Date of Award
Master of Arts (MA)
Kenneth A. Deffenbacher
This study addressed two issues in sentence memory: Hie accuracy of Gestalt representations of sentences over strictly associative accounts, and the role of syntactic, semantic, and thematic variables in characterizing the organizational basis of sentence representations. Sixteen pairs of SVO sentences were generated with a single noun shared by both member sentences of each pair. After a study period during which subjects wrote expansions of stimulus sentences, memory for the shared nouns was tested using subject-verb or object-verb cues from one or both member sentences of each pair. According to associative accounts of sentence memory, mixed cues using one component each from both members of a sentence pair should have been more effective prompts for recall of target nouns than original cues formed from only one of the sentences in a pair. In contrast to the associative predictions the original cues were more likely to prompt meaning-preserving recall of the target nouns, suggesting that theories of sentence memory should incorporate Gestalt representations. In the same experiment, the semantic, syntactic, and thematic role of the sentence components was factorially varied such that each target noun was either a case-grammar agent or object, either subject or predicate, and either given information or new information. Thediffering sentential roles of target nouns produced no reliable variations in cued recall. Potential explanations for the absence of the expected results with regard to the second issue are offered.
Brown, Tracy L., "Associative, Semantic, 'Thematic, and Syntactic Factors in the Memory Representations of Sentences" (1980). Student Work. 119.