Natalie Bleijlevens and Tanya Behne
Affiliation: Developmental Psychology, University of Göttingen, Germandy
Category: Psychology
Keywords: word learning, metacognition, mutual exclusivity, uncertainty, updating, disambiguation
Date: Tuesday 2nd of September
Time: 15:30
Location: Room 232 (232)
View the full session: Metacognition
When learning novel words, referential ambiguity is a constant part of young children’s learning environment. Despite this ambiguity, children seem to infer the referents of novel words with relative ease, e.g., by assuming novel labels to refer to unfamiliar rather than familiar objects, known as the “mutual exclusivity” effect (see Lewis et al., 2020, for a review). However, to date, it is unclear whether young children are in some way aware of the different levels of uncertainty involved in referent identification, and in how far they can effectively use this information. In two pre-registered studies, we thus assessed preschoolers’ and adults’ ability to monitor their uncertainty during referential ambiguity and update labels dependent on their initial learning context. In study 1 (https://osf.io/gmjru/), we asked 4-5-year-olds (n=82) and adults (n=70) to find the referents of novel words in contexts with maximal ambiguity (in presence of two novel objects), medium ambiguity (in presence of one novel and one known object, resolvable via mutual exclusivity), or minimal ambiguity (in presence of one novel object), and we measured explicit assessments of their own uncertainty on a 3-point scale (Hembacher & Ghetti, 2014), as well as other uncertainty behaviors (e.g., response times, social referencing, verbal questions). Afterwards, participants faced counterevidence to their previously made word-object-mappings, and needed to decide which of the mappings (learned in more vs. less ambiguous contexts) they want to update. Results show that children’s and adults’ explicit and implicit uncertainty systematically increased with the level of ambiguity in the task. Further, while high uncertainty during initial word-object-mappings increased adults’ willingness to update these links subsequently (as long as the contrast between both learning contexts was high enough), this was not the case for children. Study 2 (https://osf.io/79g8t/) thus investigates potential improvements in 5-6-year-olds’ (n=90) and adults’ (n=90) updating performance in a task with increased contrast between both initial learning contexts. Adults almost exclusively updated the most ambiguously learned word-object-mapping; data collection for children is ongoing and will be completed until September. The complete pattern of results will contribute to discussions about the mechanisms and mental processes that enable children to learn words in the face of ambiguity.