Dramatic Situations for Emergent Narrative System Authorship

Abstract

When designing and developing an emergent narrative system, one finds themselves in the difficult situation of working with low-level mechanisms while aspiring for high-level, longer-term emergent outcomes. To make things worse, the desired output is not even a concrete artifact but an ambiguous mental construct: something recognized as a story by a human. We think that dramatic situations, as conceptualized by Georges Polti, can act as a useful in-between heuristic: they are objective enough to inform the design of low-level operations, but abstract enough so as not to overdetermine the output. We have embedded 18 dramatic situations in the possibility space of our emergent narrative system and have documented three emergent stories through a constrained process. We find that identifying a dramatic situation prompts the construction of a story as it incites tracing its causes and finding out its outcome. Observing the behavior of the involved parties significantly contributes to their characterization. As dramatic situations are relatively rare in a character’s life and typically have high stakes, they are likely to be noteworthy. A potential issue is the discernable recurrence of some dramatic situation emergent patterns which might negatively impact players’ interest, a common problem of procedural generation. We hypothesize that this issue might be alleviated by the unique situatedness of players and their characters in the world as they encounter these patterns, creating special conditions of relevance.