TFPT scientific source excerpt Original repository path: experiments/theory-contracts/elliptic-moebius-source-bridge-20261026/session_coverage/supplements/CANONICAL-CUT-STATE-TIME-BRIDGE.txt Original file SHA-256: 510bb197905fb037d35dad3e73e7db839bd4ae1a19af864f020c324c6c170422 Inclusive original lines: 1-200 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 | CANONICAL CUT STATE AND TIME ON THE ACTUAL NS COVER 000002 | 26 October 2026 000003 | 000004 | Question 000005 | 000006 | The full NS cover Green field is reflection positive, but its one particle 000007 | occupation C=(1+exp(2H))^-1 is not a projection. Does the original source 000008 | problem really require a global unitary identification with the old Hardy 000009 | vacuum, or does the same geometric Dirac field already provide the local net, 000010 | state, reflection and time required by the operational seed? 000011 | 000012 | Verdict 000013 | 000014 | There is a positive and fully explicit answer inside the charged Gaussian 000015 | premise already used by NS-COVER-FULL-DIRAC-RP. Cut the actual period two NS 000016 | cover into its existing length one half cylinder. The Green covariance on one 000017 | cut face is exactly the marginal of the pure Cauchy graph state of that same 000018 | operator. The same Clifford reflection preserves the graph state and exchanges 000019 | the two faces. The boundary CAR inclusion preserves support in w. Its modular 000020 | logarithm is 2H, so the local RP, state and time data come from one geometric 000021 | datum. 000022 | 000023 | This does not revise the earlier Atlas criterion, which already allowed local 000024 | net maps and enlarged purifications. It is also not a new graph projection 000025 | theorem. The general graph theorem was already proved on 15 October. The new fact is its 000026 | exact identification with the presently verified NS Green field, its actual 000027 | length, its Pin sector exchange and its reflection. A surjective one particle 000028 | unitary from this faithful face state to the old Hardy projection is unavailable. 000029 | Such a unitary is not a premise for obtaining this particular local state and 000030 | time package. 000031 | 000032 | The construction does not yet identify this package with the old quarter Hardy 000033 | source. Channel rotation at finite aspect ratio gives an annulus amplitude with 000034 | boundary states. It does not project onto the Hardy vacuum. An exact zero mode 000035 | and partition function test below shows the difference. Thus the result closes 000036 | the internal state, time and reflection consistency of the actual NS cover, 000037 | while the local operator map carrying the old W8 field into this package remains 000038 | the next source question. The compact H dynamics is exactly periodic. This 000039 | rules out real time mixing, but the archived clustering clause does not state 000040 | that it means mixing for tau_H. It therefore remains a typing question, not a 000041 | failure of the seed axiom. 000042 | 000043 | 1. Original data and earlier theorem 000044 | 000045 | The current operator is 000046 | 000047 | D=-i(sigma1 partial_x+sigma2 partial_w), 000048 | L=i sigma1 D=partial_x+H, 000049 | H=i sigma3 partial_w, 000050 | 000051 | on w in[-1,1], with sigma1 phi(+/-1)=b phi(+/-1). The Pin lift gives 000052 | psi(x+2,w)=-psi(x,w), hence an actual Euclidean period two. The transverse 000053 | operator H is selfadjoint and has simple spectrum 000054 | 000055 | spec H={n pi/2:n in Z}. 000056 | 000057 | In particular it has one genuine zero mode. The already proved Gaussian Green 000058 | kernel is 000059 | 000060 | S(t)=exp(-tH)(1+exp(-2H))^-1, 0 CAR(H_face0 direct_sum H_face1), 000129 | a(f) -> a(0 direct_sum f). 000130 | 000131 | It preserves support exactly. R_cut sends it to the corresponding algebra on 000132 | the other face without moving the w support. The cross block X is nonlocal as 000133 | a correlation kernel, but it does not change the support of the algebra map. 000134 | For the first order H, real time has finite propagation with the given local 000135 | endpoint reflection. This is the local net content required here. It is not a 000136 | claim that every local interval is invariant under time evolution. 000137 | 000138 | The restricted quasifree state is KMS for 000139 | 000140 | a(f) -> a(exp(itH)f) 000141 | 000142 | at inverse temperature two, because 000143 | 000144 | log((1-C)/C)=2H. 000145 | 000146 | Consequently the same cut datum supplies state, cyclic standard form, 000147 | reflection and time. This is stronger than merely observing a thermal list of 000148 | occupations and weaker than demanding equality with a different global vacuum 000149 | representation. 000150 | 000151 | There is one exact time typing test. Since spec H=(pi/2)Z, 000152 | 000153 | exp(i 4H)=I. 000154 | 000155 | For example, take normalized f with proper local support and 000156 | N_f=a*(f)a(f). Faithfulness gives 0=1)(1+exp(-n pi))^2.