Connectionism: Is it a paradigm shift for psychology?

Connectionism is a method of modeling cognition as the interaction of neuron-like units. Connectionism has received a gread deal of interest and may represent a paradigm shift for psychology. The nature of a paradigm shift (Kuhn, 1970) is reviewed with respect to connectionism. The reader is provided an overview on connectionism including: an introduction to connectionist modeling, new issues it emphasizes, a brief history, its developing sociopolitical impact, theoretical impact, and empirical impact. Cautions, concerns, and enthusiasm for connectionism are expressed.     

Memory as a function of attention, level of processing, and automatization

The relationships between long-term memory (LTM) modification, attentional allocation, and type of processing are examined. Automatic/controlled processing theory (Schneider & Shiffrin, 1977) predicts that the nature and amount of controlled processing determines LTM storage and that stimuli can be automatically processed with no lasting LTM effect. Subjects performed the following: (a) an intentional learning, (b) a semantic categorization, (c) a graphic categorization, (d) a distracting digit-search while intentionally learning words, and (e) a distracting digit-search while ignoring words. Frequency judgments were more accurate in the semantic and intentional conditions than the graphic condition. Frequency judgments in the digit-search conditions were near chance. Experiment 2 extensively trained subjects to develop automatic categorization. Automatic categorization produced no frequency learning and little recognition. These results also disconfirm the Hasher and Zacks (1979) "automatic encoding" proposal regarding the nature of processing.     

Reinforcer effectiveness as a function of reinforcer rate and magnitude: a comparison of concurrent performances

Pigeons' pecks on each of two concurrently available response keys were reinforced under a variable-interval schedule that sometimes allotted food-pellet deliveries to one key and sometimes to the other. The keys differed in the number of reinforcements assigned to each and in the number of pellets delivered during each reinforcement. When the total quantity of food associated with each key during a session was constant, the proportion of responses to a key depended on the particular combinations of reinforcer rate and reinforcer magnitude scheduled on each key. A given quantity of food generated more responding on a key when it was delivered frequently in small amounts than when it was delivered infrequently in large amounts.     

Separating the effects of interreinforcement time and number of interreinforcement responses

The relative importance of interreinforcement time and interreinforcement responses was evaluated by varying each independently. To do this, a blackout was presented after each nonreinforced response under both fixed-ratio and fixed-interval schedules of reinforcement. Manipulating the blackout duration under the fixed-ratio schedule caused interreinforcement time to vary without affecting the number of interreinforcement responses. Pigeons' post-reinforcement and post-blackout response latencies were found to increase linearly with interreinforcement time. Under the fixed-interval schedule, the same blackout manipulations changed the number of interreinforcement responses without affecting interreinforcement time. Post-reinforcement and post-blackout response latencies under this condition were approximately constant. These results suggest that responding is controlled by interreinforcement time and is not influenced by the number of responses emitted between reinforcements.