One major issue in learning, the conditions issue, is concerned with the conditions under which learning occurs. For example, in order for learning to occur is it sufficient for two events to be contiguous or must they, in addition, be followed by reinforcement? A second issue, the what issue, is concerned with what is learned. For example, does learning consist of an association between a stimulus, such as a tone, and a response, such as pressing a lever, or are representations of events involved?
Violations of the conditions assumptions, decades apart, led to widespread and vigorous empirical and theoretical activity and subsequently to modification of both conditions and what assumptions. The first violation was the finding that reinforcing only some responses (partial reinforcement, PRF) produced greater resistance to extinction than reinforcing all responses (consistent reinforcement, CRF), a phenomena called the partial reinforcement extinction effect or PREE (Pearce, 2002). The PREE was seen as violating the view that forming an association between events required reinforcement
(Capaldi, 1966, 1967). The second violation of the conditions assumption occurred when it was found that presenting a stimulus X in compound with an already conditioned stimulus A and reinforcing the compound (AX) produced only weak responding to X when it was presented by itself, a phenomenon