Stories as mental representations of an agent's subjective world: A structural overview

Narrative is a universal form of human information and communication. The generative cognition of narrative will be a core element of a human-like and human-friendly autonomous agent. In the field of artificial intelligence (AI), the knowledge aspect of narrative, including script knowledge and episodic memory, has been studied from a functional perspective. However, this knowledge aspect is rooted in the close relationship between a narrative and how the human mind organizes subjective experiences into a mental representation. We use the term “story” to refer to the uniform mental representation of a piece of a subjective world containing individual meaning and a rich temporal extent. For example, stories include the following types of mental object: an episodic memory, an autobiographical memory, the contextual structure of a current situation, a prospective memory, a planned or imagined future, and a fictional or virtual story. Although narrative cognition is a very complex problem for AI, a cognitive mechanism for dynamically generating mental stories through interacting with external environments is necessary for the ultimate autonomous agent. Considering this long-term goal, in this paper, we provide a structural overview of a subjective world formed by stories as experiences of an agent. In particular, the following structural aspects of mental stories are described: a story-based context of a current situation, association of past, future, and fictional stories, association of others' mental stories, and the distinction between factual and fictitious stories.