Base-symbol occurrence projections (neutral layer) #
The generic, unsigned occurrence bookkeeping shared by the Craig engine and the Lyndon
refinement: how the base function/relation symbols of a subformula sit inside those of the whole,
their invariance under substitution and openBounds, the instConst bounds, and the atomic facts
for constEq and relInst.
Nothing here mentions inseparability, side bounds, the paired family, a consistency property, or
the Henkin completion; the file's own dependency cone stops at the generated universe. It was
extracted from PairedInsepFamily.lean so that the polarity-refined development (issue #14, Units
2-3) can consume these projections without dragging the paired/Henkin assembly into its cone --
the boundary now matches the proof architecture. Declaration names and namespaces are unchanged,
so no compatibility shims are needed.
Base-symbol component projections #
The dual of the baseFunctionsIn_*_subset union bounds: a subformula's base symbols are contained
in the whole's.
Relation occurrences are invariant under subst / openBounds (for instConst) #
Atomic base symbols #
Constant equalities have no base symbols; a relation instance has only its own relation symbol, independent of the argument constants.