Memory for language organization in concrete and abstract sentences

Subjects were presented with concrete and abstract sentences equal in lexical complexity and comprehension rating and tested on their ability (1) to recognize the words used in the sentences when presented at a 3-sec rate and (2) to recreate the initial sentences when presented with a list of the words used. Only on this second task was there a significant concrete/abstract difference. In a second experiment, there was a significant concrete/abstract difference in word recognition when a 6-sec decoding rate was used. It was suggested that the best explanation of these results could be offered by a semantic analog model of language memory.