Speeded naming, frequency and the development of the lexicon in Williams syndrome

Abstract

Atypical vocabulary has been reported as one of the most notable features of the language of adolescents and adults with Williams syndrome (WS), including use of unusual or low frequency words. Two hypotheses were identified regarding the developmental origins of this phenomenon. The intra-lexicon hypothesis views the cause in terms of domain-specific anomalies of structure or activation dynamics in the WS lexicon. The extra-lexicon hypothesis views the cause in terms of pragmatic influences, whereby individuals with WS use social engagement devices in their language in service of their “hyper-social” profile (Jones et al., 2000), and domain-general deficits to lexical-semantic representations in line with the level of learning disability present in WS. The hypotheses were evaluated using a speeded picture-naming task, in which frequency and semantic category were manipulated as implicit variables. The performance of 16 adolescents and adults with WS was compared with two individually matched control groups, one matched on chronological age (CA) and the other on receptive vocabulary age (RVA). Developmental trajectories were also constructed to assess the relationship between performance and age. Results indicated slower and less accurate naming in the WS group compared with both CA and RVA groups, but frequency and semantic category effects in line with the RVA group. Trajectories were delayed but not atypical. The data support a normal encoding of word frequency (rarity) in the WS lexicon, with rare-word usage explained as a pragmatic aspect of the WS social profile. The finding sets bounds on the domain-specificity of the complex developmental pathway between atypical genotype and atypical phenotype in WS.

Acknowledgment

This research was supported by MRC Programme Grant No. G9715642 to Annette Karmiloff-Smith and MRC Career Establishment Grant G0300188 to Michael Thomas. We would like to thank the Williams Syndrome Foundation UK. We would also like to thank John Brock and an anonymous reviewer for their helpful comments on a previous version of this article.

Notes

¹ In a pilot study of our own, we contrasted the performance of a group of children between 5 and 12 years of age who had been clinically and genetically diagnosed with WS (N=15, mean age 7;5, SD 1;8, range 5;5–11;6) on the British Picture Vocabulary Scale (Dunn, Dunn, Whetton, & Pintilie, 1982) and the naming sub-test of the BAS II (Elliott, Smith & McCulloch, 1996). One participant with WS was at floor on the BPVS and so was excluded from the analysis. Of the remainder, test ages on receptive vocabulary and naming ability were significantly correlated: linear regression: R ²=.547, F(1, 13) = 14.5, p=.003, with no significant disparity in the test ages (regression intercept not significantly different from zero, p>.5), and a gradient not significantly different from 1 (naming test age increased by 1.16 months for each month of increase in receptive vocabulary test age, with lower and upper 95% confidence intervals of 0.49 and 1.82 months respectively).

² That is, in those individuals who did not have fear of needles.

³ This is equivalent to a repeated-measures linear regression design. In contrast to a repeated measures ANOVA, in the ANCOVA, residuals are calculated from each data point to a regression line rather than to a mean.

⁴ Note that total error counts included all participants, while error proportions included only those participants who made errors on Object naming or Action naming. Hence the degrees of freedom can differ for the two.

Additional information

Notes on contributors

Michael S. C. Thomas

Correspondence should be addressed to Dr Michael Thomas, School of Psychology, Birkbeck College, University of London, Malet Street, Bloomsbury, London WC1E 7HX, UK. [email protected]