2018 Sat Poster 6454

Saturday, November 3, 2018 | Poster Session II, Metcalf Small | 3:15pm

Toddlers’ use of a third party’s gaze information in verb-action mapping
Y. Oshima-Takane, K. Boyle

Previous research has shown that children are able to use a speaker’s gaze and linguistic cues to learn novel verbs by 33 to 36 months of age (e.g., Brandon et al., 2007). However, it is not clear if children know that the third party’s eye gaze also provides important cues for the speaker’s intended referent of words. The present study investigated the age at which children begin to use gaze information from a third-party onlooker to learn novel verbs.

Thirty-two 26-month-old and 28 30-month-old French-speaking children were taught novel verbs using a modified version of the habituation method with a switch design. They were assigned to one of two conditions: Social and Non-social. Each child watched a video consisting of four trials: familiarization, control, teaching, and verb test (see Fig. 1). In the familiarization trial, two video clips were presented three times each in alternating order. Each clip showed a visual scene of an agent performing a causative action on an object (e.g., a stick being slid in and out of a cube toy) while a transitive verb sentence containing a novel verb (e.g., “Je dax le jouet”) was heard. The familiarization trial taught children verb-event pairings (e.g., ‘dax’- sliding event and ‘vop’- lifting event) using morpho-syntactic cues alone. In the control 1 trial, the same objects presented during the familiarization trial were presented on the left and right sides of the screen, but the action performed on each object was the action originally paired with the other object during the familiarization trial. The control 2 trial was identical to the control 1 with the exception of a third-party female looking straight ahead in the center of the screen. The control 1 trial determined whether one of the objects, actions, or sides was more salient or preferred than the other. The control 2 trial familiarized children to the onlooker’s face before the teaching trial. During the teaching trial, the children in the Social condition watched a video clip of the onlooker looking at the screen with the correct referent, whereas those in the Non-social condition watched a video clip of the onlooker closing her eyes while hearing the audio of one of the transitive verb sentences used in the familiarization trail. The looking times during the verb test trial was measured to determine how children interpreted the novel verbs. Children in the Social condition were expected to look at the correct referent in the verb test trial significantly longer than in the control 1 trial if they used the onlooker’s gaze information.

The results revealed that only the 30-month-old children in the Social condition looked at the correct referent significantly longer than the control trial 1 (t(13)= -2.997, p=.005, one-tailed; see Fig. 2). This finding suggests that not only do children know that an adult who is not directly involved in a conversation tends to look at the intended referent of the words produced by the speaker, but also use this information to learn novel verbs by 30 months of age.