In this article I review research and theory on the "interference paradigms" in Pavlovian
learning. In these situations (e.g., extinction, counterconditioning, and latent inhibition),
a conditioned stimulus (CS) is associated with different unconditioned stimuli (USs)
or outcomes in different phases of the experiment; retroactive interference, proactive
interference, or both are often observed. In all of the paradigms, contextual stimuli
influence performance, and when information is available, so does the passage of time.
Memories of both phases are retained, and performance may depend on which is retrieved.
Despite the similarity of the paradigms, conditioning theories tend to explain them
with separate mechanisms. They also do not provide an adequate account of the context's
role, fail to predict the effects of time, and overemphasize the role of learning
or storage deficits. By accepting 4 propositions about animal memory (i.e., contextual
stimuli guide retrieval, time is a context, different memories are differentially
dependent on context, and interference occurs at performance output), a memory retrieval
framework can provide an integrated account of context, time, and performance in the
various paradigms.