情境模型的更新：事件框架依赖假设的进一步证据

采用多指标探测范式, 从客体信息以及活动信息的角度出发, 考察事件框架内外时间、空间转换对情境模型更新的影响。结果发现：在事件框架内, 维度转换并不是情境模型更新的充分条件, 时间转换与终止的活动信息相结合引起情境模型的快速更新, 而空间转换与移除的客体信息相结合引起情境模型的快速更新; 在事件框架外, 维度转换是情境模型更新的充分条件, 维度转换会引起情境模型的更新。研究丰富了事件框架依赖假设。

The Updating of Situation Model: Further Evidence of Event Frame-Dependent Hypothesis

When comprehending narrative text, a reader forms situation models that represent occurring events. As events unfold in the text, the reader must update his or her situation model, or mental representation. Zwaan, Langston and Graesser (1995) suggested that the event is the core unit in constructing the mental representation of narrative reading. The reader establishes the mental representation of events by tracking in five dimensions: time, space, characters, causality and target. If any one of the five dimensions changes, the situation model requires updating. The temporal and spatial dimensions were the focus of the current study. A large number of studies have shown that both play important roles in constructing the situation model, yet it remains unclear whether a change in time or space is a sufficient condition for situational model updating. Xia et al. (2013) found that space shift and event shift had a close relationship. However, they did not find the combination of space shift and removing objects caused the updating of spatial situation model. He et al. (2013) further proposed and tested the Event Frame-Dependent Hypothesis to explore the effect of time shift on the updating of situation model. Results of their study confirmed the hypothesis. However, because differences between the updating of the temporal situation model and the updating of spatial situational model were not found, questions remained about the effects of the temporal and spatial situations. Therefore, in the current study we conducted four experiments to further explore the effects of temporal and spatial shift on updating of the situation model. A multi-target detection paradigm and within-subjects design were used in the study. Each experiment included eight experimental materials. 66 college students who did not participate in formal experiments rated experimental materials, and 144 additional college students participated in formal experiments. E-prime 1.1 was used to present stimuli and to collect the behavioral data (reaction time and accuracy). All statistics were analyzed with repeated measures in SPSS 13.0. Experiments 1a and 1b replicated previous study, showing that within the event frame, reaction time and accuracy between space shift and no shift conditions did not differ significantly. That is, the space shift did not cause a change of the event, and thus did not update the situation model. However, within the event frame, the combination of space shift and removing objects caused the updating of the spatial situation model. Whereas beyond the event frame, differences of reaction time and accuracy between space shift and no shift conditions were significant, the situation model updated when space shifted. In Experiments 2a and 2b, we explored the effect of different activities on temporal and spatial situation model updating within the event frame. Results showed that readers paid attention to information of activities in the updating of the temporal situation model and ignored the information of activities in the updating of spatial situational model. In sum, the event frame affects different dimensions in the updating of the situation model. Within the event frame, dimension shift was not a sufficient condition for situation model updating: It did not cause the updating of the situation model. However, time shift combined with activity information cause rapid updating of the temporal situation model. Furthermore, space shift and object information caused rapid updating of the spatial situation model. As such, whereas beyond the event frame, dimension shift is a sufficient condition of situation model updating, dimension shift, such as temporal dimension, spatial dimension, can cause the updating of situation model situation model updating. We also can infer that the five dimensions of the event-indexing model are all affected by event frame. Within and beyond event frame, the changes of dimension have different effects on situation model updating.