The well-known sampling limitation of most longitudinal corpus data can be even more consequential in the study of morphosyntactic variation in child language. An analysis of caregiver input suggests that variable use in overlapping contexts may be hard to find by solely relying on corpus data collected under the sampling procedures that are…

The first stage of combinatorial speech is better described as variable than uniform. Talk of variants obscures two different aspects of language (knowledge and use) and two different aspects of language development -- acquisition of the grammar (competence) and deployment of the grammar in speaking and listening (performance). Null subjects and…