What do they say and how? A comparative syntactic characterization of children with and without Developmental Language Disorder in European Portuguese.

Syntactic difficulties are widely reported in children with Developmental Language Disorder (DLD), yet detailed evidence from European Portuguese (EP) remains scarce. Language Sample Analysis (LSA) offers an ecologically valid approach to examining children's syntactic production; however, its application requires language-specific data and validated, multidimensional measures. This study aimed to provide a detailed syntactic comparison between children with DLD with syntactic impairment and typically developing (TD) peers acquiring EP, using a multidimensional LSA framework complemented by standardized assessments. Thirty-two children (60.22 ± 8.04 months) with DLD were individually matched to 32 TD peers (60.50 ± 5.96 months) on age, sex, and socioeconomic status. All participants produced picture-based narratives, which were segmented into C-units and coded for syntactic productivity, complexity, diversity, accuracy, and the production of specific structures. Group differences and associations among LSA measures and between standardized assessments were analyzed. Across all dimensions, children in the DLD group produced significantly shorter utterances (t(62) = 6.91, p < .001, d = 1.73), fewer multi-argument predicates, lower syntactic diversity, and less subordination than their TD peers. Accuracy differences were large (PGU: TD.96 vs. DLD .62), with error patterns dominated by omissions and structural errors. Several syntactic constructions that emerged in TD narratives were not observed in the narratives produced by children in the DLD group. LSA measures showed moderate-to-strong correlations with standardized assessment scores, supporting their convergent validity. These findings provide a detailed characterization of syntactic production in EP-speaking children with DLD with syntactic impairment, and underscore the value of multidimensional LSA for identifying clinically relevant syntactic vulnerabilities that may not be fully captured by standardized assessments.