In defence of the child innovator

This chapter confronts the two principal arguments levelled against the child-as-innovator approach to language change: (1) child innovations cannot underlie historical innovations because child innovations resolve before adulthood, when they could diffuse (e.g. Traugott and Dasher 2005; Diessel 2011), and (2) parallels must hold between child innovations and historical innovations, but parallels do not hold in the domain of morphosyntax (e.g. Diessel 2012). I argue that both parallel and oppositional alignments are predicted by the two possible innovation-types children make when solving the Mapping Problem (Clark 1977, 1993, i.a.); in short, different L1A processes underlie different types of change. I further argue that input-divergent analyses at most need to persist into the teenage years, when they can be diffused via the sociolinguistic change powerhouse of teenage peer groups (e.g. Labov 2012), and may also be reinforced and prolonged in childhood via peer-to-peer acquisitio...