Errors of multiple exponence in child English: a study of past tense formation

It is well known that children produce non-adult-like forms during language acquisition. Among these are errors where in the fashion of multiple exponence the child overtly marks an underlying feature or category more than once. In addition, children also produce errors where features that are marked fusionally with one form in the target language are marked separately with more than one form by the child. This paper is concerned with such errors in the domain of English past tense. We present a comprehensive corpus study investigating the frequencies and distribution of different error types, combining both overregularization and overtensing errors, which have previously been studied separately. We then propose an analysis based on Generalized Head Movement (Arregi & Pietraszko, 2021) and Distributed Morphology arguing that errors can be derived from two occasionally occurring underlying mistakes: negligence of secondary features and omission of obliteration. We show how these two mistakes and their interaction can account for the overall differences in error rates and distributions between different error types as well as across different verbs.