This is the most informative bit of research I’ve located so far. The speaker in this video, Espen Aarseth, is the author of “A Narrative Theory of Games” in which he proposes a framework for defining how software can be both a game and a narrative. This talk is a summary of his central ideas presented in that paper.
Among the myriad of distinctions Aarseth makes, I think the most useful is the difference between kernels and satellites with respect to narrative. Kernels are the events in the narrative which are necessary for the plots to move forward, like Cinderella meeting the prince, while satellites are the events within a narrative which have no effect on the plot, like say if Cinderella were to eat ice cream. We can think about this in terms of necessity; kernels are those events which are necessary for the narrative to proceed, while satellites are those unnecessary details which do nothing to forward the story.
Aarseth uses this distinction and a players influence over the narrative events of a game to define whether or not a piece of software constitutes a form of game, narrative or both. Generally speaking, if a person has no influence over the satellite events presented in the story then it is a piece of narrative and not a game. If the person has influence over satellite events but not kernels, then the software is both a game and a narrative form. And if the player has complete influence over both satellites and kernels then the software is strictly a game. This doesn’t include things like branching narrative models, where a player’s choice alters the narrative, but it is still a useful in principle way of determining whether or not an interactive platform can encompass both narrative and gaming elements.