Poster #113 - Not All is Forgotten: Children’s Memory for Features of a Word Learning Episode

• In Event: 3-106 - Poster Session 11
  In Poster Session: PS 11 Section - Language, Communication

Integrative Statement

Children must associate a referent and label in a rich sociolinguistic environment to facilitate word learning. Previous research has documented children’s proficiency in mapping words and facts unto referents (Carey, 2010). However, much of this work has been conducted in artificial word learning environments in which there is little exposure to information beyond a word and referent. In real-world situations, children must include a plethora of associations between words, objects, people, and the broader environmental context. Consequently, we know little about children’s associative matrices for multiple features of the environment, and their role in facilitating early word learning. The present study addressed this gap in research by investigating children’s recognition memory for associations encoded during a word learning episode. We hypothesized that children would have memory for more than just word-object mappings, and potentially have stronger memory for the contextual environment in which words are presented (Vlach & Sandhofer, 2010). To test these predictions, children (n = 57) ages 36-106 months (M = 63.67, SE = 2.39) were presented a word learning task with two phases. During the learning phase, participants were shown six novel objects across six animated storybook scenes (Fig. 1). In each scene, the storybook characters labelled a novel word twice (e.g. “Can you show me gipple?” “What is it like?” “Gipple is very comfortable to sleep on!”) and all scenes were viewed twice. The testing phase included six forced-choice recognition tests (i.e. object-word, person-object, scene-object, scene-person, scene-word, person-word; Fig. 1) containing six trials each. Interestingly, the results revealed that children were unsuccessful at remembering word-object mappings (Fig. 2). However, despite forgetting the word-object mappings, children demonstrated significant memory for associations between features of the environmental context. For example, although children did not remember the word for an object above chance level (p = .138), they did remember who labeled the object significantly above chance (p < .001). In particular, their memory for the object location was strongest. The strength of associations between words, objects, and people increased with age, yet memory for scene-object associations remained high after age 5. In sum, these results suggest that, even in cases where children do not remember word-referent associations, they possess a strong associative matrix for the word learning environment. Indeed, these results support arguments that early word learning is a contextually grounded process of determining associations, rather than a simple problem of induction.

Authors

• Melina L. Knabe, University of Wisconsin-Madison
  Presenting Author

• Haley Vlach, University of Wisconsin-Madison
  Non-Presenting Author