The unit preference strategy in theorem proving

Unit Preference and Set of Support Strategies The theorems, axioms, etc., to which the algorithm and strategies described in this paper are applied are stated in a normal form defined as follows: A literal is formed by prefixing a predicate letter to an appropriate number of arguments (constants, variables, or expressions formed with the aid of function symbols) and then perhaps writing a negation sign (-) before the predicate letter. For example: P(b, x) -P(b, x) Q(y) R(a, b, x, z, c) S are all literals if P, Q, R, and S are two-, one-, five-, and zero-place predicate letters, respectively. The predicate letter is usually thought of as standing for some n-place relation. Then the literal P(a, b), for example, is thought of as saying that the ordered pair (a, b) has the property P. The literal -P(a, b) is thought of as saying that (a, b) does not have the property P.