The Course of Representation in Memory During the Comprehension of Paragraphs

The course of the representation built in memory from a text during comprehension of paragraphs was studied by a probe technique (immediate item recognition). Three experiments showed that response times, the main variable, gradually increase as a function of the lag between the probe (a word or an atomic proposition) and its target in the text. In Experiment 3 the same result was also found with a priming technique. The absence of any kick-up over the course of time was confirmed by several additional analyses of individual data. The results were only weakly consistent with models assuming two distinct memory stores, in particular a specific “short-term memory store,” but highly consistent with models involving semantic activation and subsequent gradual deactivation.