The maintenance of single generalization inferences held in working memory

Abstract

Previous studies have found that readers are capable of drawing a single generalization inference about a single topic from five consecutive sentences of text (Ritchey, 2011; Ritchey, Barnes, Suryanarayan, & Donaldson, 2013; Ritchey & Suryanarayan, 2015). Although generalization inferences play an important role in the reading comprehension process, relatively few studies have been performed on how these inferences are constructed. Thus, the current study extends previous research by examining how a single generalization inference is processed further in working memory as readers are presented with subsequent demands for coherence. Specifically, this experiment manipulates two factors consistent with the memory-based comprehension literature (i.e. elaboration and sentence placement), and uses a resonance process to describe how sentences supporting a generalization inference are maintained in working memory. Readers should have greater difficulty comprehending elaborated text passages within a new theme condition when compared to control participants who continue to receive the same theme condition. Later placements of the target sentence within the same theme condition should help to facilitate the main effect of elaboration as participants continue to receive additional examples supporting the implied generalization inference. In contrast, later placements of the target sentence in the new theme condition should reduce the ability to maintain a generalization inference in working memory. These predictions were mostly supported by the findings, with an interaction between elaboration and placement for various generalization themes. Implications for how to best illustrate generalization themes and future research on generalization inferencing are discussed.