A few months ago, I came across an interesting idea by Scott McCloud (the originator of 24 Hour Comics challenge) called the Story Machine. Here’s an excerpt of his idea:

“Basically, it’s a big map of symbols that you randomly explore (using a four sided D-4 to move north, south, east and west) while creating stories.Writers and artists are very habit-driven, this is one way of breaking habits (the 24-hour comic was another). Contrary to the name, this isn’t a device for generating the stories themselves, just for unlocking imaginations by throwing an endless series of conceptual curve balls to stimulate unpredictable turns of thought.”

While the idea itself is pretty rudimentary, it got me wondering about the possibilities of Story Generation. Campbell, Propp, and several others had already come up with abstract models of what-makes-a-story. As I read further, I realized that the AI community had tried to tackle the problem in 60s and 70s. Roger Schank’s research group at Yale, in their attempt at understanding the processes that a human should have to understand natural language, eventually refocussed themselves to study narratives. However, only a few research groups survived in the 80s when the funding for AI research dwindled. Most AI groups started focussing on industry related problems.

However, with the explosion of computer usage, the problem of HCI (Human-Computer Interaction) naturally led to a re-emergence of Narrative Intelligence. Marc Davis and Michael Travers started working on the problem in MIT Media Lab.

Now, there are several research groups which focus on different aspects of the problem. The purpose of this blog is to act as survey/bibliography material, and also help me gain understanding enough to come up with my take for the problem.