Poster #114 - Preschoolers Flexibly Adapt Their Gaze to Seek Information That Supports Language Comprehension and Word Learning

• In Event: 1-061 - Poster Session 02
  In Poster Session: PS 02 Section - Language, Communication

Integrative Statement

Children often experience language in grounded, social interactions that provide access to multiple sources of information. But children cannot process all the available information, creating a scenario where they must decide what information to gather and at what time. In this work, we propose that children are remarkably flexible in how they allocate gaze to seek visual information that supports language-related tasks. We present evidence for our account in two domain: familiar language comprehension (E1) and novel word learning (E2).

Experiment 1 tested the hypothesis that listeners adapt their gaze to seek visual information when the auditory signal becomes less reliable. We measured the timing and accuracy of adults (n=31) and 3-5-year-old children's (n=39) eye movements during a real-time, familiar language comprehension task. Participants processed familiar sentences ("Where's the ball?") while looking at a simplified visual world with three fixation targets (Fig. 1). We manipulated the quality of the auditory signal by convolving the acoustic input with different levels of brown noise.

We analyzed Accuracy and Reaction Time (RT) of children’s first shifts from the center stimulus after the target noun. Both age groups delayed the timing of gaze shifts away from the speaker's face when processing noisy speech. This delay resulted in listeners accumulating more information from the visual signal, more accurate gaze shifts, and fewer random eye movements to the rest of the visual world (Fig. 2A). These results provide evidence that young listeners adjusted to the demands of a more challenging processing context by seeking visual information to support language understanding.

Experiment 2 (data collection is ongoing) asks whether children flexibly adapt their gaze to increase fixations to a speaker when the speaker is more likely to provide a referential cue. We measured adults (n=10) and 3-5-year-old children's (n=34) eye movements while they processed sentences containing a novel word (“dax”) and looked at a visual world with three fixation targets (Fig. 1). Participants saw a sequence of ambiguous word-learning events organized into pairs of exposure and test trials with four exposure/test pairs for each of four novel words. We manipulated the value of looks to the speaker by varying whether she provided a post-nominal gaze cue. In the Gaze condition, the speaker looked at one object after producing the label; whereas, in the Straight-ahead condition, the speaker looked straight ahead throughout the trial.

Preliminary analyses show that the presence of a gaze cue caused children and adults to allocate more fixations to the speaker, which in turn led to a higher proportion looking to the target object on exposure and test trials (Fig. 2B). This result provides preliminary support for our hypothesis that children will seek additional visual information from speakers who provide useful gaze cues, and this behavior supports the uptake of novel word-object pairings. Taken together, these two studies suggest that children are sensitive to changes in the value of fixating on a social partner for solving a diverse set of language tasks.

Authors

• Kyle MacDonald, Stanford University
  Presenting Author

• Virginia A. Marchman, Stanford University
  Non-Presenting Author

• Michael C. Frank, Stanford University
  Non-Presenting Author