Priming macropropositions: Evidence for the primacy of macropropositions in the memory for text

Subjects read paragraphs which contained explicit summarizing statements expressing a macroproposition. They were then given a word recognition test. When two words from the summarizing statement were presented on successive trials on the recognition test, responses to the second word were faster and more likely to be correct than for word pairs from a sentence from the microstructure of the text, which in turn were faster (but with equal error rates) than responses to test words from different sentences. When the summarizing statements were omitted from the paragraphs, macrowords had a higher false alarm rate and longer reaction times than thematically related distractor items, while thematically unrelated distractors produced the smallest false alarm rate and the fastest responses. These results were obtained whether or not subjects were required to write a summary before taking the recognition test. It was concluded that subjects form macrostructures during reading as an integral component of the comprehension process, and not just in response to certain task demands.