Trace-aware size measures of workspaces #
A workspace F on the carrier UnorderedTree (α ⊕ β), with Sum.inr marking a trace, is measured
by its number of components Multiset.card F, its number of accessible terms, which are the
non-root vertices Forest.numEdges F, and its number of vertices
Forest.numNodes F = card F + numEdges F. Under the trace-preserving coproduct a cut leaves a
trace leaf that is not an accessible term, so the trace-aware count discounts trace leaves:
accessibleCount T = numEdges T − traceLeafCount T on a tree, summed over a forest, and
accessibleSize F = card F + accessibleCount F. The extraction identities say how a Δ^c cut of
a lexical-rooted object splits its accessible terms between crown and trunk, one contraction per
cut.
Main definitions #
Main results #
ConnesKreimer.cutSummandsCN_accessibleCount_single,_pair: the extraction identities.
References #
- [marcolli-chomsky-berwick-2025], §1.6.1–1.6.2 (Lemma 1.6.3)
accessibleCount T = numEdges T − traceLeafCount T, the accessible terms of T that are
not traces.
Equations
- t.accessibleCount = t.numEdges - t.traceLeafCount
Instances For
External Merge adds two accessible terms.
The accessible terms of a workspace that are not traces, summed over its components.
Equations
- Forest.accessibleCount F = (Multiset.map UnorderedTree.accessibleCount F).sum
Instances For
accessibleSize F = card F + accessibleCount F, the trace-aware size of a workspace.
Equations
- Forest.accessibleSize F = F.card + Forest.accessibleCount F
Instances For
Contracting one accessible subtree Tv out of a lexical-rooted object splits its accessible
terms as accessibleCount T = accessibleCount Tv + accessibleCount (T/Tv) + 1, the +1
being the contraction itself.
Contracting two accessible subtrees adds two contractions: accessibleCount T is
accessibleCount Tv + accessibleCount Tw + accessibleCount (T/{Tv,Tw}) + 2.