Poster #110 - Concepts, access or context? A relevance account of a classic children's counting error.

• In Event: 2-110 - Poster Session 07
  In Poster Session: PS 07 Section - Language, Communication

Integrative Statement

When shown a fork broken into three pieces and asked 'how many forks are there?' children under five consistently respond with 'three forks' (Shipley & Shepperson, 1983). This error has also been shown in comparing object quantities - children consistently
judge one fork in three pieces as 'more forks' than two whole forks (Brooks et al., 2013). This error has been attributed to developmental differences in sortal concepts (Brooks et al., 2013) or the ease with which children access alternative word forms ('broken', 'whole', Srinivasan et al., 2015).

We offer an alternative, account of the error: Children (unlike adults) do not consider the whole/broken status of an object relevant to counting/comparing quantities and it is primarily an issue of relevance that prevents adultlike counting/comparing. Like Srinivasan et al. (2015) this account considers the error a pragmatic one. However, unlike Srinivasan et al. (2015) we argue there is no difficulty associated with accessing alternative word forms - rather children simply have a different set of assumptions regarding what objects are relevant in counting/comparing tasks relative to adults.

We test this hypothesis by providing four-year-olds with a vignette in which characters use the objects counted/compared. This framing highlights the relevance of whole objects since only whole objects can be properly used. The relevance condition is compared to a priming condition from Srinivasan et al. (2015) in which children are asked to describe the objects they are counting as a 'whole [object]' or a 'piece of an [object]'. If the source of children's errors is a difficulty in accessing alternative word forms they should perform better after priming with these word forms. Both conditions are also compared to a control condition where children complete the counting/comparing task with no training.

 Results show that four-year-olds are not above control performance in the priming condition. In contrast, children in the vignette condition perform at ceiling, with twenty-three out of twenty-four participants consistently counting in an adultlike manner.

Given the vignette neither alters the sortal concepts involved nor provides alternative word forms, and given that this is the first time that ceiling performance has been demonstrated in four-year-olds in this paradigm, these results both strongly support a relevance account of the Shipley and Shepperson (1983) phenomenon. Implications for our interpretations of pragmatic development and measures thereof are discussed.

Authors

• Ivan Kroupin, Harvard University
  Presenting Author

• Susan Carey, Harvard University
  Non-Presenting Author