In 1968 there was a symposium at Dalhousie University, the proceedings of which were published with the title Fundamental Issues in Associative Learning. There may have been some exaggeration in the title, but there was an extraordinary shared vision among the participants about what was deemed challenging at the time, and how to approach the challenge through an invigorated investigation of Pavlovian conditioning. Especially congruent were the presentations by Leon Kamin, Robert Rescorla, and myself. An important result of the research and theorizing that followed has been the appreciation of how “expectations” shape the basic regularities of associative learning as they play out even in the simplest instances of animal behavior.

Associative learning refers to the process by which one stimulus comes to recall the memory of another with which it has been paired There are numerous ways in which this fundamental memory process can be studied. In the procedure introduced by Ivan Pavlov with dogs, an arbitrary stimulus such as a tone or a light (referred to as a conditioned stimulus, or CS) was presented prior to providing the animal access to food (referred to as an unconditioned stimulus, or US). The indication that such pairing of CS and US caused the animal to associate the two stimuli was that the animal would come to make a conditioned response (or CR) of salivation to the CS, prior to the delivery of the US that would normally provoke such response. In my laboratory, with rabbits, we frequently used similar CSs, but employed tactile stimulation to the skin near the eye as the US, and recorded eye blinks that developed to the CS as an indication of its acquired association with the US.

An interpretive speculation by Kamin concerning one of his studies had a special impact upon my subsequent theorizing. In his study all animals were trained with a compound of two CSs (call them “A” and “X”) followed by a US (i.e., involved the sequence, AX➔US), and were subsequently tested for their CRs to one of the two CSs (X) alone. If the animals received no other training, there were substantial CRs to X. The important comparative observation was that if the animals received pre-training with A paired with the US (i.e., A➔US) before the AX➔US training, there were few or no CRs to X alone.