The construction and use of a test of syntactic discrimination are described. Fifty children (10 each at 6, 8,10,12, and 14 years of age), selected for median scholastic class performance, were tested. The number of correct syntactic discriminations and response latencies were calculated for each of the four sentence forms: active affirmative, passive affirmative, active negative, and passive negative. Comprehension of the sentences improved with age, although even the youngest children identified forms like the passive affirmative at a better-than-chance level. Age-related improvements in test performance were nonlinear and varied with sentence form. Several factors influenced test performance: the semantic relations expressed by different verbs, surface syntactic features, and the degree of syntactic complexity. The effect of a given factor varied with both the age of the child and the performance measure. After age 6, increased skill was associated with the selective allocation of processing time to sentences on the basis of syntactic complexity.
