Browse By Day
Browse By Time
Browse By Panel
Browse By Session Type
Browse By Topic Area
Infants as young as 17-months old, and perhaps younger, demonstrate a tendency to map novel words to novel objects. What are the strategies that motivate this mapping process, and do these strategies actually help children learn new words? Previous research suggests that older children and adults may utilize a logical strategy, disjunctive syllogism, to correctly infer which object is the referent of a novel word. Here, we explore the developmental trajectory of this rule and children’s ability to learn new words using this strategy.
Typically, studies have used looking time methods to understand novel word-to-object mapping. We expanded upon these methods by verifying each child’s known vocabulary prior to participation. Parents were asked to fill out a computerized version of the MacArthur CDI before arriving at the lab, and an algorithm within our assessment program generated a unique game for each child using only the reported known words in that child’s vocabulary. In addition, each known word was assigned to a category of “familiarity” using previously-reported percentages (from Wordbank; Frank, 2006) of children at each participant’s age who also knew these words. Our categories consist of Typically-Known-Words, Often-Known-Words, and Rarely-Known-Words. We used these categories as an estimate of how confident a child may be in knowing a specific label. Presentation of words from each category was balanced within the game for each child.
Thirty-nine children (21-40 months) to date have visited the lab. In our computer game, children saw a series of paired images on two different computer screens. The images could either be two known objects, or one known and one novel object. For both Known and Novel trials, the computer program displayed known images from each of the familiarity categories, ensuring a variety that could support later analyses.
On each trial, the child was asked to look at one of the images by being told either a known word (“look at the ball”) or a novel word (“look at the dax”). Children saw a variable number of trials, depending on the size and diversity of their reported known vocabulary. Shortly after the computer game, the child was presented with two of the novel objects previously seen in the game as a test of novel label learning. In another room, they were asked to point at one of the objects by name (“Do you remember which one is the dax? Can you point at the dax?”).
Eye-tracking results show that both younger and older children increased their looking to the correct object upon hearing the target word on both Known and Novel trials (Figure 1), verifying successful novel word-to-object mapping. Post-test responses indicate that mostly older children remembered the correct object for a new word.
These results suggest that children are succeeding in mapping novel words to novel objects, and that this helps them learn new words. We will continue to run a wide age range, investigating each child’s looking pattern within each trial type and perform analyses based on known word familiarity and parental report vocabulary size.