| In five experiments, I investigated how readers and listeners generate relevant contrasts in comprehending and remembering discourse. Past work has suggested that prominent words promote encoding of salient alternatives and that this benefits later memory, but it is unclear exactly which alternatives are considered or how consistent these benefits are across modalities and across individuals. Participants read or listened to discourses containing salient alternatives (e.g., Malaysia when the discourse also mentioned Indonesia). In Experiments 1 and 2, font emphasis in the initial presentation facilitated participants' ability to later reject the salient alternatives but not unmentioned items (e.g., Portuguese scientists ), generalizing past effects of contrastive pitch accents. In Experiment 3, font emphasis facilitated rejections of salient alternatives but not less plausible alternatives that were nevertheless mentioned in the discourse. Online reading time measures in Experiment 2 indicated that emphasized words did not improve performance on all trials and only benefited memory to the extent that participants devoted extra time to them, although no such relation was observed in Experiment 3. The relationship of online reading time to later memory is consistent with views of language processing in which some aspects of linguistic representations may be left underspecified because they are time- or resource-consuming to generate. Further, the effortful processing of an alternative impaired memory for the rest of the discourse in populations with more restricted online processing abilities: older adults (Experiment 4) and younger adults who have lower scores on complex span scores (Experiment 5). |
