The closed identification at an arbitrary empty label type #
The composition of two fragments is built at the label type
Fin 0 ⊕ Fin 0 and then relabelled to Fin 0. The Definition 5
value lives at the latter, the colouring recursion at the former, so
the two have to be matched across the relabel.
Both sides are the same sum of RS21 summands. At an empty label type
the chord sign is one and the label chords are empty, so any two
canonical data give the same summand; and the summand itself is
carried across a relabel by relabel_throughSummand. Together these
identify the relabelled fragment's Definition 5 value with the
constrained value downstairs, with no independence input.
At an empty label type the summand does not depend on the canonical data. The chord sign is one and the label chords are empty, so Proposition 3 equates the two signed values.
The relabelled fragment's Definition 5 value is the summand
downstairs. Both are RS21's s_h(G,H) for the same subset.
At an empty label type every subset matches every state.
At an empty label type every bijection to Fin 0 is monotone.
Equations
- RS.EdgeSubset.emptyOrderIso ee = { toEquiv := ee, map_rel_iff' := ⋯ }
Instances For
The relabelled fragment's Definition 5 partition value is the constrained value downstairs, at a monotone relabel.
The relabelled fragment's Definition 5 partition value is the constrained value downstairs. Monotonicity is automatic: there is nothing to compare.
The same identification, for a fragment presented as a relabel. Naming the relabelled fragment keeps the elaborator from having to solve for it under the relabel.