This study examined whether the process by which native Japanese speakers read narrative passages with no explicit location information and no accompanying pictures and construct a global coherent understanding of the text, i.e. a situation model, involves the construction of visuospatial representations of the passage content. In the experiment, the working memory model was adopted as its theoretical foundation and used the dual task method to set up a dot task as visuospatial load task. The construction of the situational model was measured by the presence or absence of the contradiction effect. The results suggest that Japanese native speakers can construct a situation model without constructing visuospatial representations when reading narrative passages with lower language difficulty.