The role of pragmatic context in children’s quantification of parts and wholes

Why do children sometimes include the pieces of broken things when asked to count (e.g., “Count the forks”)? In five experiments, we test two ideas: (1) that adults, but not children, exclude parts from counts by reasoning that speakers must only be interested in whole objects with functions that fulfill desired goals, and (2) that only adults exclude parts on the basis of a Gricean inference, on which it is assumed that if a speaker wanted parts to be counted, they would say so. In Experiments 1-3, we find that although children and adults are somewhat less likely to count pieces of objects when speaker goals are explicitly mentioned, large developmental differences remain unexplained. In Experiments 4-5, we provide evidence that children differ most from adults when pieces are presented in the presence of complementary parts (with which they together would form a whole), and that these pieces are significantly more likely to be conceptualized and labeled as “parts”, especially by adults. Together, these findings suggest that adults – but not children – are more likely to exclude broken parts from their counts in response to requests like “Count the forks!” when they are also able to understand that more informative requests could have been made (e.g., “Count the pieces of fork!”).