Poster #37 - Effects of Early Language Environments on Metaphorical Language Comprehension

• In Event: Poster Session 12

Abstract

Metaphorical language is difficult for preschool-aged children to understand, even when the metaphors involve familiar concepts (Billow, 1975). This difficulty with metaphors might stem from the mutual exclusivity bias—young children’s tendency to expect each concept to have only one label (Markman & Wachtel, 1988). Because metaphorical language requires children to use known words in new contexts, it may violate this bias. However, recent work demonstrates that in some situations, young children can successfully comprehend metaphors (Pouscoulous & Tomasello, 2020). The first goal of this study was to explore children’s comprehension of different types of metaphors. We tested children with perceptual metaphors, which can be resolved based on shared visual features, and abstract metaphors, which can be resolved based on cross-dimensional associations. The second goal was to explore how children’s language learning experiences influence their metaphor comprehension. Children’s expectations about how words work are calibrated as they learn language, and bilingual children exhibit a weaker mutual exclusivity bias than monolingual children. (Byers-Heinlein & Werker, 2009; Houston-Price et al.,2010). Because bilingual children already expect concepts to have more than one label, we hypothesized that they may have an advantage in comprehending metaphorical language.

In Experiment 1 (N=75 monolingual and 50 bilingual children aged 2.5-4.5 years, conducted over Zoom), children saw two cartoon pictures and were asked to choose the picture that best goes with the description provided by the experimenter (e.g., “Which one is the dog with the socks?”). Each trial showed two pictures that were identical apart from one unique feature per picture, only one of which corresponded to the metaphorical phrase (e.g., a dog with white feet versus a dog with a white bow). Overall, children performed above chance on both perceptual and abstract metaphors. However, comprehension was better for the perceptual metaphor questions compared to the abstract metaphor questions (β=0.503, p=0.017) and there was a slight trend for bilingual children to perform better than monolingual children (β=-0.342, p=0.053; Figure 2). In this sample, we noticed that many children aged 2.5-3.5 years failed our attention and vocabulary test questions. In an exploratory analysis of children aged 3.5-4.5 years, bilingual children did outperform monolingual children (β=-0.46, p=.048). The effect of metaphor type on performance was more robust amongst these children as well (β=0.825, p = 0.007).

Experiment 2 (N=70 monolingual and 30 bilingual children aged 3.5-4.5 years, preregistered N=66 per language group) is an ongoing replication being conducted on Lookit. In contrast with Experiment 1, preliminary analyses suggest performance is equivalent for both types of metaphor questions (β =0.217, p=0.574) and for monolingual and bilingual children (β =-0.369, p=0.304). Taken together, these data provide new insight into children’s abilities to understand metaphorical language as well as how different language learning environments influence children’s ability to flexibly use familiar words in new contexts.

Authors

• Sahana Sridhar, University of Washington
  Presenting Author

• Taylor Petersen, University of Washington

• Ariel Starr, University of Washington