Semantic and Pragmatic Abilities Can Be Spared in Italian Children with SLI

ABSTRACT

Specific language impairment (SLI) is a heterogeneous disorder affecting various aspects of language. While most studies have investigated impairments in the domain of syntax and morphosyntax, little is known about compositional semantics and the process of deriving pragmatic meanings in SLI. We selected a group of sixteen monolingual Italian-speaking children with SLI (mean-age 7;4) with a severe morphosyntactic deficit at the receptive and expressive level. We tested their comprehension of quantified sentences including all and some in order to establish whether they were competent with the logical and pragmatic meanings of these quantifiers. Children performed as well as their typically developing controls in understanding logical meanings. In comprehending pragmatic meanings, they obtained lower scores than age-matched controls but they were not different from language-matched children. However, differences in this ability correlated positively with age and with the ability to understand simple sentences in the SLI group. This suggests that aspects of the syntactic component might be involved in the development of this ability and that, despite their severe morphosyntactic deficits, children with SLI might catch up with their peers in deriving pragmatic meanings.

Notes

¹ Informativity is a crucial notion in the process of SI derivation. Consider the sentence (1’), in which some appears in the antecedent of a conditional sentence; in this configuration, entailment relations are reversed: (1’) If some of the apples are in the box, then we have something to eat.In this case, no SI is derived, provided that (1’) does not certainly mean that “if some but not all the apples are in the box we have something to eat” (cf. Chierchia, 2013 for a discussion).

² Interestingly, although children scored at least 2 SDs below the mean score for their age on the TCGB, their raw scores on the test negatively correlated with their age (r = −0.545, p = 0.029), indicating that the number of errors in comprehension decreases with age (the raw score for the TCGB indicates the number of errors). An analysis of the types of errors showed that while errors occurred across different sentence types for younger subjects, older subjects mainly produced errors with relative clauses and passive constructions. This suggests that the comprehension of syntactically simple sentences not involving movements might still be under development and that the comprehension of syntactically complex sentences, requiring sophisticated morphosyntactic abilities, might not.

³ Correlations are calculated on raw scores and not on Z-scores (or SDs, standardized scores, percentiles). Z-scores express deviations from mean scores of children of the same age. In children with different ages a Z-score might be associated with different raw scores and might express different levels of grammar proficiency.