There were some typos in the Preference and Value formulas, which I fixed.
I wanted to encompass the idea that many times when we talk about World States, what we're really talking about is the State of a particular Agent's World Model, as opposed to the abstract World State itself. I'm suggesting that whenever we subscript a World State with a particular Agent name, we are talking about the World Model State relative to that Agent. I propagated this change throughout Value and Preference as well.