You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The hypothesis [H] is the result of our encoding of the LL universal
216
-
quantifier as the (dependent type) constructor [forall] in Coq. In the
217
-
usual sequent calculi presentation, that hypothesis is stronger:
218
-
#<i>for any fresh variable x, there is a proof of height n</i>#. Then,
219
-
it suffices to generalize H with a fresh variable, and then, from
220
-
alpha conversion, we can use exactly the same fresh variable to
221
-
conclude the goal.
222
-
214
+
The hypothesis [H] results in inductive proofs where the principal
215
+
formula is (the LL universal quantifier} [F{ FX}]. Note that we cannot
216
+
conclude the goal [G] from [H] since our hypothesis is weaker than the
217
+
similar one in pencil/paper proofs. More precisely, in a paper proof,
218
+
we can generalize [H] with a fresh variable [x]. Then, there exists
219
+
[n] s.t. [n |- Gamma, Subst Fx x]. By using the substitution lemma,
220
+
for any [y], it must be the case [n |- Gamma, Subst Fx y] and we can
221
+
easily conclude the goal [G].
223
222
*)
224
223
Axiom ax_subs_prop: forall B L M FX (P:nat -> Prop), (forall x : Term, exists n : nat, (P n) /\ n |-F- B; L; UP (Subst FX x :: M)) -> exists n, (P n) /\ forall x, n |-F- B; L; UP (Subst FX x :: M) .
0 commit comments