Early knowledge of small number word meanings in toddlers

• In Event: 3-226 - Number Knowledge Before, During, and After Children’s Mastery of Counting

Integrative Statement

By many accounts, learning the meanings of number words is a lengthy and difficult process (Fuson, 1988). According to prominent theories, children understand “one” around two years, understand “two” around three years, and higher numbers after age four (Wynn, 1992). Despite the protracted trajectory of number acquisition revealed by behavioral tasks, some indications suggest that children’s entry into number knowledge begins much earlier. Children as young as 18 months are sensitive to counting violations (Slaughter et al., 2011). In dual-marked languages like Slovene, children learn some number words much earlier (Almoammer et al, 2013). Diary studies suggest that young children’s production of the word “two” is generally correct (e.g., Mix, 2002). These findings raise the possibility that younger children know more about number word meanings than previously thought, and that conventional measures, like Give-a-Number, underestimate children’s knowledge. Here, we report on behaviorally simpler implicit looking methods to investigate early number knowledge.

The first line of studies used a preferential looking (PL) method to probe early number knowledge in 14-30 month-olds. We hypothesized that if children understood the words “one”, “two,” or “three”, they would look longer to the matching set with that number of objects, when the paired visual displays contained one, two, or three objects (different set sizes in each pair). In three studies (combined N> 400), we have found no indication that children looked to the matching visual target. Preliminary analyses suggest that children’s preferential looking behavior does not reveal earlier knowledge than Give-N. As a group, only children who pass Give-N for a set size also show increased looking to that set size when labeled.

A conceptually related investigation was designed using violation-of-expectancy (VOE) instead of preferential looking. 18-22 month-olds participated in an initial study (N=16). On each trial, a curtain was raised to reveal a live experimenter and an occluder. The experimenter labeled the numerosity (“Look! Two sheep!”) before revealing the display. The label matched the number of objects on half the trials and mismatched on the other half. We hypothesized that if children understood the labeled word, they would look longer to the mismatch. A repeated-measures ANOVA on looking time revealed a significant interaction between Label and Set Size (p=.036), due to significantly longer looking to three objects than two objects after hearing the label “two” (p=.007; Figure 1). A pre-registered replication is ongoing, using a tablet-based presentation similar to the PL studies but with a VOE paradigm. Given the differing results from the two methods, it is of interest to determine which methods reveal, or failing to reveal, earlier number knowledge.

The VOE results suggest that children as young as 18-22 months show some comprehension of “two” and possibly other small number words, potentially calling into question canonical interpretations of children’s entry into number word meanings. Together, the results from these studies provide an opportunity to refine theories about the nature of children’s initial entry into number word knowledge.

Authors

• Anna Shusterman, Wesleyan University
  Presenting Author

• Pierina Cheung, Nanyang Technological University, National Institute of Education
  Non-Presenting Author

• Katie Wagner, UCSD
  Non-Presenting Author

• Sona Kumar, Boston University
  Non-Presenting Author

• David Barner, University of California, San Diego
  Non-Presenting Author