0.12/0.12 % Problem : theBenchmark.p : TPTP v0.0.0. Released v0.0.0. 0.12/0.13 % Command : twee %s --tstp --casc --quiet --explain-encoding --conditional-encoding if --smaller --drop-non-horn 0.12/0.33 % Computer : n004.cluster.edu 0.12/0.33 % Model : x86_64 x86_64 0.12/0.33 % CPU : Intel(R) Xeon(R) CPU E5-2620 v4 @ 2.10GHz 0.12/0.33 % Memory : 8042.1875MB 0.12/0.33 % OS : Linux 3.10.0-693.el7.x86_64 0.12/0.33 % CPULimit : 960 0.12/0.33 % WCLimit : 120 0.12/0.33 % DateTime : Thu Jul 2 07:16:58 EDT 2020 0.12/0.34 % CPUTime : 2.22/0.67 % SZS status Theorem 2.22/0.67 2.22/0.67 % SZS output start Proof 2.22/0.67 Take the following subset of the input axioms: 2.22/0.67 fof(less_property, axiom, ![X, Y]: ((X!=Y & ~less(Y, X)) <=> less(X, Y))). 2.22/0.67 fof(something_not_n12, conjecture, ?[X]: X!=n12). 2.22/0.67 2.22/0.67 Now clausify the problem and encode Horn clauses using encoding 3 of 2.22/0.67 http://www.cse.chalmers.se/~nicsma/papers/horn.pdf. 2.22/0.67 We repeatedly replace C & s=t => u=v by the two clauses: 2.22/0.67 fresh(y, y, x1...xn) = u 2.22/0.67 C => fresh(s, t, x1...xn) = v 2.22/0.67 where fresh is a fresh function symbol and x1..xn are the free 2.22/0.67 variables of u and v. 2.22/0.67 A predicate p(X) is encoded as p(X)=true (this is sound, because the 2.22/0.67 input problem has no model of domain size 1). 2.22/0.67 2.22/0.67 The encoding turns the above axioms into the following unit equations and goals: 2.22/0.67 2.22/0.67 Axiom 1 (something_not_n12): X = n12. 2.22/0.67 2.22/0.67 Lemma 2: X = ?. 2.22/0.67 Proof: 2.22/0.67 X 2.22/0.67 = { by axiom 1 (something_not_n12) } 2.22/0.67 n12 2.22/0.67 = { by axiom 1 (something_not_n12) } 2.22/0.67 ? 2.22/0.67 2.22/0.67 Goal 1 (true_equals_false): true = false. 2.22/0.67 Proof: 2.22/0.67 true 2.22/0.67 = { by lemma 2 } 2.22/0.67 ? 2.22/0.67 = { by lemma 2 } 2.22/0.67 false 2.22/0.67 % SZS output end Proof 2.22/0.67 2.22/0.67 RESULT: Theorem (the conjecture is true). 2.22/0.68 EOF