Effects of a Stimulus' Valence and Utility on the Observing Response of Tufted Capuchins

Pigeons will step on a treadle to produce a discriminative stimulus indicating whether key pecking will produce food. Learning theorists attribute this “observing response” to the conditioned reinforcing power that accrues to the stimulus through its occasional pairing with food. Subsequently, cognitive theorists argued treadle stepping is due to the stimulus’ informative value because it informs about the likelihood of reinforcement. One view of the literature supports the notion that the information-theory hypothesis explains human and other primate observing behavior, whereas a conditioned-reinforcement interpretation is appropriate for pigeons and rats. Here, it is proposed that emission of observing responses is maintained by whether a discriminative stimulus (a) signals reward is proximal or not (positive or negative valence), and (b) enables greater or lesser response efficiency (large or small utility). Eight capuchins were trained in four observing-response experiments that varied a discriminative stimulus’ valence and utility. If cognitive theorists are correct that observing responses serve to reduce uncertainty about reinforcement, stimuli predicting reward is imminent (good news) or not (bad news) should support observing behavior in all experiments because they inform equally well. However, if the conditioned-reinforcement thesis is correct, good news should support more observing than bad news except when the utility of that bad news is large. The results support the conditioned-reinforcement hypothesis and suggest that observing behavior can be explained as due to learning-theory conceptions regardless of species’ phyletic level.