Inferring What Others Know: The Role of a Speaker’s Epistemic Credentials

• In Event: 3-236 - Linguistic and Cognitive Representations of Information Access

Integrative Statement

Children appeal to a range of considerations when reasoning about others as sources of information (e.g., familiarity, Corriveau & Harris, 2009; age, Jaswal & Neely, 2006; morality, Landrum, Mills, & Johnston, 2013; accuracy, Pasquini, Corriveau, Koenig, & Harris, 2007). The relation between a claim and the speaker’s epistemic credentials is especially relevant for children’s learning. For example, when learning new semantic information, children favor sources with a history of relevant linguistic or conceptual expertise but privilege perceptual access when learning episodic information (Brosseau-Liard & Birch, 2011). Although a body of research suggests that children are sensitive to a speaker’s epistemic bases (e.g., Kondrad & Jaswal, 2011; Stephens & Koenig, 2015), much of this research includes children from relatively homogenous, upper-socioeconomic (SES) backgrounds. Here we investigated these questions in a predominantly low-income sample.
One recent study reported individual differences in 3-year-olds’ source evaluations that were linked to parents’ authoritarianism (Reifen Tagar, Federico, Lyons, Ludeke, & Koenig, 2014). Children whose parents endorsed more authoritarian parenting values (i.e., respect for elders, obedience) were more discriminating between competent and incompetent sources in their learning decisions.
In the current study, we investigated children’s source evaluations in 280 low-SES preschoolers (M = 45.2 mos; SD = 2.5 m). Children were presented with semantically-competent (e.g., “chair” – [chair]) and -incompetent (e.g., “ball” – [shoe]) informants. At test, they selected which informant to ask and endorse for semantic (e.g., a novel animal label) as well as episodic (e.g., the unseen contents of a box) information. To examine the link between mothers’ authoritarian predisposition and children’s learning decisions, mothers were asked to select the more important quality for a child to have from a series of 4 paired qualities (e.g., obedience or self-reliance, Reifen Tagar et al., 2014).
Overall, children systematically selected the more competent source for novel semantic, t(279) = 7.51, p < .001, and episodic information, t(279) = 4.79, p < .001. They also systematically endorsed the novel semantic information offered by the competent source, t(278) = 5.89, p < .001. However, they endorsed the episodic information offered by each source at rates no different from chance (chance = .50), t(279) = 1.32, p = .19. Also, children from high authoritarian families (i.e., endorsed the authoritarian value on 75% -100% of trials) were less likely to ask the competent speaker for semantic information (M = .64, SD = .38) than children from less-authoritative mothers (M = .73, SD = .33), t(239.7) = 2.01, p < .05.
These results suggest that young, low-SES children flexibly draw inferences about speakers who display more or less semantic competence as evident in their circumscribed patterns of information-seeking. Children’s endorsement of new semantic information from a speaker seen to have semantic competence did not extend to the new episodic information she offered. Additionally, children from low authoritarian families might be especially sensitive to the nature and scope of the knowledge offered by a semantically-accurate speaker, choosing to ask her more often for semantic information than children with more authoritarian mothers.

Authors

• Katherine Ridge, University of Minnesota
  Presenting Author

• Fred Rogosch, University of Rochester
  Non-Presenting Author

• Dante Cicchetti, University of Minnesota
  Non-Presenting Author

• Melissa A Koenig, Institute of Child Development, University of Minnesota
  Non-Presenting Author