Teasing Apart Explanations of a Developmental Delay in Binding: Experimental Evidence from the Comparison of SLI and Williams Syndrome

abstract

This study investigates the knowledge of binding in 21 English-speaking children with SLI, aged 6;08–16;05, compared to 21 children with WS, language- and age-matched, and 21 language-matched control children, aged 4–7;10. Our results demonstrate no difficulties in the interpretation of reflexive or personal pronouns in SLI, revealing an intact knowledge of reflexive binding, as well as spared pragmatic functioning. Children with WS, however, show difficulties with their interpretation of pronouns, accepting the local binding of a pronoun, indicating impaired pragmatics. Not surprisingly, our language-matched typical controls, aged 4–7;08 (M = 5;03), showed a classic pattern of the Apparent Delay of Principle B Effect (ADPBE). In view of reported pragmatic but not memory deficits in WS, we interpret our results as consistent with the pragmatic deficit explanation but not the memory deficit explanation of the ADPBE.

Acknowledgments

We thank all our participants, their families and schools, and the Williams Syndrome Association. We also thank undergraduate and postgraduate students from the Wexler Language Lab at MIT and the UCL Speech and Language Sciences program for their help in collecting the data.

Funding

This research was supported by the Anne and Paul Marcus Family Foundation; the Brain Development and Disorders Project (BDDP) Postdoctoral Award (to A.P.); and the Brain Infrastructure Grant Program to the Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, from the Simons Initiative on Autism.

Notes

¹ See Leonard (2014) for a review of accounts of SLI focusing on specific deficits in linguistic knowledge (e.g., Rice & Wexler 1996) and others that focus on processing limitations, general (e.g., Kail 1994), or specific (e.g., Tallal & Piercy 1973).

² The standard assumption in the literature is that pragmatic difficulties are not a hallmark of SLI: language impaired children who show pragmatic difficulties are considered a separate population, in between SLI and autism. The terms often used in the literature are semantic-pragmatic syndrome (Rapin & Allen 1983), semantic-pragmatic disorder (Bishop & Rosenbloom 1987) or pragmatic language impairment (Bishop 2000). See, however, Bishop, Chan et al. (2000) for a discussion of a subgroup of children with SLI in their sample whose conversational responsiveness lagged behind their language skills.

³ In addition to NWR, studies show that children with SLI perform poorer on other measures of phonological short-term memory, such as digit recall (e.g., Briscoe & Rankin 2009) and word recall (e.g., Lum et al. 2012).

⁴ C-command: Node A in a phrase-marker c-commands node B if the lowest node that dominates A also dominates B.

⁵ Though early results relying on the original methodology of TVJ have recently been disputed (Conroy et al. 2009), the disparity between children’s comprehension of reflexives versus pronouns is a robust result reported across a range of languages (see Guasti 2002, for a comprehensive review), and with different methods (see Akker et al. 2012, for picture selection task).

⁶ The stimuli were separated into “match” or “mismatch,” where the required answers to the match question was “yes,” and the required answer to the mismatch question was “no.” Research shows that participants generally perform better on “match” than on “mismatch” conditions, thus we summarize van der Lely & Stollwerck’s results only on the more difficult “mismatch” experimental conditions.

⁷ Children with SLI were worse than control children on five out of eight conditions: Name–Pronoun match and Name−Pronoun mismatch; Quantifier−Reflexive match and Quantifier−Reflexive mismatch; as well as Quantifier−Pronoun mismatch.

⁸ Their stimuli involved no quantifiers but included an additional mismatch-only condition showing an incorrect agent performing the correct action.

⁹ Some researchers use the cutoff point of 85 for standard scores on tests of nonverbal reasoning (Tomblin et al. 1997) while others use the cutoff point of 80 (Bishop, Bright et al. 2000; Norbury, Bishop & Briscoe 2002). We follow the latter researchers in including the cutoff point of 80, based on findings of Tomblin & Zhang (1999), who show no difference in the patterns of language deficits in language-impaired children with nonverbal IQ of 85 and above, compared to those with an IQ of less than 85.

¹⁰ Seven children were excluded for not reaching the threshold on the test of nonverbal reasoning and 13 children had TROG-2 scores that were within the unimpaired range (SS between 81 and 106). Three children were excluded because they were bilingual: One of the children with an unimpaired TROG score was also bilingual, and two further children who otherwise fulfilled the criteria were bilingual.

¹¹ The table does not provide details of CELF scores, since only children with SLI were tested on this assessment as part of our inclusion criteria. In the selected sample, 10 children were tested by our research team on CELF and the remaining 11 were tested by SLT teams at the child’s school, who confirmed that the children met our inclusion criteria on this assessment.

¹² To establish the exact effects of age in the acquisition of binding in SLI and any the resolution of ADPBE, future studies may include a larger number of participants under at least the age 8.