Categorizing and quantifying parts of objects: A window into how children and adults interpret count nouns and their reference
Count nouns such as ‘cup’ and ‘ball’ are among the first words that children correctly interpret and produce. One might, then, reason that children have the same understanding about what these terms refer to (or what falls under their extension) as adults do. However, experimental results seem to indicate otherwise. A curious finding documented in a number of studies since Shipley and Shepperson (1990) is that young children seem to treat discrete parts of objects as if they were whole objects. For example, when asked to count a set of forks (some parts of forks and some whole forks), or touch every fork, children include partial forks in the same way as they do the whole forks. Adults robustly do not. Different proposals have attempted to account for this divergent pattern, implicating conceptual representations, comparison among lexical alternatives, or the computation of pragmatic inferences tied to the context at hand or speaker goals. In this talk, I present a set of experiments with children and adults, the results of which argue in favor of developmental continuity, in that children and adults actually share the same conceptual representations of objects and semantic representation of count nouns, but differ in their ability to rapidly recruit contextual information (e.g., speaker goals) and measure out fractions of objects along more fine-grained scales.
Joint work with Athulya Aravind (MIT)