Reliability of retrieval from semantic memory: Noun meanings

The purpose of this experiment was to determine how reliably college students could retrieve from memory semantic information about commonly used nouns. In two recall sessions separated by 1 week, subjects gave definitions for abstract nouns, category labels, and concrete nouns. It was found that more propositions were given in definitions of concrete nouns (7.13) than in definitions of category labels (6.73) or abstract nouns (4.13). The within-subject reliability of retrieval was.55 for concrete nouns,.46 for category labels, and.43 for abstract nouns. The between-subjects reliability was.29 for concrete nouns,.20 for category labels, and.17 for abstract nouns. Discussed are the implications of these data for the hypothesis that the single word represents the unit of meaning.