Poster #53 - Experience Comparing Similar Events Helps Japanese- and English-speaking Children Learn Verbs

• In Event: Poster Session 18

Abstract

Learning verbs is critical to learning one’s native language. Some studies suggest young children have difficulty learning new verbs (Imai et al., 2005; 2008), but these included only a single learning event and test trial type. Two new studies add additional comparison trials and an additional test trial type; children in both language groups demonstrate better verb learning than seen in prior studies.

In two studies, children were shown events linked to four novel verbs. Half were assigned to a Similar First condition (two events with similar objects were shown before an event with varied objects) or an All Varied condition (all had differing objects). High similarity comparisons are predicted to be helpful in both structure mapping (e.g., Gentner & Namy, 2006) and statistical learning theories (e.g., Yu & Smith, 2007). After watching three video events on a screen, they saw two test trials and were asked to point to one of two scenes to extend the verb. One test trial was predicted to be easier because it showed all new objects in both scenes. A second test trial was predicted to be more difficult because the distractor event showed the same object from learning trials with a new action, which can be hard for children to ignore.

In Study 1, 2½-, 3½- and 4½-year-old Japanese-speaking children (total n=170; n= 43 2.5s; n= 48 3.5s, n= 48 4.5s) saw these events presented by a Japanese researcher in their preschools in Japan. A multivariate ANOVA with Age (3: 2, 3, 4) and Condition (2: similar first, all varied) as between-subjects factors, and test trial (prop.easy, prop.difficult) as a within-subjects factor, shows a main effect of Age (p< .001) and Age x Condition interaction (p< .005). 3.5-year-old children performed well in the Similar First condition across both test trial types, and struggled in the All Varied condition, suggesting experience seeing similar events was useful for verb learning (see Fig. 1).

In Study 2, English-speaking children in the US (total n=93 to date; n= 25 2.5s; n= 31 3.5s; n= 37 4.5s) were tested on the same stimuli. The same analysis applied here reveals a main effect of Age, Test trial, and a Test trial x Condition interaction (ps< .03). As in Study 1, by 3.5 years, they were able to perform well in both test trials only in the Similar First condition (ps< .005) (see Fig. 2). As in Study 1, their performance increased with age.

Thus, results in both languages show children benefit from seeing events they can compare, and both show that events with high similarity early in learning can facilitate verb learning. Thus, these studies show an important mechanism for verb learning that may be useful across languages. Links to key theories in verb learning, including structure mapping and statistical learning, will be discussed.

Authors

• Jane B. Childers, Trinity University - San Antonio
  Presenting Author

• Masata Ohba, Keio University

• Akari Notsu, Keio University

• Marissa Young, Trinity University

• Leah Marsh, Trinity University - San Antonio

• Mutsumi Imai, Keio University