i omega^3 lambda_3 a^3 + i omega lambda_4 c + - omega^2 lambda_5 a c - omega^2 lambda_6 c^2. + +The covariance from (1) gives exactly + + Var(lambda.J_0) = sum_{n in S_K} q_n |P_lambda(n)|^2. (2) + +This formula includes both members of each mode pair and the constant. +A basepoint phase has modulus one and does not change (2). + +If the variance vanishes, every P_lambda(n) vanishes because all q_n>0. +As a polynomial in the original two frequency coordinates, P_lambda has +degree at most three in each coordinate. For every fixed second coordinate +on the grid it has at least five distinct roots in its first coordinate; +all coefficients in the first coordinate vanish on the second grid. +Applying the same root argument in the second coordinate shows that +P_lambda is identically zero. The invertible frame change then gives an +identity in the independent variables a,c. Its homogeneous cubic, quadratic, +linear and constant terms force, respectively, + + lambda_3=0; + lambda_2=lambda_5=lambda_6=0; + lambda_1=lambda_4=0; + lambda_0=0. + +Hence Gamma_0=Cov(J_0) is positive definite. Its entries depend continuously +on the frame. Compactness of O(2) gives a positive minimum of its least +eigenvalue. This argument preserves the actual directional covariance; +it never replaces a square-torus covariance by an isotropic one. + +In particular the six-row subvector U_0 obtained by dropping F_vv has +positive covariance. The Schur complement proves + + Var(F_vv | U_0)>0. (3) + +Under regression U_0=(b,0,0,12k,0,0), the six conditioned coordinates are +deterministic. It would be false to assert positive covariance of all seven +rows under that conditioned law. Equations (2)-(3) concern unconditional +covariance and the remaining conditional scalar, respectively. + +## 3. The literal divided differences and their contact limit + +Use P's row order, with A_r the transverse curvature at the left pin: + + U_r = ( + [F(M)+F(S)]/2, + [F(S)-F(M)]/r, + [F_u(S)-F_u(M)]/r, + (6/r^2)[F_u(M)+F_u(S)-2(F(S)-F(M))/r], + [F_v(M)+F_v(S)]/2, + [F_v(S)-F_v(M)]/r), + A_r = F_vv(M), J_r=(U_r,A_r). (4) + +The exact target is + + v_r=(b-k r^3/2, -k r^2, 0, 12k, 0, 0). (5) + +The original six observations and U_r are related by an invertible linear +map for r>0, with absolute determinant 12 r^-5. No approximate target or +missing factor 12 is used. + +Every difference row has an integral representation. With -1/2<=s<=1/2, + + (U_r)_1 = integral F_u(x+r s u) ds, + (U_r)_2 = integral F_uu(x+r s u) ds, + (U_r)_5 = integral F_uv(x+r s u) ds, + (U_r)_3 = 6 integral (1/4-s^2) F_uuu(x+r s u) ds. (6) + +The fourth identity follows by integrating twice by parts: + + integral_M^S (t+r/2)(r/2-t) F_uuu(x+t u) dt + = r[F_u(M)+F_u(S)]-2[F(S)-F(M)], + +where t is measured from x. Its kernel has integral one and is nonnegative. +The other two U rows are endpoint averages. Thus J_r converges to J_0 in +mean square uniformly over basepoints and frames. For example, + + |(U_r)_3-F_uuu(x)| <= (3r/16) ||F||_{C^4,dir}. + +The other rows follow from the fundamental theorem of calculus and endpoint +bounds. Here a coordinate C^q norm and its directional version differ by +fixed finite constants. + +More explicitly, with xi the finite standard Gaussian coefficient vector, +write J_r=L_r(x,u,v) xi. Equation (6) gives a continuous extension L_0 at r=0 +and uniform convergence on the compact set of basepoints and frames. +Consequently Gamma_r=L_r L_r^T converges uniformly to Gamma_0. +There exist r_*>0 and 0=1, + + sup_{r,x,frame,b in B,k in [k_-,k_+]} + E_Q (1+||F||_{C^q})^p < infinity. (8) + +Write mu_r(z)=E_Q F(z), a_r=E_Q A_r, and s_r^2=Var_Q A_r. +The Gaussian regression decomposition is an actual whole-field identity: + + F(z)=mu_r(z)+B_r(z)(A_r-a_r)+g_r(z), + B_r(z)=Cov_Q(F(z),A_r)/s_r^2. (9) + +The centered residual coefficient vector is orthogonal to the scalar +coefficient vector producing A_r-a_r. Joint Gaussianity makes the entire +residual coefficient vector, and hence the smooth field g_r, independent +of A_r. This is not a claim that its internal derivatives or evaluations +are independent. + +Equation (7), the finite expansion and bounded derivative cross-covariances +bound B_r and all fixed-order derivatives uniformly. All C^q moments of g_r +are uniformly bounded. The law of A_r has variance bounded above and away +from zero and bounded mean, so on the compact mark rectangle + + density_Q(A)<=C exp(-c A^2). + +For a fixed coordinate norm, J=1+||g_r||_{C^4} is independent of A_r, +has every finite moment uniformly bounded, and (9) yields + + M_3,M_4 <= C(J+|A_r|), (10) + +where M_j bounds directional order-j derivatives on the local cylinder. +No inverse field-Hessian moment or division by a normalizer is used. + +## 5. Compact-mark endpoint normalizer + +Define H_i as the physical Hessian in frame coordinates at i=M,S, +and let + + alpha_i=F_uu(i)/r, beta_i=F_uv(i)/r, A_i=F_vv(i), + H_i=[[r alpha_i,r beta_i],[r beta_i,A_i]], + W_r=|det H_M| |det H_S| + 1{H_M negative definite} + 1{H_S has exactly one negative eigenvalue}, + Z_r=E_Q W_r. (11) + +All matrices are physical field Hessians, not observation covariances. +The exact pins and integral remainder identities give + + |alpha_i|,|beta_i|<=C M_3, + |A_S-A_M|<=r M_3, + |alpha_M+6k|,|alpha_S-6k|<=r M_4/2, + det H_i/r=alpha_i A_i-r beta_i^2. (12) + +These are P's endpoint identities specialized to dimension two. The two +gradient pins make the interval averages of F_uu and F_uv zero; subtracting +each average bounds its endpoint by the derivative Lipschitz bound. +The weighted third-derivative identity from (6) and target 12k give the +two axial limits with the displayed error. No simultaneous vector Rolle +zero is assumed. + +Let A_0 have the conditional law of F_vv(x) given +U_0=(b,0,0,12k,0,0). Its variance is positive by (3). +Conditional means and covariances of A_r converge uniformly on compact marks. +One can couple the finite conditional Gaussian coefficient vectors using +their covariance square roots and a common standard Gaussian vector. +Continuity of positive semidefinite square roots gives uniform L^p convergence +for each finite p. Equations (8),(12) then show that the congruence-scaled +Hessians converge to + + diag(-6k,A_0), diag(6k,A_0). + +Specifically the scaled matrix is diag(r^-1/2,1) H_i diag(r^-1/2,1); +its off-diagonal entry is sqrt(r) beta_i, as required by E1. On k>=k_->0, the limit is +nonsingular except when A_0=0. The uniform bounded Gaussian density gives +sup P(|A_0|<=epsilon)<=C epsilon. The type indicators therefore converge +uniformly in probability, and the limit of their product is 1{A_0<0}. +Uniform integrability follows from (8),(12). Hence + + Z_r/r^2 -> z_0(b,k,u,v) + =36 k^2 E[A_0^2 1{A_0<0}], (13) + +uniformly on the compact mark rectangle and frames. z_0 is continuous +and strictly positive; compactness yields a positive minimum. After reducing +r_* for this rectangle, Z_r/r^2 has a positive uniform lower bound. + +The positive normalizer in this conclusion is compact-mark and small-r. +It neither certifies a specified band nor supplies a globally uniform +inverse-Z bound as b,k become unbounded. + +## 6. The unnormalized all-mark envelope + +The radius and covariance bounds in (7) do not depend on b,k. +For every b in R and k>0, equation (5), with r<=1, gives + + |v_r|<=C(|b|+k), |v_r|^2>=c(b^2+k^2). + +For the lower bound, b^2<=2(b-k r^3/2)^2+k^2/2 and the fourth +coordinate is 12k. Thus the six-dimensional Gaussian pin density satisfies + + pi_r(v_r)<=C exp[-c(b^2+k^2)]. + +The conditional coefficient mean is O(|b|+k) and its centered covariance +is bounded by I, so the finite-expansion proof yields + + E_Q (1+||F||_{C^3})^p <= C_p(1+|b|+k)^p. + +Using only the gradient-average bounds in (12), r<=1, and +|A_i|<=C||F||_{C^2}, each |det H_i|/r is at most +C(1+||F||_{C^3})^2. Therefore + + 0<=12 pi_r(v_r) Z_r/r^2 + <=C(1+|b|+k)^4 exp[-c(b^2+k^2)] =: H(b,k). (14) + +The product k^-2/3 H(b,k) is integrable over R x (0,infinity). +Near zero the exponent -2/3 is greater than -1; at infinity Gaussian +tails absorb the polynomial. This is exactly the unnormalized +typed-candidate amplitude envelope of P (13.4), with d=2. +It is not an estimate of a conditioned selection probability. +No division by k, by Z_r or by the transverse Hessian occurs in its proof. + +## 7. Two distinct sites and the off-diagonal interface + +Let x!=y as points of X. The unconditioned vector + + (F(x),grad F(x),F(y),grad F(y)) + +has positive covariance. To prove it, translate x to zero and choose a +coordinate in which x,y differ modulo L, say the first. Put +t=exp(i omega(y_1-x_1)), so t!=1. A zero-variance real combination would +annihilate every mode and, for each fixed n_2, have the form + + (a+b n_1)+(c+d n_1)t^n_1=0 (15) + +on the mode grid. Its coefficients include the nonzero phase from n_2 +and the linear dependence on n_2 of the transverse gradient coefficient. + +Four consecutive integers n_1 suffice to distinguish these four sequences. +Indeed the matrix with columns 1,j,t^j,j t^j, j=0,1,2,3, has determinant +t(t-1)^4!=0. Starting at another integer only gives invertible shifts +and nonzero factors. K>=2 supplies such a consecutive block. +Thus a=b=c=d=0 for each fixed n_2. Taking two different n_2 values +then forces both value and both transverse derivative coefficients to +vanish, as well as the two axial coefficients. The combination is zero. + +The same argument in the other coordinate covers a difference only there. +This is a full two-site first-jet proof, not a sampled covariance check. +For torus distance at least a fixed rho>0, compactness yields a uniform +positive covariance floor. Conditional finite coefficient moments then +grow polynomially in the original height target (b,b-ell), and its +density decays like exp(-c b^2) for 0=3 the bounded argument also proves independence of both full +two-site jets through order two: the two polynomials in n_1 then have +degree at most two, requiring six consecutive frequencies. The confluent +Vandermonde matrix for roots 1,t, each of multiplicity three, is invertible. +Three different n_2 values force every remaining polynomial coefficient +to vanish. This extension must not be asserted for K=2; Section 9 gives +a counterexample. + +## 8. The local jet coefficient functional + +The parity of the covariance in (1) makes odd-order jets independent +of even-order jets. Let + + G=(F_u,F_v), V=(F_uu,F_uv), T=F_uuu, A=F_vv, + tau_u^2=Var(T|G=0), s_u^2=Var(A|V=0). + +The seven-row covariance proves tau_u^2>0 and s_u^2>0. +Both are continuous in the frame. Conditional A given V=0 is a centered +scalar Gaussian, so its negative-curvature cone moment is exactly + + D_u=E[A^2 1{A<0}|V=0]=s_u^2/2. (16) + +The local coefficient functional obtained from P's exact contact ledger is + + C_loc = Gamma(7/6)/(24^(1/3) sqrt(pi)) + integral_{S^1} p_G(0) p_V(0) + tau_u^(4/3) (s_u^2/2) d sigma(u). (17) + +Every factor is finite, continuous and positive, so 00, k>0, + + A_M<=-a, + r M_3<=a/2, r M_4<=k, r M_3^2<=a k. (19) + +Then A_S<=-a/2. For A_i<0 the Schur pivot of H_i is + + s_i=r alpha_i+r^2 beta_i^2/(-A_i), + det H_i=A_i s_i. + +At M, (12) and (19) imply + + s_M/r<=-6k+k/2+r M_3^2/(4a)<=-21k/4. + +At S the same bounds give s_S/r>=6k-k/2=11k/2. +Thus H_M is strictly negative definite and H_S has one negative eigenvalue; +both physical Hessian determinants are nonzero. More explicitly, + + det H_M >= (21/4) a k r, + -det H_S >= (11/4) a k r. (20) + +This uses literal physical derivatives of the conditioned field and actual +criticality from the pins. Fixed positive a,k windows and finite derivative +bounds supply a sufficient small-r band through (19). They supply no +r-independent physical Hessian floor. Exact source-mode encoding, derivative +identification and a Lean application of C210 remain separate formal work. + +Height/gradient pins and negative transverse curvature alone do not suffice. +For fixed r,k>0, let z=x+r/2 and consider the deterministic local polynomial + + f(x,y)=b+(20k/r^2)(z^5/5-r z^4/4)-y^2/2. (21) + +Its axial derivative is (20k/r^2) z^3(z-r). At M, z=0, +f=b and grad f=0; the leading axial term is -(5k/r)z^4. +Thus M is a strict degenerate local maximum, with f_xx(M)=0 and +det H_M=0 despite A_M=-1. At S, z=r, f=b-k r^3, +grad f=0 and f_xx(S)=20k r, so S is a saddle with A_S=-1. +The fourth-derivative bound grows like 1/r and (19) fails. +This is a deterministic local falsifier of the proposed implication, +not a positive-probability counterexample for any catalogue Gaussian law. + +Even the exact generic cubic fold forbids a uniform ordinary Morse-chart +radius about M as r goes to zero. Take + + g(x,y)=b+2k x^3-(3/2)k r^2 x-k r^3/2-y^2/2. (22) + +Its two critical points are M,S, with H_M=diag(-6k r,-1) +and H_S=diag(6k r,-1). With z=x+r/2 measured from M, + + g(x,0)-b=-z^2(3k r-2k z). + +The exact axial Morse coordinate v=z sqrt(3k r-2k z) has derivative +zero at z=r. More generally, a diffeomorphic chart transforming g-b +to a nondegenerate quadratic form has only one critical point in its +domain, so it cannot contain S. Any physical ball about M contained +in that domain therefore has radius at most r. + +For the Hessian normalization H_pullback=2J used by C210, choose the +canonical positive axial scaling P_soft=(3k r)^(-1/2). The other +critical point has normalized axial distance sqrt(3k) r^(3/2). +Any normalized source ball contained in a Morse-chart domain has +radius at most that distance. These bounds are for this explicit +canonical normalization and example; they do not evaluate C210's +returned radius or prove a bound for every permissible normalizer. +They demonstrate that a nonzero finite-r hdet cannot justify a +uniform ordinary Morse chart through the fold collision. +A fold-family chart with an explicit rescaling and parameter domain +is a different interface, still to be proved. diff --git a/experiments/universality/fold_isolation_diagonal/experiments/universality/iid_short_bar_process/PROOF.md b/experiments/universality/fold_isolation_diagonal/experiments/universality/iid_short_bar_process/PROOF.md new file mode 100644 index 000000000..cd7ef8b8a --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/experiments/universality/iid_short_bar_process/PROOF.md @@ -0,0 +1,618 @@ +# Marked iid-copy Poisson limit for actual short H0 bars + +Mathematical source by OpenAI/Codex universality_review,8October2026; +repository composition and current-source binding by OpenAI/Codex root, +task01a11cbf-068d-7102-861e-75814e715c98, standing owner authorization, +exact seven-path claim6070584604. Initial base is +main2c84170712c59d9de580c172815bd30bac5d93cd. Mathematical Sections1-8 +are byte-identical to the frozen complete process preparation. The +[review record](REVIEW.md) binds original and current exact-byte verdicts. + +The result superposes independent whole copies of the fixed torus law. +Its actual weighted cluster-avoidance proof is a substantive extra step +beyond the first moment. It supplies no spatial expanding-domain law, +regional witness closure, global second-factorial bound, numerical window, +sampler certificate, Lean proof or scientific promotion. All named +performers are AI agents with substantial source and route exposure; +organizational independence is zero. + +## 1. Scope: fixed law, empty one-field limit, nonempty independent-copy limit + +Fix an integer d>=2 and a side length L>0 before taking any limit. Set +X=R^d/(L Z^d), with its flat metric and volume L^d. The actual centered, +variance-one stationary Gaussian field has covariance + + K_L(z)=sum_(j in Z^d) exp(-|z+Lj|^2/2) + /sum_(j in Z^d) exp(-|Lj|^2/2). (1) + +Its exact Fourier weights are q_n proportional to +exp[-(2pi/L)^2 |n|^2/2] on every n in Z^d. It has its random constant +mode and the smooth version of IA below. It is neither a finite-cutoff +catalogue law nor a spatially unbounded field. + +Use ordinary finite superlevel H0 bars, with decreasing level and the +elder rule. Exclude the single essential global-maximum class. On the +Morse/distinct-critical-value locus, a finite bar has a unique younger +maximum M and its unique merging index-(d-1) saddle S, with + + b=F(M), ell=F(M)-F(S)>0. + +Let N_t be the number of these bars with 00. (2) + +Consequently, a single field's process on any fixed compact set of +rescaled lifetimes tends to the empty process, not a nondegenerate +Poisson process. Indeed P(N_(Tt)>0)<=E N_(Tt)->0 for each fixed T. +More strongly, almost every one field has finitely many finite bars, +all with positive lifetimes, so N_(Tt)=0 eventually along every t->0. +If there are no finite bars this assertion is immediate. Rescaling +lifetimes by t alone makes existing bars escape to infinity. + +The nondegenerate regime considered here is instead **iid superposition +of independent whole field copies on this same fixed torus**. Let +t_n->0 and let n be the number of copies, with + + n t_n^(2/3) -> lambda, 0infinity/lifetime limit. + +The new substantive input beyond (2) is the weighted cluster-avoidance +estimate + + E[N_(Tt) 1{N_(Tt)>=2}] = o(t^(2/3)) for every fixed T. (4) + +Section 4 proves (4) from the same physical contact coupling and global +critical-point isolation. A first moment alone does not imply (4); +Section 8 gives an explicit counterexample. We do not replace (4) by +an unproved second-factorial assertion. + +## 2. Borel bar counts and precise marks + +Let G be the open locus of C4 functions on X which are Morse with +distinct critical values. The actual law gives G probability one by +IA's finite-block/tail incidence proof, with Sard before Rice. The +intrinsic Borel elder selector sigma(M,S,f), assigned zero outside G, +is the one in ST Section 2: the death threshold is the rational-level +open-superlevel connectivity supremum to a point of higher birth. +Morse handle attachment proves its unique actual H0 once-counting. + +For completeness, N_t is a legitimate joint Borel mark in (f,t), not +an assumed measurable random root count. Around any f in G, implicit +critical branches enumerate all its finitely many critical points; +a compact gradient floor excludes additional roots outside their +neighborhoods. These branches are continuous in the C4 parameter and +their heights are continuous. The separable C4 function space has a +countable cover of such neighborhoods. On each, sum the finitely many +jointly Borel expressions + + sigma(p_i(f),p_j(f),f) + 1{00, u=h(M,S)/r in S^(d-1), k=ell/r^3>0. (5) + +The cut-locus convention will not affect the limit. In the short spatial +band rx at fixed scaled lifetime and k. We retain the actual birth mark b +and physical gap variable k in the point process; omitting k is a later +projection, not a change of conditioning. +An orthonormal completion of u can be chosen Borel, with continuous +choices on local sphere charts. Transverse orthogonal changes only +rotate the contact rows with absolute determinant one and conjugate A; +the density, determinant/type weights and full-field pin law are +unchanged. No globally continuous frame over the sphere is required. + +Define the finite random point measure + + Xi_t=sum_(finite bars) delta_(ell/t, b, k, M, u) (7) + +on E=(0,infinity)_a x R_b x (0,infinity)_k x X_x x S^(d-1). +The a coordinate denotes rescaled lifetime. E is locally compact and +second countable. Xi_t is zero on G's complement. In particular its +restriction to 00,u,x, under its actual common-field +coupling, + + F_r -> F_0 globally in C4, + V_r -> V_0=36k^2(det A_0)^2 1{A_0<0}, + E_Qr[V_r sigma_r] -> z_0(b,k,u), (9) + +and for all b in R,k>0 and common 0=ta/r_0^3} + [12 pi_r(v_r)/(3 k^(2/3))] + E_Qr[V_r sigma_r] + h(a,b,k,x-r u/2,u) da db dk dx d sigma(u), + r=(ta/k)^(1/3), 0h(a,b,k,x,u). +The absolute integrand is bounded by + + ||h||infinity a^(-1/3) H(b,k)/(3k^(2/3)), 0 integral_E h d eta, (13) + d eta=4 a^(-1/3) k^(-2/3) pi_0(v_0) z_0(b,k,u) + da db dk dx d sigma(u). (14) + +For compact-supported continuous h, (13) is vague mean-measure +convergence. The same argument for the full lifetime window h=1{a<=T} +gives its total mass, and (12) controls the noncompact b,k tails and +the a->0 tail. Hence the finite measures restricted to a<=T are tight +and converge weakly whenever their boundary at a=T is treated by a +continuous approximation. Eta has no atom at T. + +Integrating k defines + + beta(b,u)=4 integral_0^infinity + k^(-2/3) pi_0(v_0) z_0(b,k,u) dk, + d eta_(a,b,x,u)=a^(-1/3) beta(b,u) + da db dx d sigma(u). (15) + +The b,u integral of beta is c_(d,L), so + + eta{00, A_0<0 and the literal cubic axial +derivative is 12k>0. IA/ST's local vector ridge has positive third +derivative, exact zeros at the pinned M and S, and no other local +critical point. Away from the contact neighborhood, implicit branches +persist from the finitely many Morse roots, and a compact gradient +floor excludes new roots. Therefore for all sufficiently small r, +the entire critical set of F_r consists of the pinned M,S and those +old branches, with their old distinct limiting heights. + +The actual elder-transfer result makes M,S a finite actual paired bar, +not merely a typed candidate. Each maximum supplies at most one birth +bar, and each merging saddle kills at most one bar. Thus no other bar +can use M or S as an endpoint. Every other finite bar uses two old +branches. If there are at least two old branches, the minimum of their +finitely many distinct limiting height differences is positive. Their +perturbed differences are consequently bounded below by half that +minimum for all sufficiently small r. With fewer old branches, no +other finite bar can have two old endpoints. The pairing of old +branches need not be combinatorially fixed for this argument: every +possible two-old-endpoint gap has a positive limiting lower bound. +The essential class is excluded throughout. + +Now fix a in (0,T], b in R, k>0, u and x, and put +r_t=(ta/k)^(1/3). Along this actual canonical conditional coupling, +the pinned bar has lifetime k r_t^3=ta<=Tt. On V_0>0 the preceding +endpoint fact proves + + N_(Tt)(F_(r_t))=1 eventually as t->0. (18) + +The realization-dependent lower gap need not have an integrable +inverse, an explicit deterministic cutoff, or a uniform probability +rate. On V_0=0, V_(r_t)->0 regardless of the count or selector. +Consequently, for every fixed a,b,k,u,x, + + V_(r_t) sigma_(r_t) + 1{N_(Tt)(F_(r_t))>=2} ->0 almost surely, + E_Q_(r_t)[V_(r_t) sigma_(r_t) + 1{N_(Tt)>=2}] ->0. (19) + +The second line follows from the same polynomial field-norm domination +as (9), at these fixed marks. This is a statement about the specified +contact conditional coupling, not about independence of two contacts. +No disintegration of unconditional Morse genericity to all pin targets +is used; IA proves the required limiting contact genericity separately. + +Apply COUNT to the actual bar selector with the additional nonnegative +Borel mark 1{N_(Tt)>=2}. Its left side is exactly +E[N_(Tt) 1{N_(Tt)>=2}]. In (11) replace E[V_r sigma_r] by + + E[V_r sigma_r 1{N_(Tt)>=2}]. + +This modification is permitted for the full-field global mark by IA's +all-Borel finite-tail area formula. Only one Hessian determinant +product enters. The integrand's pointwise limit is zero by (19), and +its unnormalized absolute bound is (12) with ||h||infinity=1. Dominated +convergence on a in (0,T], all b,k, frames and space gives + + t^(-2/3) E[N_(Tt) 1{N_(Tt)>=2}]_near ->0. (20) + +The far **anchor** contribution counts actual far bars with the same +bounded indicator; it is bounded by their unconditional expected +number O(t). It does not multiply the far count by another count. +Combining it with (20) proves (4). + +In particular P(N_(Tt)>=2)<=E[N_(Tt)1{N_(Tt)>=2}]/2 +=o(t^(2/3)). Also E N_(Tt)-P(N_(Tt)=1) +=E[N_(Tt)1{N_(Tt)>=2}]=o(t^(2/3)), so + + P(N_(Tt)=1) ~ (3/2)L^d c_(d,L) T^(2/3) t^(2/3). (21) + +Equation (4) is deliberately not the assertion +E[N_(Tt)(N_(Tt)-1)]=o(t^(2/3)). The factor N_(Tt)-1 is unbounded; +no proof here bounds it by the existing single-contact polynomial +envelope. Rare high multiplicities could distinguish the estimates. +The process theorem needs (4), and (4) has just been derived without +that stronger factorial premise. + +## 5. Laplace functional of iid superposition + +Take n independent copies of the full actual law and their measures +Xi_(t_n)^(j), all marked in the same torus X. Put + + S_n=sum_(j=1,...,n) Xi_(t_n)^(j). (22) + +Let f>=0 be continuous with compact support in E. Its lifetime support +is contained in 0=2}. (23) + +Thus (4) and (13) imply + + q_t=t^(2/3) [integral_E (1-exp(-f)) d eta+o(1)]. (24) + +Only independence of whole copies is used in the next step: + + E exp(-S_n f)=(1-q_(t_n))^n + ->exp[-lambda integral_E(1-exp(-f))d eta]. (25) + +Indeed n q_(t_n)->lambda eta(1-exp(-f)), and +n q_(t_n)^2=O(n t_n^(4/3))->0. The right side is the Laplace +functional of the Poisson random measure with intensity lambda eta. + +Here is an elementary finite-window description of the process +convergence, so (25) is not mistaken for a first-moment assertion. +Restrict to a<=T. By (4), the probability that any of the n copies +contributes two or more points is at most + + n P(N_(Tt_n)>=2)=o(n t_n^(2/3))->0. + +Censor each copy which contributes two or more points, replacing its +measure by zero. The censored copies remain independent, each a Bernoulli +single point, and their superposition differs from the original with +probability tending to zero. The singleton intensity differs from its mean measure +in total mass by at most E[N_(Tt)1{N_(Tt)>=2}], so its conditional +singleton mark law converges weakly to eta restricted to a<=T, +normalized by m_T=(3/2)L^d c_(d,L)T^(2/3). Its success probability +is m_T t^(2/3)+o(t^(2/3)). The binomial number of successful copies +therefore converges to Poisson(lambda m_T), and their independent +marks converge to the normalized limiting mark law. This proves +finite-window marked Poisson convergence. Equivalently, apply (25) +to linear combinations of tests in that window. Tightness of the +marks follows from the all-mark domination in (12), including small +a and the b,k tails; counts are tight from their bounded expectations. + +Increasing T yields vague point-process convergence on E, or, +equivalently, the usual jointly Poisson counts on disjoint relatively +compact eta-continuity sets with the displayed intensity. No new +point-process literature result beyond the elementary binomial-to-Poisson +limit and the standard characterization by these marked finite-window +laws is needed. All calculations above directly identify that law. + +Projecting out k, or out b,k,u, gives the corresponding limits provided +the full bounded-lifetime window is used with the tail domination just +proved. In particular the rescaled lifetime process has intensity + + lambda L^d c_(d,L) a^(-1/3) da, + +and its number of points with 00. On its compact center/frame domain, +these ranks yield a joint contact covariance floor. For both radii +r_i<=r_*(eta_0), continuity keeps a common positive floor for the +joint U_(r_1),U_(r_2). Their Gaussian targets have a joint coercive +bound in (b_1,k_1,b_2,k_2). Regression supplies polynomial conditional +field moments; the product of the two normalized determinant weights +is bounded by a field-norm polynomial of degree 4d. Thus there is a +joint integrable Gaussian mark envelope on this separated band. +Each fold has the same r_i dr_i ledger. Applying both substitutions +ell_i=k_i r_i^3 and integrating 0 eta on each bounded-lifetime + window, with finite mass/tight tails. +* Weighted cluster avoidance: E[N_(Tt)1{N_(Tt)>=2}]=o(s(t)) for each T. + +Under those two interfaces and n s(t)->lambda, (23)-(25) prove the +Poisson theorem for any iid random point measures. If the second +interface were retained as an assumption, this would only be a +conditional Laplace theorem. For the actual law (1), Section 4 derives +it from the source-bound CONTACT, conditional global genericity, +actual elder transfer, all-Borel COUNT and ENVELOPE, with FAR for the +anchor complement. These are not supplied by covariance positivity +alone. The unbounded set of birth/gap marks is handled by the actual +H bound, rather than a normalized compact-mark probability claim. + +The pointwise isolation proof does not require a common deterministic +gap for all realizations or an all-target samplewise generic event. +It fixes a,b,k,u,x first, uses that fixed contact law's probability-one +genericity, then integrates its weighted zero limit. All-mark dominated +convergence is the precise quantifier change. The proof uses the +canonical Q_r at each target, so it does not inherit a conclusion +from an unspecified conditional version on null fibers. + +## 8. Explicit first-moment counterexample + +This is an abstract comparison process, not a counterexample to the +actual Gaussian law. On a mark window with finite nonzero measure eta +of mass m, let P=eta/m. With probability m t^(2/3)/2, produce exactly +two independent P marks; otherwise produce no points. For small t +this is a valid random finite point measure Z_t, and + + E Z_t=t^(2/3) eta, + E[N_t 1{N_t>=2}]=m t^(2/3). + +Its first moment is exactly the same rare-event scale, but its iid +superposition under n t^(2/3)->lambda has Laplace functional + + exp{-(lambda m/2) + [1-(integral exp(-f) dP)^2]}. + +This is a Poisson process of two-point clusters, not the Poisson +random measure with exponent -lambda integral(1-exp(-f))d eta. +In particular its total count takes only even values. It also has +second factorial of order t^(2/3). The example demonstrates why the +new physical isolation/weighted-cluster argument is needed and why +the leading lifetime intensity by itself cannot identify a process law. + +## 9. Existing lanes and residual scope + +The actual PAPER1_NOTE lines 61-69 and the research README/completion +audit were consumed before this derivation. They distinguish the +regional shrinking additional-witness problem, rejected candidates, +replacement bars and actual bars, and explicitly warn that a global +factorial bound alone does not discharge a regional collision estimate. +This preparation does not close or rename that owner lane. Its random +count is all finite actual H0 bars of one unconditioned fixed field; +its asymptotic conditioning is the one-bar contact Palm kernel. The +refined-selector issue main#229 remains separate with its own radius, +mark, weighting and confinement requirements. + +There is no expanding-domain process claim, covariance-only universality, +same-field spatial Poisson assertion, higher-homology result, all factorial +moment estimate, selected-density continuity theorem or quantitative +convergence rate. The realization-dependent old-height separation gives +qualitative dominated convergence, not a certified numerical t window. +No finite spectral/grid experiment measures the infinite field exactly. +No seed or held-out field is consumed. Source digests and exposed AI +technical checks do not establish external uptake, priority, publication +acceptance, blind confirmation or a compiled formal theorem. + +The substantive advance is the explicit marked mean law14, the physical +weighted cluster-avoidance estimate4, and the marked iid-copy limit25 for +this fixed actual law. The current source bindings and separately recorded +verdicts below distinguish original preparations from incorporated proofs. + +## 10. Current source bindings, historical exposure and residual scope + +The current analytic dependencies are: + +| Input | Current repository source | UTF-8 bytes / lines | SHA256 | +| --- | --- | --- | --- | +| IA | [Exact periodized Gaussian H0](../periodized_gaussian_h0/PROOF.md) | 34629 / 637 | c8bd489ca3fbb3760f7b2d53e7385131ac5336c51ba983e00e4d540fe2f00aaa | +| ST | [Regular-fold structural theorem](../fold_structural_universality/PROOF.md) | 43927 / 804 | 3d952a906553d578876456539a8f0dcd64aedd45099e8394cbe9db99866ec531 | +| S24 | [Actual physical-law source](../../../formal/sources/side24_v1/PROOF.md) | 10272 / 216 | c06daccc4ba4b9168522b9888b76a7d599934fc3b91bd753ee5d492262917769 | + +IA preserves mathematical Sections1-8 of original34281B/639lines +SHAcd0abd19cfb0bdd398142c136729d3bf96724939b3e774a3773a683310078a7f, +common-core SHA430b123f42162279f391b02e8320823d8a82f8cab6e86700aede72e8ab05b6cc. +ST preserves Sections1-10 of corrected43761B/813lines +SHAb077e2465174cadf78f2df120a80992534f99b838e18b5ba15fed250f4c28cca, +common-core SHAe30bdc85cfdd2eebfbe26037ea00b5aa90257aadff541c4a16682208025c6323. +The original process author fully read those original saved sources and +reread IA's actual all-Borel,contact genericity,selection,envelope and +complement arguments, plus ST's Borel selector. The current rebind is +root's separately checked composition; original read extents are not +relabeled as fresh reads of later wrapper bytes. + +Original process preparation31220B/627lines +SHAe69927b62c25f152b6e977fd409ff8890c86b8b43bf92336912c33d292fd2280 +remains frozen in ordinary coordinator custody. Root read all627 actual +lines, challenged endpoint exhaustion/Borel counting, and checked the +actual incorporation difference. The earlier benchmark route attack was +bounded and prefile; it did not review that saved source. REVIEW.md +records the fresh full original and current-source verdicts separately. +A byte-preserved mathematical core does not automatically rebind a +whole-file verdict to new operational source text. + +The original process derivation also recorded prior full reads of +SB25439B/501lines SHA5f529e5ed2632c90819746848146e3a24acda7e888cd3c9342df76edcbcbb7a3, +ET21138B/418lines SHA1cac9b525831b7e1f37f0d0561d3231fa73c8713c4c92283dca1d5b51b8a67ce, +and S24c06 at historicalroot6541/blob44b66f04f89fcd87383b3603fa69f1feb64cdddd. +SB preserves the density-version distinction; ET's original finite d2 +law is not substituted for IA's actual higher-dimensional adaptation. +Current [SB](../finite_h0_lifetime/PROOF.md)5ea9 and +[ET](../finite_elder_transfer/PROOF.md)c9e4 preserve their mathematical +Sections2-8 and2-13 with their own complete source reviews. These are +history/scope inputs; IA's actual hypotheses and coupling drive the new +process argument. + +The existing [PAPER1_NOTE](../../../docs/PAPER1_NOTE.md),6322B/69lines +SHA161ba87433879ee77ade047f1516d8c919fa3097812577cac54d7fe978a7ad63, +was previously fully read, with50-69 reread by the process author. +The existing [research plan](../../../docs/research-translation/20260930/README.md), +17156B/127lines SHAfcf43e5929b175ed9e9b046b9e92123d200c6cbff689d3b68196b3cca55228c7, +was read through95 with1-64 reread. The existing +[completion audit](../../../docs/research-translation/20260930/COMPLETION_AUDIT.md), +21094B/386lines SHA f6e557bb15706d614407daf160a8e44efce6d47775a3fbf078fff0a296093f89, +was read through140 with1-118 reread. Those bounded context reads do not +accept their historical imports, retire their nodes or transfer ownership. +The regional shrinking additional-witness and refined-selector main229 +lanes remain distinct. Root independently read PAPER1/S24 in full and +checked the actual current source identities; no new primary-paper or +compiler claim follows from this composition. + +IA records inherited bounded primary statements and the explicit classical +Sard,nonnegative-Borel area-formula,Morse/handle/critical-free-flow imports. +Gaussian regression,Tonelli/Fubini,smooth Fourier moments and DCT are +consumed with IA's proved actual hypotheses. The process calculation11-25, +including the new physical cluster estimate, is derived here rather than +attributed to a publication that did not state it. + +The stated fixed-d,L iid-copy theorem requires no knowingly unresolved +cluster assumption beyond its source-bound analytic proofs and named +imports. Coalescing-center/global factorial and spatial/regional collision +questions remain unproved here. Changed dependency bytes require fresh +source binding and review. No blind seed,experiment,numerical coefficient, +formal target,model label or scientific status is changed. diff --git a/experiments/universality/fold_isolation_diagonal/finite_fourier_count_controls.py b/experiments/universality/fold_isolation_diagonal/finite_fourier_count_controls.py new file mode 100644 index 000000000..eacf1d785 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/finite_fourier_count_controls.py @@ -0,0 +1,136 @@ +"""Exact finite algebra controls for the critical-count diagonal route. + +No Gaussian samples, approximate roots, gcd oracle, Rice calculation, or process test. +The all-K factor and genericity arguments are mathematical proofs in the report. +""" +from fractions import Fraction as Q +import hashlib +import json +from pathlib import Path + + +def clean(p): + return {e: v for e, v in p.items() if v} + + +def add(*ps): + out = {} + for p in ps: + for e, v in p.items(): + out[e] = out.get(e, Q(0)) + v + return clean(out) + + +def scale(a, p): + return clean({e: a*v for e, v in p.items()}) + + +def mul(p, q): + out = {} + for e, a in p.items(): + for f, b in q.items(): + ef = tuple(x+y for x, y in zip(e, f)) + out[ef] = out.get(ef, Q(0)) + a*b + return clean(out) + + +def homogenize(p, d): + assert all(sum(e) <= d for e in p) + return {e+(d-sum(e),): c for e, c in p.items()} + + +def witness(k, c): + # P=W(Z^2-1)+c(Z^2 W^2-1), Q=Z(W^2-1)+c(Z^2 W^2-1). + p = {(2*k, k): Q(1), (0, k): Q(-1), + (2*k, 2*k): c, (0, 0): -c} + q = {(k, 2*k): Q(1), (k, 0): Q(-1), + (2*k, 2*k): c, (0, 0): -c} + return clean(p), clean(q) + + +def run(): + checks = {} + for k in range(1, 9): + for c in (Q(1), Q(-2), Q(3, 5)): + p, q = witness(k, c) + rhs = mul({(k, 0): Q(1), (0, k): Q(-1)}, + {(k, k): Q(1), (0, 0): Q(1)}) + assert add(p, scale(-1, q)) == rhs + assert max(map(sum, p)) == max(map(sum, q)) == 4*k + for poly in (p, q): + ph = homogenize(poly, 4*k) + infinity = {e: a for e, a in ph.items() if e[2] == 0} + assert infinity == {(2*k, 2*k, 0): c} + # Recover P,Q from the actual three cosine modes' Laurent derivatives. + f = {(k, 0): Q(1, 2), (-k, 0): Q(1, 2), + (0, k): Q(1, 2), (0, -k): Q(1, 2), + (k, k): c/2, (-k, -k): c/2} + for axis, expected in ((0, p), (1, q)): + actual = clean({(e[0]+k, e[1]+k): Q(2*e[axis], k)*a + for e, a in f.items()}) + assert actual == expected + checks["lifted_witness_difference_and_actual_gradient_normalization_K1_to_K8"] = "PASS" + checks["full_degree_4K_and_no_homogenizing_variable_factor"] = "PASS" + + # All line factors z-zeta*w with zeta^K=1 give Z=W. + for c in (Q(1), Q(-2), Q(3, 5)): + p, _ = witness(1, c) + on_line = {} + for (i, j), a in p.items(): + on_line[(i+j,)] = on_line.get((i+j,), Q(0)) + a + assert clean(on_line) == {(3,): Q(1), (1,): Q(-1), + (4,): c, (0,): -c} + # All hyperbola factors zw-rho with rho^K=-1 give Z=-W^-1. + on_hyperbola = {} + for (i, j), a in p.items(): + e = (j-i,) + on_hyperbola[e] = on_hyperbola.get(e, Q(0)) + a*(-1)**i + assert clean(on_hyperbola) == {(-1,): Q(1), (1,): Q(-1)} + checks["nonzero_line_and_hyperbola_restrictions"] = "PASS" + + # The c=0 witness is affine-coprime by factor inspection, but homogenizes badly. + for k in range(1, 9): + p, q = witness(k, Q(0)) + assert max(map(sum, p)) == max(map(sum, q)) == 3*k + assert min(e[2] for e in homogenize(p, 4*k)) == k + assert min(e[2] for e in homogenize(q, 4*k)) == k + checks["false_degree_3K_projective_witness_has_common_T_power"] = "PASS" + + # Small radii relative to an upper cutoff do not control the actual-center ratio. + eta, rho, actual_s, r = Q(1, 10), Q(1, 4), Q(1, 10000), Q(1, 1000) + assert r <= rho*eta and r > rho*actual_s + checks["actual_center_ratio_gap_preserved"] = "PASS" + + # Gaussian tail exponent after logarithmic cutoff c0=cstar/64. + cstar = Q(5, 17) + c0 = cstar/64 + exponent = cstar/(16*c0) + assert exponent == 4 and exponent-Q(4, 3) == Q(8, 3) + checks["log_cutoff_normalized_exponent"] = "PASS" + + frozen = { + "QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md": + "827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30", + "fold_isolation_controls.py": + "13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5", + "FOLD_ISOLATION_CONTROLS.json": + "ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf", + } + root = Path(__file__).resolve().parent + for name, expected in frozen.items(): + assert hashlib.sha256((root/name).read_bytes()).hexdigest() == expected + checks["three_previous_isolation_files_unchanged"] = "PASS" + return { + "classification": "EXACT_FINITE_ALGEBRA_CONTROLS_ONLY", + "arithmetic": "Python Fraction; no floating arithmetic", + "sampling": "NONE", "joint_Rice_density": "NOT_USED", + "genericity_proof": "CONVENTIONAL_REPORT_ARGUMENT_NOT_MACHINE_VERIFIED", + "process_claim": "NONE", "scientific_effect": "NONE", + "checks": checks, + "normalization": {"chosen_tail_exponent": str(exponent), + "after_t_minus_4_over_3": str(exponent-Q(4, 3))}, + } + + +if __name__ == "__main__": + print(json.dumps(run(), indent=2)) diff --git a/experiments/universality/fold_isolation_diagonal/fold_isolation_controls.py b/experiments/universality/fold_isolation_diagonal/fold_isolation_controls.py new file mode 100644 index 000000000..7426866b8 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/fold_isolation_controls.py @@ -0,0 +1,156 @@ +"""Exact finite algebra controls for the local isolation proposal; no field sampler.""" +from fractions import Fraction as Q +from math import factorial +import json + + +def clean(p): + return {ij: c for ij, c in p.items() if c} + + +def add(*ps): + out = {} + for p in ps: + for ij, c in p.items(): + out[ij] = out.get(ij, Q(0)) + c + return clean(out) + + +def scale(c, p): + return clean({ij: c * x for ij, x in p.items()}) + + +def mul(p, q): + out = {} + for (i, j), a in p.items(): + for (k, l), b in q.items(): + ij = (i + k, j + l) + out[ij] = out.get(ij, Q(0)) + a * b + return clean(out) + + +def power(p, n): + out = {(0, 0): Q(1)} + for _ in range(n): + out = mul(out, p) + return out + + +def derivative(p, axis, order=1): + for _ in range(order): + out = {} + for ij, c in p.items(): + d = list(ij) + if d[axis]: + out[tuple(d[:axis] + [d[axis] - 1] + d[axis + 1:])] = c * d[axis] + p = out + return p + + +def restrict(p, h): + return add(*(scale(c, mul({(i, 0): Q(1)}, power(h, j))) + for (i, j), c in p.items())) + + +def evaluate(p, x, y=Q(0)): + return sum(c * x**i * y**j for (i, j), c in p.items()) + + +def run(): + x = {(1, 0): Q(1)} + y = {(0, 1): Q(1)} + one = {(0, 0): Q(1)} + # Prescribe a nontrivial exact stationary ridge with all cubic mixed terms present. + h = add(scale(Q(1, 7), power(x, 2)), scale(Q(1, 11), power(x, 3))) + f = add(scale(-2, power(x, 2)), scale(Q(3, 7), power(x, 3)), + scale(Q(-1, 11), power(x, 4)), scale(Q(1, 13), power(x, 5))) + a = add(scale(2, one), scale(Q(1, 5), x), scale(Q(1, 17), power(x, 2))) + b = add(scale(Q(1, 3), one), scale(Q(-1, 13), x)) + q = add(y, scale(-1, h)) + field = add(f, scale(Q(-1, 2), mul(a, power(q, 2))), + scale(Q(1, 6), mul(b, power(q, 3)))) + hp = derivative(h, 0) + assert restrict(derivative(field, 1), h) == {} + assert restrict(field, h) == f + rhs = add(restrict(derivative(field, 0, 3), h), + scale(3, mul(restrict(derivative(derivative(field, 0, 2), 1), h), hp)), + scale(3, mul(restrict(derivative(derivative(field, 0), 1, 2), h), power(hp, 2))), + mul(restrict(derivative(field, 1, 3), h), power(hp, 3))) + assert rhs == derivative(f, 0, 3) + + # Endpoint-distance normalization: r=2delta, k=1/12, not a half-distance mark. + delta, av = Q(2, 7), Q(3) + cubic = add(scale(Q(1, 6), power(x, 3)), + scale(-delta**2 / 2, x), scale(-av / 2, power(y, 2))) + ell = evaluate(cubic, -delta) - evaluate(cubic, delta) + r = 2 * delta + k = ell / r**3 + assert k == Q(1, 12) + assert evaluate(derivative(cubic, 0, 3), -delta) == 12 * k + + # Positive k does not imply a nonzero endpoint cubic without an r*C4 condition. + rr, kk = Q(2, 11), Q(5, 7) + quartic = add(scale(kk / rr, power(x, 4)), + scale(-2 * kk * rr, power(x, 2)), scale(-av / 2, power(y, 2))) + assert evaluate(derivative(quartic, 0), 0) == 0 + assert evaluate(derivative(quartic, 0), rr) == 0 + assert evaluate(derivative(quartic, 0, 2), 0) < 0 + assert evaluate(derivative(quartic, 0, 2), rr) > 0 + assert evaluate(derivative(quartic, 0, 3), 0) == 0 + qgap = evaluate(quartic, 0) - evaluate(quartic, rr) + assert qgap == kk * rr**3 + assert rr * evaluate(derivative(quartic, 0, 4), 0) / 2 == 12 * kk + + # The reduced cubic changes at the B^2/a scale even when D4F=0. + aa, tau, mixed, mu = Q(3), Q(2), Q(7), Q(1, 10) + cross = add(scale(-mu / 2, power(x, 2)), scale(tau / 6, power(x, 3)), + scale(mixed / 2, mul(power(x, 2), y)), scale(-aa / 2, power(y, 2))) + hh = scale(mixed / (2 * aa), power(x, 2)) + reduced = restrict(cross, hh) + expected = add(scale(tau, one), scale(3 * mixed**2 / aa, x)) + assert derivative(reduced, 0, 3) == expected + assert all(derivative(derivative(cross, 0, j), 1, 4-j) == {} for j in range(5)) + xloss = -aa * tau / (3 * mixed**2) + assert evaluate(expected, xloss) == 0 + + # A cubic fold has no guaranteed full enclosing height barrier. + radius, mm, tt = Q(1, 10), Q(1, 100), Q(1) + escape = add(scale(-mm / 2, power(x, 2)), scale(tt / 6, power(x, 3)), + scale(-av / 2, power(y, 2))) + assert evaluate(escape, radius) > evaluate(escape, 0) + + # Two maxima can approach only as the derivative bounds deteriorate in this control. + sep, target = Q(1, 17), Q(3, 5) + amp = target / (12 * sep) + doublewell = add(scale(-amp, mul(power(x, 2), power(add(x, scale(-sep, one)), 2))), + scale(-av / 2, power(y, 2))) + for point in (Q(0), sep): + assert evaluate(derivative(doublewell, 0), point) == 0 + assert evaluate(derivative(doublewell, 0, 2), point) < 0 + assert abs(evaluate(derivative(doublewell, 0, 3), point)) == target + assert abs(evaluate(derivative(doublewell, 0, 4), 0)) == 2 * target / sep + + # Rational checks of every numerical margin used by the chart argument. + p = Q(1, 8) + assert 3*p + 3*p*p + p**3 <= 4*p + assert Q(31, 32)**2 / 4 >= Q(1, 8) + assert Q(1, 2) - Q(1, 64) == Q(31, 64) + assert Q(16, 32) == Q(1, 2) + return {"classification": "EXACT_FINITE_ALGEBRA_CONTROLS_ONLY", + "arithmetic": "Python Fraction; no floating arithmetic", + "sampling": "NONE", "process_claim": "NONE", "scientific_effect": "NONE", + "checks": {"all_mixed_reduced_third_derivative_identity": "PASS", + "endpoint_distance_and_factor12": "PASS", + "quartic_positive_mark_zero_endpoint_cubic_counterexample": "PASS", + "mixed_cubic_Bsquared_over_a_counterexample": "PASS", + "false_full_height_barrier_counterexample": "PASS", + "two_maxima_derivative_scale_control": "PASS", + "chart_margin_constants": "PASS"}, + "exact_values": {"cubic_k": str(k), "mixed_cubic_sign_loss_x": str(xloss), + "quartic_gap": str(qgap), + "horizontal_escape_height": str(evaluate(escape, radius)), + "two_maxima_separation": str(sep)}} + + +if __name__ == "__main__": + print(json.dumps(run(), indent=2)) diff --git a/experiments/universality/fold_isolation_diagonal/jet-factorization-primary-sources/SOURCE_IDENTITY.json b/experiments/universality/fold_isolation_diagonal/jet-factorization-primary-sources/SOURCE_IDENTITY.json new file mode 100644 index 000000000..99264e07d --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/jet-factorization-primary-sources/SOURCE_IDENTITY.json @@ -0,0 +1,7 @@ +{ + "url": "https://webusers.imj-prg.fr/~bernard.teissier/documents/Lojasiewicz_2012.pdf", + "path": "work/core-theorem-audit-20261015/jet-factorization-primary-sources/teissier-lojasiewicz-2012.pdf", + "bytes": 325166, + "sha256": "cb88d0410e5c0a629f85592864809b987612663593a04dd8ea2928502cac6873", + "consumed": "Theorem 3.1 compact real-analytic distance inequality only, not the entire paper or its cited original proof" +} diff --git a/experiments/universality/fold_isolation_diagonal/jet_factorization_obstruction_controls.py b/experiments/universality/fold_isolation_diagonal/jet_factorization_obstruction_controls.py new file mode 100644 index 000000000..98e9c19b2 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/jet_factorization_obstruction_controls.py @@ -0,0 +1,130 @@ +"""Exact finite controls for coherent positive spectra and kernel bookkeeping. + +No Gaussian sampling, eigenvalue approximation, Poisson simulation, or numerical +Rice integral. The all-frame, stability, and analytic kernel arguments are in the report. +""" +from fractions import Fraction as Q +from math import factorial +from pathlib import Path +import hashlib +import json + + +ROWS = ((0, 0), (1, 0), (2, 0), (3, 0), (0, 1), (1, 1), (0, 2)) + + +def covariance(weights, scale): + out = [] + for p in ROWS: + row = [] + for q in ROWS: + phase = (sum(p)-sum(q)) % 4 + if phase % 2: + row.append(Q(0)) + continue + sign = 1 if phase == 0 else -1 + row.append(sign*sum(a*(Q(i, scale)**(p[0]+q[0])) + *(Q(j, scale)**(p[1]+q[1])) + for (i, j), a in weights.items())) + out.append(row) + return out + + +def ldl_pivots(matrix): + n = len(matrix) + low = [[Q(int(i == j)) for j in range(n)] for i in range(n)] + diagonal = [] + for i in range(n): + value = matrix[i][i]-sum(low[i][j]**2*diagonal[j] for j in range(i)) + assert value > 0 + diagonal.append(value) + for k in range(i+1, n): + low[k][i] = (matrix[k][i]-sum(low[k][j]*low[i][j]*diagonal[j] + for j in range(i)))/value + return diagonal + + +def run(): + checks = {} + base = {(i, j): Q(1, 25) for i in range(-2, 3) for j in range(-2, 3)} + base_cov = covariance(base, 1) + pivots = ldl_pivots(base_cov) + assert sum(base.values()) == 1 + checks["fixed_base_axial_seven_jet_exact_LDL_positive"] = "PASS" + moment_rows = [] + for n in (4, 5, 8, 12): + k = 2*n + eps = Q(1, n**12) + size = (4*n+1)**2 + fill = {(i, j): Q(1, size) for i in range(-k, k+1) for j in range(-k, k+1)} + weights = {ij: eps*a for ij, a in fill.items()} + for (i, j), a in base.items(): + weights[(n*i, n*j)] += (1-eps)*a + assert len(weights) == size and all(a > 0 for a in weights.values()) + assert sum(weights.values()) == 1 + m12 = sum(a*Q(i*i+j*j, n*n)**6 for (i, j), a in weights.items()) + assert m12 <= 8**6 == 262144 + actual_cov = covariance(weights, n) + fill_cov = covariance(fill, n) + for i in range(7): + for j in range(7): + assert actual_cov[i][j] == (1-eps)*base_cov[i][j]+eps*fill_cov[i][j] + ldl_pivots(actual_cov) + # Complete periods plus one zero residue imply the exact resonant fill covariance. + residues = [0]*n + for frequency in range(-2*n, 2*n+1): + residues[frequency % n] += 1 + assert residues == [5]+[4]*(n-1) + corr = 1-eps+eps/Q(4*n+1) + small_value_eigenvalue = 1-corr + assert small_value_eigenvalue == eps*(1-Q(1, 4*n+1)) < eps + moment_rows.append({"N": n, "M12": str(m12), + "two_value_small_eigenvalue": str(small_value_eigenvalue)}) + checks["positive_full_square_normalized_uniform_physical_M12"] = "PASS" + checks["single_jet_mixture_identity_and_resonant_two_value_collapse"] = "PASS" + checks["resonant_uniform_fill_geometric_sum"] = "PASS" + + # A dominant DC mode plus positive fill leaves every derivative variance O(eps). + for eps in (Q(1, 10), Q(1, 100), Q(1, 10000)): + assert eps*8 <= 8*eps + checks["DC_single_jet_failure_bookkeeping"] = "PASS" + + # Candidate cap for K=2N; transfer error uses P(different)<=N^-10. + for n in (4, 8, 16): + cap = (16*(2*n)**2)**2 + assert cap == 4096*n**4 + assert (cap+4096*n*n)*Q(1, n**10) <= Q(8192, n**6) + assert Q(1, n**12)*(4*n+1)**2 <= Q(25, n**10) + checks["small_fill_C4_dimension_ledger_and_expectation_transfer_cap"] = "PASS" + + # One fold: spatial r dr, height r^3, observation 12r^-5, Hessian r^2. + assert 1+3-5+2 == 1 + # Two folds: joint U dimension 12; normalized physical determinant degree eight. + assert 12//2+8 == 14 + assert 8+8 == 16 and 6+9 == 15 and 6+10 == 16 and 16 >= 15 + assert Q(1, 3)/2 == Q(1, 6) + checks["ordered_fold_Jacobian_and_middle_far_kernel_exponents"] = "PASS" + + frozen = { + "UNIFORM_LOCAL_NORM_DIAGONAL_PROPOSAL.md": "8e1eefca9888d43824a9fd9caca34c44d2de4bf8e30c668b474fa2097491c76b", + "uniform_local_norm_controls.py": "4b84755b771598ed34193c075718a41da447719f90a2ff5c6a17d48d0abc3dc4", + "UNIFORM_LOCAL_NORM_CONTROLS.json": "5d6cdb2d4925c6ba0b1b545d7a5aca1977655c1d18aa1ea29a82273f30845ece", + "QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md": "827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30", + "FINITE_FOURIER_CRITICAL_COUNT_DIAGONAL_ROUTE.md": "621cee96854aa440498cecf5452a8d1e23f90a4458779c3286c18eba106fe138", + "NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW.md": "16152427f7013822ac1bba81142429800334e6bf0f354a6439b23a1997bc4524", + } + root = Path(__file__).resolve().parent + for name, expected in frozen.items(): + assert hashlib.sha256((root/name).read_bytes()).hexdigest() == expected + checks["consumed_frozen_files_unchanged"] = "PASS" + return {"classification": "EXACT_FINITE_ALGEBRA_CONTROLS_ONLY", + "arithmetic": "Python Fraction and integers; no floating arithmetic", + "Gaussian_sampling": "NONE", "process_simulation": "NONE", + "scientific_effect": "NONE", "checks": checks, + "base_axial_LDL_pivots": [str(x) for x in pivots], + "finite_family_cases": moment_rows, + "scope": "Finite checks support arithmetic; all-frame positivity, actual process obstruction, and covariance-kernel estimates require the conventional report proofs."} + + +if __name__ == "__main__": + print(json.dumps(run(), indent=2)) diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/FINAL_REVIEW_INTEGRITY.json b/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/FINAL_REVIEW_INTEGRITY.json new file mode 100644 index 000000000..186155745 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/FINAL_REVIEW_INTEGRITY.json @@ -0,0 +1,20 @@ +{ + "checked_at_utc": "2026-10-16T02:01:55.192633+00:00", + "state": "REVIEW_AND_ACTUAL_EXECUTION_RECORDS_VERIFIED", + "verdict": "ACCEPT_SCOPED", + "frozen_new_author_files_unchanged": 3, + "frozen_previous_isolation_files_unchanged": 3, + "total_consumed_inputs_verified_unchanged": 12, + "check_processes_exit_zero": 2, + "author_output_exact_frozen_bytes": 849, + "independent_finite_cases": 226, + "review_manifest_files_verified": 11, + "report_bytes": 21880, + "report_sha256": "16152427f7013822ac1bba81142429800334e6bf0f354a6439b23a1997bc4524", + "manifest_bytes": 4392, + "manifest_sha256": "b4cd749e3d552f9d943d6116ac77288d971c616cbd022f881049bb42ffa5e7d0", + "remote_or_production_mutations": 0, + "source_adoption": "NONE", + "scientific_acceptance": "NONE", + "organizational_independence_credit": 0 +} diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/PRIMARY_IMPORT_AUDIT.json b/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/PRIMARY_IMPORT_AUDIT.json new file mode 100644 index 000000000..48b59245f --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/PRIMARY_IMPORT_AUDIT.json @@ -0,0 +1,49 @@ +{ + "reviewer": "OpenAI/Codex /root/c261a_local_runtime_20261016", + "checked_at_utc": "2026-10-16T02:00:14.209079+00:00", + "saved_source_manifest_sha256": "2c42d92a4732f09a58aebd9b9ef8672b51b672ab99f3bbdad61681a0dcf4f504", + "saved_source_bytes_verified": [ + { + "path": "work/core-theorem-audit-20261015/critical-count-primary-sources/guth-lect13.pdf", + "url": "https://math.mit.edu/~lguth/PolyMethod/lect13.pdf", + "bytes": 83539, + "sha256": "92ed0fc0f940179871a25f3a244eb181aaa0176ee5738ac1f8f68a8bacd072a9" + }, + { + "path": "work/core-theorem-audit-20261015/critical-count-primary-sources/stacks-01WC.html", + "url": "https://stacks.math.columbia.edu/tag/01WC", + "bytes": 15419, + "sha256": "08a39f68a8159e34e436df9c4b8f5e8886f58b3c0d5163c15610e12aef302737" + }, + { + "path": "work/core-theorem-audit-20261015/critical-count-primary-sources/stacks-01W6.html", + "url": "https://stacks.math.columbia.edu/tag/01W6", + "bytes": 14844, + "sha256": "060b8c179f815d59977f8a31057b5f5f9e1f2104351d46c33f1fb6d9ed95a202" + } + ], + "primary_sources_read_live": [ + { + "url": "https://math.mit.edu/~lguth/PolyMethod/lect13.pdf", + "scope": "complete six-page lecture; Theorem0.1 and its direct coprime affine inequality proof are consumed" + }, + { + "url": "https://stacks.math.columbia.edu/tag/01WC", + "scope": "full named lemma and proof; locally projective morphisms are proper" + }, + { + "url": "https://stacks.math.columbia.edu/tag/01W6", + "scope": "full named lemma and proof; proper source and separated target imply closed image" + }, + { + "url": "https://math.mit.edu/classes/18.721/ag-may31-2021.pdf", + "scope": "Section7.8.1 and its note on relatively prime reducible forms; Section7.8 proof, PDFpages178-179; not full216pagebook; no saved byte identity claimed" + }, + { + "url": "https://shelah.logic.at/files/94751/947.pdf", + "scope": "Section1 definitions and analytic-set universal-measurability conclusion, printedpages174-175; not fullpaper; no saved byte identity claimed" + } + ], + "mathematical_imports": "named classical results, not new formal proofs", + "additional_agent_crosscheck": "UNAVAILABLE: followup dispatch returned agent thread limit reached; none participated in this review" +} diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/author-controls.stderr b/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/author-controls.stderr new file mode 100644 index 000000000..e69de29bb diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/author-controls.stdout b/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/author-controls.stdout new file mode 100644 index 000000000..2f85a9034 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/author-controls.stdout @@ -0,0 +1,22 @@ +{ + "classification": "EXACT_FINITE_ALGEBRA_CONTROLS_ONLY", + "arithmetic": "Python Fraction; no floating arithmetic", + "sampling": "NONE", + "joint_Rice_density": "NOT_USED", + "genericity_proof": "CONVENTIONAL_REPORT_ARGUMENT_NOT_MACHINE_VERIFIED", + "process_claim": "NONE", + "scientific_effect": "NONE", + "checks": { + "lifted_witness_difference_and_actual_gradient_normalization_K1_to_K8": "PASS", + "full_degree_4K_and_no_homogenizing_variable_factor": "PASS", + "nonzero_line_and_hyperbola_restrictions": "PASS", + "false_degree_3K_projective_witness_has_common_T_power": "PASS", + "actual_center_ratio_gap_preserved": "PASS", + "log_cutoff_normalized_exponent": "PASS", + "three_previous_isolation_files_unchanged": "PASS" + }, + "normalization": { + "chosen_tail_exponent": "4", + "after_t_minus_4_over_3": "8/3" + } +} diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/independent-controls.stderr b/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/independent-controls.stderr new file mode 100644 index 000000000..e69de29bb diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/independent-controls.stdout b/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/independent-controls.stdout new file mode 100644 index 000000000..6038646cd --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/independent-controls.stdout @@ -0,0 +1,20 @@ +{ + "state": "INDEPENDENT_EXACT_FINITE_CONTROLS_ONLY", + "all_checks": "PASS", + "arithmetic": "Python Fraction; no field samples or approximate roots", + "counts": { + "lifted_factor_restrictions_full_degree_and_real_parameter_witness": 36, + "failed_lower_degree_witness_common_homogenizing_power": 24, + "ordered_factorial_count_and_shared_role_incidence_cases": 36, + "fresh_outer_cutoff_geometry_including_zero_center_separation": 12, + "rational_arbitrary_orientation_cutoff_geometry": 48, + "shared_maximum_three_distinct_critical_points_geometry": 1, + "torus_seam_short_lift_control": 1, + "fresh_cutoff_constant_and_radius_contradiction_cases": 64, + "literal_logarithmic_and_cumulative_normalization_exponents": 4 + }, + "total_cases": 226, + "all_K_coprimeness_projective_closedness_and_analytic_lemma": "REVIEWED_CONVENTIONALLY_NOT_MECHANICALLY_PROVED", + "gaussian_conditioning_or_Rice_integrals": "NOT_USED", + "scientific_acceptance": "NONE" +} diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/review_checks.py b/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/review_checks.py new file mode 100644 index 000000000..4b2405bbb --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/review_checks.py @@ -0,0 +1,154 @@ +"""Independent finite algebra and counting controls; no sampled fields or analytic proof automation.""" +from fractions import Fraction as Q +from itertools import product +import json + + +def merge(items): + result={} + for key,value in items: + result[key]=result.get(key,Q(0))+value + return {key:value for key,value in result.items() if value} + + +def multiply(a,b): + return merge(((i+k,j+l),x*y) for (i,j),x in a.items() for (k,l),y in b.items()) + + +def gradient_from_laurent(modes,K,axis): + # Differentiate z^m w^n, divide out i*omega, clear the Laurent monomial. + return merge(((m+K,n+K),Q((m,n)[axis])*value) for (m,n),value in modes.items()) + + +def on_line(poly,K): + # z=zeta*w with zeta^K=1: keep the zeta power as a formal quotient coordinate. + return merge(((i%K,i+j),v) for (i,j),v in poly.items()) + + +def on_hyperbola(poly,K): + # z=rho/w with rho^K=-1: exact quotient in rho and Laurent w. + return merge(((i%K,j-i),v*(-1)**(i//K)) for (i,j),v in poly.items()) + + +def run(): + counts={} + ncases=0 + for K,c in product(range(1,13),[Q(-3,7),Q(1),Q(5,2)]): + modes={(K,0):Q(1,2),(-K,0):Q(1,2),(0,K):Q(1,2),(0,-K):Q(1,2), + (K,K):c/2,(-K,-K):c/2} + assert all(modes[(-m,-n)]==v for (m,n),v in modes.items()) + P={key:2*v/K for key,v in gradient_from_laurent(modes,K,0).items()} + R={key:2*v/K for key,v in gradient_from_laurent(modes,K,1).items()} + difference=merge(list(P.items())+[(key,-v) for key,v in R.items()]) + factored=multiply({(K,0):Q(1),(0,K):Q(-1)},{(K,K):Q(1),(0,0):Q(1)}) + assert difference==factored + assert max(i+j for i,j in P)==max(i+j for i,j in R)==4*K + assert on_line(P,K)=={(0,3*K):Q(1),(0,K):Q(-1),(0,4*K):c,(0,0):-c} + assert on_hyperbola(P,K)=={(0,-K):Q(1),(0,K):Q(-1)} + for poly in (P,R): + homogenized={(i,j,4*K-i-j):v for (i,j),v in poly.items()} + assert min(t for i,j,t in homogenized)==0 + assert {key:v for key,v in homogenized.items() if key[2]==0}=={(2*K,2*K,0):c} + # A rational real Gaussian-coordinate witness for a valid positive full-square law. + N=(2*K+1)**2; H=N + q=Q(1,2*H*H);qzero=1-(N-1)*q + assert q>0 and qzero>0 and qzero+(N-1)*q==1 + assert Q(H,1)/H==1 and (c*H)/H==c + ncases+=1 + counts['lifted_factor_restrictions_full_degree_and_real_parameter_witness']=ncases + + for K in range(1,13): + modes={(K,0):Q(1,2),(-K,0):Q(1,2),(0,K):Q(1,2),(0,-K):Q(1,2)} + for axis in (0,1): + poly=gradient_from_laurent(modes,K,axis) + assert max(i+j for i,j in poly)==3*K + assert min(4*K-i-j for i,j in poly)==K + counts['failed_lower_degree_witness_common_homogenizing_power']=24 + + # Endpoint identities count endpoint pairs once; every ordered pair configuration is distinct. + total=0 + for maxima,saddles in product(range(1,7),range(1,7)): + crit=maxima+saddles + candidates=[(('M',i),('S',j)) for i in range(maxima) for j in range(saddles)] + ordered=[(p,q) for p in candidates for q in candidates if p!=q] + assert len(candidates)==maxima*saddles<=crit**2 + assert len(ordered)==len(candidates)*(len(candidates)-1)<=crit**4 + shared_M=sum(p[0]==q[0] for p,q in ordered) + shared_S=sum(p[1]==q[1] for p,q in ordered) + assert shared_M==maxima*saddles*(saddles-1) + assert shared_S==saddles*maxima*(maxima-1) + presentations=[(p,sign) for p in candidates for sign in (-1,1)] + assert len({p for p,sign in presentations})==len(candidates) + assert all(p[0]!=q[1] and p[1]!=q[0] for p,q in ordered) + total+=1 + counts['ordered_factorial_count_and_shared_role_incidence_cases']=total + + # New cutoff permits coincidence and arbitrarily smaller actual center separation. + L=Q(1);R0=L/16;eta0=R0/4;rho=Q(3,4);eta=eta0 + r1=eta/2;r2=eta/3 + for s in [Q(0),eta/10000,eta]: + for u1,u2 in product([-1,1],repeat=2): + M1=-r1*u1/2;M2=s-r2*u2/2 + S1=r1*u1/2;S2=s+r2*u2/2 + assert r1<=rho*eta and r2<=rho*eta + assert abs(M2-M1)<=(1+rho)*eta + assert max(abs(M1),abs(M2),abs(S1),abs(S2))rho*(eta/10000) + shared_M=Q(0);S1=r1;S2=-r2 + assert S1!=S2 and abs(S1-shared_M)<=rho*eta and abs(S2-shared_M)<=rho*eta + assert abs((shared_M+S1)/2-(shared_M+S2)/2)<=eta + counts['shared_maximum_three_distinct_critical_points_geometry']=1 + + # Periodic seam: choose common short lifts rather than subtracting canonical [0,L) coordinates. + z1=L-eta/4; z2_mod=eta/4; z2_lift=L+eta/4 + assert abs(z2_lift-z1)==eta/2 + assert abs(z2_mod-z1)>eta + counts['torus_seam_short_lift_control']=1 + + # Scalar contradiction constants for the new bounding-cutoff hypothesis. + ncases=0 + for a,km,b3,b4,rho,eta0 in product([Q(1,3),Q(2)],[Q(1,7),Q(3)], + [Q(1,2),Q(4)],[Q(1,5),Q(2)], + [Q(1,4),Q(3,4)],[Q(1,100),Q(1,1000)]): + R0=Q(1,16);tau=6*km + kap=min(R0,a/(32*b3),tau/(32*(b4+b3*b3/a))) + D0=b4/2+3*b3*b3/a + cs=min(a*a/(rho*rho*b3*b3*eta0),6*km/(rho*D0),kap/(1+rho)) + assert cs>0 + assert rho*rho*b3*b3*cs*eta0<=a*a + assert rho*D0*cs<=6*km + assert kap/cs>=1+rho + # Choose a coefficient norm and cutoff strictly inside the contradiction premise. + W=Q(2);cutoff=min(eta0,cs/(2*W*W)) + assert W*W(1+rho)*cutoff + ncases+=1 + counts['fresh_cutoff_constant_and_radius_contradiction_cases']=ncases + + for cs in [Q(1,101),Q(3,7),Q(19),Q(1000,3)]: + c0=cs/64 + assert cs/(16*c0)==4 + assert cs/(16*c0)-Q(4,3)==Q(8,3) + assert 4*c0/cs==Q(1,16) + counts['literal_logarithmic_and_cumulative_normalization_exponents']=4 + return {'state':'INDEPENDENT_EXACT_FINITE_CONTROLS_ONLY','all_checks':'PASS', + 'arithmetic':'Python Fraction; no field samples or approximate roots', + 'counts':counts,'total_cases':sum(counts.values()), + 'all_K_coprimeness_projective_closedness_and_analytic_lemma':'REVIEWED_CONVENTIONALLY_NOT_MECHANICALLY_PROVED', + 'gaussian_conditioning_or_Rice_integrals':'NOT_USED','scientific_acceptance':'NONE'} + + +if __name__=='__main__': + print(json.dumps(run(),indent=2)) diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/run_checks.py b/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/run_checks.py new file mode 100644 index 000000000..b9858b4e2 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-finite-count-checks/run_checks.py @@ -0,0 +1,58 @@ +import datetime +import hashlib +import json +import pathlib +import subprocess +import sys +import time + +HERE=pathlib.Path(__file__).resolve().parent +ROOT=HERE.parent +EXPECTED={ + 'FINITE_FOURIER_CRITICAL_COUNT_DIAGONAL_ROUTE.md':(25699,'621cee96854aa440498cecf5452a8d1e23f90a4458779c3286c18eba106fe138'), + 'finite_fourier_count_controls.py':(5153,'fd6f29739280a55017945526cb75fb46ebc80d86c012ff4a1ff65baa404c31de'), + 'FINITE_FOURIER_COUNT_CONTROLS.json':(849,'bbe147fdf409a3e37b450d4cbcd5d2a927140f95bd3ce0f36cafe16861123bfd'), + 'QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md':(19138,'827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30'), + 'fold_isolation_controls.py':(6308,'13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5'), + 'FOLD_ISOLATION_CONTROLS.json':(795,'ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf'), + 'critical-count-primary-sources/SOURCE_IDENTITIES.json':(796,'2c42d92a4732f09a58aebd9b9ef8672b51b672ab99f3bbdad61681a0dcf4f504'), + 'critical-count-primary-sources/guth-lect13.pdf':(83539,'92ed0fc0f940179871a25f3a244eb181aaa0176ee5738ac1f8f68a8bacd072a9'), + 'critical-count-primary-sources/stacks-01WC.html':(15419,'08a39f68a8159e34e436df9c4b8f5e8886f58b3c0d5163c15610e12aef302737'), + 'critical-count-primary-sources/stacks-01W6.html':(14844,'060b8c179f815d59977f8a31057b5f5f9e1f2104351d46c33f1fb6d9ed95a202'), + 'experiments/universality/finite_gaussian_contact/PROOF.md':(25367,'7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474'), + 'NONAUTHOR_FOLD_ISOLATION_REVIEW.md':(18409,'f06c61e3f3e60c795a1a14e824640f1689d4143dba4f6698cada7a41929eec90')} + +def inventory(): + result=[] + for n,(size,expected) in EXPECTED.items(): + b=(ROOT/n).read_bytes();digest=hashlib.sha256(b).hexdigest() + assert len(b)==size and digest==expected,n + result.append({'path':n,'bytes':len(b),'sha256':digest}) + return result + +before=inventory();runs=[] +for label,script in [('author-controls',ROOT/'finite_fourier_count_controls.py'),('independent-controls',HERE/'review_checks.py')]: + argv=[sys.executable,'-B',str(script)] + start=datetime.datetime.now(datetime.timezone.utc).isoformat();tick=time.perf_counter() + r=subprocess.run(argv,cwd=ROOT,stdout=subprocess.PIPE,stderr=subprocess.PIPE) + elapsed=time.perf_counter()-tick + (HERE/(label+'.stdout')).write_bytes(r.stdout);(HERE/(label+'.stderr')).write_bytes(r.stderr) + data={'command_argv':argv,'cwd':str(ROOT),'started_at_utc':start, + 'finished_at_utc':datetime.datetime.now(datetime.timezone.utc).isoformat(), + 'wall_seconds':elapsed,'exit_code':r.returncode,'stdout':label+'.stdout','stderr':label+'.stderr', + 'stdout_sha256':hashlib.sha256(r.stdout).hexdigest(),'stderr_sha256':hashlib.sha256(r.stderr).hexdigest()} + (HERE/(label+'.meta.json')).write_text(json.dumps(data,indent=2)+'\n') + assert r.returncode==0,(label,r.stderr.decode()) + output=json.loads(r.stdout) + if label=='author-controls': + assert r.stdout==(ROOT/'FINITE_FOURIER_COUNT_CONTROLS.json').read_bytes() + data['exact_frozen_output_match']=True + else: + assert output['all_checks']=='PASS';data['independent_finite_cases']=output['total_cases'] + runs.append(data) +after=inventory();assert after==before +result={'state':'EXACT_CONTROL_REPLAY_AND_SEPARATE_FINITE_CHECKS_ONLY','performer':'OpenAI/Codex /root/c261a_local_runtime_20261016', + 'source_exposure':'full new frozen proposal, control source/output and named primary imports read; prior exact chart consumed', + 'organization_independence_credit':0,'interpreter_version':sys.version,'inputs_before':before,'inputs_after':after, + 'all_12_inputs_unchanged':True,'runs':runs,'production_or_remote_actions':0,'scientific_acceptance':'NONE'} +(HERE/'CHECK_RUN_SUMMARY.json').write_text(json.dumps(result,indent=2)+'\n');print(json.dumps(result,indent=2)) diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/FINAL_REVIEW_INTEGRITY.json b/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/FINAL_REVIEW_INTEGRITY.json new file mode 100644 index 000000000..ef1029327 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/FINAL_REVIEW_INTEGRITY.json @@ -0,0 +1,18 @@ +{ + "checked_at_utc": "2026-10-16T01:51:07.108101+00:00", + "state": "ALL_REVIEW_INPUTS_AND_RECORDED_CHECK_OUTPUTS_VERIFIED", + "verdict": "ACCEPT_SCOPED", + "frozen_author_files_verified_unchanged": 3, + "actual_check_processes_exit_zero": 2, + "independent_finite_cases": 255, + "review_manifest_files_verified": 11, + "report_bytes": 18409, + "report_sha256": "f06c61e3f3e60c795a1a14e824640f1689d4143dba4f6698cada7a41929eec90", + "manifest_bytes": 2834, + "manifest_sha256": "3beb504fbd6bb216cfc8ba50c420a20e0782b5e77f11a1ca62ac7eea77c8fe05", + "remote_writes": 0, + "production_or_author_file_edits": 0, + "lean_or_formal_execution": false, + "organizational_independence_credit": 0, + "scientific_acceptance": "NONE" +} diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/author-controls.stderr b/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/author-controls.stderr new file mode 100644 index 000000000..e69de29bb diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/author-controls.stdout b/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/author-controls.stdout new file mode 100644 index 000000000..7f51cec43 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/author-controls.stdout @@ -0,0 +1,23 @@ +{ + "classification": "EXACT_FINITE_ALGEBRA_CONTROLS_ONLY", + "arithmetic": "Python Fraction; no floating arithmetic", + "sampling": "NONE", + "process_claim": "NONE", + "scientific_effect": "NONE", + "checks": { + "all_mixed_reduced_third_derivative_identity": "PASS", + "endpoint_distance_and_factor12": "PASS", + "quartic_positive_mark_zero_endpoint_cubic_counterexample": "PASS", + "mixed_cubic_Bsquared_over_a_counterexample": "PASS", + "false_full_height_barrier_counterexample": "PASS", + "two_maxima_derivative_scale_control": "PASS", + "chart_margin_constants": "PASS" + }, + "exact_values": { + "cubic_k": "1/12", + "mixed_cubic_sign_loss_x": "-2/49", + "quartic_gap": "40/9317", + "horizontal_escape_height": "7/60000", + "two_maxima_separation": "1/17" + } +} diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/gaussian-crosscheck/GAUSSIAN_OPERATOR_TAIL_CROSSCHECK_20261016T014803Z.md b/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/gaussian-crosscheck/GAUSSIAN_OPERATOR_TAIL_CROSSCHECK_20261016T014803Z.md new file mode 100644 index 000000000..eb4c19b65 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/gaussian-crosscheck/GAUSSIAN_OPERATOR_TAIL_CROSSCHECK_20261016T014803Z.md @@ -0,0 +1,153 @@ +# Independent Gaussian operator and occurrence-tail cross-check + +Reviewer: OpenAI/Codex `/root/c261a_local_runtime_20261016/gaussian_tail_crosscheck`, dispatched by OpenAI/Codex `/root/c261a_local_runtime_20261016`. Organizational-independence credit: **0**. Source fully exposed. Local conventional-mathematics review only. No production or remote writes, Lean result, global-bar result, process result, or scientific acceptance is asserted. + +## Verdict and exact scope + +**ACCEPT_SCOPED** for the finite coefficient/operator estimates, arbitrary-angle mark conversion, norm dichotomy, and unconditional occurrence tails in proposal Sections 4–5, conditional on the deterministic isolation chart in Section 2. I found no error requiring an amendment to those statements as presently scoped. + +**ACCEPT_SCOPED** for extending those occurrence tails to the canonical Gaussian law under any fixed deterministic **zero-target linear conditioning**. This includes singular observation covariance, and needs no uniform positive covariance eigenvalue. The precise law and its limits are stated below; the proposal does not presently state this extension. + +**AMEND before any expanded claim** that the same constants automatically cover nonzero targets without conditional-mean control, adaptive selection of the conditioning locations, ordinary densities for rank-deficient observations, or determinant/count-weighted expectations. Those assertions do not follow from the occurrence proof. These are boundaries of an extension, not defects in the present unconditional result. + +I have not re-reviewed the old collinear calculation. I independently checked the Gaussian and arbitrary-angle parts against the actual finite-source formula. + +## Sources read and frozen identities + +The complete proposal was read: `QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md`, 19,138 bytes, SHA-256 `827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30`. + +The complete bound source was read: `experiments/universality/finite_gaussian_contact/PROOF.md`, 25,367 bytes, SHA-256 `7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474`. This agrees with the identity quoted in the proposal. Its source law is lines 41–57; literal endpoint convention is lines 68–76; divided differences and target are lines 145–186; coefficient regression is lines 205–224; physical Hessian weights are lines 253–307; finite-support/rank boundaries are lines 425–477. + +The three frozen proposal/control files were only read. Their identities observed before the note was written were: + +- proposal: 19,138 bytes, SHA-256 `827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30`; +- `fold_isolation_controls.py`: 6,308 bytes, SHA-256 `13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5`; +- `FOLD_ISOLATION_CONTROLS.json`: 795 bytes, SHA-256 `ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf`. + +## 1. Actual finite law and multilinear operator bound + +The bound source, lines 41–57, has independent standard real coordinates + +`Z=(C_0,(C_n,S_n)_(n in P_K)) in R^N`, `N=(2K+1)^2`, + +and `sqrt(2 q_n)[C_n cos(omega n.x)+S_n sin(omega n.x)]` for each nonzero pair representative. Here `K>=2`, `omega=2pi/L`, `q_n=q_-n>0`, and `sum q_n=1`. + +For arbitrary unit derivative slots `v_1,...,v_j`, set `p_n=product_l (omega n.v_l)`. The two row coefficients belonging to mode pair `n` have squared sum exactly `2 q_n p_n^2`, since differentiation only shifts the sine/cosine phase. Therefore the squared norm of the entire row is + +`sum_(n in S_K) q_n product_l (omega n.v_l)^2 <= sum_(n in S_K) q_n (omega |n|)^(2j) = beta_j^2`. + +Cauchy–Schwarz applies for **every** location and every collection of unit slots with the same bound. Taking the supremum subsequently proves the proposal's operator estimate (15), with no coordinate-to-operator conversion factor and no rotation-invariance assumption. Including the constant coefficient in `||Z||` is conservative. The positivity of `beta_3,beta_4` follows from the source's positive nonzero modes. + +The simple frequency bound is `beta_j <= (2pi sqrt(2) K/L)^j`. Thus there is no hidden independence from `K,L`. The proposal correctly fixes the law; the prefactor also involves `N`. Its later claim does not use any lower covariance eigenvalue, support-rank theorem, or independent-copy replacement. + +## 2. Endpoint normalization and arbitrary direction + +The source's distance is `r=|S-M|`, with `M=z-r u/2`, `S=z+r u/2`, not a half-distance (source lines 68–76). At the pinned critical endpoints the fourth observation has target `12k` (lines 149–171). Changing from the midpoint variable to distance `s` measured from `M` gives exactly + +`12k = integral_0^r [6s(r-s)/r^3] D^3F(M+s u)[u,u,u] ds`. + +This positive kernel has mass one and first moment `r/2`, so an operator fourth-derivative bound gives `|F_uuu(M)-12k| <= Cr/2`. The endpoint estimate in (9) and the factor 12 are correct. + +The vector Taylor formula in (10) gives `||H u|| <= Br/2`. Since `H` is negative definite and a pair-transverse Rayleigh quotient is at most `-a`, its hard eigenvalue has absolute value at least `a`. Its smallest singular value, hence its soft eigenvalue in absolute value, is at most `Br/2`. Under `rB<=a` the two eigenspaces are distinct. For the hard unit vector `v_h`, `t=|u.v_h|<=Br/(2a)<=1/2`. Choose the soft unit vector `e` with `e.u>=0`. Then + +`||e-u||^2 = 2(1-sqrt(1-t^2)) = 2t^2/(1+sqrt(1-t^2)) <= 2t^2`. + +A three-term multilinear telescoping expansion yields + +`|D^3F[e,e,e]-D^3F[u,u,u]| <= 3B||e-u|| <= (3sqrt(2)/2) B^2r/a <= 3B^2r/a`. + +Thus (11) is conservative, and (12)–(13) indeed imply `F_eee(M)>=6k_-`. No upper mark bound or assumption that the pair frame diagonalizes the Hessian entered this argument. The negative-definite anchor, the pair-segment derivative bound, and the two displayed derivative error conditions are essential. + +## 3. Radius, norm dichotomy, and constants + +With `W=1+||Z||>=1`, `B=beta_3 W`, `C=beta_4 W`, the radius terms satisfy + +`R0 >= R0/W^2`, + +`a/(32B) >= [a/(32beta_3)]/W^2`, + +`tau/[32(C+B^2/a)] >= {tau/[32(beta_4+beta_3^2/a)]}/W^2`. + +Hence (16), `R>=kappa/W^2`, is valid. If another maximum is at Euclidean distance less than `R`, each of its coordinates in the first maximum's eigenframe has absolute value less than `R`; it is in the chart's square. This step has no relative-angle restriction. + +For the pair formulation, let the **actual** center distance be `s`. The triangle inequality gives `delta<=s+(r_1+r_2)/2<=(1+rho)s`. Assume `W^2=kappa/W^2>kappa s/c_*>=(1+rho)s>=delta`, + +a contradiction. Each term of (18) has the stated role, and no missing factor of two occurs. The second pair's marks are stronger hypotheses than the contradiction needs; it suffices that its maximum is distinct. This harmless redundancy does not invalidate the statement. + +For a cutoff event with `00`, `00`. There is no claimed uniform positive lower bound as `a` or `k_-` vanishes. In particular `c_*` contains an `a^2` term and a `k_-` term; neither dependence may be discarded. The embedded-lift assumptions in the proposal keep the Euclidean comparisons legitimate on the torus. + +## 4. Unconditional occurrence probability + +Put `c=kappa` or `c=c_*` and let `t` denote the appropriate distance cutoff. The deterministic result gives `W^2>=c/t`. If `t<=c/4`, then `T=sqrt(c/t)>=2`, so + +`||Z|| >=T-1>=T/2`, and therefore `||Z||^2>=c/(4t)`. + +The exact finite Gaussian moment is + +`E exp(||Z||^2/4)=2^(N/2)`. + +Markov's inequality gives `P(E_t)<=2^(N/2) exp[-c/(16t)]`, precisely (17) or (20). The deterministic inclusion also establishes the outer-probability version if event measurability has not yet been proved. There is no counting factor here, and existence over arbitrary locations or angles requires no union bound because the norm obstruction is simultaneous in all of them. + +## 5. Zero-target linear conditioning, including singular observations + +Let `A:R^N -> R^m` be any deterministic real linear map, possibly dependent on fixed geometry or a deterministic scale, with rank `q`. Use the **canonical Gaussian conditional law on the support of `AZ`**. The zero target always lies in that support. If `P` is orthogonal projection onto `ker A` and `G~N(0,I_N)`, this conditional law is + +`Z | {AZ=0} = P G` in distribution. + +This statement uses the Gaussian disintegration, not a positive-probability interpretation of the equality event and not an arbitrary choice of a regular conditional version at a null target. It remains valid when `AA^T` is singular. The exact exponential moment is + +`E[exp(||Z||^2/4) | AZ=0] =2^((N-q)/2) <=2^(N/2)`. + +Every deterministic derivative and configuration implication above remains true on `ker A`. Consequently + +`P(E_t | AZ=0) <=2^((N-q)/2) exp[-c/(16t)]`, `t<=c/4`. + +This is uniform over all such deterministic zero-target maps if the original `2^(N/2)` prefactor is retained. Rank changes and small nonzero observation eigenvalues cause no failure of this tail statement. The conclusion also holds for conditional outer probability. + +Zero gradient pins at fixed sites are an example. Adding a redundant zero observation does not require a new covariance lower bound for this tail alone. It still does **not** provide a nonsingular Lebesgue density of the observations, a valid Rice Jacobian, any normalizer, or a determinant/count-weighted estimate. The finite source explicitly reserves enlarged conditioning lists and finite-support rank control for those other uses (lines 425–477). + +## 6. Nonzero targets and adaptive conditioning: limits of the extension + +For a nonzero target `y`, first require `y in im A`; otherwise the canonical supported conditional law at that target is not defined. For a supported target let `m=A^dagger y`, the minimum-norm solution, so + +`Z | {AZ=y} =m+P G`, with `m` orthogonal to `ker A`. + +Thus + +`E[exp(||Z||^2/4) | AZ=y] = exp(||m||^2/4) 2^((N-q)/2)`, + +and the same deterministic obstruction yields the valid estimate + +`P(E_t | AZ=y) <=2^((N-q)/2) exp[||m||^2/4-c/(16t)]`, `t<=c/4`. + +A uniform bound on `||m||` therefore suffices for a uniform constant. Merely bounding the target or allowing singular conditioning does not suffice. For example, `A_epsilon z=epsilon z_1`, `y=1`, fixes `z_1=1/epsilon` and makes the norm arbitrarily large as `epsilon->0`. This is a counterexample to transferring the centered radial tail uniformly over bounded nonzero targets; it is not a constructed two-pair event in this finite Fourier law. + +The bound source does supply mean control for its **specific** `U_r=v_r` conditioning when `r<=r_*` and `b,k` lie in a compact rectangle: source (7) has a uniform covariance lower bound and lines 211–224 use it. Explicitly, + +`||m||^2=v_r^T (L_U L_U^T)^(-1) v_r <= ||v_r||^2/m_source`. + +For `r<=1`, `|b|<=b_max`, `k<=k_+`, one may use `||v_r||^2<=2b_max^2+(291/2)k_+^2`. That source-bound case admits the shifted-mean estimate. It does not establish uniform mean control for arbitrary four-site height targets, unbounded birth/cubic marks, or enlarged rank-degenerate lists. Source lines 305–307 expressly limit its inverse-normalizer result too. + +Finally, the observation map must be fixed before the field draw, or its conditional Gaussian law must be independently justified. Selecting critical sites or configurations from the realized field generally changes the conditioning law; it is not licensed just by writing an apparently linear observation at the selected sites. External random geometry independent of `Z` can be handled by first conditioning on that geometry. Field-dependent selection, Palm/Rice reweighting, and factorial counts require separate arguments. + +## 7. Measurability + +The proposal correctly offers outer probability and so has no measurability gap in its stated tail. For the usual literal configuration event here, a Borel formulation is also available. The field, its derivatives, endpoint height gaps, transverse curvatures, and Hessian eigenvalues are jointly continuous in `Z` and the finite-dimensional torus/frame/radius variables. Negative definiteness of an anchored critical point is precisely nondegenerate local maximality. + +Use compact torus/frame parameter spaces and the stated short-lift domain. Exhaust strict conditions with `||Z||<=n`, radii and positive/distinct distances at least `1/n`, and maximum Hessian eigenvalues at most `-1/n`; an open patch condition can likewise be exhausted by compact subsets. All remaining equalities and weak inequalities are closed. Projection of each resulting compact configuration set to the finite coefficient space is compact. A countable union over the exhaustion and a finite/countable chart atlas gives a Borel event. Alternatively, its Borel configuration set has analytic projection and hence is universally measurable under the completed Gaussian law. Either route is enough; no count measurability or selected-bar measurability is inferred. + +## Scope of this review + +These are analytic, finite-dimensional source checks. No Gaussian draws, persistence computations, source edits, remote updates, Lean compilation, kernel audit, or new global mathematical admission occurred. The occurrence tail is valid within its fixed positive curvature/cubic windows and short-pair geometry. The organizational independence remains zero. diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/independent-controls.stderr b/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/independent-controls.stderr new file mode 100644 index 000000000..e69de29bb diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/independent-controls.stdout b/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/independent-controls.stdout new file mode 100644 index 000000000..59e929ff3 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/independent-controls.stdout @@ -0,0 +1,24 @@ +{ + "state": "INDEPENDENT_EXACT_FINITE_CHECKS_ONLY", + "all_checks": "PASS", + "counts": { + "mixed_chain_rule_jet_cancellations": 96, + "general_kernel_identity_and_endpoint_mark": 21, + "closed_x_face_saddles_both_signs": 2, + "hard_pivot_schur_classifications": 18, + "retained_false_barrier_and_D4zero_obstruction": 2, + "finite_spectral_multilinear_rows": 16, + "inverse_W_squared_radius_inequalities": 96, + "singular_centered_regression_projector_MGF_checks": 4 + }, + "total_cases": 255, + "boundary_second_roots": [ + "1/32", + "-1/32" + ], + "actual_radius_full_height_escape": "113/24576000", + "mixed_cubic_sign_loss": "-2/49", + "sampling": "NONE", + "analytic_proof_mechanically_verified": false, + "formal_or_scientific_acceptance": "NONE" +} diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/review_checks.py b/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/review_checks.py new file mode 100644 index 000000000..b5dc8fe7e --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/review_checks.py @@ -0,0 +1,175 @@ +"""Independent finite algebra attacks; these are not an analytic proof or field experiment.""" +from fractions import Fraction as Q +from itertools import product +import json + + +def d_at(coeffs, x, order=1): + ans = Q(0) + for j, a in enumerate(coeffs): + if j >= order: + factor = 1 + for k in range(order): + factor *= j-k + ans += a * factor * x**(j-order) + return ans + + +def value(coeffs, x): + return d_at(coeffs, x, 0) + + +def kernel_third(coeffs, r): + # Integrate the monomials exactly against 6s(r-s)/r^3. + ans = Q(0) + for j, a in enumerate(coeffs): + if j >= 3: + factor = j*(j-1)*(j-2)*a + ans += 6*factor/r**3 * (r*r**(j-1)/Q(j-1)-r**j/Q(j)) + return ans + + +def det(matrix): + a = [list(row) for row in matrix] + ans = Q(1) + for i in range(len(a)): + pivot = next((j for j in range(i,len(a)) if a[j][i]),None) + if pivot is None: + return Q(0) + if pivot != i: + a[i],a[pivot]=a[pivot],a[i] + ans=-ans + v=a[i][i] + ans*=v + for j in range(i+1,len(a)): + c=a[j][i]/v + for k in range(i+1,len(a)): + a[j][k]-=c*a[i][k] + return ans + + +def run(): + counts = {} + # Independent jet cancellation: start with the unreduced differentiated f''. + count=0 + for A,p,xxx,xxy,xyy,yyy in product([Q(-1),Q(-7,3)], [Q(-1,8),Q(0),Q(1,11)], + [Q(2,3),Q(-3,5)], [Q(5,7),Q(-2)], + [Q(-4,9),Q(3)], [Q(7,11),Q(-5)]): + xy=-A*p + hpp=-(xxy+2*xyy*p+yyy*p*p)/A + direct=xxx+2*xxy*p+xyy*p*p+xy*hpp + assert direct==xxx+3*xxy*p+3*xyy*p*p+yyy*p**3 + count+=1 + counts['mixed_chain_rule_jet_cancellations']=count + + # Check integration by parts before imposing endpoint criticality, then impose it. + count=0 + for degree,r in product(range(3,10),[Q(1,7),Q(2,3),Q(3,2)]): + coeffs=[Q((-1)**j*(j+1),j+2) for j in range(degree+1)] + lhs=kernel_third(coeffs,r) + rhs=6/r**3*(r*(d_at(coeffs,0)+d_at(coeffs,r))-2*(value(coeffs,r)-value(coeffs,0))) + assert lhs==rhs + coeffs[1]=Q(0) + coeffs[2]=-sum(Q(j)*coeffs[j]*r**(j-2) for j in range(3,len(coeffs)))/2 + assert d_at(coeffs,0)==0 and d_at(coeffs,r)==0 + k=(value(coeffs,0)-value(coeffs,r))/r**3 + assert kernel_third(coeffs,r)==12*k + count+=1 + counts['general_kernel_identity_and_endpoint_mark']=count + + # Closed-square endpoint roots: the proposed radius itself, for both cubic signs. + R=Q(1,32); mu=R/2; a=Q(1) + boundary=[] + for tau in [Q(1),Q(-1)]: + F=[Q(0),Q(0),-mu/2,tau/6] + second=2*mu/tau + assert abs(second)==R + assert d_at(F,0)==0 and d_at(F,second)==0 + assert d_at(F,0,2)==-mu<0 and d_at(F,second,2)==mu>0 + assert min(Q(1),a/32,Q(1)/(32*(Q(0)+Q(1)/a)))==R + assert -a*d_at(F,second,2)<0 + boundary.append(str(second)) + counts['closed_x_face_saddles_both_signs']=2 + + # Negative hard pivot: Schur sign, including nonzero off-diagonal Hessians. + count=0 + for A,b,s in product([Q(-1),Q(-2,3),Q(-11)], [Q(0),Q(5),Q(-7,13)], [Q(-1,9),Q(2,7)]): + xx=s+b*b/A + assert xx-b*b/A==s + assert xx*A-b*b==A*s + if s<0: + assert xx<0 and xx*A-b*b>0 + else: + assert xx*A-b*b<0 + count+=1 + counts['hard_pivot_schur_classifications']=count + + # Full boundary height barrier fails even at the actual stated chart radius. + escape=[Q(0),Q(0),-Q(1,1000)/2,Q(1,6)] + assert value(escape,R)>value(escape,0) + # At a stationary mixed-cubic ridge the reduced cubic is tau+3b^2*x/a. + aa,tau,b=Q(3),Q(2),Q(7) + reduced=[Q(0),Q(0),-Q(1,10)/2,tau/6,b*b/(8*aa)] + xloss=-aa*tau/(3*b*b) + assert d_at(reduced,xloss,3)==0 + assert d_at(reduced,xloss-Q(1,100),3)<0 + assert d_at(reduced,xloss+Q(1,100),3)>0 + counts['retained_false_barrier_and_D4zero_obstruction']=2 + + # Exact anisotropic spectral row/operator bound at arbitrary rational frames. + vectors=[(Q(3,5),Q(4,5)),(Q(12,13),Q(5,13)),(Q(-5,13),Q(12,13)),(Q(1),Q(0))] + count=0 + for K,j in product([2,3],[3,4]): + modes=[(n,m) for n in range(-K,K+1) for m in range(-K,K+1)] + raw={(n,m):Q(1+n*n+2*m*m) for n,m in modes} + total=sum(raw.values());weights={n:w/total for n,w in raw.items()} + beta2=sum(weights[n]*(n[0]*n[0]+n[1]*n[1])**j for n in modes) + for offset in range(4): + slots=[vectors[(offset+k)%4] for k in range(j)] + row2=Q(0) + for n in modes: + factor=Q(1) + for v in slots: + assert v[0]*v[0]+v[1]*v[1]==1 + projection=n[0]*v[0]+n[1]*v[1] + assert projection**2 <= n[0]*n[0]+n[1]*n[1] + factor*=projection**2 + row2+=weights[n]*factor + assert row2<=beta2 + count+=1 + counts['finite_spectral_multilinear_rows']=count + + # Radius lower bound is a scalar algebraic inequality for every W>=1. + count=0 + for R0,a,tau,b3,b4,W in product([Q(1,16),Q(2)], [Q(1,9),Q(3)], + [Q(1,11),Q(4)], [Q(1,3),Q(7)], + [Q(0),Q(5)], [Q(1),Q(3,2),Q(9)]): + B,C=b3*W,b4*W + R=min(R0,a/(32*B),tau/(32*(C+B*B/a))) + kappa=min(R0,a/(32*b3),tau/(32*(b4+b3*b3/a))) + assert R>=kappa/(W*W) + count+=1 + counts['inverse_W_squared_radius_inequalities']=count + + # Zero-target regression projectors can be singular and need no covariance floor. + projectors=[([[Q(0)]],0),([[Q(1)]],1), + ([[Q(1,2),Q(1,2)],[Q(1,2),Q(1,2)]],1), + ([[Q(int(i==j))-Q(1,3) for j in range(3)] for i in range(3)],2)] + for P,rank in projectors: + n=len(P) + for i,j in product(range(n),repeat=2): + assert P[i][j]==P[j][i] + assert sum(P[i][k]*P[k][j] for k in range(n))==P[i][j] + # det(I-P/2)^(-1/2)=2^(rank/2), the Gaussian MGF at t=1/4. + assert det([[Q(int(i==j))-P[i][j]/2 for j in range(n)] for i in range(n)])==Q(1,2**rank) + counts['singular_centered_regression_projector_MGF_checks']=len(projectors) + return {'state':'INDEPENDENT_EXACT_FINITE_CHECKS_ONLY','all_checks':'PASS', + 'counts':counts,'total_cases':sum(counts.values()), + 'boundary_second_roots':boundary,'actual_radius_full_height_escape':str(value(escape,Q(1,32))), + 'mixed_cubic_sign_loss':str(xloss), + 'sampling':'NONE','analytic_proof_mechanically_verified':False, + 'formal_or_scientific_acceptance':'NONE'} + + +if __name__=='__main__': + print(json.dumps(run(),indent=2)) diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/run_checks.py b/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/run_checks.py new file mode 100644 index 000000000..360f73ec6 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-fold-isolation-checks/run_checks.py @@ -0,0 +1,58 @@ +import datetime +import hashlib +import json +import pathlib +import subprocess +import sys +import time + +HERE=pathlib.Path(__file__).resolve().parent +SOURCE=HERE.parent +EXPECTED={ + 'QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md':(19138,'827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30'), + 'fold_isolation_controls.py':(6308,'13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5'), + 'FOLD_ISOLATION_CONTROLS.json':(795,'ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf')} + +def inventory(): + rows=[] + for name,(size,digest) in EXPECTED.items(): + b=(SOURCE/name).read_bytes() + actual=hashlib.sha256(b).hexdigest() + assert len(b)==size and actual==digest,name + rows.append({'path':str(SOURCE/name),'bytes':len(b),'sha256':actual}) + return rows + +before=inventory() +runs=[] +for label,path in [('author-controls',SOURCE/'fold_isolation_controls.py'),('independent-controls',HERE/'review_checks.py')]: + argv=[sys.executable,'-B',str(path)] + start=datetime.datetime.now(datetime.timezone.utc).isoformat() + tick=time.perf_counter() + result=subprocess.run(argv,cwd=SOURCE,stdout=subprocess.PIPE,stderr=subprocess.PIPE) + elapsed=time.perf_counter()-tick + (HERE/(label+'.stdout')).write_bytes(result.stdout) + (HERE/(label+'.stderr')).write_bytes(result.stderr) + record={'command_argv':argv,'cwd':str(SOURCE),'started_at_utc':start, + 'finished_at_utc':datetime.datetime.now(datetime.timezone.utc).isoformat(), + 'wall_seconds':elapsed,'exit_code':result.returncode, + 'stdout':label+'.stdout','stdout_sha256':hashlib.sha256(result.stdout).hexdigest(), + 'stderr':label+'.stderr','stderr_sha256':hashlib.sha256(result.stderr).hexdigest()} + (HERE/(label+'.meta.json')).write_text(json.dumps(record,indent=2)+'\n') + assert result.returncode==0,(label,result.stderr.decode()) + payload=json.loads(result.stdout) + if label=='author-controls': + assert result.stdout==(SOURCE/'FOLD_ISOLATION_CONTROLS.json').read_bytes() + record['exact_frozen_output_match']=True + else: + assert payload['all_checks']=='PASS' + record['finite_cases']=payload['total_cases'] + runs.append(record) +after=inventory() +assert after==before +summary={'state':'LOCAL_REPLAY_AND_INDEPENDENT_FINITE_CONTROLS_VERIFIED_NO_PROOF_ACCEPTANCE', + 'performer':'OpenAI/Codex /root/c261a_local_runtime_20261016', + 'organization_independence_credit':0,'exposure':'frozen proposal and source read; nonauthor of reviewed bytes', + 'interpreter_version':sys.version,'frozen_inputs_before':before,'frozen_inputs_after':after, + 'runs':runs,'frozen_inputs_unchanged':True} +(HERE/'CHECK_RUN_SUMMARY.json').write_text(json.dumps(summary,indent=2)+'\n') +print(json.dumps(summary,indent=2)) diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/FROZEN_INPUT_INTEGRITY.json b/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/FROZEN_INPUT_INTEGRITY.json new file mode 100644 index 000000000..10b4fef1f --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/FROZEN_INPUT_INTEGRITY.json @@ -0,0 +1,223 @@ +{ + "accepted_local_norm_manifest_closure": { + "inputs": 14, + "matches": true, + "outputs": 11 + }, + "frozen_inputs": [ + { + "bytes": 23286, + "path": "COUNTERFAMILY_NONAUTHOR_REVIEW.md", + "sha256": "26b1b1a9a1e978edd036ee2d1e244e8b8b2089d3c6a48edb94d78d37577cc2c5" + }, + { + "bytes": 4458, + "path": "COUNTERFAMILY_NONAUTHOR_REVIEW_MANIFEST.json", + "sha256": "614f84154b75505c0004f70ca2e81d974c9dec49a9f5c64c422b63a787d82a56" + }, + { + "bytes": 849, + "path": "FINITE_FOURIER_COUNT_CONTROLS.json", + "sha256": "bbe147fdf409a3e37b450d4cbcd5d2a927140f95bd3ce0f36cafe16861123bfd" + }, + { + "bytes": 25699, + "path": "FINITE_FOURIER_CRITICAL_COUNT_DIAGONAL_ROUTE.md", + "sha256": "621cee96854aa440498cecf5452a8d1e23f90a4458779c3286c18eba106fe138" + }, + { + "bytes": 795, + "path": "FOLD_ISOLATION_CONTROLS.json", + "sha256": "ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf" + }, + { + "bytes": 1618, + "path": "JET_FACTORIZATION_OBSTRUCTION_CONTROLS.json", + "sha256": "1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764" + }, + { + "bytes": 34627, + "path": "JET_NONDEGENERACY_FACTORIZATION_OBSTRUCTION.md", + "sha256": "0dfdddfbcac47b2fd07449f4152b7f5d935f92aed15272fb8333ed9cc60b5766" + }, + { + "bytes": 21880, + "path": "NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW.md", + "sha256": "16152427f7013822ac1bba81142429800334e6bf0f354a6439b23a1997bc4524" + }, + { + "bytes": 18409, + "path": "NONAUTHOR_FOLD_ISOLATION_REVIEW.md", + "sha256": "f06c61e3f3e60c795a1a14e824640f1689d4143dba4f6698cada7a41929eec90" + }, + { + "bytes": 23887, + "path": "NONAUTHOR_UNIFORM_LOCAL_NORM_REVIEW.md", + "sha256": "ff552dfacadede53405530a219302c1163dca8ba97d518f02d407970d3de3c1c" + }, + { + "bytes": 4763, + "path": "NONAUTHOR_UNIFORM_LOCAL_NORM_REVIEW_MANIFEST.json", + "sha256": "1ccd38374adcf4931b87abeacdefaa64990080a9b3b98835a7c140bde2f505a7" + }, + { + "bytes": 19138, + "path": "QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md", + "sha256": "827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30" + }, + { + "bytes": 1790, + "path": "UNIFORM_LOCAL_NORM_CONTROLS.json", + "sha256": "5d6cdb2d4925c6ba0b1b545d7a5aca1977655c1d18aa1ea29a82273f30845ece" + }, + { + "bytes": 26687, + "path": "UNIFORM_LOCAL_NORM_DIAGONAL_PROPOSAL.md", + "sha256": "8e1eefca9888d43824a9fd9caca34c44d2de4bf8e30c668b474fa2097491c76b" + }, + { + "bytes": 2560, + "path": "counterfamily-nonauthor-checks-20261016/EXECUTION_RECORD.json", + "sha256": "8ae808eed6b7f8fec67435a3b97c8422a08212e7f18ababbced3b367380e04fb" + }, + { + "bytes": 0, + "path": "counterfamily-nonauthor-checks-20261016/author_controls.stderr", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 1618, + "path": "counterfamily-nonauthor-checks-20261016/author_controls.stdout.json", + "sha256": "1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764" + }, + { + "bytes": 8080, + "path": "counterfamily-nonauthor-checks-20261016/counterfamily_checks.py", + "sha256": "0718a7f0e415f75110b5b08b5850fdc5e76c601e515ee64bb0770ceaa21834e9" + }, + { + "bytes": 0, + "path": "counterfamily-nonauthor-checks-20261016/independent_controls.stderr", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 12359, + "path": "counterfamily-nonauthor-checks-20261016/independent_controls.stdout.json", + "sha256": "3c53ed54756c911292180c5686750de9a47cf4a54c56cc0d2616155c3f5af35b" + }, + { + "bytes": 29190, + "path": "experiments/universality/finite_candidate_lifetime/PROOF.md", + "sha256": "ed31f562f0ffaaf6d4bd2ceb1ae3d234798e983a55d55c6983b128fce27133f5" + }, + { + "bytes": 25367, + "path": "experiments/universality/finite_gaussian_contact/PROOF.md", + "sha256": "7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474" + }, + { + "bytes": 30935, + "path": "experiments/universality/iid_short_bar_process/PROOF.md", + "sha256": "320bc988c468bea4938bb777a2d05952c6ab8e0a384f33725b50de847aeec1eb" + }, + { + "bytes": 5153, + "path": "finite_fourier_count_controls.py", + "sha256": "fd6f29739280a55017945526cb75fb46ebc80d86c012ff4a1ff65baa404c31de" + }, + { + "bytes": 6308, + "path": "fold_isolation_controls.py", + "sha256": "13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5" + }, + { + "bytes": 427, + "path": "jet-factorization-primary-sources/SOURCE_IDENTITY.json", + "sha256": "683fb8d29a62c9e73f01a93bf3b288c1e7189f050da4b77eb831bb063ac88492" + }, + { + "bytes": 325166, + "path": "jet-factorization-primary-sources/teissier-lojasiewicz-2012.pdf", + "sha256": "cb88d0410e5c0a629f85592864809b987612663593a04dd8ea2928502cac6873" + }, + { + "bytes": 5818, + "path": "jet_factorization_obstruction_controls.py", + "sha256": "5b1e8fc18a6e32e3a312a253e3e4d1110bb674385f3d49447a45e78ad4df541f" + }, + { + "bytes": 7279, + "path": "nonauthor-uniform-local-norm-checks/CHECK_RUN_SUMMARY.json", + "sha256": "480e1e4786933f2201a6dadb93ee720dc21b632c5a749a91ea9323f1fc506364" + }, + { + "bytes": 1419, + "path": "nonauthor-uniform-local-norm-checks/PRIMARY_IMPORT_AUDIT.json", + "sha256": "98c521f7c773e1338b2ed6bac74ebba0c3f146cdc0b989abc7dd67390b07239b" + }, + { + "bytes": 715, + "path": "nonauthor-uniform-local-norm-checks/author-controls.meta.json", + "sha256": "0e8fd7097e06de071dd233b8f6a8ff7623d370e77d78421d7b11a88b9deffac8" + }, + { + "bytes": 0, + "path": "nonauthor-uniform-local-norm-checks/author-controls.stderr", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 1790, + "path": "nonauthor-uniform-local-norm-checks/author-controls.stdout", + "sha256": "5d6cdb2d4925c6ba0b1b545d7a5aca1977655c1d18aa1ea29a82273f30845ece" + }, + { + "bytes": 738, + "path": "nonauthor-uniform-local-norm-checks/independent-controls.meta.json", + "sha256": "c88b17ff52d1eb833679da97a4cb93a2f6068345c785b8e24c8a989ac2c55a34" + }, + { + "bytes": 0, + "path": "nonauthor-uniform-local-norm-checks/independent-controls.stderr", + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + { + "bytes": 1439, + "path": "nonauthor-uniform-local-norm-checks/independent-controls.stdout", + "sha256": "bd553b02c66ad4cf761700f53043ac4b03f7371b5e15aa03ad08b4d4feef2fce" + }, + { + "bytes": 8199, + "path": "nonauthor-uniform-local-norm-checks/review_checks.py", + "sha256": "d764331f77125a27f7f6a6809a813e5b4a2cacc65f66d0dc889300347074f8c5" + }, + { + "bytes": 4309, + "path": "nonauthor-uniform-local-norm-checks/run_checks.py", + "sha256": "98f86202f1ac2a15a18058a7b8cd9a49bbea180f9ab416536b224d6967f639af" + }, + { + "bytes": 443, + "path": "uniform-local-norm-primary-sources/SOURCE_IDENTITY.json", + "sha256": "92bf5297b309e947f42ee1dc837487775e7341bc7c1051e493899963ff85891a" + }, + { + "bytes": 940975, + "path": "uniform-local-norm-primary-sources/simon-lecs-on-pde.pdf", + "sha256": "e1f1b2f51d2558aa467a8c064fe82e7b9e157ad28c9c75f01e38596cde656870" + }, + { + "bytes": 6486, + "path": "uniform_local_norm_controls.py", + "sha256": "4b84755b771598ed34193c075718a41da447719f90a2ff5c6a17d48d0abc3dc4" + } + ], + "helper_manifest_closure": { + "matches": true, + "outputs": 7, + "source_inputs": 11 + }, + "root_relative_paths_only": true, + "verification": "All listed sizes and SHA-256 identities read directly from retained local bytes; no prior controls rerun.", + "verified_frozen_input_count": 41, + "verified_utc": "2026-10-16T02:38:13.687091+00:00" +} diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/INDEPENDENT_KERNEL_CONTROLS.json b/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/INDEPENDENT_KERNEL_CONTROLS.json new file mode 100644 index 000000000..85f09a1dc --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/INDEPENDENT_KERNEL_CONTROLS.json @@ -0,0 +1,152 @@ +{ + "checks": { + "canonical_conditioning_projector_with_singular_extension": "PASS", + "density_weight_radial_volume_exponent_ledger": "PASS", + "fourteen_dimensional_gaussian_interpolation_derivative": "PASS", + "frozen_author_source_identities_unchanged": "PASS", + "literal_six_row_transform_and_twelve_row_block_jacobian": "PASS", + "physical_hessian_congruence_type_and_contact_normalization": "PASS", + "regular_endpoint_integral_rows_degrees_3_through_8": "PASS" + }, + "classification": "INDEPENDENT_EXACT_FINITE_KERNEL_CONTROLS_ONLY", + "frozen_source": { + "JET_FACTORIZATION_OBSTRUCTION_CONTROLS.json": { + "bytes": 1618, + "sha256": "1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764" + }, + "JET_NONDEGENERACY_FACTORIZATION_OBSTRUCTION.md": { + "bytes": 34627, + "sha256": "0dfdddfbcac47b2fd07449f4152b7f5d935f92aed15272fb8333ed9cc60b5766" + }, + "jet_factorization_obstruction_controls.py": { + "bytes": 5818, + "sha256": "5b1e8fc18a6e32e3a312a253e3e4d1110bb674385f3d49447a45e78ad4df541f" + } + }, + "limitations": [ + "No Gaussian draws or floating eigenvalues", + "Finite controls do not prove uniform analytic estimates", + "No Rice integral, process simulation, Lean or hosted execution", + "No scientific or provider-independent approval" + ], + "negative_controls_rejected": { + "missing_12_in_each_fold": true, + "omitted_nonzero_observation_pin_detected": true, + "wrong_r_instead_of_sqrt_r_offdiagonal": true, + "zero_value_crosscovariance_does_not_factor_local_A_weight": true + }, + "principal": "OpenAI/Codex nonauthor; source-exposed; same provider; org credit 0", + "records": { + "density_derivative": [ + { + "log_density_derivative": "19/45", + "rho": "0" + }, + { + "log_density_derivative": "-88840744/281302875", + "rho": "1/8" + }, + { + "log_density_derivative": "-20014868/15946875", + "rho": "1/4" + }, + { + "log_density_derivative": "-3664996/637875", + "rho": "1/2" + }, + { + "log_density_derivative": "2255123/672000", + "rho": "-1/3" + } + ], + "hessians": [ + { + "contact_pair_weight": "2025/16", + "normalized_pair_weight": "256271/2025", + "r": "1/10" + }, + { + "contact_pair_weight": "2025/16", + "normalized_pair_weight": "16402211/129600", + "r": "1/20" + }, + { + "contact_pair_weight": "2025/16", + "normalized_pair_weight": "12814417/101250", + "r": "1/50" + } + ], + "projection": [ + { + "coefficient_dimension": 16, + "conditional_extension_covariance_determinant": "0", + "r": "0" + }, + { + "coefficient_dimension": 16, + "conditional_extension_covariance_determinant": "0", + "r": "1/16" + }, + { + "coefficient_dimension": 16, + "conditional_extension_covariance_determinant": "0", + "r": "1/64" + }, + { + "coefficient_dimension": 20, + "conditional_extension_covariance_determinant": "0", + "r": "0" + }, + { + "coefficient_dimension": 20, + "conditional_extension_covariance_determinant": "0", + "r": "1/16" + }, + { + "coefficient_dimension": 20, + "conditional_extension_covariance_determinant": "0", + "r": "1/64" + }, + { + "coefficient_dimension": 28, + "conditional_extension_covariance_determinant": "0", + "r": "0" + }, + { + "coefficient_dimension": 28, + "conditional_extension_covariance_determinant": "0", + "r": "1/16" + }, + { + "coefficient_dimension": 28, + "conditional_extension_covariance_determinant": "0", + "r": "1/64" + } + ], + "scalar_only_failure": { + "joint_A2_negative_A2_negative": "3/2", + "joint_A_is_shared": true, + "product_marginals": "1/4", + "scope": "negative example for scalar-only inference; cross-jet rho=1, outside lemma", + "value_crosscovariance": "0" + }, + "transform": [ + { + "det_abs": "384", + "r": "1/2" + }, + { + "det_abs": "2916", + "r": "1/3" + }, + { + "det_abs": "12500/81", + "r": "3/5" + }, + { + "det_abs": "393216", + "r": "1/8" + } + ] + } +} diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/PRIMARY_IMPORT_RECEIPT.json b/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/PRIMARY_IMPORT_RECEIPT.json new file mode 100644 index 000000000..454fdef5c --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/PRIMARY_IMPORT_RECEIPT.json @@ -0,0 +1,13 @@ +{ + "bytes": 586327, + "consumed_scope": "Theorem 1 and Corollary 4 only, together giving a local finite positive distance exponent for an analytic determinant. No full paper or original Lojasiewicz proof claimed.", + "sha256": "b11ceb8df55335260ed8eaef4a71406cfa592565869ced020f757354124913c1", + "teissier_saved_pdf": { + "bytes": 325166, + "consumed_scope": "Theorem 3.1 on PDF page 7, rendered and visually inspected. Printed exponent >=0; the standard positive exponent in the nonempty-zero-set application is made explicit by Feehan.", + "relative_path": "../jet-factorization-primary-sources/teissier-lojasiewicz-2012.pdf", + "sha256": "cb88d0410e5c0a629f85592864809b987612663593a04dd8ea2928502cac6873" + }, + "url": "https://arxiv.org/html/1708.09775v6", + "version": "arXiv:1708.09775v6, 11 June 2018" +} diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/PROCESS_RUN_RECEIPT.json b/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/PROCESS_RUN_RECEIPT.json new file mode 100644 index 000000000..737812579 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/PROCESS_RUN_RECEIPT.json @@ -0,0 +1,37 @@ +{ + "python_version": "Python 3.12.14", + "runs": [ + { + "label": "author", + "command": [ + "python3", + "-B", + "jet_factorization_obstruction_controls.py" + ], + "started_utc": "2026-10-16T02:35:18.553011+00:00", + "wall_seconds": 0.49629570869728923, + "returncode": 0, + "stdout_bytes": 1618, + "stdout_sha256": "1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764", + "stderr_bytes": 0, + "stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "reproduces_frozen_json_byte_for_byte": true + }, + { + "label": "independent", + "command": [ + "python3", + "-B", + "nonauthor-jet-factorization-review/independent_kernel_controls.py" + ], + "started_utc": "2026-10-16T02:36:01.939405+00:00", + "wall_seconds": 0.3060152088291943, + "returncode": 0, + "stdout_bytes": 4385, + "stdout_sha256": "1e94c994a8111bd1b741d83e8404a242109a69ac5331cd07eddf8b0930a46330", + "stderr_bytes": 0, + "stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "fixture_refinement": "Nonzero exact cross-quadratic term 19/45 exercises the Gaussian derivative sign." + } + ] +} diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/author_stderr.txt b/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/author_stderr.txt new file mode 100644 index 000000000..e69de29bb diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/author_stdout.json b/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/author_stdout.json new file mode 100644 index 000000000..ce40a1f8c --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/author_stdout.json @@ -0,0 +1,49 @@ +{ + "classification": "EXACT_FINITE_ALGEBRA_CONTROLS_ONLY", + "arithmetic": "Python Fraction and integers; no floating arithmetic", + "Gaussian_sampling": "NONE", + "process_simulation": "NONE", + "scientific_effect": "NONE", + "checks": { + "fixed_base_axial_seven_jet_exact_LDL_positive": "PASS", + "positive_full_square_normalized_uniform_physical_M12": "PASS", + "single_jet_mixture_identity_and_resonant_two_value_collapse": "PASS", + "resonant_uniform_fill_geometric_sum": "PASS", + "DC_single_jet_failure_bookkeeping": "PASS", + "small_fill_C4_dimension_ledger_and_expectation_transfer_cap": "PASS", + "ordered_fold_Jacobian_and_middle_far_kernel_exponents": "PASS", + "consumed_frozen_files_unchanged": "PASS" + }, + "base_axial_LDL_pivots": [ + "1", + "2", + "14/5", + "72/25", + "2", + "4", + "14/5" + ], + "finite_family_cases": [ + { + "N": 4, + "M12": "837542830378721313/17592186044416", + "two_value_small_eigenvalue": "1/17825792" + }, + { + "N": 5, + "M12": "1251428171060188861397852/26285648345947265625", + "two_value_small_eigenvalue": "4/1025390625" + }, + { + "N": 8, + "M12": "695556060602940566395768283/14609821306377964879872", + "two_value_small_eigenvalue": "1/70866960384" + }, + { + "N": 12, + "M12": "3863598447035119394611266277217/81153031520128153386221568", + "two_value_small_eigenvalue": "1/9101852540928" + } + ], + "scope": "Finite checks support arithmetic; all-frame positivity, actual process obstruction, and covariance-kernel estimates require the conventional report proofs." +} diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/independent_kernel_controls.py b/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/independent_kernel_controls.py new file mode 100644 index 000000000..f88907d91 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/independent_kernel_controls.py @@ -0,0 +1,285 @@ +"""Independent exact controls for the scoped nonauthor kernel review. + +These are finite algebra checks, not a numerical proof of Gaussian integral +bounds, asymptotic mixing, analytic distance inequalities, or a process limit. +No author file is written. Run with Python -B to avoid source-side caches. +""" +from fractions import Fraction as Q +from math import comb, factorial +from pathlib import Path +import hashlib +import json + + +def tr(a): + return list(map(list, zip(*a))) + + +def eye(n): + return [[Q(i == j) for j in range(n)] for i in range(n)] + + +def matmul(a, b): + bt = tr(b) + return [[sum(x*y for x, y in zip(row, col)) for col in bt] for row in a] + + +def add(a, b, scale=Q(1)): + return [[x+scale*y for x, y in zip(ra, rb)] for ra, rb in zip(a, b)] + + +def det(a): + a = [list(map(Q, row)) for row in a] + ans = Q(1) + for i in range(len(a)): + p = next((p for p in range(i, len(a)) if a[p][i]), None) + if p is None: + return Q(0) + if p != i: + a[p], a[i] = a[i], a[p] + ans = -ans + value = a[i][i] + ans *= value + for j in range(i+1, len(a)): + ratio = a[j][i]/value + for k in range(i+1, len(a)): + a[j][k] -= ratio*a[i][k] + return ans + + +def inv(a): + n = len(a) + a = [list(map(Q, row))+unit for row, unit in zip(a, eye(n))] + for i in range(n): + p = next(p for p in range(i, n) if a[p][i]) + a[p], a[i] = a[i], a[p] + value = a[i][i] + a[i] = [x/value for x in a[i]] + for j in range(n): + if j != i: + ratio = a[j][i] + a[j] = [x-ratio*y for x, y in zip(a[j], a[i])] + return [row[n:] for row in a] + + +def col(v): + return [[Q(x)] for x in v] + + +def trace(a): + return sum(a[i][i] for i in range(len(a))) + + +def quadratic(a, v): + return matmul(tr(col(v)), matmul(a, col(v)))[0][0] + + +def derivative(p, order=1): + for _ in range(order): + p = [(i+1)*p[i+1] for i in range(len(p)-1)] + return p + + +def evaluate(p, x): + return sum(a*x**i for i, a in enumerate(p)) + + +def compose_affine(p, a, b): + out = [Q(0)]*len(p) + for i, c in enumerate(p): + for j in range(i+1): + out[j] += c*comb(i, j)*a**(i-j)*b**j + return out + + +def integral_01(p): + return sum(c/Q(i+1) for i, c in enumerate(p)) + + +def run(): + checks, records, negatives = {}, {}, {} + # Literal source transform: [FM,FuM,FvM,FS,FuS,FvS] -> six U rows. + transform_records = [] + for r in (Q(1, 2), Q(1, 3), Q(3, 5), Q(1, 8)): + t = [[Q(0)]*6 for _ in range(6)] + t[0][0] = t[0][3] = Q(1, 2) + t[1][0], t[1][3] = -1/r, 1/r + t[2][1], t[2][4] = -1/r, 1/r + t[3][0], t[3][3] = 12/r**3, -12/r**3 + t[3][1] = t[3][4] = 6/r**2 + t[4][2] = t[4][5] = Q(1, 2) + t[5][2], t[5][5] = -1/r, 1/r + actual = abs(det(t)) + assert actual == 12/r**5 + b, k = Q(2, 7), Q(4, 3) + original = [b, 0, 0, b-k*r**3, 0, 0] + target = [b-k*r**3/2, -k*r**2, 0, 12*k, 0, 0] + assert matmul(t, col(original)) == col(target) + block = [[Q(0)]*12 for _ in range(12)] + for block_index in range(2): + for i in range(6): + for j in range(6): + block[6*block_index+i][6*block_index+j] = t[i][j] + assert abs(det(block)) == 144/r**10 + transform_records.append({'r':str(r), 'det_abs':str(actual)}) + negatives['missing_12_in_each_fold'] = actual != 1/r**5 + assert negatives['missing_12_in_each_fold'] + checks['literal_six_row_transform_and_twelve_row_block_jacobian'] = 'PASS' + records['transform'] = transform_records + + # Endpoint extensions are regular integral rows, without gradient conditioning. + for degree in range(3, 9): + p = [Q((-1)**i*(i+2), i+1) for i in range(degree+1)] + for r in (Q(1, 2), Q(2, 7)): + a = -r/2 + p1, p2, p3 = derivative(p), derivative(p, 2), derivative(p, 3) + slope = (evaluate(p1, a+r)-evaluate(p1, a))/r + alpha_m = (evaluate(p2, a)-slope)/r + alpha_s = (evaluate(p2, a+r)-slope)/r + q = compose_affine(p3, a, r) + weighted_s = [Q(0)]+q + assert alpha_m == -integral_01(q)+integral_01(weighted_s) + assert alpha_s == integral_01(weighted_s) + checks['regular_endpoint_integral_rows_degrees_3_through_8'] = 'PASS' + + # A critical pair with nonzero mixed Hessian entries; exact polynomial jets. + hessian_records = [] + for r in (Q(1, 10), Q(1, 20), Q(1, 50)): + b, k, a, c, q = Q(1, 7), Q(3, 2), Q(-5, 4), Q(2, 3), Q(1, 5) + # F(x,y)=b+2kx^3-(3/2)kr^2x-kr^3/2 + # +Ay^2/2+c y(x^2-r^2/4)+q(x^2-r^2/4)^2. + f = [b-k*r**3/2+q*r**4/16, -Q(3, 2)*k*r**2, + -q*r**2/2, 2*k, q] + assert evaluate(f, -r/2) == b + assert evaluate(f, r/2) == b-k*r**3 + product_normalized = Q(1) + for sign in (-1, 1): + x = sign*r/2 + assert evaluate(derivative(f), x) == 0 + assert c*(x*x-r*r/4) == 0 + alpha = evaluate(derivative(f, 2), x)/r + beta = 2*c*x/r + assert alpha == sign*6*k+2*q*r and beta == sign*c + physical_det = (r*alpha)*a-(r*beta)**2 + assert physical_det/r == alpha*a-r*beta**2 + negatives['wrong_r_instead_of_sqrt_r_offdiagonal'] = ( + physical_det/r != alpha*a-r*r*beta**2) + if sign == -1: + assert r*alpha < 0 and physical_det > 0 + max_weight = physical_det/r + else: + assert physical_det < 0 + saddle_weight = -physical_det/r + product_normalized *= physical_det/r + assert negatives['wrong_r_instead_of_sqrt_r_offdiagonal'] + assert max_weight*saddle_weight == -product_normalized + hessian_records.append({'r':str(r), 'normalized_pair_weight':str(-product_normalized), + 'contact_pair_weight':str(36*k*k*a*a)}) + checks['physical_hessian_congruence_type_and_contact_normalization'] = 'PASS' + records['hessians'] = hessian_records + + # Full 14-dimensional log-density derivative, compared to a block closed form. + derivative_records = [] + for rho in (Q(0), Q(1, 8), Q(1, 4), Q(1, 2), Q(-1, 3)): + d = [Q(i+1) for i in range(7)] + sigma = [[Q(0)]*14 for _ in range(14)] + direction = [[Q(0)]*14 for _ in range(14)] + for i, di in enumerate(d): + sigma[i][i] = sigma[7+i][7+i] = di + sigma[i][7+i] = sigma[7+i][i] = rho*di + direction[i][7+i] = direction[7+i][i] = di + si = inv(sigma) + y1 = [Q((-1)**i*(i+1), 9) for i in range(7)] + y2 = [Q(i*i-2, 5) for i in range(7)] + y = y1+y2 + generic = (quadratic(matmul(si, matmul(direction, si)), y) + -trace(matmul(si, direction)))/2 + aa = sum(x*x/di for x, di in zip(y1, d)) + bb = sum(x*x/di for x, di in zip(y2, d)) + cc = sum(x*z/di for x, z, di in zip(y1, y2, d)) + assert cc == Q(19, 45) # exercise the cross-quadratic term and its sign + closed = 7*rho/(1-rho*rho)+(cc*(1+rho*rho)-rho*(aa+bb))/(1-rho*rho)**2 + assert generic == closed + assert det(sigma) == factorial(7)**2*(1-rho*rho)**7 + assert matmul(sigma, si) == eye(14) + derivative_records.append({'rho':str(rho), 'log_density_derivative':str(generic)}) + checks['fourteen_dimensional_gaussian_interpolation_derivative'] = 'PASS' + records['density_derivative'] = derivative_records + # Independent value coordinates but shared A gives a polynomial local failure. + shared_half_fourth, product_half_seconds = Q(3, 2), Q(1, 4) + assert shared_half_fourth != product_half_seconds + negatives['zero_value_crosscovariance_does_not_factor_local_A_weight'] = True + records['scalar_only_failure'] = { + 'value_crosscovariance':'0', 'joint_A_is_shared':True, + 'joint_A2_negative_A2_negative':str(shared_half_fourth), + 'product_marginals':str(product_half_seconds), + 'scope':'negative example for scalar-only inference; cross-jet rho=1, outside lemma'} + + # Canonical coupling with fixed 12 pins, changing coefficient dimension, + # and deliberately singular conditional extension covariance. + projection_records = [] + for n in (16, 20, 28): + for r in (Q(0), Q(1, 16), Q(1, 64)): + l = [[Q(0)]*n for _ in range(12)] + for i in range(12): + l[i][i] = Q(12+i, 24) + l[i][12+i%(n-12)] = r*Q((i%3)-1, 3) + gram = matmul(l, tr(l)) + v = [Q(0)]*12 + v[0], v[3], v[6], v[9] = Q(1, 7), Q(12), Q(-2, 9), Q(18) + c = matmul(tr(l), matmul(inv(gram), col(v))) + p = add(eye(n), matmul(tr(l), matmul(inv(gram), l)), Q(-1)) + assert matmul(l, c) == col(v) + assert matmul(l, p) == [[Q(0)]*n for _ in range(12)] + assert p == tr(p) and matmul(p, p) == p + assert matmul(p, c) == [[Q(0)] for _ in range(n)] + e = [[Q((j%4)-1, 10) for j in range(n)]] + e = e+e # duplicate rows: conditional E covariance is singular. + ecov = matmul(e, matmul(p, tr(e))) + assert det(ecov) == 0 + assert ecov[0][0] >= 0 + altered_v = list(v) + altered_v[9] = 0 + wrong_c = matmul(tr(l), matmul(inv(gram), col(altered_v))) + assert matmul(l, wrong_c) != col(v) + projection_records.append({'coefficient_dimension':n, 'r':str(r), + 'conditional_extension_covariance_determinant':'0'}) + checks['canonical_conditioning_projector_with_singular_extension'] = 'PASS' + negatives['omitted_nonzero_observation_pin_detected'] = True + records['projection'] = projection_records + + # Exact exponent ledger, reported separately from the conventional estimates. + assert Q(12, 2) == 6 + assert 6+8 == 14 and 8+8 == 16 and 6+9 == 15 and 6+10 == 16 + assert Q(1, 3)/2 == Q(1, 6) + assert 1+3-5+2 == 1 + assert Q(4, 3)-Q(2, 3) == Q(2, 3) + checks['density_weight_radial_volume_exponent_ledger'] = 'PASS' + root = Path(__file__).resolve().parents[1] + frozen = { + 'JET_NONDEGENERACY_FACTORIZATION_OBSTRUCTION.md': + (34627, '0dfdddfbcac47b2fd07449f4152b7f5d935f92aed15272fb8333ed9cc60b5766'), + 'jet_factorization_obstruction_controls.py': + (5818, '5b1e8fc18a6e32e3a312a253e3e4d1110bb674385f3d49447a45e78ad4df541f'), + 'JET_FACTORIZATION_OBSTRUCTION_CONTROLS.json': + (1618, '1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764'), + } + source_identity = {} + for name, (size, sha) in frozen.items(): + b = (root/name).read_bytes() + assert len(b) == size and hashlib.sha256(b).hexdigest() == sha + source_identity[name] = {'bytes':size, 'sha256':sha} + checks['frozen_author_source_identities_unchanged'] = 'PASS' + assert all(negatives.values()) + return {'classification':'INDEPENDENT_EXACT_FINITE_KERNEL_CONTROLS_ONLY', + 'principal':'OpenAI/Codex nonauthor; source-exposed; same provider; org credit 0', + 'checks':checks, 'negative_controls_rejected':negatives, + 'records':records, 'frozen_source':source_identity, + 'limitations':['No Gaussian draws or floating eigenvalues', + 'Finite controls do not prove uniform analytic estimates', + 'No Rice integral, process simulation, Lean or hosted execution', + 'No scientific or provider-independent approval']} + + +if __name__ == '__main__': + print(json.dumps(run(), indent=2, sort_keys=True)) diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/independent_stderr.txt b/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/independent_stderr.txt new file mode 100644 index 000000000..e69de29bb diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/independent_stdout.json b/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/independent_stdout.json new file mode 100644 index 000000000..85f09a1dc --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-jet-factorization-review/independent_stdout.json @@ -0,0 +1,152 @@ +{ + "checks": { + "canonical_conditioning_projector_with_singular_extension": "PASS", + "density_weight_radial_volume_exponent_ledger": "PASS", + "fourteen_dimensional_gaussian_interpolation_derivative": "PASS", + "frozen_author_source_identities_unchanged": "PASS", + "literal_six_row_transform_and_twelve_row_block_jacobian": "PASS", + "physical_hessian_congruence_type_and_contact_normalization": "PASS", + "regular_endpoint_integral_rows_degrees_3_through_8": "PASS" + }, + "classification": "INDEPENDENT_EXACT_FINITE_KERNEL_CONTROLS_ONLY", + "frozen_source": { + "JET_FACTORIZATION_OBSTRUCTION_CONTROLS.json": { + "bytes": 1618, + "sha256": "1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764" + }, + "JET_NONDEGENERACY_FACTORIZATION_OBSTRUCTION.md": { + "bytes": 34627, + "sha256": "0dfdddfbcac47b2fd07449f4152b7f5d935f92aed15272fb8333ed9cc60b5766" + }, + "jet_factorization_obstruction_controls.py": { + "bytes": 5818, + "sha256": "5b1e8fc18a6e32e3a312a253e3e4d1110bb674385f3d49447a45e78ad4df541f" + } + }, + "limitations": [ + "No Gaussian draws or floating eigenvalues", + "Finite controls do not prove uniform analytic estimates", + "No Rice integral, process simulation, Lean or hosted execution", + "No scientific or provider-independent approval" + ], + "negative_controls_rejected": { + "missing_12_in_each_fold": true, + "omitted_nonzero_observation_pin_detected": true, + "wrong_r_instead_of_sqrt_r_offdiagonal": true, + "zero_value_crosscovariance_does_not_factor_local_A_weight": true + }, + "principal": "OpenAI/Codex nonauthor; source-exposed; same provider; org credit 0", + "records": { + "density_derivative": [ + { + "log_density_derivative": "19/45", + "rho": "0" + }, + { + "log_density_derivative": "-88840744/281302875", + "rho": "1/8" + }, + { + "log_density_derivative": "-20014868/15946875", + "rho": "1/4" + }, + { + "log_density_derivative": "-3664996/637875", + "rho": "1/2" + }, + { + "log_density_derivative": "2255123/672000", + "rho": "-1/3" + } + ], + "hessians": [ + { + "contact_pair_weight": "2025/16", + "normalized_pair_weight": "256271/2025", + "r": "1/10" + }, + { + "contact_pair_weight": "2025/16", + "normalized_pair_weight": "16402211/129600", + "r": "1/20" + }, + { + "contact_pair_weight": "2025/16", + "normalized_pair_weight": "12814417/101250", + "r": "1/50" + } + ], + "projection": [ + { + "coefficient_dimension": 16, + "conditional_extension_covariance_determinant": "0", + "r": "0" + }, + { + "coefficient_dimension": 16, + "conditional_extension_covariance_determinant": "0", + "r": "1/16" + }, + { + "coefficient_dimension": 16, + "conditional_extension_covariance_determinant": "0", + "r": "1/64" + }, + { + "coefficient_dimension": 20, + "conditional_extension_covariance_determinant": "0", + "r": "0" + }, + { + "coefficient_dimension": 20, + "conditional_extension_covariance_determinant": "0", + "r": "1/16" + }, + { + "coefficient_dimension": 20, + "conditional_extension_covariance_determinant": "0", + "r": "1/64" + }, + { + "coefficient_dimension": 28, + "conditional_extension_covariance_determinant": "0", + "r": "0" + }, + { + "coefficient_dimension": 28, + "conditional_extension_covariance_determinant": "0", + "r": "1/16" + }, + { + "coefficient_dimension": 28, + "conditional_extension_covariance_determinant": "0", + "r": "1/64" + } + ], + "scalar_only_failure": { + "joint_A2_negative_A2_negative": "3/2", + "joint_A_is_shared": true, + "product_marginals": "1/4", + "scope": "negative example for scalar-only inference; cross-jet rho=1, outside lemma", + "value_crosscovariance": "0" + }, + "transform": [ + { + "det_abs": "384", + "r": "1/2" + }, + { + "det_abs": "2916", + "r": "1/3" + }, + { + "det_abs": "12500/81", + "r": "3/5" + }, + { + "det_abs": "393216", + "r": "1/8" + } + ] + } +} diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/FINAL_REVIEW_INTEGRITY.json b/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/FINAL_REVIEW_INTEGRITY.json new file mode 100644 index 000000000..a58b21ccd --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/FINAL_REVIEW_INTEGRITY.json @@ -0,0 +1,22 @@ +{ + "checked_at_utc": "2026-10-16T02:18:44.730512+00:00", + "state": "REVIEW_INPUTS_AND_ACTUAL_CONTROL_RECORDS_VERIFIED", + "verdict": "ACCEPT_SCOPED", + "new_frozen_author_files_unchanged": 3, + "previous_author_files_and_reviews_unchanged": 8, + "total_consumed_inputs_verified_unchanged": 14, + "actual_check_processes_exit_zero": 2, + "author_output_exact_frozen_bytes": 1790, + "independent_finite_cases": 2132, + "unequal_or_singular_PSD_spectra_checked": 2001, + "previous_controls_rerun": false, + "manifest_files_verified": 11, + "report_bytes": 23887, + "report_sha256": "ff552dfacadede53405530a219302c1163dca8ba97d518f02d407970d3de3c1c", + "manifest_bytes": 4763, + "manifest_sha256": "1ccd38374adcf4931b87abeacdefaa64990080a9b3b98835a7c140bde2f505a7", + "production_or_remote_mutations": 0, + "formal_execution": false, + "source_adoption_or_scientific_promotion": "NONE", + "organizational_independence_credit": 0 +} diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/PRIMARY_IMPORT_AUDIT.json b/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/PRIMARY_IMPORT_AUDIT.json new file mode 100644 index 000000000..cfa99edd1 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/PRIMARY_IMPORT_AUDIT.json @@ -0,0 +1,17 @@ +{ + "performer": "OpenAI/Codex /root/c261a_local_runtime_20261016", + "read_scope_recorded_at_utc": "2026-10-16T02:16:52.363424+00:00", + "source_manifest_bytes": 443, + "source_manifest_sha256": "92bf5297b309e947f42ee1dc837487775e7341bc7c1051e493899963ff85891a", + "saved_primary_pdf_verified": { + "url": "https://math.stanford.edu/~lms/lecs-on-pde.pdf", + "path": "work/core-theorem-audit-20261015/uniform-local-norm-primary-sources/simon-lecs-on-pde.pdf", + "bytes": 940975, + "sha256": "e1f1b2f51d2558aa467a8c064fe82e7b9e157ad28c9c75f01e38596cde656870", + "consumed": "Lecture 5 Fourier derivative, inversion, Parseval/Plancherel properties (a)-(h), PDF page 26; Sobolev statement/proof on PDF pages 25 and 27; not the entire book" + }, + "direct_primary_read_scope": "Leon Simon Lectures on PDE, Lecture5 Fourier properties(a)-(h), unitary derivative/inversion/Parseval/Plancherel identities and the displayed Cauchy-Schwarz Sobolev proof; bounded mathematical import only, not the full118-page PDF", + "exact_embedding_constants": "independently derived from the proposal unitary H6 norm and explicit beta integrals; no unnamed uniform Sobolev constant imported", + "old_lemmas": "consumed at exact accepted source/review hashes; no duplicate first review and no old-control rerun", + "review_lineage": "Root and core agent authors; this separate actual OpenAI/Codex nonauthor, source exposed, org0" +} diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/author-controls.stderr b/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/author-controls.stderr new file mode 100644 index 000000000..e69de29bb diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/author-controls.stdout b/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/author-controls.stdout new file mode 100644 index 000000000..cc6abf25c --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/author-controls.stdout @@ -0,0 +1,64 @@ +{ + "classification": "EXACT_FINITE_ALGEBRA_CONTROLS_ONLY", + "arithmetic": "Python Fraction and integers; no floating arithmetic", + "Gaussian_field_sampling": "NONE", + "Rice_density": "NOT_USED", + "process_claim": "NONE", + "scientific_effect": "NONE", + "checks": { + "Fourier_integral_and_rational_operator_constants": "PASS", + "separate_M0_M12_lower_moment_bound_and_DC_falsifier": "PASS", + "explicit_cutoff_derivative_and_H6_trace_recurrence": "PASS", + "multinomial_H6_weights_and_Leibniz_weight_sums": "PASS", + "dimension_free_finite_Gaussian_product_margin": "PASS", + "fixed_cover_rotated_chart_and_endpoint_margins": "PASS", + "growing_K_L_log_cutoff_normalization": "PASS", + "growing_lifetime_cap_radius_failure_control": "PASS", + "six_prior_author_files_and_two_reviews_unchanged": "PASS" + }, + "Gaussian_product_cases": 256, + "constants": { + "A_j": [ + 1, + 4, + 310, + 66942, + 28414348, + 19964924740, + 20971147093146 + ], + "E_j": [ + 8, + 512, + 105216, + 36388608, + 18277303808, + 12499913351680, + 11322431891459328 + ], + "H_j": [ + 8, + 544, + 111792, + 38662896, + 19419635296, + 13281157936160, + 12030083884675536 + ], + "d_j_upper": [ + 6, + 591872, + 24994902528, + 2989639054213632, + 754244470059298015232, + 352778312250451501271091200, + 289445836544660069899559993774592 + ], + "A_theta": "430299201784277045424464881586844996", + "A_theta_decimal_digits": 36, + "embedding_p3": "1/15", + "embedding_p4": "2/15", + "normalized_t_exponent": "8/3" + }, + "proof_status": "Analytic norm, trace, cover, isolation and asymptotic arguments are conventional report proofs; not machine verified" +} diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/independent-controls.stderr b/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/independent-controls.stderr new file mode 100644 index 000000000..e69de29bb diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/independent-controls.stdout b/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/independent-controls.stdout new file mode 100644 index 000000000..db557d4e4 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/independent-controls.stdout @@ -0,0 +1,35 @@ +{ + "state": "INDEPENDENT_EXACT_FINITE_CONTROLS_ONLY", + "all_checks": "PASS", + "arithmetic": "Fraction and integers; no sampled fields or numerical Sobolev integrals", + "counts": { + "closed_cutoff_coefficients_and_reindexed_trace_identity": 8, + "physical_H6_multinomial_weights": 28, + "beta_integral_and_operator_majorant_checks": 7, + "unequal_and_singular_PSD_MGF_product_cases": 2001, + "maximally_correlated_Leibniz_weight_cases": 28, + "all_lower_physical_moments_from_mass_and_M12": 42, + "fixed_physical_cover_and_rotation_parameter_cases": 12, + "periodic_seam_nearest_grid_lift": 1, + "growing_K_L_and_lifetime_normalization_identities": 5 + }, + "total_finite_cases": 2132, + "negative_controls": { + "trace_without_PSD_exponential_moment_failure": "PRESERVED", + "dropping_weight_sum_as_if_derivatives_independent": "REJECTED", + "M12_alone_does_not_bound_DC_cutoff_trace": "PRESERVED", + "finite_smoothness_without_uniform_M12": "PRESERVED", + "log_K_of_order_t_minus_half_does_not_guarantee_radius_inclusion": "PRESERVED", + "Tgap_equal_one_over_t_keeps_radius_upper_bound_constant": "PRESERVED" + }, + "independent_A_theta": "430299201784277045424464881586844996", + "I_j_divided_by_pi": [ + "1/5", + "1/20", + "1/30", + "1/20", + "1/5" + ], + "continuous_and_asymptotic_proof_status": "CONVENTIONAL_REVIEW_NOT_MACHINE_VERIFICATION", + "formal_or_scientific_promotion": "NONE" +} diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/review_checks.py b/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/review_checks.py new file mode 100644 index 000000000..9fa5a181b --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/review_checks.py @@ -0,0 +1,181 @@ +"""Independent exact finite controls for the uniform local norm review. + +No fields are sampled. Continuous cutoff, Fourier, Gaussian, and asymptotic arguments +are reviewed mathematically, not proved by these finite controls. +""" +from fractions import Fraction as Q +from itertools import product +from math import comb, factorial, ceil +from pathlib import Path +import json + + +def poly_mul(a,b,degree): + out=[Q(0)]*(degree+1) + for i,x in enumerate(a): + for j,y in enumerate(b): + if i+j<=degree: + out[i+j]+=x*y + return out + + +def weak_compositions(total,length): + if length==1: + yield (total,) + else: + for first in range(total+1): + for rest in weak_compositions(total-first,length-1): + yield (first,)+rest + + +def run(): + counts={};negative={} + # Independent closed derivative coefficients (unsigned Lah numbers with alternating signs), + # checked against differentiation rather than copied from the author's sparse recurrence. + L=[{0:1}] + for j in range(1,7): + L.append({j+k:(-1)**(j-k)*factorial(j)//factorial(k)*comb(j-1,k-1) for k in range(1,j+1)}) + for j in range(6): + direct={} + for power,coefficient in L[j].items(): + direct[power+2]=direct.get(power+2,0)+coefficient + if power: + direct[power+1]=direct.get(power+1,0)-power*coefficient + assert {m:c for m,c in direct.items() if c}==L[j+1] + aa=[sum(abs(c)*max(1,m**m) for m,c in polynomial.items()) for polynomial in L] + # E(t)=8/(1-16 sum_(j>=1) A_j t^j/j!) as a truncated exponential generating series. + base=[Q(0)]+[Q(16*aa[j],factorial(j)) for j in range(1,7)] + power=[Q(1)]+[Q(0)]*6;reciprocal=[Q(0)]*7 + for k in range(7): + reciprocal=[x+8*y for x,y in zip(reciprocal,power)] + power=poly_mul(power,base,6) + ee_raw=[reciprocal[j]*factorial(j) for j in range(7)] + assert all(x.denominator==1 for x in ee_raw) + ee=[int(x) for x in ee_raw] + gg_coeff=poly_mul([Q(aa[j],factorial(j)) for j in range(7)],reciprocal,6) + gg_raw=[gg_coeff[j]*factorial(j) for j in range(7)] + assert all(x.denominator==1 for x in gg_raw) + gg=[int(x) for x in gg_raw] + dd=[6]+[2*gg[j]**2 for j in range(1,7)] + # Sum first over alpha-gamma, using its multinomial expansion (1+2+2)^(6-|gamma|). + atheta=sum(factorial(6)//(factorial(i)*factorial(j)*factorial(6-i-j)) + *2**(i+j)*5**(6-i-j)*dd[i]*dd[j] + for i in range(7) for j in range(7-i)) + frozen=json.loads((Path(__file__).resolve().parent.parent/'UNIFORM_LOCAL_NORM_CONTROLS.json').read_text())['constants'] + assert aa==frozen['A_j'] and ee==frozen['E_j'] and gg==frozen['H_j'] and dd==frozen['d_j_upper'] + assert atheta==int(frozen['A_theta'])==430299201784277045424464881586844996 + counts['closed_cutoff_coefficients_and_reindexed_trace_identity']=8 + + # Polynomial expansion of (1+x+y)^6, independent of a direct multinomial loop. + weights={(0,0):1} + for _ in range(6): + out={} + for (i,j),c in weights.items(): + for shift in [(0,0),(1,0),(0,1)]: + key=(i+shift[0],j+shift[1]);out[key]=out.get(key,0)+c + weights=out + assert sum(weights.values())==3**6 + assert all(w==factorial(6)//(factorial(i)*factorial(j)*factorial(6-i-j)) for (i,j),w in weights.items()) + counts['physical_H6_multinomial_weights']=len(weights) + + # Exact beta integrals via expanding v^j(1-v)^(4-j) on [0,1]. + integrals=[] + for j in range(5): + integral=sum(Q((-1)**k*comb(4-j,k),j+k+1) for k in range(5-j)) + assert integral==Q(factorial(j)*factorial(4-j),factorial(5)) + integrals.append(str(integral)) + assert Q(1,15)**2>Q(1,240) and Q(2,15)**2>Q(1,60) + counts['beta_integral_and_operator_majorant_checks']=7 + + # Include all unequal, rank-deficient and rank-one spectra of a small rational trace budget. + ncases=0 + for dimension in range(1,6): + for total in range(9): + for parts in weak_compositions(total,dimension): + aj=[Q(p,16) for p in parts] + margin=Q(1) + for a in aj: + margin*=1-a + assert margin>=1-sum(aj)>=Q(1,2) + ncases+=1 + assert ncases==2001 + assert 1-Q(1,2)==Q(1,2) # rank one lambda=T attains the sqrt(2) bound. + counts['unequal_and_singular_PSD_MGF_product_cases']=ncases + # Trace without PSD is insufficient: a negative eigenvalue can conceal a divergent positive one. + eigenvalues=[Q(3),Q(-2)];T=Q(1) + assert sum(eigenvalues)==T and eigenvalues[0]>=2*T + negative['trace_without_PSD_exponential_moment_failure']='PRESERVED' + + # Weighted Cauchy is needed when all Gaussian summands are the same variable. + count=0 + for i,j in weights: + c=[comb(i,u)*comb(j,v) for u in range(i+1) for v in range(j+1)] + assert sum(c)==2**(i+j) + assert sum(c)**2==2**(i+j)*sum(c) + if i+j: + assert sum(c)**2>sum(c) + count+=1 + counts['maximally_correlated_Leibniz_weight_cases']=count + negative['dropping_weight_sum_as_if_derivatives_independent']='REJECTED' + + count=0 + for r,j in product([Q(0),Q(1,100),Q(1,2),Q(1),Q(2),Q(100)],range(7)): + assert r**(2*j)<=1+r**12;count+=1 + counts['all_lower_physical_moments_from_mass_and_M12']=count + # A constant Gaussian mode has no positive spectral moment but grows the cutoff trace. + for amplitude in [Q(1),Q(10),Q(1000)]: + b=Q(1,16) + dc_M12=Q(0) + trace_lower=16*b*b*amplitude*amplitude + assert dc_M12==0 and trace_lower>0 + negative['M12_alone_does_not_bound_DC_cutoff_trace']='PRESERVED' + # Normalized high-frequency mass: fixed L and mass imply sixth derivative variance grows K^12. + for K in [2,4,8,16]: + mass=Q(1,3) + assert mass*K**12 > mass*K**8 + negative['finite_smoothness_without_uniform_M12']='PRESERVED' + + # Physical cover, support embedding, arbitrary rational rotations and a periodic seam. + rotations=[(Q(1),Q(0)),(Q(3,5),Q(4,5)),(Q(5,13),Q(12,13)),(Q(-12,13),Q(5,13))] + count=0 + for Lmin in [Q(1,10),Q(1),Q(16),Q(100)]: + b=min(Q(1),Lmin/16) + for length in [Lmin,Lmin+Q(1,7),10*Lmin]: + n=ceil(length/b) + assert 6*bb/2 + counts['periodic_seam_nearest_grid_lift']=1 + + # Abstract logarithmic coefficients, without evaluating exponential samples. + assert 8*Q(1,3)+2*Q(1,3)+4==Q(22,3) + assert Q(4)-Q(4,3)==Q(8,3) + assert 16**4==65536 + for p,q,u in [(8,2,Q(4)),(9,2,Q(3,2)),(8,3,Q(2))]: + assert 8-p<=0 and 2-q<=0 and u-Q(4,3)>0 + counts['growing_K_L_and_lifetime_normalization_identities']=5 + assert Q(1,2)-Q(1,3)==Q(1,6)>0 + negative['log_K_of_order_t_minus_half_does_not_guarantee_radius_inclusion']='PRESERVED' + assert Q(1,3)-Q(1,3)==0 + negative['Tgap_equal_one_over_t_keeps_radius_upper_bound_constant']='PRESERVED' + return {'state':'INDEPENDENT_EXACT_FINITE_CONTROLS_ONLY','all_checks':'PASS', + 'arithmetic':'Fraction and integers; no sampled fields or numerical Sobolev integrals', + 'counts':counts,'total_finite_cases':sum(counts.values()),'negative_controls':negative, + 'independent_A_theta':str(atheta),'I_j_divided_by_pi':integrals, + 'continuous_and_asymptotic_proof_status':'CONVENTIONAL_REVIEW_NOT_MACHINE_VERIFICATION', + 'formal_or_scientific_promotion':'NONE'} + + +if __name__=='__main__': + print(json.dumps(run(),indent=2)) diff --git a/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/run_checks.py b/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/run_checks.py new file mode 100644 index 000000000..e98b4bf2b --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/nonauthor-uniform-local-norm-checks/run_checks.py @@ -0,0 +1,64 @@ +import datetime +import hashlib +import json +import pathlib +import subprocess +import sys +import time + +HERE=pathlib.Path(__file__).resolve().parent +ROOT=HERE.parent +EXPECTED={ + 'UNIFORM_LOCAL_NORM_DIAGONAL_PROPOSAL.md':(26687,'8e1eefca9888d43824a9fd9caca34c44d2de4bf8e30c668b474fa2097491c76b'), + 'uniform_local_norm_controls.py':(6486,'4b84755b771598ed34193c075718a41da447719f90a2ff5c6a17d48d0abc3dc4'), + 'UNIFORM_LOCAL_NORM_CONTROLS.json':(1790,'5d6cdb2d4925c6ba0b1b545d7a5aca1977655c1d18aa1ea29a82273f30845ece'), + 'QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md':(19138,'827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30'), + 'fold_isolation_controls.py':(6308,'13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5'), + 'FOLD_ISOLATION_CONTROLS.json':(795,'ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf'), + 'FINITE_FOURIER_CRITICAL_COUNT_DIAGONAL_ROUTE.md':(25699,'621cee96854aa440498cecf5452a8d1e23f90a4458779c3286c18eba106fe138'), + 'finite_fourier_count_controls.py':(5153,'fd6f29739280a55017945526cb75fb46ebc80d86c012ff4a1ff65baa404c31de'), + 'FINITE_FOURIER_COUNT_CONTROLS.json':(849,'bbe147fdf409a3e37b450d4cbcd5d2a927140f95bd3ce0f36cafe16861123bfd'), + 'NONAUTHOR_FOLD_ISOLATION_REVIEW.md':(18409,'f06c61e3f3e60c795a1a14e824640f1689d4143dba4f6698cada7a41929eec90'), + 'NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW.md':(21880,'16152427f7013822ac1bba81142429800334e6bf0f354a6439b23a1997bc4524'), + 'uniform-local-norm-primary-sources/SOURCE_IDENTITY.json':(443,'92bf5297b309e947f42ee1dc837487775e7341bc7c1051e493899963ff85891a'), + 'uniform-local-norm-primary-sources/simon-lecs-on-pde.pdf':(940975,'e1f1b2f51d2558aa467a8c064fe82e7b9e157ad28c9c75f01e38596cde656870'), + 'experiments/universality/finite_gaussian_contact/PROOF.md':(25367,'7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474')} + +def inventory(): + rows=[] + for name,(size,expected) in EXPECTED.items(): + b=(ROOT/name).read_bytes();digest=hashlib.sha256(b).hexdigest() + assert len(b)==size and digest==expected,name + rows.append({'path':name,'bytes':len(b),'sha256':digest}) + return rows + +before=inventory();runs=[] +for label,script in [('author-controls',ROOT/'uniform_local_norm_controls.py'),('independent-controls',HERE/'review_checks.py')]: + argv=[sys.executable,'-B',str(script)] + start=datetime.datetime.now(datetime.timezone.utc).isoformat();tick=time.perf_counter() + r=subprocess.run(argv,cwd=ROOT,stdout=subprocess.PIPE,stderr=subprocess.PIPE) + elapsed=time.perf_counter()-tick + (HERE/(label+'.stdout')).write_bytes(r.stdout);(HERE/(label+'.stderr')).write_bytes(r.stderr) + record={'command_argv':argv,'cwd':str(ROOT),'started_at_utc':start, + 'finished_at_utc':datetime.datetime.now(datetime.timezone.utc).isoformat(), + 'wall_seconds':elapsed,'exit_code':r.returncode,'stdout':label+'.stdout','stderr':label+'.stderr', + 'stdout_sha256':hashlib.sha256(r.stdout).hexdigest(),'stderr_sha256':hashlib.sha256(r.stderr).hexdigest()} + (HERE/(label+'.meta.json')).write_text(json.dumps(record,indent=2)+'\n') + assert r.returncode==0,(label,r.stderr.decode()) + output=json.loads(r.stdout) + if label=='author-controls': + assert r.stdout==(ROOT/'UNIFORM_LOCAL_NORM_CONTROLS.json').read_bytes() + record['exact_frozen_output_match']=True + else: + assert output['all_checks']=='PASS' + record['independent_finite_cases']=output['total_finite_cases'] + runs.append(record) +after=inventory();assert before==after +result={'state':'NEW_CONTROLS_REPLAY_AND_INDEPENDENT_FINITE_CHECKS_ONLY','performer':'OpenAI/Codex /root/c261a_local_runtime_20261016', + 'source_exposure':'new proposal and controls read fully; bounded Fourier/Plancherel import read; prior two exact lemmas consumed, not re-reviewed', + 'organizational_independence_credit':0,'interpreter_version':sys.version,'inputs_before':before,'inputs_after':after, + 'all_14_inputs_unchanged':True,'runs':runs,'previous_controls_rerun':False, + 'production_remote_mutations_or_formal_execution':0,'primary_source_read_only_browsing':True, + 'source_adoption_or_scientific_promotion':'NONE'} +(HERE/'CHECK_RUN_SUMMARY.json').write_text(json.dumps(result,indent=2)+'\n') +print(json.dumps(result,indent=2)) diff --git a/experiments/universality/fold_isolation_diagonal/publication-projections/counterfamily-nonauthor-checks-20261016/EXECUTION_RECORD.json b/experiments/universality/fold_isolation_diagonal/publication-projections/counterfamily-nonauthor-checks-20261016/EXECUTION_RECORD.json new file mode 100644 index 000000000..9aa197651 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/publication-projections/counterfamily-nonauthor-checks-20261016/EXECUTION_RECORD.json @@ -0,0 +1,83 @@ +{ + "original_bytes": 2560, + "original_disposition": "WITHHELD_LOCAL_CAPTURE_PATHS", + "original_relative_path": "counterfamily-nonauthor-checks-20261016/EXECUTION_RECORD.json", + "original_sha256": "8ae808eed6b7f8fec67435a3b97c8422a08212e7f18ababbced3b367380e04fb", + "path_transform": "Absolute original capture-root prefix replaced with ; all nonpath values preserved. No new execution or original-byte verification is supplied by this derivative.", + "projected_record": { + "after": { + "JET_FACTORIZATION_OBSTRUCTION_CONTROLS.json": { + "bytes": 1618, + "sha256": "1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764" + }, + "JET_NONDEGENERACY_FACTORIZATION_OBSTRUCTION.md": { + "bytes": 34627, + "sha256": "0dfdddfbcac47b2fd07449f4152b7f5d935f92aed15272fb8333ed9cc60b5766" + }, + "jet_factorization_obstruction_controls.py": { + "bytes": 5818, + "sha256": "5b1e8fc18a6e32e3a312a253e3e4d1110bb674385f3d49447a45e78ad4df541f" + } + }, + "before": { + "JET_FACTORIZATION_OBSTRUCTION_CONTROLS.json": { + "bytes": 1618, + "sha256": "1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764" + }, + "JET_NONDEGENERACY_FACTORIZATION_OBSTRUCTION.md": { + "bytes": 34627, + "sha256": "0dfdddfbcac47b2fd07449f4152b7f5d935f92aed15272fb8333ed9cc60b5766" + }, + "jet_factorization_obstruction_controls.py": { + "bytes": 5818, + "sha256": "5b1e8fc18a6e32e3a312a253e3e4d1110bb674385f3d49447a45e78ad4df541f" + } + }, + "commands": [ + { + "argv": [ + "python3", + "-B", + "jet_factorization_obstruction_controls.py" + ], + "cwd": "", + "end_utc": "2026-10-16T02:33:00.849744+00:00", + "exit_code": 0, + "label": "author_controls", + "start_utc": "2026-10-16T02:32:59.641490+00:00", + "stderr": { + "bytes": 0, + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + "stdout": { + "bytes": 1618, + "sha256": "1d79c4f420d8a007638eff98a0b93ffed74e5b1acbac60c2f7777c2cf012e764" + }, + "wall_seconds": 1.2081267500000001 + }, + { + "argv": [ + "python3", + "-B", + "/counterfamily-nonauthor-checks-20261016/counterfamily_checks.py" + ], + "cwd": "", + "end_utc": "2026-10-16T02:33:01.179511+00:00", + "exit_code": 0, + "label": "independent_controls", + "start_utc": "2026-10-16T02:33:00.849850+00:00", + "stderr": { + "bytes": 0, + "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" + }, + "stdout": { + "bytes": 12359, + "sha256": "3c53ed54756c911292180c5686750de9a47cf4a54c56cc0d2616155c3f5af35b" + }, + "wall_seconds": 0.32966637499999996 + } + ], + "frozen_author_files_unchanged": true + }, + "projection_kind": "PUBLICATION_PATH_ROLE_DERIVATIVE_NOT_ORIGINAL" +} diff --git a/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-finite-count-checks/CHECK_RUN_SUMMARY.json b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-finite-count-checks/CHECK_RUN_SUMMARY.json new file mode 100644 index 000000000..c242f2b92 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-finite-count-checks/CHECK_RUN_SUMMARY.json @@ -0,0 +1,178 @@ +{ + "original_bytes": 6479, + "original_disposition": "WITHHELD_LOCAL_CAPTURE_PATHS", + "original_relative_path": "nonauthor-finite-count-checks/CHECK_RUN_SUMMARY.json", + "original_sha256": "5faeb21e56f042394de6772e54081fd16eded69b42abd89109773d5193898132", + "path_transform": "Absolute original capture-root prefix replaced with ; all nonpath values preserved. No new execution or original-byte verification is supplied by this derivative.", + "projected_record": { + "all_12_inputs_unchanged": true, + "inputs_after": [ + { + "bytes": 25699, + "path": "FINITE_FOURIER_CRITICAL_COUNT_DIAGONAL_ROUTE.md", + "sha256": "621cee96854aa440498cecf5452a8d1e23f90a4458779c3286c18eba106fe138" + }, + { + "bytes": 5153, + "path": "finite_fourier_count_controls.py", + "sha256": "fd6f29739280a55017945526cb75fb46ebc80d86c012ff4a1ff65baa404c31de" + }, + { + "bytes": 849, + "path": "FINITE_FOURIER_COUNT_CONTROLS.json", + "sha256": "bbe147fdf409a3e37b450d4cbcd5d2a927140f95bd3ce0f36cafe16861123bfd" + }, + { + "bytes": 19138, + "path": "QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md", + "sha256": "827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30" + }, + { + "bytes": 6308, + "path": "fold_isolation_controls.py", + "sha256": "13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5" + }, + { + "bytes": 795, + "path": "FOLD_ISOLATION_CONTROLS.json", + "sha256": "ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf" + }, + { + "bytes": 796, + "path": "critical-count-primary-sources/SOURCE_IDENTITIES.json", + "sha256": "2c42d92a4732f09a58aebd9b9ef8672b51b672ab99f3bbdad61681a0dcf4f504" + }, + { + "bytes": 83539, + "path": "critical-count-primary-sources/guth-lect13.pdf", + "sha256": "92ed0fc0f940179871a25f3a244eb181aaa0176ee5738ac1f8f68a8bacd072a9" + }, + { + "bytes": 15419, + "path": "critical-count-primary-sources/stacks-01WC.html", + "sha256": "08a39f68a8159e34e436df9c4b8f5e8886f58b3c0d5163c15610e12aef302737" + }, + { + "bytes": 14844, + "path": "critical-count-primary-sources/stacks-01W6.html", + "sha256": "060b8c179f815d59977f8a31057b5f5f9e1f2104351d46c33f1fb6d9ed95a202" + }, + { + "bytes": 25367, + "path": "experiments/universality/finite_gaussian_contact/PROOF.md", + "sha256": "7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474" + }, + { + "bytes": 18409, + "path": "NONAUTHOR_FOLD_ISOLATION_REVIEW.md", + "sha256": "f06c61e3f3e60c795a1a14e824640f1689d4143dba4f6698cada7a41929eec90" + } + ], + "inputs_before": [ + { + "bytes": 25699, + "path": "FINITE_FOURIER_CRITICAL_COUNT_DIAGONAL_ROUTE.md", + "sha256": "621cee96854aa440498cecf5452a8d1e23f90a4458779c3286c18eba106fe138" + }, + { + "bytes": 5153, + "path": "finite_fourier_count_controls.py", + "sha256": "fd6f29739280a55017945526cb75fb46ebc80d86c012ff4a1ff65baa404c31de" + }, + { + "bytes": 849, + "path": "FINITE_FOURIER_COUNT_CONTROLS.json", + "sha256": "bbe147fdf409a3e37b450d4cbcd5d2a927140f95bd3ce0f36cafe16861123bfd" + }, + { + "bytes": 19138, + "path": "QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md", + "sha256": "827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30" + }, + { + "bytes": 6308, + "path": "fold_isolation_controls.py", + "sha256": "13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5" + }, + { + "bytes": 795, + "path": "FOLD_ISOLATION_CONTROLS.json", + "sha256": "ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf" + }, + { + "bytes": 796, + "path": "critical-count-primary-sources/SOURCE_IDENTITIES.json", + "sha256": "2c42d92a4732f09a58aebd9b9ef8672b51b672ab99f3bbdad61681a0dcf4f504" + }, + { + "bytes": 83539, + "path": "critical-count-primary-sources/guth-lect13.pdf", + "sha256": "92ed0fc0f940179871a25f3a244eb181aaa0176ee5738ac1f8f68a8bacd072a9" + }, + { + "bytes": 15419, + "path": "critical-count-primary-sources/stacks-01WC.html", + "sha256": "08a39f68a8159e34e436df9c4b8f5e8886f58b3c0d5163c15610e12aef302737" + }, + { + "bytes": 14844, + "path": "critical-count-primary-sources/stacks-01W6.html", + "sha256": "060b8c179f815d59977f8a31057b5f5f9e1f2104351d46c33f1fb6d9ed95a202" + }, + { + "bytes": 25367, + "path": "experiments/universality/finite_gaussian_contact/PROOF.md", + "sha256": "7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474" + }, + { + "bytes": 18409, + "path": "NONAUTHOR_FOLD_ISOLATION_REVIEW.md", + "sha256": "f06c61e3f3e60c795a1a14e824640f1689d4143dba4f6698cada7a41929eec90" + } + ], + "interpreter_version": "3.9.6 (default, Dec 2 2025, 07:27:58) \n[Clang 17.0.0 (clang-1700.6.3.2)]", + "organization_independence_credit": 0, + "performer": "OpenAI/Codex /root/c261a_local_runtime_20261016", + "production_or_remote_actions": 0, + "runs": [ + { + "command_argv": [ + "/Library/Developer/CommandLineTools/usr/bin/python3", + "-B", + "/finite_fourier_count_controls.py" + ], + "cwd": "", + "exact_frozen_output_match": true, + "exit_code": 0, + "finished_at_utc": "2026-10-16T01:59:41.129068+00:00", + "started_at_utc": "2026-10-16T01:59:41.103430+00:00", + "stderr": "author-controls.stderr", + "stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "stdout": "author-controls.stdout", + "stdout_sha256": "bbe147fdf409a3e37b450d4cbcd5d2a927140f95bd3ce0f36cafe16861123bfd", + "wall_seconds": 0.025243334000000003 + }, + { + "command_argv": [ + "/Library/Developer/CommandLineTools/usr/bin/python3", + "-B", + "/nonauthor-finite-count-checks/review_checks.py" + ], + "cwd": "", + "exit_code": 0, + "finished_at_utc": "2026-10-16T01:59:41.158468+00:00", + "independent_finite_cases": 226, + "started_at_utc": "2026-10-16T01:59:41.129282+00:00", + "stderr": "independent-controls.stderr", + "stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "stdout": "independent-controls.stdout", + "stdout_sha256": "c2fd4842cc1bc9ba9e7a947d8630a256386064bd63e94b724ffa699258ae7046", + "wall_seconds": 0.028933707999999995 + } + ], + "scientific_acceptance": "NONE", + "source_exposure": "full new frozen proposal, control source/output and named primary imports read; prior exact chart consumed", + "state": "EXACT_CONTROL_REPLAY_AND_SEPARATE_FINITE_CHECKS_ONLY" + }, + "projection_kind": "PUBLICATION_PATH_ROLE_DERIVATIVE_NOT_ORIGINAL" +} diff --git a/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-finite-count-checks/author-controls.meta.json b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-finite-count-checks/author-controls.meta.json new file mode 100644 index 000000000..c3effd75a --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-finite-count-checks/author-controls.meta.json @@ -0,0 +1,24 @@ +{ + "original_bytes": 717, + "original_disposition": "WITHHELD_LOCAL_CAPTURE_PATHS", + "original_relative_path": "nonauthor-finite-count-checks/author-controls.meta.json", + "original_sha256": "07fa0d98d0f407f912444426ffdacaf57f5b0f51928f52be197937ef00bd1463", + "path_transform": "Absolute original capture-root prefix replaced with ; all nonpath values preserved. No new execution or original-byte verification is supplied by this derivative.", + "projected_record": { + "command_argv": [ + "/Library/Developer/CommandLineTools/usr/bin/python3", + "-B", + "/finite_fourier_count_controls.py" + ], + "cwd": "", + "exit_code": 0, + "finished_at_utc": "2026-10-16T01:59:41.129068+00:00", + "started_at_utc": "2026-10-16T01:59:41.103430+00:00", + "stderr": "author-controls.stderr", + "stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "stdout": "author-controls.stdout", + "stdout_sha256": "bbe147fdf409a3e37b450d4cbcd5d2a927140f95bd3ce0f36cafe16861123bfd", + "wall_seconds": 0.025243334000000003 + }, + "projection_kind": "PUBLICATION_PATH_ROLE_DERIVATIVE_NOT_ORIGINAL" +} diff --git a/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-finite-count-checks/independent-controls.meta.json b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-finite-count-checks/independent-controls.meta.json new file mode 100644 index 000000000..747342d1c --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-finite-count-checks/independent-controls.meta.json @@ -0,0 +1,24 @@ +{ + "original_bytes": 741, + "original_disposition": "WITHHELD_LOCAL_CAPTURE_PATHS", + "original_relative_path": "nonauthor-finite-count-checks/independent-controls.meta.json", + "original_sha256": "49e1a99d70a759a0e2e108b1a0e40303fbd05d6f3c6afb8ad45384abfebd4edf", + "path_transform": "Absolute original capture-root prefix replaced with ; all nonpath values preserved. No new execution or original-byte verification is supplied by this derivative.", + "projected_record": { + "command_argv": [ + "/Library/Developer/CommandLineTools/usr/bin/python3", + "-B", + "/nonauthor-finite-count-checks/review_checks.py" + ], + "cwd": "", + "exit_code": 0, + "finished_at_utc": "2026-10-16T01:59:41.158468+00:00", + "started_at_utc": "2026-10-16T01:59:41.129282+00:00", + "stderr": "independent-controls.stderr", + "stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "stdout": "independent-controls.stdout", + "stdout_sha256": "c2fd4842cc1bc9ba9e7a947d8630a256386064bd63e94b724ffa699258ae7046", + "wall_seconds": 0.028933707999999995 + }, + "projection_kind": "PUBLICATION_PATH_ROLE_DERIVATIVE_NOT_ORIGINAL" +} diff --git a/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-fold-isolation-checks/CHECK_RUN_SUMMARY.json b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-fold-isolation-checks/CHECK_RUN_SUMMARY.json new file mode 100644 index 000000000..68d4622d5 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-fold-isolation-checks/CHECK_RUN_SUMMARY.json @@ -0,0 +1,86 @@ +{ + "original_bytes": 3567, + "original_disposition": "WITHHELD_LOCAL_CAPTURE_PATHS", + "original_relative_path": "nonauthor-fold-isolation-checks/CHECK_RUN_SUMMARY.json", + "original_sha256": "638503d7560f87424e55cb3f118ae63b3bbc355e01350a86e40076b155e7429d", + "path_transform": "Absolute original capture-root prefix replaced with ; all nonpath values preserved. No new execution or original-byte verification is supplied by this derivative.", + "projected_record": { + "exposure": "frozen proposal and source read; nonauthor of reviewed bytes", + "frozen_inputs_after": [ + { + "bytes": 19138, + "path": "/QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md", + "sha256": "827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30" + }, + { + "bytes": 6308, + "path": "/fold_isolation_controls.py", + "sha256": "13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5" + }, + { + "bytes": 795, + "path": "/FOLD_ISOLATION_CONTROLS.json", + "sha256": "ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf" + } + ], + "frozen_inputs_before": [ + { + "bytes": 19138, + "path": "/QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md", + "sha256": "827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30" + }, + { + "bytes": 6308, + "path": "/fold_isolation_controls.py", + "sha256": "13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5" + }, + { + "bytes": 795, + "path": "/FOLD_ISOLATION_CONTROLS.json", + "sha256": "ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf" + } + ], + "frozen_inputs_unchanged": true, + "interpreter_version": "3.9.6 (default, Dec 2 2025, 07:27:58) \n[Clang 17.0.0 (clang-1700.6.3.2)]", + "organization_independence_credit": 0, + "performer": "OpenAI/Codex /root/c261a_local_runtime_20261016", + "runs": [ + { + "command_argv": [ + "/Library/Developer/CommandLineTools/usr/bin/python3", + "-B", + "/fold_isolation_controls.py" + ], + "cwd": "", + "exact_frozen_output_match": true, + "exit_code": 0, + "finished_at_utc": "2026-10-16T01:48:33.913456+00:00", + "started_at_utc": "2026-10-16T01:48:33.890108+00:00", + "stderr": "author-controls.stderr", + "stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "stdout": "author-controls.stdout", + "stdout_sha256": "ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf", + "wall_seconds": 0.022967584 + }, + { + "command_argv": [ + "/Library/Developer/CommandLineTools/usr/bin/python3", + "-B", + "/nonauthor-fold-isolation-checks/review_checks.py" + ], + "cwd": "", + "exit_code": 0, + "finished_at_utc": "2026-10-16T01:48:33.958915+00:00", + "finite_cases": 255, + "started_at_utc": "2026-10-16T01:48:33.913680+00:00", + "stderr": "independent-controls.stderr", + "stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "stdout": "independent-controls.stdout", + "stdout_sha256": "4b1eb60b2f44d9e67caf3469dbba06d90fa29d7d32a74cb6fe46488b988bd8ec", + "wall_seconds": 0.044948958 + } + ], + "state": "LOCAL_REPLAY_AND_INDEPENDENT_FINITE_CONTROLS_VERIFIED_NO_PROOF_ACCEPTANCE" + }, + "projection_kind": "PUBLICATION_PATH_ROLE_DERIVATIVE_NOT_ORIGINAL" +} diff --git a/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-fold-isolation-checks/author-controls.meta.json b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-fold-isolation-checks/author-controls.meta.json new file mode 100644 index 000000000..a7b31b07a --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-fold-isolation-checks/author-controls.meta.json @@ -0,0 +1,24 @@ +{ + "original_bytes": 702, + "original_disposition": "WITHHELD_LOCAL_CAPTURE_PATHS", + "original_relative_path": "nonauthor-fold-isolation-checks/author-controls.meta.json", + "original_sha256": "932a196891241aba910293fbdf326b116e8bb341dfe4f3d0f37158aae2a9d05f", + "path_transform": "Absolute original capture-root prefix replaced with ; all nonpath values preserved. No new execution or original-byte verification is supplied by this derivative.", + "projected_record": { + "command_argv": [ + "/Library/Developer/CommandLineTools/usr/bin/python3", + "-B", + "/fold_isolation_controls.py" + ], + "cwd": "", + "exit_code": 0, + "finished_at_utc": "2026-10-16T01:48:33.913456+00:00", + "started_at_utc": "2026-10-16T01:48:33.890108+00:00", + "stderr": "author-controls.stderr", + "stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "stdout": "author-controls.stdout", + "stdout_sha256": "ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf", + "wall_seconds": 0.022967584 + }, + "projection_kind": "PUBLICATION_PATH_ROLE_DERIVATIVE_NOT_ORIGINAL" +} diff --git a/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-fold-isolation-checks/independent-controls.meta.json b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-fold-isolation-checks/independent-controls.meta.json new file mode 100644 index 000000000..5344cbc6e --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-fold-isolation-checks/independent-controls.meta.json @@ -0,0 +1,24 @@ +{ + "original_bytes": 734, + "original_disposition": "WITHHELD_LOCAL_CAPTURE_PATHS", + "original_relative_path": "nonauthor-fold-isolation-checks/independent-controls.meta.json", + "original_sha256": "3606ca471c372e5f86b936c842a40eccd5b7ee9c17471f062872742bc6b156cc", + "path_transform": "Absolute original capture-root prefix replaced with ; all nonpath values preserved. No new execution or original-byte verification is supplied by this derivative.", + "projected_record": { + "command_argv": [ + "/Library/Developer/CommandLineTools/usr/bin/python3", + "-B", + "/nonauthor-fold-isolation-checks/review_checks.py" + ], + "cwd": "", + "exit_code": 0, + "finished_at_utc": "2026-10-16T01:48:33.958915+00:00", + "started_at_utc": "2026-10-16T01:48:33.913680+00:00", + "stderr": "independent-controls.stderr", + "stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "stdout": "independent-controls.stdout", + "stdout_sha256": "4b1eb60b2f44d9e67caf3469dbba06d90fa29d7d32a74cb6fe46488b988bd8ec", + "wall_seconds": 0.044948958 + }, + "projection_kind": "PUBLICATION_PATH_ROLE_DERIVATIVE_NOT_ORIGINAL" +} diff --git a/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-uniform-local-norm-checks/CHECK_RUN_SUMMARY.json b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-uniform-local-norm-checks/CHECK_RUN_SUMMARY.json new file mode 100644 index 000000000..99cbcb3b7 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-uniform-local-norm-checks/CHECK_RUN_SUMMARY.json @@ -0,0 +1,200 @@ +{ + "original_bytes": 7279, + "original_disposition": "WITHHELD_LOCAL_CAPTURE_PATHS", + "original_relative_path": "nonauthor-uniform-local-norm-checks/CHECK_RUN_SUMMARY.json", + "original_sha256": "480e1e4786933f2201a6dadb93ee720dc21b632c5a749a91ea9323f1fc506364", + "path_transform": "Absolute original capture-root prefix replaced with ; all nonpath values preserved. No new execution or original-byte verification is supplied by this derivative.", + "projected_record": { + "all_14_inputs_unchanged": true, + "inputs_after": [ + { + "bytes": 26687, + "path": "UNIFORM_LOCAL_NORM_DIAGONAL_PROPOSAL.md", + "sha256": "8e1eefca9888d43824a9fd9caca34c44d2de4bf8e30c668b474fa2097491c76b" + }, + { + "bytes": 6486, + "path": "uniform_local_norm_controls.py", + "sha256": "4b84755b771598ed34193c075718a41da447719f90a2ff5c6a17d48d0abc3dc4" + }, + { + "bytes": 1790, + "path": "UNIFORM_LOCAL_NORM_CONTROLS.json", + "sha256": "5d6cdb2d4925c6ba0b1b545d7a5aca1977655c1d18aa1ea29a82273f30845ece" + }, + { + "bytes": 19138, + "path": "QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md", + "sha256": "827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30" + }, + { + "bytes": 6308, + "path": "fold_isolation_controls.py", + "sha256": "13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5" + }, + { + "bytes": 795, + "path": "FOLD_ISOLATION_CONTROLS.json", + "sha256": "ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf" + }, + { + "bytes": 25699, + "path": "FINITE_FOURIER_CRITICAL_COUNT_DIAGONAL_ROUTE.md", + "sha256": "621cee96854aa440498cecf5452a8d1e23f90a4458779c3286c18eba106fe138" + }, + { + "bytes": 5153, + "path": "finite_fourier_count_controls.py", + "sha256": "fd6f29739280a55017945526cb75fb46ebc80d86c012ff4a1ff65baa404c31de" + }, + { + "bytes": 849, + "path": "FINITE_FOURIER_COUNT_CONTROLS.json", + "sha256": "bbe147fdf409a3e37b450d4cbcd5d2a927140f95bd3ce0f36cafe16861123bfd" + }, + { + "bytes": 18409, + "path": "NONAUTHOR_FOLD_ISOLATION_REVIEW.md", + "sha256": "f06c61e3f3e60c795a1a14e824640f1689d4143dba4f6698cada7a41929eec90" + }, + { + "bytes": 21880, + "path": "NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW.md", + "sha256": "16152427f7013822ac1bba81142429800334e6bf0f354a6439b23a1997bc4524" + }, + { + "bytes": 443, + "path": "uniform-local-norm-primary-sources/SOURCE_IDENTITY.json", + "sha256": "92bf5297b309e947f42ee1dc837487775e7341bc7c1051e493899963ff85891a" + }, + { + "bytes": 940975, + "path": "uniform-local-norm-primary-sources/simon-lecs-on-pde.pdf", + "sha256": "e1f1b2f51d2558aa467a8c064fe82e7b9e157ad28c9c75f01e38596cde656870" + }, + { + "bytes": 25367, + "path": "experiments/universality/finite_gaussian_contact/PROOF.md", + "sha256": "7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474" + } + ], + "inputs_before": [ + { + "bytes": 26687, + "path": "UNIFORM_LOCAL_NORM_DIAGONAL_PROPOSAL.md", + "sha256": "8e1eefca9888d43824a9fd9caca34c44d2de4bf8e30c668b474fa2097491c76b" + }, + { + "bytes": 6486, + "path": "uniform_local_norm_controls.py", + "sha256": "4b84755b771598ed34193c075718a41da447719f90a2ff5c6a17d48d0abc3dc4" + }, + { + "bytes": 1790, + "path": "UNIFORM_LOCAL_NORM_CONTROLS.json", + "sha256": "5d6cdb2d4925c6ba0b1b545d7a5aca1977655c1d18aa1ea29a82273f30845ece" + }, + { + "bytes": 19138, + "path": "QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md", + "sha256": "827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30" + }, + { + "bytes": 6308, + "path": "fold_isolation_controls.py", + "sha256": "13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5" + }, + { + "bytes": 795, + "path": "FOLD_ISOLATION_CONTROLS.json", + "sha256": "ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf" + }, + { + "bytes": 25699, + "path": "FINITE_FOURIER_CRITICAL_COUNT_DIAGONAL_ROUTE.md", + "sha256": "621cee96854aa440498cecf5452a8d1e23f90a4458779c3286c18eba106fe138" + }, + { + "bytes": 5153, + "path": "finite_fourier_count_controls.py", + "sha256": "fd6f29739280a55017945526cb75fb46ebc80d86c012ff4a1ff65baa404c31de" + }, + { + "bytes": 849, + "path": "FINITE_FOURIER_COUNT_CONTROLS.json", + "sha256": "bbe147fdf409a3e37b450d4cbcd5d2a927140f95bd3ce0f36cafe16861123bfd" + }, + { + "bytes": 18409, + "path": "NONAUTHOR_FOLD_ISOLATION_REVIEW.md", + "sha256": "f06c61e3f3e60c795a1a14e824640f1689d4143dba4f6698cada7a41929eec90" + }, + { + "bytes": 21880, + "path": "NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW.md", + "sha256": "16152427f7013822ac1bba81142429800334e6bf0f354a6439b23a1997bc4524" + }, + { + "bytes": 443, + "path": "uniform-local-norm-primary-sources/SOURCE_IDENTITY.json", + "sha256": "92bf5297b309e947f42ee1dc837487775e7341bc7c1051e493899963ff85891a" + }, + { + "bytes": 940975, + "path": "uniform-local-norm-primary-sources/simon-lecs-on-pde.pdf", + "sha256": "e1f1b2f51d2558aa467a8c064fe82e7b9e157ad28c9c75f01e38596cde656870" + }, + { + "bytes": 25367, + "path": "experiments/universality/finite_gaussian_contact/PROOF.md", + "sha256": "7fea2656e11d986e1a5d6bfb0a05c8421bd93262d28051b8cfef05320595f474" + } + ], + "interpreter_version": "3.9.6 (default, Dec 2 2025, 07:27:58) \n[Clang 17.0.0 (clang-1700.6.3.2)]", + "organizational_independence_credit": 0, + "performer": "OpenAI/Codex /root/c261a_local_runtime_20261016", + "previous_controls_rerun": false, + "primary_source_read_only_browsing": true, + "production_remote_mutations_or_formal_execution": 0, + "runs": [ + { + "command_argv": [ + "/Library/Developer/CommandLineTools/usr/bin/python3", + "-B", + "/uniform_local_norm_controls.py" + ], + "cwd": "", + "exact_frozen_output_match": true, + "exit_code": 0, + "finished_at_utc": "2026-10-16T02:16:05.127350+00:00", + "started_at_utc": "2026-10-16T02:16:05.072614+00:00", + "stderr": "author-controls.stderr", + "stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "stdout": "author-controls.stdout", + "stdout_sha256": "5d6cdb2d4925c6ba0b1b545d7a5aca1977655c1d18aa1ea29a82273f30845ece", + "wall_seconds": 0.054072959000000004 + }, + { + "command_argv": [ + "/Library/Developer/CommandLineTools/usr/bin/python3", + "-B", + "/nonauthor-uniform-local-norm-checks/review_checks.py" + ], + "cwd": "", + "exit_code": 0, + "finished_at_utc": "2026-10-16T02:16:05.197781+00:00", + "independent_finite_cases": 2132, + "started_at_utc": "2026-10-16T02:16:05.127654+00:00", + "stderr": "independent-controls.stderr", + "stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "stdout": "independent-controls.stdout", + "stdout_sha256": "bd553b02c66ad4cf761700f53043ac4b03f7371b5e15aa03ad08b4d4feef2fce", + "wall_seconds": 0.069728208 + } + ], + "source_adoption_or_scientific_promotion": "NONE", + "source_exposure": "new proposal and controls read fully; bounded Fourier/Plancherel import read; prior two exact lemmas consumed, not re-reviewed", + "state": "NEW_CONTROLS_REPLAY_AND_INDEPENDENT_FINITE_CHECKS_ONLY" + }, + "projection_kind": "PUBLICATION_PATH_ROLE_DERIVATIVE_NOT_ORIGINAL" +} diff --git a/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-uniform-local-norm-checks/author-controls.meta.json b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-uniform-local-norm-checks/author-controls.meta.json new file mode 100644 index 000000000..bc8a24a4a --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-uniform-local-norm-checks/author-controls.meta.json @@ -0,0 +1,24 @@ +{ + "original_bytes": 715, + "original_disposition": "WITHHELD_LOCAL_CAPTURE_PATHS", + "original_relative_path": "nonauthor-uniform-local-norm-checks/author-controls.meta.json", + "original_sha256": "0e8fd7097e06de071dd233b8f6a8ff7623d370e77d78421d7b11a88b9deffac8", + "path_transform": "Absolute original capture-root prefix replaced with ; all nonpath values preserved. No new execution or original-byte verification is supplied by this derivative.", + "projected_record": { + "command_argv": [ + "/Library/Developer/CommandLineTools/usr/bin/python3", + "-B", + "/uniform_local_norm_controls.py" + ], + "cwd": "", + "exit_code": 0, + "finished_at_utc": "2026-10-16T02:16:05.127350+00:00", + "started_at_utc": "2026-10-16T02:16:05.072614+00:00", + "stderr": "author-controls.stderr", + "stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "stdout": "author-controls.stdout", + "stdout_sha256": "5d6cdb2d4925c6ba0b1b545d7a5aca1977655c1d18aa1ea29a82273f30845ece", + "wall_seconds": 0.054072959000000004 + }, + "projection_kind": "PUBLICATION_PATH_ROLE_DERIVATIVE_NOT_ORIGINAL" +} diff --git a/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-uniform-local-norm-checks/independent-controls.meta.json b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-uniform-local-norm-checks/independent-controls.meta.json new file mode 100644 index 000000000..9bab082e1 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/publication-projections/nonauthor-uniform-local-norm-checks/independent-controls.meta.json @@ -0,0 +1,24 @@ +{ + "original_bytes": 738, + "original_disposition": "WITHHELD_LOCAL_CAPTURE_PATHS", + "original_relative_path": "nonauthor-uniform-local-norm-checks/independent-controls.meta.json", + "original_sha256": "c88b17ff52d1eb833679da97a4cb93a2f6068345c785b8e24c8a989ac2c55a34", + "path_transform": "Absolute original capture-root prefix replaced with ; all nonpath values preserved. No new execution or original-byte verification is supplied by this derivative.", + "projected_record": { + "command_argv": [ + "/Library/Developer/CommandLineTools/usr/bin/python3", + "-B", + "/nonauthor-uniform-local-norm-checks/review_checks.py" + ], + "cwd": "", + "exit_code": 0, + "finished_at_utc": "2026-10-16T02:16:05.197781+00:00", + "started_at_utc": "2026-10-16T02:16:05.127654+00:00", + "stderr": "independent-controls.stderr", + "stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "stdout": "independent-controls.stdout", + "stdout_sha256": "bd553b02c66ad4cf761700f53043ac4b03f7371b5e15aa03ad08b4d4feef2fce", + "wall_seconds": 0.069728208 + }, + "projection_kind": "PUBLICATION_PATH_ROLE_DERIVATIVE_NOT_ORIGINAL" +} diff --git a/experiments/universality/fold_isolation_diagonal/uniform-local-norm-primary-sources/SOURCE_IDENTITY.json b/experiments/universality/fold_isolation_diagonal/uniform-local-norm-primary-sources/SOURCE_IDENTITY.json new file mode 100644 index 000000000..1d930d5d3 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/uniform-local-norm-primary-sources/SOURCE_IDENTITY.json @@ -0,0 +1,7 @@ +{ + "url": "https://math.stanford.edu/~lms/lecs-on-pde.pdf", + "path": "work/core-theorem-audit-20261015/uniform-local-norm-primary-sources/simon-lecs-on-pde.pdf", + "bytes": 940975, + "sha256": "e1f1b2f51d2558aa467a8c064fe82e7b9e157ad28c9c75f01e38596cde656870", + "consumed": "Lecture 5 Fourier derivative, inversion, Parseval/Plancherel properties (a)-(h), PDF page 26; Sobolev statement/proof on PDF pages 25 and 27; not the entire book" +} diff --git a/experiments/universality/fold_isolation_diagonal/uniform_local_norm_controls.py b/experiments/universality/fold_isolation_diagonal/uniform_local_norm_controls.py new file mode 100644 index 000000000..08c6b10f8 --- /dev/null +++ b/experiments/universality/fold_isolation_diagonal/uniform_local_norm_controls.py @@ -0,0 +1,142 @@ +"""Exact finite controls for the uniform local-norm diagonal proposal. + +No Fourier-field samples, numerical Sobolev integrals, root counts, or process test. +""" +from fractions import Fraction as Q +from math import comb, factorial, ceil +from pathlib import Path +import hashlib +import json + + +def cutoff_constants(): + # d^j/dt^j exp(-1/t) = exp(-u) L_j(u), u=1/t. + # L_(j+1)=u^2(L_j-L_j'). Keep sparse integer coefficient dictionaries. + polys = [{0: 1}] + for j in range(6): + old = polys[-1] + new = {} + for m, a in old.items(): + new[m+2] = new.get(m+2, 0) + a + if m: + new[m+1] = new.get(m+1, 0) - m*a + polys.append({m: a for m, a in new.items() if a}) + assert polys[:3] == [{0: 1}, {2: 1}, {4: 1, 3: -2}] + aa = [sum(abs(a)*max(1, m**m) for m, a in poly.items()) for poly in polys] + ee = [8] + for n in range(1, 7): + ee.append(8*sum(comb(n, j)*2*aa[j]*ee[n-j] for j in range(1, n+1))) + hh = [sum(comb(n, j)*aa[j]*ee[n-j] for j in range(n+1)) for n in range(7)] + dd = [6] + [2*hh[j]**2 for j in range(1, 7)] + atheta = 0 + weight_sum = 0 + for i in range(7): + for j in range(7-i): + order = i+j + weight = factorial(6)//(factorial(6-order)*factorial(i)*factorial(j)) + weight_sum += weight + assert sum(comb(i, u)*comb(j, v) + for u in range(i+1) for v in range(j+1)) == 2**order + inner = sum(comb(i, u)*comb(j, v)*dd[u]*dd[v] + for u in range(i+1) for v in range(j+1)) + atheta += weight*2**order*inner + assert weight_sum == 3**6 + return aa, ee, hh, dd, atheta + + +def run(): + checks = {} + integrals = [Q(factorial(j)*factorial(4-j), factorial(5)) for j in range(5)] + assert integrals == [Q(1, 5), Q(1, 20), Q(1, 30), Q(1, 20), Q(1, 5)] + # d3^2=1/(80pi), d4^2=1/(20pi); pi>3 suffices for these rational majorants. + assert Q(1, 15)**2 > Q(1, 240) + assert Q(2, 15)**2 > Q(1, 60) + checks["Fourier_integral_and_rational_operator_constants"] = "PASS" + + for radius in (Q(0), Q(1, 4), Q(1), Q(3)): + for j in range(7): + assert radius**(2*j) <= 1+radius**12 + # Adding a DC variance A^2 changes M0 but not any positive spectral moment. + for amplitude in (Q(1), Q(10), Q(100)): + m0, m12 = amplitude**2+1, Q(1) + assert m12 == 1 and m0 >= amplitude**2 + checks["separate_M0_M12_lower_moment_bound_and_DC_falsifier"] = "PASS" + + # e <= sum_0^4 1/n! + (1/5!)/(1-1/6) < 11/4, hence e^2<8. + ebound = sum(Q(1, factorial(n)) for n in range(5)) + Q(1, factorial(5))/(1-Q(1, 6)) + assert ebound < Q(11, 4) and Q(11, 4)**2 < 8 + aa, ee, hh, dd, atheta = cutoff_constants() + checks["explicit_cutoff_derivative_and_H6_trace_recurrence"] = "PASS" + checks["multinomial_H6_weights_and_Leibniz_weight_sums"] = "PASS" + + # The finite product inequality holds without a dimension-dependent prefactor. + cases = 0 + for n in range(1, 65): + for mass in (Q(0), Q(1, 8), Q(1, 4), Q(1, 2)): + rows = [mass/n]*n + prod = Q(1) + for x in rows: + prod *= 1-x + assert prod >= 1-sum(rows) >= Q(1, 2) + cases += 1 + assert (1-Q(1, 2)) == Q(1, 2) + checks["dimension_free_finite_Gaussian_product_margin"] = "PASS" + + # Rotation and cover margins are exact inequalities, not angle sampling. + assert Q(2) < Q(3, 2)**2 # 1/2+sqrt(2)<2. + for lmin in (Q(1, 10), Q(1), Q(10), Q(100)): + b = min(Q(1), lmin/16) + assert 6*b < lmin + for length in (lmin, 2*lmin, 10*lmin): + n = ceil(length/b) + assert length/n <= b + assert n*n <= (1+length)**2/b**2 + eta0 = b/4 + for rho in (Q(1, 10), Q(1, 4), Q(9, 10)): + assert (1+rho)*eta0 <= b/2 + assert b/2+rho*eta0 < 2*b + checks["fixed_cover_rotated_chart_and_endpoint_margins"] = "PASS" + + # The selected logarithmic cutoff exactly cancels K^8 and (1+L)^2. + p, q, u = Q(8), Q(2), Q(4) + assert p-8 == q-2 == 0 and u-Q(4, 3) == Q(8, 3) + assert 16**4 == 65536 + checks["growing_K_L_log_cutoff_normalization"] = "PASS" + + # Moving the lifetime-band cap to T(t)=1/t makes rbar constant, not o(eta). + rho = Q(1, 4) + for n in (10, 100, 1000): + eta, rbar = Q(1, n), Q(1) + assert rbar > rho*eta + checks["growing_lifetime_cap_radius_failure_control"] = "PASS" + + frozen = { + "QUANTITATIVE_FOLD_ISOLATION_PROPOSAL.md": "827c82953845eb12005542ff6a4e1c99886b68d1ea9f1aeae16f981622ac6a30", + "fold_isolation_controls.py": "13fae24ace9b3e8afb26d115f1a03acaec9ea72b0953a74ac54f2f1fd1fd46f5", + "FOLD_ISOLATION_CONTROLS.json": "ccaf141441c3f890970c8e168c478a1f98f6eec8791aba002ffacfef06333bcf", + "FINITE_FOURIER_CRITICAL_COUNT_DIAGONAL_ROUTE.md": "621cee96854aa440498cecf5452a8d1e23f90a4458779c3286c18eba106fe138", + "finite_fourier_count_controls.py": "fd6f29739280a55017945526cb75fb46ebc80d86c012ff4a1ff65baa404c31de", + "FINITE_FOURIER_COUNT_CONTROLS.json": "bbe147fdf409a3e37b450d4cbcd5d2a927140f95bd3ce0f36cafe16861123bfd", + "NONAUTHOR_FOLD_ISOLATION_REVIEW.md": "f06c61e3f3e60c795a1a14e824640f1689d4143dba4f6698cada7a41929eec90", + "NONAUTHOR_FINITE_COUNT_DIAGONAL_REVIEW.md": "16152427f7013822ac1bba81142429800334e6bf0f354a6439b23a1997bc4524", + } + base = Path(__file__).resolve().parent + for name, expected in frozen.items(): + assert hashlib.sha256((base/name).read_bytes()).hexdigest() == expected + checks["six_prior_author_files_and_two_reviews_unchanged"] = "PASS" + return { + "classification": "EXACT_FINITE_ALGEBRA_CONTROLS_ONLY", + "arithmetic": "Python Fraction and integers; no floating arithmetic", + "Gaussian_field_sampling": "NONE", "Rice_density": "NOT_USED", + "process_claim": "NONE", "scientific_effect": "NONE", + "checks": checks, "Gaussian_product_cases": cases, + "constants": {"A_j": aa, "E_j": ee, "H_j": hh, "d_j_upper": dd, + "A_theta": str(atheta), "A_theta_decimal_digits": len(str(atheta)), + "embedding_p3": "1/15", "embedding_p4": "2/15", + "normalized_t_exponent": "8/3"}, + "proof_status": "Analytic norm, trace, cover, isolation and asymptotic arguments are conventional report proofs; not machine verified", + } + + +if __name__ == "__main__": + print(json.dumps(run(), indent=2))