Previous studies on L2 learner's story telling have shown that the learners are unable to tell stories from a fixed viewpoint unlike native speakers. Most of the studies employ the picture description task as the method of verbal elicitation, which may have affected the results. Therefore, the present study examines the validity of this picture description task by comparing it with the personal story-telling task. Native Japanese speakers and Chinese advanced learners of Japanese narrate on ten topics about their past experiences and tell stories using three comic strips. The results show both native speakers and learners describe their past experiences from the speaker's viewpoint, though some inappropriate use of viewpoint expressions are observed in Chinese learners' narratives. On the other hand, the viewpoints in the data elicited from the picture description task are not unified. The findings indicate that the nature of task affects the use of viewpoint expressions.