TFPT scientific source excerpt Original repository path: experiments/theory-contracts/tfpt-unified-source-derivation-20261029/aps_source_review/CLAUDE28-CUT-FIRST-REVIEW.txt Original file SHA-256: e7a8c9f509bfc5d57e4d3bf57cc8f9cb5c6e694f58988281f039cc26d43340f3 Inclusive original lines: 1-138 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 | [OMITTED: non-scientific conversation or execution metadata] 000002 | [OMITTED: non-scientific conversation or execution metadata] 000003 | [OMITTED: non-scientific conversation or execution metadata] 000004 | [OMITTED: non-scientific conversation or execution metadata] 000005 | [OMITTED: non-scientific conversation or execution metadata] 000006 | [OMITTED: non-scientific conversation or execution metadata] 000007 | 000008 | VERDICT 000009 | The anti-inversion formula is correct. The claimed uniqueness of the 000010 | q-compatible mark-avoiding E0 cut is false: BOTH z->i bar(z) and 000011 | z->-i bar(z) qualify at the original Fuchs/curve level, and the supplied 000012 | quarter rotation conjugates them. Thus those premises select neither 000013 | one cut nor a physical sewing/Pin phase. For the explicitly named point 000014 | map z->i^k bar(z), the constant coefficient map is U^k conjugation; 000015 | U^-k is wrong for odd k in that convention. 000016 | 000017 | 1. COMPLETE FINITE CLASS, NOT JUST THE k=1 EXAMPLE 000018 | 000019 | Let M={1,i,-1,-i}, and retain its original circle |z|=1. All holomorphic 000020 | Mobius maps preserving M are precisely 000021 | 000022 | r_k(z)=i^k z, s_k(z)=i^k/z, k=0,1,2,3. 000023 | 000024 | Their restrictions have the eight D4 permutations of the marked circle. 000025 | A circle-preserving Mobius map preserves or reverses cyclic order, hence 000026 | has at most these eight possible permutations; three points determine 000027 | the map uniquely, and the displayed maps realize all eight. 000028 | 000029 | Composing with ordinary conjugation gives every anti-Mobius map: 000030 | 000031 | sigma_k(z)=i^k bar(z), rho_k(z)=i^k/bar(z). 000032 | 000033 | There are sixteen sphere maps in total, with only eight distinct circle 000034 | restrictions: the central equatorial involution rho_0 acts identically 000035 | on the circle. Equivalently the full sphere stabilizer is D4 x C2. 000036 | Every map taking the four marks to themselves already preserves their 000037 | unique generalized circle. Thus no further anti-Mobius maps are omitted. 000038 | 000039 | All four sigma_k are antiinvolutions. Their fixed line is 000040 | arg z=k pi/4 modulo pi. For k=0,2 this line hits two marks. For k=1,3 000041 | its circle intersection consists of two unmarked points. In particular 000042 | 000043 | sigma_1=i bar(z), sigma_3=-i bar(z) 000044 | 000045 | are TWO distinct mark-avoiding real cuts. Rotation r_1 conjugates 000046 | sigma_1 to sigma_3. In particular they are equivalent under the already 000047 | supplied original quarter symmetry if no further marking breaks it. 000048 | 000049 | For completeness rho_k^2=r_(2k). The only antiinvolutions of this form 000050 | are rho_0 (fixed circle containing all marks) and rho_2 (no fixed points). 000051 | rho_1,rho_3 have order4, so they are not real involutions defining cuts. 000052 | The holomorphic s_1,s_3 have the same unmarked fixed-point pairs on the 000053 | circle, but are not antiinvolutions of the sphere. 000054 | 000055 | 2. q COMPATIBILITY AND WHETHER ZERO MUST BE FIXED 000056 | 000057 | The original curve has div z=Dq-Dp. Maps r_k and sigma_k fix 0 and 000058 | infinity separately and preserve the q/p packets; maps s_k and rho_k 000059 | exchange 0 and infinity and exchange the packets. Consequently retention 000060 | of the selected q packet leaves four holomorphic rotations and four 000061 | antiinvolutions sigma_k. Of their real cuts exactly sigma_1 AND sigma_3 000062 | avoid the marks. There is no surviving unique candidate. 000063 | 000064 | Fixing0 is not forced by the four marks or the circle alone: s_k,rho_k 000065 | are explicit counterexamples. It follows here from the ADDITIONAL 000066 | retained q-versus-p marking. This is legitimate when that ordered packet 000067 | is part of the original source data; it must not be advertised as a 000068 | consequence of the unlabelled four-point set. 000069 | 000070 | On C the two anti maps are Q tau0 and Q^3 tau0. They differ by the 000071 | hyperelliptic involution Q^2, which fixes all branch points. Both fix 000072 | each q_a and preserve the divisor classes S0 and L_a. Mere preservation 000073 | of these classes, even with the flat coefficient lifts below, still 000074 | does not identify a graded Pin or physical CAR reflection. 000075 | 000076 | 3. DIRECT ORIGINAL ODE TEST, INCLUDING THE CONSTANT-MATRIX SIGN 000077 | 000078 | Retain the real rational original A from ORIGINAL-FUCHS-ALGEBRAIC-FRAME: 000079 | 000080 | A=1/(z^4-1)[[2z^3,a z^2,0],[a,z^3,b z],[0,b z,z^3]], 000081 | a=2sqrt(2)/3,b=sqrt(5)/3,U=diag(1,i,-i), Psi'=+A Psi. 000082 | 000083 | For rho_0(z)=1/bar(z) the anti-pulled coefficient is 000084 | 000085 | -z^-2 A(1/z)=A(z)^t-diag(2,1,1)/z. 000086 | 000087 | Every entry of this identity was checked exactly. This is a dual-type 000088 | identity with a degree term, not the statement that rho_0 preserves the 000089 | same holomorphic flat connection. It agrees with the old q/p exchange 000090 | and does not evade the existing rational inversion obstruction in 000091 | D4-CONNECTION-LIFT.txt. 000092 | 000093 | Now specify the convention rather than infer a sign from a label. For 000094 | the bundle POINT map 000095 | 000096 | j_k:(z,v)->(i^k bar(z), U^k bar(v)), 000097 | 000098 | the covariance i^k A(i^k z)=U^k A(z)U^-k proves that horizontal 000099 | solutions map to horizontal solutions. Also j_k^2=I since 000100 | U^k overline(U^k)=I. Thus all four sigma_k carry actual constant flat 000101 | Real structures. For k=1 and3 BOTH are available at this operator level. 000102 | 000103 | Equivalently, on sections the anti-pulled coefficient before multiplication 000104 | by the fiber matrix is B_k=i^-k A(i^-k z)=U^-k A(z)U^k. 000105 | Conjugating B_k by U^k returns A. Conjugating by U^-k does not for odd k. 000106 | The exact four-case check returned true for U^k for all k, and false for 000107 | U^-k at k=1,3. U^-k can describe the oppositely indexed base reflection 000108 | or an explicitly changed transport convention; it is not the above map. 000109 | 000110 | The known algebraic fundamental matrix G supplies an independent 000111 | differential trivialization, so this test is not based merely on equal 000112 | monodromy orders. The two anti maps are conjugate by the actual flat 000113 | rotation (z,v)->(iz,Uv), just as their base maps are. 000114 | 000115 | 4. WHAT WOULD REALLY BE NEEDED TO SELECT A PHYSICAL CUT 000116 | 000117 | This finite/operator classification supplies a decisive counterexample 000118 | to uniqueness from the reported premises. An original ordered-face map, 000119 | Pin lift, full charged Phi coupling, actual boundary domain or state could 000120 | break the remaining symmetry. Such a datum must be read and intertwined; 000121 | it does not follow from flat holomorphic coefficients or cut avoidance. 000122 | No such extra selection is supplied by the asserted uniqueness argument. 000123 | 000124 | The prior source-selected spin quarter phase and its overall sign are 000125 | retained. They have their existing spin-lift derivation; this report does 000126 | not refute that derivation. The failed cut-uniqueness claim cannot serve 000127 | as a new derivation of a physical xi sewing phase. Geometric Real maps, 000128 | coefficient conjugations, Pin lifts, Phi, oriented faces Gamma+/Gamma- 000129 | and the selfdual CAR C0 remain different typed operations until the 000130 | actual source map is exhibited. No assertion that the full Seed lacks 000131 | its original reflection, state or time is made. 000132 | 000133 | Sources reused: September28 original Fuchs/curve section203-340; 000134 | C4-DESCENT-FINITE-MONODROMY.txt; D4-CONNECTION-LIFT.txt; 000135 | ORIGINAL-FUCHS-ALGEBRAIC-FRAME.txt; C3-REAL-QUARTER-COMPATIBILITY-REVIEW.txt. 000136 | Targeted Atlas lookup also retained premise 000137 | full-original-four-mark-inversion-and-equatorial-real-slots and the old 000138 | reviewed distinction between linear duality and Real reflection.