TFPT scientific source excerpt
Original repository path: experiments/theory-contracts/tfpt-unified-source-derivation-20261029/aps_source_review/WHOLE-SCALE-ENERGY-RECONSTRUCTION-PROOF.txt
Original file SHA-256: 9693f9051ae803cb2e750035b934cd99d22044efbd5430cb592eb7ad47b2a96f
Inclusive original lines: 1-130
Line numbers below refer to the original source.
This bounded excerpt is source evidence; it is not a new proof/checker run.
Conversation metadata and personal execution details are omitted where indicated.

000001 | WHOLE ORIGINAL SCALE, ENERGY RESPONSE AND RELATIVE SPECTRAL FLOW
000002 | 31 October 2026. R040/R041. Existing source, Y/MR, normal and reference retained.
000003 | 
000004 | 1. Original calls and exact question
000005 | Archive45 metrology001 defines Dgeo=Dref+Drel=D, f=exp(-sigma/2),
000006 |  Ksigma=f Dgeo f+Brel,
000007 |  Zrel(s)=Tr_rel(|Ksigma|^-2s-|Dref|^-2s)+i pi DeltaEta/2.
000008 | Its ED is Res1 and Gamma4 is FP0, not logdet. Boundary001 starts instead
000009 | with the WHOLE Dplus and B_Sigma, then constructs its Cauchy data and double.
000010 | QFT002 defines Dadm on RanPstr; no equality Dadm=Ksigma is silently added.
000011 | The normal constant coupled source, full96 Higgs Phi=L(H)+DM, the original
000012 | quarter sectors, fixed APS inverse and full two-face graph already exist.
000013 | The earlier massless conformal-trace and R033-R039 results are REUSED.
000014 | New task: compute the actual whole scale transport and its energy response,
000015 | then the first exact recovery of an entire normal scale-profile class.
000016 | No new geometry, Y, scale value, normal, reference, time or state is selected.
000017 | 
000018 | 2. Whole sandwich, including the original H insertion
000019 | First use the existing full inner realization Ksigma=f K0 f, K0=D_H.
000020 | K0 contains ALL original Phi(H) and DM already. On a compact given branch,
000021 | f is smooth positive bounded with bounded inverse. The transported closed
000022 | selfadjoint domain is Dom Ksigma=f^-1 Dom K0. This is a congruence, not a
000023 | unitary conjugation. The corresponding maximal local solution spaces obey
000024 |  Ksigma u=0 iff K0(fu)=0.
000025 | Thus ker Ksigma=f^-1 ker K0 and the closed nullity is unchanged.
000026 | This algebra applies to the whole variable matrix Phi; it is not restricted
000027 | to massless fields. A different fresh APS(Ksigma) rule is not substituted.
000028 | 
000029 | Under the already given conformal spin/density interpretation, the physical
000030 | metric is ghat=exp(2sigma)g and the WHOLE potential is Phihat=exp(-sigma)Phi,
000031 | including the supplied Majorana coefficient. This is R033 field transport,
000032 | not a new numerical Majorana mass. In the common bulk Hilbert space the
000033 | correct fixed-boundary normal trace is
000034 |  Tsigma=f_boundary Tr.
000035 | The solution usigma=f^-1 u0 has Tsigma usigma=Tr u0 on BOTH labelled faces.
000036 | Hence the same transported closed Cauchy subspace and its ORTHOGONAL
000037 | Calderon projector are identical. With sigma supported in the interior,
000038 | f_boundary=1 and even the raw labelled boundary data agree.
000039 | This extends the earlier explicitly massless theorem to the existing full
000040 | Higgs sandwich. Orthogonality uses the transported boundary metric; an
000041 | untransported graph metric or arbitrary Seeley projection is not substituted.
000042 | The GIVEN conditional quasifree boundary state has the same transported
000043 | bounded CAR words, adjoints and face marginals. No invariance of the full
000044 | H-field integration law, absolute Z[0], determinant, newly defined stress,
000045 | physical clock or mixed dynamical source is inferred. Those data remain.
000046 | 
000047 | 3. Exact energy-dependent reader of the SAME operator
000048 | At every z, with w=f^-2=exp(sigma),
000049 |  Ksigma-z=f(K0-z w)f,
000050 |  Gsigma(z)=f^-1 (K0-z w)^-1 f^-1.                    (1)
000051 | For nonreal z the inverse exists for selfadjoint Ksigma, with its original
000052 | domain. A varying weight cannot be replaced by a scalar energy shift.
000053 | Writing Rplus=(K0-i omega w)^-1 and Rminus=(K0+i omega w)^-1 gives
000054 |  (Ksigma^2+omega^2)^-1
000055 |   =[Gsigma(i omega)-Gsigma(-i omega)]/(2i omega)
000056 |   =f^-1 Rplus w Rminus f^-1.                       (2)
000057 | The ordering is essential; K0 and w need not commute. Rplus w Rminus
000058 | is also Rminus w Rplus by the same resolvent difference identity.
000059 | The original reference has its OWN full reader and is subtracted under
000060 | the original trace/continuation. Zero modes and eta are retained.
000061 | 
000062 | If the actual K0 is invertible, G0=K0^-1 gives the norm-convergent expansion
000063 |  (K0-z w)^-1=sum_(n>=0) z^n (G0 w)^n G0,
000064 |  |z|<1/(||G0|| ||w||).                              (3)
000065 | This is an exact operator construction, not a candidate scan. The existing
000066 | constant whole quarter fixed-APS source has ||G0||<=4 when its given normal
000067 | N is unitary. Only on THAT same domain/class does the guaranteed radius
000068 | become exp(-max sigma)/4. No such bound is inferred for arbitrary variableH.
000069 | Outside this disk the nonreal resolvent still exists, but not by this series.
000070 | 
000071 | 4. Whole normal profile can be reconstructed from the energy derivative
000072 | Use the ACTUAL homogeneous coupled collar D0=N(partial_r+B),0<=r<=ell,
000073 |  N*=-N, N^2=-I, {N,B}=0, B*=B,
000074 | with its already fixed normal and whole96 tangential coefficient.
000075 | Here sigma=sigma(r), while the original B and finite Phi are constant in r
000076 | and theta. This is a stated subfamily of the original scale, not a newly
000077 | chosen normal model. Let v=f u. The full homogeneous equation becomes
000078 |  v'= -B v-z w(r) N v.
000079 | Let T_sigma(z) be its full two-face transfer, from0 toell. Duhamel gives
000080 |  T_sigma(0)=exp(-ell B),
000081 |  T_sigma'(0)=-int_0^ell exp(-(ell-r)B) w(r) N exp(-rB) dr
000082 |             =-N int_0^ell w(r) exp((ell-2r)B) dr.   (4)
000083 | Zero energy alone contains no w. The first ENERGY derivative does.
000084 | For a known unit eigenvector v_n with Bv_n=E_n v_n,E_n>0,
000085 |  M_n=exp(-ell E_n)<v_n,N T_sigma'(0)v_n>
000086 |     =int_0^ell w(r)exp(-2r E_n)dr.                 (5)
000087 | All original channels remain; any one existing positive channel suffices
000088 | for scalar w. No charged states or other vertices are removed.
000089 | 
000090 | For an actual quarter alpha=1/4 or3/4 and any one fixed eigenvalue mu^2 of
000091 | the unchanged full Phi^2, E_n=sqrt((n+alpha)^2+mu^2), n>=0. These are
000092 | strictly increasing and E_n~n. Set a=exp(-2ell)>0, x=exp(-2r),
000093 |  lambda_n=E_n-E_0. Then lambda_0=0 and
000094 |  M_n=(1/2)int_a^1 w(-log(x)/2)x^(E_0-1)x^lambda_n dx. (6)
000095 | The nonzero lambdas have inf>0 and sum1/lambda_n=infinity. The classical
000096 | Muntz theorem therefore makes span{1,x^lambda_1,...} dense in C[0,1].
000097 | Continuous extension from[a,1] gives density there too. If two continuous
000098 | positive profiles give ALL the same M_n, their weighted difference
000099 | annihilates this dense space and hence every continuous function. Testing
000100 | its own difference proves it vanishes. Thus w, and sigma=log w, are UNIQUE
000101 | in this entire normal-dependent class. This is an infinite-class theorem.
000102 | Primary method: Borwein and Erdelyi, Full Muntz Theorem, pages1-3,
000103 | https://people.tamu.edu/~terdelyi/papers-online/Fullmuntz.pdf .
000104 | The theorem is cited, not newly proved or Lean-formalized here.
000105 | 
000106 | There is also a constructive approximation, with no physical metric chosen.
000107 | In L2([a,1],x^(E0-1)dx/2), project w onto q_n=x^lambda_n. Its Gram is
000108 |  Gij=[1-a^(E_i+E_j-E0)]/[2(E_i+E_j-E0)].             (7)
000109 | It is positive definite because distinct powers are linearly independent.
000110 | Solving G c=M gives w_N=sum_(j=0)^N c_j q_j. Density proves w_N->w in
000111 | this L2 norm. This is functional inversion, not an action/measure choice.
000112 | Finite approximants need not be positive; no finite-data uniqueness, noise
000113 | stability, uniform inverse bound, or uniform log-error estimate is claimed.
000114 | The EXACT existing flat source gives M_n=(1-exp(-2ell E_n))/(2E_n)=G_n0.
000115 | It therefore yields c=(1,0,...), w=1 uniquely in this stated class.
000116 | This recovers its given flat representative; it does not select a physical
000117 | flat geometry or choose a new scale in the theory.
000118 | 
000119 | Full orthogonal two-face graph data determine T(z)=G21(z)G11(z)^-1 modewise.
000120 | At nonreal z the orthogonal projector need not be holomorphic; its real-z
000121 | energy derivative at0 suffices here. No bounded inverse of G11 uniformly
000122 | over all modes is claimed. Normalized P(0) alone does NOT supply T'(0).
000123 | The original full Dplus/continued source must actually provide this energy
000124 | response before the reconstruction is applied to another original branch.
000125 | Nor does this identify Dadm with Ksigma or recover theta-dependent scales.
000126 | 
000127 | 5. Exact spectral flow separates ED, FP0 and the determinant
000128 | Let h=delta sigma be a smooth inner variation, with the transported fixed
000129 | domain above and the ordinary cyclic trace on its actual module. Write
000130 |  K=Ksigma,L=K^2, deltaK=-{h,K}/2.