Abstract: Situational model is one of the most intensively investigated topics on discourse comprehension. Up to date, it is widely agreed that narrative comprehension includes mental representation that is based on text description. Event-indexing model explains situational model in detail. The former model proposes that events are the core units of the mental representation, which are defined in five situational dimensions: time, space, character, causality and intentionality. Space is an important dimension for constructing situational model, and a number of studies have shown that readers track spatial information while reading texts. In addition, some studies provided evidence supporting that readers use spatial shift to segment the consecutive narrative. Radvansky and Copeland (2010) explored the influence of spatial shift on the processes of situational model updated in narrative reading by multi-index detection paradigm, which integrated the previous study on spatial situational model updating. However, after scrutinizing their experiment, we found that they did not separate the factors spatial shift and event shift, which actually are not consistent. Based on the above consideration, event segmentation and probe words paradigms were adopted to investigate the function of event shift and spatial shift in narrative reading by four experiments. Experiment 1a repeated a previous study, exploring the effect of spatial shift on situational model updating and reading time, and further verified the rationality of using multi-index detection paradigm in our study. Experiment 1b investigated the influence of spatial shift on situational model updating under the separation of spatial shift and event shift. Experiment 2a studied the effect of spatial shift on event segmentation, and further verified the rationality of using event segmentation paradigm in our study. Experiment 2b explored the influence of spatial shift on event segmentation under the separation of spatial shift and event shift, which can verify the affect of spatial shift to situational model updating. The results showed that the separation of spatial shift and event shift did not cause the updating of situational model. Event unit was the core unit of constructing the narrative’s mental representation. Only when the event shift was identified could spatial shift cause readers to update situational model.

Key words: situation model, spatial shift, event segmentation