The roles of expectation and syntactic knowledge in second and fourth graders' comprehension of complex sentences.

Abstract

Second and fourth graders were tested, using a picture selection test, for comprehension of complex sentences. Pictures were withheld from subjects until after they showed that they understood the input sentence. Sentences were clefts and pseudoclefts with relative clauses minus markers and auxiliaries. Sentences were either passive or active and their noun-verb-noun relations were either according to children's expectations or contrary to expectations. Fourth graders performed significantly better than second graders; actives were significantly easier than passives; and sentences that were according to expectation (typicals) were significantly easier than sentences that were contrary to expectation (atypicals). There was no significant difference in performances on cleft and pseudocleft sentences. Comparisons were made that investigated the interplay of the two determinants of difficulty (sentence voice and typicalness of NVN relations). Typical passives were not significantly easier than atypical passives; nor were typical actives significantly easier than atypical actives. In addition, active atypicals were not significantly easier than passive atypicals; nor were active typicals significantly easier than passive typicals. However, one comparison of combinations of syntactic and expectancy related factors was significant: that which compared passive atypicals with active typicals. When both the syntactic and the expectancy-related determinants of difficulty are in phase, there is a significant difference in second and fourth graders' understanding of cleft and pseudocleft sentences with relative clauses minus markers and auxiliaries. It was concluded that the interplay of both syntactic and non-syntactic factors must be considered in any characterization of children's sentence comprehension ability.