Generic interpretations of possessive recursion in English-speaking children

Abstract

Two-part s-possessives such as the dad’s kid’s bike admit at least two distinct interpretations: the dad has a kid who has a bike, or the dad has a bike that is made for kids. We propose that the former interpretation derives from recursively embedding DP-possessives, and the latter from representing kid’s bike as a generic NP-possessive. Accordingly, in the right context, two-part s-possessives are fully ambiguous for adults between ‘recursive’ and ‘generic’ readings. These readings can be disambiguated syntactically. Consider the difference in meaning when we insert a relative clause and extract the constituent kid’s bike — the kid’s bike that is the dad’s — versus when we extract the head noun bike — the bike that is the dad’s kid’s. Our story-based experiment demonstrates that 4-7-year-olds (N=79) and adults (N=68) strongly favor (~80%) the generic interpretation of phrases like the kid’s bike that is the dad’s, as the A-over-A constraint blocks the extraction of a DP-possessive out of a recursive DP. Similarly, adults show a strong preference (~80%) for recursive interpretations of phrases like the bike that is the dad’s kid’s, as the A-over-A constraint blocks the extraction of the head noun bike out of the generic NP-possessive kid’s bike. However, 4-5-year-olds admit generic readings of these recursive phrases 54% of the time; it is not until 6 or 7 years that children show an adult-like preference for the recursive interpretation (~80%). These data support two complementary claims. First, that recursive possessives are acquired late on account of their syntax, and second that children, like adults, represent generic possessives under a different syntactic node than regular possessives.