\DOC ITAUT_TAC \TYPE {ITAUT_TAC : tactic} \SYNOPSIS Simple intuitionistic logic prover. \DESCRIBE The tactic {ITAUT} attempts to prove the goal using a basic tableau-type proof search for intuitionistic first-order logic. The restriction to intuitionistic logic means that no principles such as the ``law of the excluded middle'' or ``law of double negation'' are used. \FAILURE May fail if the goal is unprovable, e.g. for purely propositional problems. For unsolvable problems with quantifiers it usually just loops. \EXAMPLE Suppose we try to prove the logical equivalence of ``contraposition'', already embedded in the pre-proved theorem {CONTRAPOS_THM}: { # g `!p q. (p ==> q) <=> (~q ==> ~p)`;; } \noindent by splitting it into two subgoals: { # e(REPEAT GEN_TAC THEN EQ_TAC);; val it : goalstack = 2 subgoals (2 total) `(~q ==> ~p) ==> p ==> q` `(p ==> q) ==> ~q ==> ~p` } \noindent The first subgoal (printed at the bottom) can be solved by {ITAUT_TAC}, indicating that it's intuitionistically valid: { # e ITAUT_TAC;; ... val it : goalstack = 1 subgoal (1 total) `(~q ==> ~p) ==> p ==> q` } \noindent but the other one isn't, though it is solvable by full classical logic: { # e(MESON_TAC[]);; val it : goalstack = No subgoals } \COMMENTS Normally, first-order reasoning should be performed by {MESON_TAC[]}, which is much more powerful, complete for all classical logic, and handles equality. The function {ITAUT_TAC} is mainly for ``bootstrapping'' purposes. Nevertheless it may sometimes be intellectually interesting to see that certain logical formulas are provable intuitionistically. \SEEALSO ITAUT, MESON_TAC. \ENDDOC