TFPT scientific source excerpt
Original repository path: experiments/theory-contracts/tfpt-unified-source-derivation-20261029/aps_source_review/LOSSLESS-CARRIER-ODD-READER-DICTIONARY-REVIEW.txt
Original file SHA-256: b6f457488f84a06f5d97288c36210fc06ee2d94eb709ed90118e0fb27e40b767
Inclusive original lines: 1-146
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 | LOSSLESS DICTIONARY FOR THE GIVEN COUPLED CARRIER SOURCE
000002 | Independent targeted review, 30 October 2026. No Atlas or original edit.
000003 | 
000004 | VERDICT
000005 | The new check_current_odd_reader.py follows the actual typed chain
000006 | 
000007 |   B_s -> G(B_s) -> E* G_22(B_s) E -> v258 logit -> Pi_odd^gamma
000008 | 
000009 | and its formulas are correct. Its strongest natural continuation is an exact
000010 | lossless dictionary inside the already given smooth coupled operator class.
000011 | The full face covariance retains h. The odd reader retains Phi. It does not
000012 | retain the even connection A, so reconstructing h from the odd reader alone
000013 | requires A to be kept as known source data.
000014 | 
000015 | 1. Targeted checker review
000016 | The checker uses the actual v252 full_D, grading and KO6 swap. Its 192 by 192
000017 | operator
000018 | 
000019 |   B_s=Z tensor p-S tensor Phi_s
000020 | 
000021 | is the scalar-mode instance of the original doubled coefficient. The displayed
000022 | 384 by 384 matrix is the complete two-face projector G(B_s), and the selected
000023 | block is genuinely its second face. Since E reduces B_s,
000024 | 
000025 |   E*G_22(B_s)E=(I+exp(2 ell h_s))^-1,
000026 |   h_s=p gamma-i gamma Phi_s.
000027 | 
000028 | The script does not commute a logarithm through a nonreducing face map. It
000029 | first takes the already labeled face and then uses the reducing E chart.
000030 | 
000031 | The exact SymPy checks retain symbolic momentum p and source parameter s.
000032 | The numerical route then executes the actual B192, G384 and v258 induce call.
000033 | The recorded maximum reader residual 1.90e-13 and central-difference jet
000034 | residual 9.05e-10 agree with the exact identities. Its explicit eigenvalue
000035 | guard also makes v258's numerical clipping inactive in every tested case.
000036 | 
000037 | Two scope points should remain visible. The numerical run is a finite set of
000038 | Fourier modes, while the whole-circle logarithm is justified analytically by
000039 | spectral calculus. The symbolic test uses scalar p and therefore does not by
000040 | itself test a noncommuting smooth connection; the theorem below does and does
000041 | not need [A,Phi]=0.
000042 | 
000043 | 2. Exact smooth operator class
000044 | Let
000045 | 
000046 |   H=L2(S1,C^96),
000047 |   gamma=gamma*=gamma^-1,
000048 |   A=-i partial_theta+a(theta)
000049 | 
000050 | with A selfadjoint on H1, a smooth bounded Hermitian coefficient and
000051 | [A,gamma]=0. Let Phi(theta)=Phi(theta)* be smooth and bounded with
000052 | {gamma,Phi}=0. Put
000053 | 
000054 |   h=A gamma-i gamma Phi.                                (1)
000055 | 
000056 | The bounded Phi term is selfadjoint, so Kato-Rellich makes h selfadjoint on
000057 | Dom(A)=H1. The constant gamma preserves this domain. No commutator condition
000058 | [A,Phi]=0 is used here. That commutator first matters in h^2, not in the
000059 | reader.
000060 | 
000061 | For ell>0 define the actual second face
000062 | 
000063 |   C(h)=(I+exp(2 ell h))^-1.                             (2)
000064 | 
000065 | C and I-C are injective. Their inverses are generally unbounded on the whole
000066 | circle. Borel functional calculus nevertheless gives on Dom(h)
000067 | 
000068 |   L(C):=log((I-C)C^-1)=2 ell h.                         (3)
000069 | 
000070 | With Pi_even(T)=(T+gamma T gamma)/2 and
000071 | Pi_odd(T)=(T-gamma T gamma)/2, equations (1)-(3) give
000072 | 
000073 |   Pi_even L(C)=2 ell A gamma,
000074 |   Pi_odd  L(C)=-2 ell i gamma Phi.                      (4)
000075 | 
000076 | The odd difference in (4), initially written on H1, extends to the displayed
000077 | bounded multiplication operator. This remains true for [A,Phi] nonzero.
000078 | 
000079 | 3. The precise lossless maps
000080 | Fix gamma, ell>0 and a nonzero declared mu_geo. Define the original odd reader
000081 | 
000082 |   R_Phi=mu_geo Pi_odd L(C).
000083 | 
000084 | Then
000085 | 
000086 |   R_Phi=-2 ell mu_geo i gamma Phi
000087 |        =2 ell mu_geo U_gamma Phi U_gamma*,              (5)
000088 | 
000089 | where U_gamma=exp(-i pi gamma/4). Its inverse is exact:
000090 | 
000091 |   Phi=(i gamma R_Phi)/(2 ell mu_geo)
000092 |      =U_gamma* R_Phi U_gamma/(2 ell mu_geo).             (6)
000093 | 
000094 | Thus Phi <-> R_Phi is one-to-one for fixed ell and mu_geo. This statement
000095 | does not select their values.
000096 | 
000097 | The reader R_Phi intentionally deletes A. With A retained as known source
000098 | data, one reconstructs
000099 | 
000100 |   h=A gamma+R_Phi/(2 ell mu_geo),                        (7)
000101 | 
000102 | then C(h), the complete two-face G(h), and through E plus the conjugate C0
000103 | chart the original G(B_Phi). Without the retained A, equation (7) is not a
000104 | reader-only reconstruction.
000105 | 
000106 | The full covariance contains still more information than R_Phi. From C alone
000107 | one obtains L(C), and (4) reconstructs both coefficients:
000108 | 
000109 |   A=(Pi_even L(C)) gamma/(2 ell),
000110 |   Phi=i gamma Pi_odd L(C)/(2 ell).                       (8)
000111 | 
000112 | Hence, within this declared operator class, the full face C, h and the pair
000113 | (A,Phi) are equivalent data once gamma and ell are fixed. The complete pure
000114 | graph G is also equivalent because its second face is C and its other blocks
000115 | are fixed spectral functions of h.
000116 | 
000117 | Finally, the given G is a quasifree covariance. It determines all finite CAR
000118 | word expectations by Wick's rule, and the already checked unitary carrier and
000119 | selfdual charts transport those words. This is mathematical losslessness for
000120 | the given quasifree source. It is not an original theorem selecting this
000121 | source, Phi, A, ell or mu_geo, and it adds no interaction or physical field
000122 | coupling.
000123 | 
000124 | 4. The direct reconstruction loop selects no Yukawa pattern
000125 | If one first removes the fixed U_gamma comparison but does not divide by the
000126 | known scale, feeding the reader output back as a new coefficient gives simply
000127 | 
000128 |   F(Phi)=c Phi,  c=2 ell mu_geo.
000129 | 
000130 | Its fixed points are all odd Phi when c=1, and only Phi=0 when c is not 1.
000131 | The latter need not lie in the affine family Phi_s=(1+s)D_Y+D_MR when the
000132 | fixed Majorana contribution is nonzero. If instead one uses the exact inverse
000133 | (6), the feedback is the identity and again fixes every Phi. Therefore this
000134 | lossless reconstruction loop supplies no preference among Yukawa matrices and
000135 | no useful genetic search objective by itself. This conclusion is restricted
000136 | to direct feedback through this reader. External original constraints, other
000137 | fields or an actual action may still select a coefficient; no claim that a
000138 | noncommuting compression is the only possible generation mechanism follows.
000139 | 
000140 | SOURCES REVIEWED
000141 | aps_source_review/check_current_odd_reader.py;
000142 | aps_source_review/CURRENT-ODD-READER-CHECK.json;
000143 | aps_source_review/CURRENT-CARRIER-ODD-READER-WHOLE-SOURCE.txt;
000144 | graded_boundary_functor/FINITE-COUPLED.txt;
000145 | verification/v252_full_finite_triple.py;
000146 | verification/v258_dirac_covariance_induction.py.