Necessary and sufficient factors in classical conditioning.

The issue of necessary and sufficient factors (pairing-contiguity vs. contingency-correlation) in classical (Pavlovian) excitatory conditioning is examined: first, in terms of definitional (logical) and manipulational requirements of "necessary" and "sufficient"; second, in terms of Boolean logic test models indicating experimental and control manipulations in tests of pairing and contingency as necessary and sufficient factors; and, third, by a selective review of reference experiments showing appropriate experimental and control manipulations of pairing and contingency indicated in the Boolean logic test models. Results of examination show pairing-contiguity as the sole necessary and sufficient factor for excitatory conditioning, while contingency-correlation is conceptualized as a modulating factor controlling minimal-maximal effects of pairing-contiguity. Reservations and diagnostic experiments are indicated to assess effects of uncontrolled conditioned stimulus--unconditioned stimulus (--CS--US) probability characteristics (e.g., p (CS--US)/p (--CS--US) in truly random (TR) schedule manipulations). Similar analysis of conditioned inhibition reveals insufficient evidence to support a choice among current alternatives.