Lemmas and tactics are introduced for induction on the rank of
an expression (`inverse image induction'). One deficiency of Nuprl's
type theory is that in some situations (notably when proving well-formedness
goals) lemmas are unusable, and instead one must resort to using
tactics that reproduce equivalent sequences of primitive rules.
The theorem
Thm* r:(TTProp), S:Type, f:(ST).
Thm* WellFnd{i}(T;x,y.r(x,y)) WellFnd{i}(S;x,y.r(f(x),f(y)))
provides a template for such a tactic
that mirrors the
Thm* r:(TTProp), S:Type, f:(ST).
Thm* WellFnd{i}(T;x,y.r(x,y)) WellFnd{i}(S;x,y.r(f(x),f(y)))
lemma.
About: