DEMOS

Spiegel, C., Halberda, J. (submitted). Fast-mapping abilities in two-year-olds: referent-selection and referent-retention of multiple novel labels.

Abstract

We investigated 2-year old’s (N=72) ability to infer the referent of a novel noun from a single ambiguous exposure, and their ability to retain this mapping over time. On critical trials, images of a known and a novel object appeared on two computer screens and a voice asked the child to “point at the [e.g., glark].” Following label onset, children had only 3 seconds in which to infer the correct referent, point at it, and store this new word-object mapping. Children saw 6 novel label trials and 22 known label trials during a 10-minute learning phase. Each novel object appeared only once during the study and no feedback was given. In a final post-test trial, all 6 previously labeled novel objects appeared and children were asked to identify one (e.g., Can you find the “glark”?). To succeed, children needed to initially map the novel labels correctly and retain these mappings over the course of the study. Overall, children successfully identified the target object on this retention trial. We conclude that 2-year-olds can fast-map a novel noun from a single 3-second ambiguous exposure and can retain multiple mappings over the course of a 10-minute experiment.

Methods

Children viewed 24 ambiguous labeling trials in which objects were presented side by side on separate monitors and a voice asked children to "Point at the [e.g. brush]". Over the course of the 5 min. study, children saw 36 familiar objects and 12 novel objects, 6 of which were given a novel label. Children had 3 seconds to identify the target, point to it, and store it in memory before the next trial began. After viewing 24 ambiguous labeling trials, children saw a single post-test trial in which we tested retention of a randomly chosen novel label (label asked for was counterbalanced across conditions). During this post-test, children saw all 6 of the previously labeled novel objects and were asked to identify one of them.

Methods Demo: Double-click the movie on the right to watch the demo
Results: Percent of children correctly choosing the novel target object at retention (chance=16.67%). The x-axis indicates when the novel object appeared during referent-selection (e.g., position 1 is early in the study, position 6 is late). Overall, children successfully remembered the novel labels.