Gamasutra.com has an interesting feature, called Modeling Opinion Flow in Humans Using Boids Algorithm & Social Network Analysis
Similarily to pathfinding, you can model people and influences as nodes with vertices representing the "weight" or degree of normalized influence between them. Then the actions of the player (or other major events) - and most importantly; how those events are communicated - can influence how popular opinion in general and the actual opinions of individual npcs will fluctuate over time.
"At the end of the day in our game universe, one supports King John, supports King Richard, or doesn’t particularly support anyone."
Like the author says, its not the Matrix yet - but a step forward in modelling and adapting networks to opinions.
Unfourtunately the supporting "site" is just a "hello world" page, rather than actual links for sourcecode and downloads. Hopefully it will be up over the weekend, since the article was dated the 26th.
EDIT: the java code and a collection of Excel files is now up
Of course - the challenge of finding a timed "network" remains - from whom do the people hear what news when and how does that affect their interpertation of it (is it the bard that has been travelling with Richard - or the wounded soldier fighting for John - or the neighbours daughter?)
