A label-coordinate bookkeeping for entangled pairs: what co-location can and cannot explain

We formalize and evaluate the conjecture that the non-locality of an entangled pair's state update belongs to the laboratory's description: in a chart adapted to the pair, built from a label coordinate and label-dependent shears, the partners are co-located while negativity-entangled. The central result is negative and exact: a rule acting where the partners meet can produce every one-way-no-signaling correlation, so co-location explains neither the Tsirelson bound nor no-signaling; the correlation strength is imported from quantum mechanics, and the one axiom the geometry can be given, order-independence, narrows the class exactly to the no-signaling polytope. The mechanism is stated in four postulates (a locality-graph rule, a state-driven-component rule, a transfer rule for swapping, and an invariant interval clock for a Lorentz-covariant variant in 1+1 dimensions) and proved empirically equivalent to the nonrelativistic quantum mechanics of distinguishable massive particles for spin instruments that do not couple spin to position. Three stronger readings are excluded by the no-communication theorem, by gravity and collider data, and by fine-tuning; a massive-particle test of the first, needing of order 10^2 to 10^3 trials, is proposed. The covariant variant orders a pair's measurements by a strict Lorentz-invariant partial order carried by events, not charts, undefined on a tie set the symmetric Bell configuration occupies; swapped pairs are at constant separation, not co-located. It differs empirically only from dynamical-collapse models, predicting zero collapse noise as every no-collapse account does. The claim is validity, not correctness: a state-dependent bookkeeping that locates and orders the update, at stated costs.