Pavlovian occasion-setting, attention and observing

Six pigeons were exposed to a procedure in which they could peck a white key to "observe" a 6-s keylight (CS) that was immediately followed by food 50% of the time. Another keylight (occasion-setter) was correlated with food 100% of the time, and illuminated at some point prior to food. The interval separating the onset of the occasion-setter from the opportunity to observe the CS was manipulated. The CS's temporal and probabilistic relationship to food, as well as the occasion-setter's probabilistic relationship to food, remained unchanged. Subjects were exposed to a set of conditions each of which arranged a different interval separating the occasion-setter from the opportunity to observe the CS. When the interval separating the occasion-setter from the CS was large, subjects pecked at higher rates to observe the CS following the onset of the occasion-setter than during its absence. The differential observing corresponded with higher rates of CS pecking on occasion-setting trials. Also, different occasion-setting presentations (trace and continuous) established differential observing. When the interval separating the occasion-setter presentations from the opportunities to observe the CS was reduced, rates of pecking to observe the CS, and rates of pecking the CS, were lower or became undifferentiated. The results support an interpretation of occasion-setting as a procedure which implicitly arranges for the differential reinforcement of "attending" to the CS