Poster #30 - Domain-specific and Domain-general Lexical Ambiguity and Their Relationship with Early Math

• In Event: 2-163 - Poster Session 08
  In Poster Session: PS 08 Section - Cognitive Processes

Integrative Statement

Number words provide unique challenges to children (e.g. 2 eggs, 2 dozen eggs, 2 o’clock all have the number word “two” but refer to different quantities). Moreover, an abundance of literature suggests situational and linguistic contexts affect the interpretation of word meanings. We hypothesize that an individual’s sensitivity to these contexts may be informative to individual differences in how children interpret number words and their developing number concepts.
Number words themselves may be ambiguous in a way that is specific to number concepts (domain-specific lexical ambiguity). Alternatively, number word interpretation may function like other polysemous words rather than be influenced specifically by numerical properties per se (domain-general lexical ambiguity). The purpose of the present study was to evaluate performance across measures of lexical ambiguity that conform to either domain-specific or domain-general “contextual sensitivity” constructs. We hypothesized that domain-specific sensitivity to lexical ambiguity would be related to early math ability, but that domain-general sensitivity to lexical ambiguity may at least partially mediate this relationship.
Children between K and 2nd grade were recruited from public schools in a large city in the Midwestern United States. The present study reports on the results of 162 children who completed the study to date. Domain-specific lexical ambiguity was assessed with the investigator-designed Numerical Ambiguity Interpretation Test (NAIT), a forced choice task set in a story context. Children were taught during a warm up that “I don’t know” was an acceptable and expected response on some trials. Only children that correctly responded “I don’t know” at least once during the warm up were included in these analyses (n=136). Similarly, domain-general lexical ambiguity was assessed with the investigator-designed Homonym Interpretation Test (HIT), a six alternative forced choice task also presented in a story context. Table 1 contains examples of items from these measures. General mathematics ability was assessed with a standardized assessment, Test of Early Math Ability (TEMA).
First, NAIT and HIT were examined separately for their relation with math ability. Previous work suggests that having the quantities physically visible inoculates against a large number word bias compared to when one quantity is not visible. The current study results (see Table 2) suggest that only when one quantity was not visible was performance correlated with math ability, and this is only true for 1st and 2nd grades. Similarly, HIT performance was only significantly correlated with math ability at 2nd grade. Thus, 2nd grade is the only grade where the mediation model becomes a viable explanation of child responses. However, NAIT invisible quantity items (panels C and D in Table 1) and HIT pseudo-homonyms are uncorrelated with each other in 2nd grade, r = 0.02 and r = 0.14.
Together, the results of the current study point to domain-specific and domain-general lexical ambiguity skills being perhaps more robust predictors of mathematics skills in 2nd grade as compared to Kindergarten children. Moreover, sensitivity within domain-specific and domain-general contexts were distinct and act as separable independent correlates of early math performance.

Authors

• Sarah Lukowski, University of Minnesota
  Presenting Author

• Michele Mazzocco, University of Minnesota
  Non-Presenting Author

• Rachel E Olson, University of Minnesota
  Non-Presenting Author