On the learning of morphological rules: An experimental analogy

Three experiments were conducted using modifications of a paradigm used by Esper (1925) . Using the study-test procedure, Ss learned 16 paired associates in which the stimuli were 2-digit numbers and the responses were 2-letter pairs. Each single digit was associated with a letter to form a four-by-four matrix of 2-digit 2-letter stimulus-response pairs. Sixteen college Ss were presented 12 of the 16 pairs (Exp. I); 12 of the 16 pairs plus 4 irregular pairs (Exp. II); or 12 of the 16 pairs plus 2 irregular pairs (Exp. III), and were tested with the 16 stimuli in all cases. The irregular pairs were presented three times as often (Exp. II) or twice as often (Exp. III) as the regular pairs. It was observed that the omitted pairs were learned quickly after the regularities of the presented pairs were learned; the irregular pairs were learned more rapidly than the regular pairs, and errors on the irregular pairs took the form of overgeneralization of the regular responses. It was assumed that the experimental conditions were analogous to the learning of verb inflection by children and the results were remarkably similar to the behavior of young children acquiring the morphological rules of verb inflection.