Overregularization of the past tense in English

The phenomenon of overregularization of the past tense in English has been thought to be an indication of children's rule-based behavior in language. To examine this phenomenon, the transcripts of a longitudinal case study of one child's language production and archival longitudinal data for seven other children were examined for errors in past tense production. Such errors, which occur when the context calls for the past tense to be used, include neglecting to mark a verb for tense, misapplying the regular past tense ending (-ed) to an irregular verb ( e.g., breaked), and incorrectly marking an irregular verb twice for the past tense (e.g., broked). The frequency of these errors was compared with the frequency of correct regular and irregular past tense usage. The analyses examined several predictions of a dual mechanism theory. This theory states that there are two processes used by the child to create past tense forms: the regular rule (add -ed) is applied to regular verb forms, and irregular past tense forms are retrieved from memory through a network of associations between the irregular stem form and the correct past tense form. The majority of the results of the analyses of the different past tense forms supported the dual mechanism theory for English. Consistent with this theory, overregularization errors remained at a low rate over time for all the children. As also predicted by the dual mechanism theory, verbs with lower frequency in English had higher overregularization rates than high frequency verbs. In addition, verbs that were more recently acquired were more likely to be overregularized than those verbs that had been present in the child's vocabulary for a longer period of time. The dual mechanism theory, however, also predicts that overregularization errors should be preceded by the use of unmarked stems more often than by correct irregular forms; this prediction was not supported. Finally, it is suggested that, in the future, analyses of the effect of the child's own usage on his or her later productions, compared with the effect of parental usage on the same productions, be examined.