Categories of Word Comprehension in Toddlers with ASD or Typical Development: An Extension of Beckage, Smith, and Hills (2011)

Poster Presentation
Thursday, May 10, 2018: 11:30 AM-1:30 PM
Hall Grote Zaal (de Doelen ICC Rotterdam)
S. Kover1, D. A. Fein2 and L. R. Naigles2, (1)University of Washington, Seattle, WA, (2)Psychological Sciences, University of Connecticut, Storrs, CT

In concert with social and environmental influences, vocabulary development and outcomes hinge on the contents of a child’s lexicon—what has already been learned influences what is learned next (Perry & Samuelson, 2011; Samuelson & Smith, 1999; Stokes, 2010). In particular, semantic relatedness of children’s lexicons predicts learning, with delayed learners showing less relatedness among words (Beckage, Smith, & Hills, 2010). In typical development, a lure-of-the-associates model suggests that new words will be semantically related to previously acquired words (Hills et al., 2010). Beckage et al. (2011) proposed that, in contrast to that pattern, delayed learners preferentially acquire “oddballs”: words not semantically related to known words. We extend this hypothesis to ASD, in comparison to typically developing (TD) children matched on receptive vocabulary.


We characterize the semantic relatedness of vocabulary items acquired by children with ASD or TD with two questions: (1) Despite similar vocabulary size, do toddlers with ASD or TD acquire words with different category patterns? and (2) Are toddlers with ASD more likely than TD toddlers to understand “oddballs” (i.e., words isolated within categories)?


Toddlers with ASD (n=26; M age=32 months) and TD toddlers (n=26; M age=19 months) were matched pairwise on parent-reported receptive vocabulary (MacArthur-Bates Communicative Development Inventories [CDI]-Words&Gestures) within a margin of 10 words: M=111, SD=72/75, respectively. Groups did not differ on number of words understood, p=.991.

Words understood were tallied according to the 18 CDI categories (e.g., Clothing, Furniture) and calculated as proportions of children’s receptive vocabularies. Extending Beckage et al., oddballs were identified as isolated words understood within a category in which at least one word was understood: 1-2 words understood were oddballs; more than 2 words within a category were considered associative. For each category, we tested: (1) group differences in counts and proportions of words understood and (2) group differences in proportions of children whose words in a given category were oddballs.


(1) Relative to TD toddlers, toddlers with ASD understood fewer pronouns, t(50)=2.21, p=.031. Using proportions, toddlers with ASD produced more sound effects, but lower proportions of time (e.g., day), description (e.g., big), and pronoun words, ts(50)>2.0, ps<.050. See Figure 1.

(2) The proportion of children with oddball words differed between groups for two categories. Fewer children with ASD (2 of 24) than TD (10 of 24) had oddball clothing words, p=.017. More children with ASD (7 of 22) than TD (1 of 23) had oddball description words, p=.022. See Figure 2.


The findings point to many similarities, but also some differences in the relatedness of words in the lexicons of children with ASD and TD. Further semantic differences may emerge over developmental time: young children with ASD show typical semantic processing (Ellis Weismer et al., 2016; Rescorla, 2013), yet semantics is an area of weakness later in childhood (Ellawadi et al., 2016; Norbury et al., 2010). The current study has implications for theories of lexical acquisition in ASD, including whether words with certain functions (e.g., adjectives), rather than semantic properties, are vulnerable to delay.