Children's knowledge of free choice inferences

This paper presents experimental results showing that four-year-old Mandarin- speaking children draw free choice inferences from disjunctive statements, though they are not able to compute inferences of exclusivity for disjunctive statements or other scalar implicatures. The findings connect to those of Chemla & Bott (under review) who report differences in how adults process free choice inferences versus scalar implicatures and, prima facie, the findings pose a challenge to treatments that attempt to unify inferences of both kinds. Instead, the findings appear to favour accounts that invoke different analyses for each kind of inference, such as Zimmerman 2000a, Geurts 2005, and Barker 2010. The results, however, also support the recent approach in the experimental literature which attributes children’s failures to compute scalar implicatures to a difficulty with alternatives: children may lack the lexical knowledge of alternatives, or these implicatures impose such a high processing cost that children are unable to handle the alternatives necessary to compute them (Gualmini, Crain, Meroni, Chierchia & Guasti 2001 Chierchia, Crain, Guasti & Thornton 2001 Reinhart 2006; Barner, Brooks & Bale 2011; Singh, Wexler, Astle, Kamawar & Fox 2012). If accessing alternatives is the source of children’s difficulty, then they would be expected to perform better if the requisite alternatives are made explicit, as sub-strings of the asserted sentences. This is exactly what we found. Children were able to compute free choice inferences based on alternatives that were made explicit in the assertion, but children were unable to compute ‘regular’ scalar implicatures arising from alternatives lacking this property. We discuss the implications of these findings for the debate about the relationship between free choice inferences and scalar implicatures and children’s knowledge of alternatives.