Poster #125 - Turkish- and English-speaking Children’s Belief Revision Based on the Reliability of the Counterargument

• In Event: Poster Session 09

Abstract

Children revise their beliefs depending on the reliability of the counter-evidence offered to them. Research with English- and German-speaking children showed that 5-year-olds revised their beliefs more often when they were given a contradicting claim based on first-hand evidence compared to second-hand evidence, while 3-year-olds were less likely to change their minds (Köymen & Tomasello, 2018; Schleihauf et al., 2022). Unlike English or German, in which evidential marking is optional, in languages such as Turkish, the obligatory use of evidential markers provides information about the information’s reliability, e.g., whether the speaker’s knowledge is based on direct- or indirect observation. When reporting a past event, Turkish speakers must choose one of two evidential markers: -DI for the direct observation; –mIş for the indirect information or an inference. The source information that comes with the obligatory use of evidentials might provide advantages for Turkish-speaking children in their evaluations of the reliability of the information (Aydın & Ceci, 2009). In this preregistered online study, we investigate whether Turkish- and English-speaking 3- and 5-year-olds revise their beliefs given counter-evidence varying in degrees of reliability (based on direct evidence, indirect evidence, or inference).
Three- and 5-year-old Turkish-speaking (based in Turkey) and English-speaking (based in Canada) children (preregistered N=144, data collection ongoing, 90% complete) were tasked with finding the needs of a novel animal, e.g., what it eats. At the beginning of each trial, they were asked to choose a food item for the animal between a plausible (apples) and an implausible item (rocks). Children predominantly chose the plausible item (apples). Then, an informant offered evidence in favor of the implausible item. In the direct-observation condition, the informant said, “I saw it eating rocks, this animal ate rocks” [Ben taş yerken gördüm (-DI), bu hayvan taş yedi (-DI)]. In the indirect-observation condition, the informant said, “Someone saw it eating rocks, this animal ate rocks”[Biri taş yerken görmüş (-mIş), bu hayvan taş yemiş (-mIş)]. In the inference condition, the informant said “I saw rocks on its plate, this animal ate rocks”[Ben tabağında taş gördüm (-DI), bu hayvan taş yemiş (-mIş). Then, children were asked again which of the two items they think the animal eats to see whether they revised their initial beliefs.
We predict that children will be more likely to revise their beliefs in the direct-observation condition compared to the indirect-observation and inference conditions. We expect the condition difference will be higher in Turkish-speaking children (especially for 3-year-olds). We will also explore whether children revise their beliefs more often in the inference condition than in the indirect-observation condition.
We will use a Generalized Linear Mixed Model with binomial error distribution in our analysis. Our model will include the main predictors of age (3 vs. 5), condition (within-subjects), linguistic group (Turkish speakers vs. English speakers), their interactions; control predictors of trial order, and gender; and the random intercept of the child.
With this, we aim to shed light on the role of linguistic properties and reliability cues on children’s belief revision.

Authors

• F. Ece Ozkan, University of Toronto
  Presenting Author

• Samuel Ronfard, University of Toronto

• Bahar Koymen, University of Manchester