TFPT scientific source excerpt Original repository path: experiments/theory-contracts/elliptic-moebius-source-bridge-20261026/SOURCE-SELECTED-THETA-SPIN-INDEPENDENT-REVIEW.txt Original file SHA-256: ab9a1e2a12a2b25175834f2140d3e17664e470c59ca55da204264f5ebb9cb54c Inclusive original lines: 1-84 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 | INDEPENDENT REVIEW: SOURCE SELECTED THETA SPIN 000002 | 2026-10-26 000003 | 000004 | Reviewed sources 000005 | 000006 | SOURCE-SELECTED-THETA-SPIN.txt 000007 | SHA-256 cf4478191ccd75d62a2f049a54193541009bed24877df9cbe0ed9f942d07d717 000008 | 000009 | HYPERELLIPTIC-SPIN-QUARTER.txt 000010 | SHA-256 68163e1664b76d9cbd2b28ffe8314acb9fefec78a14a4ce9b557d4e8240138eb 000011 | 000012 | VERDICT 000013 | 000014 | ACCEPTED. The new result closes the theta-characteristic and half-turn-lift 000015 | choices from the already marked q|p curve data. It does not close the module 000016 | map to the native Gaussian E8 clock. 000017 | 000018 | 1. THETA CHARACTERISTIC AND PARITY 000019 | 000020 | For every Weierstrass point r on C, 2r is a canonical divisor. Hence 000021 | 000022 | S_0=O_C(q_0+q_1+q_2-K_C) 000023 | 000024 | has degree one and 2S_0 is isomorphic to K_C. Moreover 000025 | 000026 | div((w^3+1)/eta)=q_0+q_1+q_2-p_0-p_1-p_2, 000027 | 000028 | so exchanging the q and p sides leaves S_0 unchanged. 000029 | 000030 | The even-parity argument is valid. An effective degree-one theta 000031 | characteristic on a genus-two hyperelliptic curve is O_C(r) for a Weierstrass 000032 | point r. If S_0 were one of these six classes, use the q expression when r is 000033 | a q point and the equivalent p expression when r is a p point. Cancelling r 000034 | and replacing K_C by twice another point in the same triple would make the 000035 | difference of two distinct points principal. That would define a degree-one 000036 | map to P1, impossible for a genus-two curve. Thus h^0(S_0)=0 and S_0 is even. 000037 | 000038 | 2. UNIQUENESS FROM THE CUBIC SOURCE SYMMETRY 000039 | 000040 | The automorphism c cycles each of the q and p triples, so it fixes S_0. On 000041 | 000042 | Jac(C)[2]=V_q direct-sum V_p 000043 | 000044 | its restriction to either two-dimensional F2 plane has order three and no 000045 | nonzero fixed vector. Therefore Jac(C)[2]^c=0. Since theta characteristics 000046 | form an affine Jac(C)[2] torsor, S_0 is the unique c-invariant theta 000047 | characteristic. This uses the already selected source action rather than an 000048 | additional theta marking. 000049 | 000050 | 3. Q LIFT AND PULLBACK CONVENTION 000051 | 000052 | The marked map Q preserves S_0 and fixes q_0. At q_0, eta is a local 000053 | coordinate and 000054 | 000055 | Q^*eta=i eta, Q^*deta=i deta. 000056 | 000057 | For a square-root-compatible map alpha:Q^*S_0->S_0, its fiber scalar lambda at 000058 | q_0 therefore obeys lambda^2=i. The fourth pullback iterate has value 000059 | lambda^4=-1 at the fixed point and is a global scalar, proving 000060 | 000061 | alpha^4=-I, alpha^8=I. 000062 | 000063 | The composition 000064 | 000065 | beta=alpha composed with Q^*alpha:(Q^2)^*S_0->S_0 000066 | 000067 | is type-correct. Replacing alpha by -alpha leaves beta unchanged. In the 000068 | stated pullback convention its fiber value at q_0 is lambda^2=+i, and 000069 | beta^2=alpha^4=-I. Thus Q selects the sign-independent order-four projective 000070 | spin lift of h=Q^2. This is not an honest C2-linearization, whose cocycle 000071 | square would be +I. 000072 | 000073 | 4. SCOPE 000074 | 000075 | The proof now selects the even spin line S_0, the Q lift up to an overall sign 000076 | that cancels in beta, and the induced local +i phase of beta. It does not construct an 000077 | equivariant identification of the resulting theta or spin module with the 000078 | native four-dimensional Gaussian E8 module or its central generator J. The 000079 | separate obstruction to the literal identification with the Gaussian code 000080 | quotient V=L/(1+i)L therefore remains in force. That quotient labels Pauli 000081 | contexts, not Pauli translations. It is not an obstruction to the correctly 000082 | typed theta target Jac(C)[2] -> N64/mu4. For that target the source cubic and 000083 | the frozen native sigma both have zero-dimensional fixed space; the marked 000084 | symplectic and Heisenberg intertwiner remains to be constructed.