Exploration Facilitates Retention during Novel Noun Generalization

• In Event: Cascading Consequences of Object Play on Cognition

Abstract

When building a lexicon, children must map and retain word-referent pairs and generalize the labels to novel instances. Two-year-olds are capable of this, successfully mapping a novel label to a novel object in a fast-mapping task (Carey, 1978) and reliably extending novel names to solid objects by their overall similarity in shape as opposed to other dimensions in a novel noun generalization (NNG) task (Landau et al., 1988). That is, word learning is supported by the combination of multiple unique processes (mapping, retaining, and generalizing). Part of this process (retention) improves dramatically when children interact with the novel objects before a fast-mapping task (Kucker & Samuelson, 2012), but it is unknown what impact such interaction has on retention or generalization in NNG.
To address these questions, 134 17-30-month-olds participated in a canonical NNG task either with or without pre-familiarization. In the pre-familiarization condition, participants interacted with an exemplar, a shape-match item, and a material-match item before every trial. The experimenter then presented the exemplar and labeled it three times (“This is my kiv.”) and asked children to select either the shape-match or material-match (“Can you get your kiv?”). In the no familiarization condition, trials were conducted in the same manner, except that children did not interact with the objects before any trial. After the NNG task, children were tested on their retention of the labels for the exemplars. Parents completed the MacArthur Bates Communicative Development Inventory as a measure of productive vocabulary (Fenson et al., 1994). Children’s choices to shape during generalization, their retention responses, and their vocabulary size were analyzed.

Overall, children in both conditions chose the shape-matching item during generalization. Children in the pre-familiarization condition also retained at above chance levels, t(186)=3.38, p=.001, whereas the no familiarization condition did not, t(207)=1.16, p=.25. Furthermore, consistent with prior work, children’s generalization abilities were predicted by their vocabulary, especially the number of count nouns, B=.001, t=2.36, p=.019, but as hypothesized, generalization was not impacted by condition, B=-.05, t=-1.24, p=.22, or interaction of condition and vocabulary, B=.001, t=1.11, p=.27 (see Figure 1). However, retention performance was predicted by an interaction of condition and count noun vocabulary, B=.001, t=2.48, p=.015; there was a positive correlation between vocabulary and retention in the pre-familiarization, but performance was at chance regardless of vocabulary in the no-familiarization condition (Figure 2).

Collectively, these data suggest that exploring stimuli before a trial significantly improves retention without impacting generalization. That is, embodied experiences, such as physically interacting with novel toys, helps children build long-lasting connections, supporting learning. These results are consistent with previous research by Kucker and Samuelson (2012) where familiarization with objects before a referent selection task aided retention of word-referent mappings. Critically, the present results show that robust word-referent mappings can be made in the context of a generalization task, but that this mapping depends on first being able to explore the objects. Consequently, these results provide further evidence for the role of object exploration in early word learning and emphasize the importance embodiment plays in development generally.

Authors

• Megan G Lorenz, Augustana College
  Presenting Author

• Sarah C Kucker, Southern Methodist University
  Non-Presenting Author