SigmaLayer/TensorSector: weaving the tensor product into the ledger #
Category: 7-SigmaLayer (the projective-sector layer (Paper C)).
How the composite/tensor structure sits in the base. The honest split is:
Derived (theorems, NOT posits). The finite tensor product
ℂ^{NA} ⊗ ℂ^{NB}is a genuine, constructible object: it is the projective sector on the PRODUCT indexFin NA × Fin NB, identified withFin (NA · NB)byfinProdFinEquiv(tensorIndexEquiv). On it the local operator algebra is real: Alice'sU ⊗ Iand Bob'sI ⊗ Qcommute (aliceOp_bobOp_commute), and Alice cannot signal to Bob (the operator-level no-signallingCSD.Empirical.QM.NoCommunication.no_communication). None of this is assumed; it follows from the Kronecker algebra.Bridge B6 (dim
= NA·NB) — now DERIVABLE, still a structure field here. The reconstruction claim that a composite's projective sector IS this tensor sector withdim = NA · NBisCompositeSector.tensor_dimension, a named field. As of 2026-07-17 it is no longer parked: the abstract theoremCSD.SigmaLayer.compositeAlgReconstruction(SigmaLayer/TensorReconstruction.lean) PROVES that commuting local algebrasM_m, M_nthat GENERATE a composite𝒜force𝒜 ≃ₐ M_m ⊗ M_n, andCSD.SigmaLayer.composite_dim_eqderivesk = m·nfor𝒜 = M_k— i.e. B6's dimension relation is a THEOREM under locality + generation, and the interface now HAS the constructor:CSD.SigmaLayer.CompositeSector.ofReconstructionbuilds aCompositeSectorwhosetensor_dimensionfield is FILLED bycomposite_dim_eq(derived from commuting, generating local embeddings). So B6 is dischargeable — the by-hand entangled tier (singlet, GHZ, CGLMP onFin 2 × Fin 2) still supplies the field directly per instance, but no longer must.
So tensors ARE woven in: the tensor product, its no-signalling algebra, the composite-is-tensor
reconstruction (compositeAlgReconstruction), AND the B6-discharging constructor
(CompositeSector.ofReconstruction) are all theorems/tools.
The finite tensor index equivalence. Fin NA × Fin NB ≃ Fin (NA · NB) (finProdFinEquiv): the
joint projective sector CP^{NA·NB−1} is genuinely the tensor product ℂ^{NA} ⊗ ℂ^{NB}, realised on the
product index, not posited.
Equations
Instances For
The composite dimension is the tensor-product index. For any composite sector, the posited
dimension relation NA · NB = Njoint (B6) transports the tensor index equivalence to
Fin NA × Fin NB ≃ Fin Njoint: the joint sector index IS the pair of party indices. This exhibits the
B6 posit as exactly the tensor-product identification.
Equations
- C.tensorIndex = (CSD.SigmaLayer.tensorIndexEquiv NA NB).trans (finCongr ⋯)
Instances For
The local operator algebra commutes (derived). Alice's U ⊗ I and Bob's I ⊗ Q commute on the
bipartite tensor sector: (U ⊗ I)(I ⊗ Q) = (I ⊗ Q)(U ⊗ I), both equal to U ⊗ Q. This is the algebraic
root of no-signalling, and it is a theorem of the Kronecker algebra, not a postulate.
Operator-level no-signalling on the tensor sector (re-exposed). Bob's expectation of I ⊗ Q is
unchanged by Alice's local unitary U ⊗ I: Alice cannot signal to Bob. This is the operator-form
companion of the singlet marginal HasNoSignalling (T15), on the genuine tensor sector. From
NoCommunication.no_communication.