Author:

Abstract:

The differences in mental abilities between human and non-human animals have aroused human curiosity from antiquity. At the end of the 19th century, researchers started to realize that knowledge about an animal’s learning history was of crucial importance for deriving conclusions about the mental capacities subserving its behavior. In tandem, the question was raised whether the mechanisms underlying learning would be different for humans than for non-human animals. Until now, a consensus on this issue has not been reached. The goal of the first part of the current research was to investigate in rodents the mechanisms underlying blocking, a phenomenon that can be accounted for by association-formation models as well as by inferential reasoning models of learning. The involvement of inference making in blocking is suggested by effects of (sub)additive pretraining on blocking in humans and rats. However, models have been proposed that put those observations within reach of association-formation accounts. Importantly, such accounts would have difficulties explaining subadditivity posttraining effects in rats. To investigate whether blocking is attenuated by subadditivity posttraining, we aimed to establish a blocking procedure that would yield a strong blocking effect. However, over fifteen experiments we failed to observe a robust blocking effect. We discuss the implications of our failures for the canonical status of blocking and the psychological literature in general. The second goal of this research was to explore to what extent other forms of seemingly rule-based behavior in rats rely on the operation of inferential reasoning processes. Animals are capable of learning complex discriminations that instantiate overarching rules. However, this need not imply that they actually represent those rules. We investigated the generalization of patterning rules in rats, pigeons and humans. All rat and pigeon subjects generalized on the basis of similarity, whereas about half of the human subjects generalized on the basis of rules. In follow-up experiments, we attempted to test for generalization in rats (1) after training on multiple examples and (2) after training on instances of a different rule. However, rats failed to learn the initial discrimination problems in those experiments, preventing us from testing our substantive hypotheses. The third goal of this project was to investigate the potential contribution of three cognitive factors (intelligence, systemizing and visual processing) to individual differences in learning and generalization of patterning problems observed in humans. Only intelligence seemed to influence learning and generalization, a findings that calls for a re-evaluation of previous claims regarding the cognitive processes involved in the learning of patterning discriminations in humans.