The Linguistic Ability of Semantic Integration: Evidence from Proposition Entailment in Chinese Williams Syndrome

This study investigates the hypothesis of selective impairment in meaning relative to form in the language processing of individuals with Williams Syndrome. It has been known that individuals with WS have spared grammatical knowledge even with mental retardation (average IQ of 55 or below) and poor cognition. Past research also has shown that individuals with WS have preserved normal verbal working memory and such intact verbal ability has been thought to be responsible for their relatively good language performance (Wang & Bellugi, 1994; Jerrold, Baddely, & Hewes, 1999; Vicari, Brizzolara, Carlesimo, Pezzini, & Volterra, 1996; Vicari, Carlesimo, Brizzolara, and Pezzini, 1996; Karmiloff-Smith, Grant, Berthoud, Davies, Howlin, & Udwin, 1997; Robinson, Mervis, & Robinson, 2003; Laing, E., Grant, J., Thomas, M. S .C. & Karmiloff-Smith, A., in press). With a good verbal memory but a deficit in cognitive ability, individuals with WS are hypothesized to rely heavily on verbal working memory in learning language. This may explain the finding that the grammatical knowledge of WS individuals is strong while their semantic understanding might be weak (Zukowski, 2001; Grant, Valian, and Karmiloff-Smith, 2002). In other words, these individuals might have selective dissociation of form from meaning in linguistic ability. In order to examine this issue, a Bransford & Franks’ recognition paradigm (1972) was employed. Participants were trained to implicitly learn a series of sentences which were composed of different events from various superset sentences, and later to make recognition judgments about each sentence as to whether it had been heard before. They were asked to assign a recognition confidence rating value for each judgment. Two sets of experiments were conducted. The first set involved experiments in which recognition score and confidence ratings of new and old sentences were compared across the unimpaired and individuals with WS (Experiments I & II). According to form and meaning selective impairment hypothesis, it was expected that individuals with WS would show a high rejection rate for all new sentences and high hit rate on all old sentences. On the other hand, the unimpaired would show a high false positive rate to all new sentences and high hit rate to all old sentences. The results showed that individuals with WS performed similarly to the unimpaired, i.e. their chronological age-matched normal controls. In the second set of experiments, new and scrambled sentences were compared (Experiments III & IV). It was expected that individuals with WS would have a high rejection rate for all new sentences and scrambled sentences, whereas the unimpaired would show a high false positive rate to all new sentences and high rejection rate for scrambled sentences. Again, the results showed that individuals with WS performed similarly to the unimpaired. In conclusion, individuals with WS showed spontaneous proposition integration in semantics like the unimpaired. 1. Memory Ability of Individuals with Williams Syndrome Wang and Bellugi (1994) first demonstrated the dissociation between verbal and visual-spatial short-term memory on individuals with Williams syndrome. They used digit span as the test for verbal short-term memory and corsi block as a measure for visual-spatial short-term memory. They demonstrated a double dissociation on short term storage for phonological and for visual-spatial information compared with individuals with Down syndrome. Jarrold, Baddeley, and Hewes, (1999) reexamined this finding with more careful control groups. In Wang and Bellugi, they matched full IQ scores of WS and DS individuals. However, Jarrold et al. argued that this match held a confound. The dissociation, they argued, might result from the deficient verbal ability of DS individuals and the impaired visual-spatial ability of WS individuals. Thus, Jarrold et al. took both verbal and non-verbal IQ on these two genetic groups and covaried out the effect of any differences in these measures. Meanwhile, they recruited moderate learning disability individuals as a control group. Their results replicated those of Wang and Bellugi. Individuals with WS performed better on digit span tasks than on spatial tasks. On the contrary, individuals with DS showed the reverse pattern. Vicari, Brizzolara, Carlesimo, Pezzini, & Volterra (1996) also confirmed this dissociation. They further pursued the issue to see if there is a selective impairment within verbal working memory. They used immediate recall tasks testing individuals with WS and normal mental-age matched controls. They reasoned that if the phonological competence of individuals with WS is intact while semantic competence is deficient, then spared short-term and impaired long-term systems in working memory is to be expected. The results confirmed their predictions. When compared with normal controls, individuals with WS showed a significant difference in recall rate for primacy effect items, but no difference in recency effect when compared with normal controls. These results indicate a clear dissociation between