Murder Mysteries: The White Whale of Narrative Generation?

Abstract

The field of narrative generation appears to have a particular fascination with the genre of murder mysteries, perhaps because the genre’s defining characteristics include a clear, and predefined structure, a reliance on logic, and an immense popularity. However, while there have been several attempts to generate a satisfactory mystery story, most such efforts never made it beyond an initial prototype or idea. In this article we investigate what may draw researchers towards this genre, and then disseminate why the very properties that make murder mysteries such an appealing target also make it a particularly challenging domain. For the apparent rigid scaffolding structure provided by the genre, we discuss its flexibility; for the apparent hard grounding in logic we illuminate the deficiency of such; and for the seemingly clear communication maxims prescribed by written “rules” for the genre we explore which ones must be broken. Our goal is to show the actual scale of the difficulty of the problem, and which challenges have yet to be addressed. While it may seem that our focus is overly narrow, we will also explore the implications for narrative generation more broadly.

Sabre: A Narrative Planner Supporting Intention and Deep Theory of Mind

“…Supporting theory of mind enables new stories (Shirvani, Ware, and Farrell 2017), but adds significant cost above an intention-only planner. Sabre does not support uncertainty, which may be necessary for some stories, like murder mysteries (Eger 2020).…”