Verbatim memory and gist extraction in elementary school children with impaired language skills.

Two experiments were conducted to determine which of three proposals (memory limitations, inferential reasoning deficits, or task-related interference) best accounts for gist-extraction deficits observed in children with impaired language skills (LI group) relative to age-matched children developing language normally (NL group). Sixteen children between the ages of 8;0 and 10; 11 (years;months) from each group participated in each experiment. Experiment 1 utilized a comprehension paradigm (instructions focused on meaning) to investigate children's recognition of true and false premises, paraphrases, and inferences in short passages. Experiment 2 utilized a sentence-verification paradigm (instructions focused on verbatim form) to investigate children's ability to recognize presented premises in these same passages, and differentiate them from nonpresented items. In both experiments, passages were read aloud to each child, half of the passages were accompanied by pictures, and immediate and delayed recognition testing was conducted. In Experiment 1, previously documented inference-recognition deficits were associated with the auditory-only presentation of passage information. Gistextraction deficits involving inaccurate recognition of true paraphrases and acceptance of false foils were also observed. However, when pictures accompanied passages, the LI group's overall recognition of gist improved significantly relative to the auditory-only condition, and between-group differences were eliminated. Forgetting rates did not differ significantly as a function of language group in either the picture or no-picture condition. In Experiment 2, there were no significant between-group differences in the recognition of presented premises, with both groups erroneously recognizing nonpresented true sentences more frequently than false sentences (gist intrusions). Nevertheless, the LI group was significantly less likely than the NL group to differentiate presented premises from nonpresented items. Findings from both experiments indicated that gistextraction deficits observed in LI groups were not associated with either memory limitations or deficits in inferential reasoning. Instead, in line with fuzzy-trace theory (e.g., Brainerd & Reyna, 1990), difficulties these children had in extracting meaning from linguistic input presented auditorily appeared to interfere with (i.e., degrade) the simultaneous processing of verbatim input associated with presented premises. Comparable interference effects were not observed for children developing language normally