The following is a series of answers regarding the background of the Threefold Model and its assumptions. It was inspired by a long thread on the RPGnet forums, entitled "A summary of rgfa theory" (July 2006). Many of the questions could be answered from the original RGFA FAQ files, and my Threefold Model Page. However, there were some important points.
1) Is the Threefold Model only about personal intent, and if so, why?
The Threefold Model is about the process of decision-making. Now, decisions can be made either individually or socially. For example, I can just declare game actions without consulting or explaining to the rest of the group. In that case, then the extent of the Threefold is only what is inside my head. However, I can also propose something and the group can discuss it. Social decisions can be classified according to the Threefold by what arguments went into them. For example, suppose an argument based on realism is rejected but one based on genre is accepted. That suggests a Dramatist process.
One of the central points of rgfa discussion was coining the term "group contract" for both formal and informal agreements among the group. As we used the term, all agreements of play were part of group contract -- including the mechanics as well as agreements like "Don't feed the cat" or "Call in advance if you're late". I've noticed in later usage people tend to used the term "Social Contract" to mean only non-mechanical issues -- which implies that the system and mechanics are not social agreements, when of course they are.
One thing which bugs me is the idea that there is a natural distinction between "social contract", "system", and "content" of a game -- or worse that the natural way of that is the one true way. A game can have printed rules which cover virtually anything which is in the social contract. For example, the length of a game session is usually left up to the players to decide. However, there are a few games like Puppetland which specify session length. Similarly, many people assume that the game rules provide a character generation system and the players create characters. However, a number of games instead provide characters, like Timelord or the Parlor Larp series.
What goes into game decisions includes different input: the game rules and the game advice, as well as the reasoning of the player.
2) Shouldn't the Threefold Model cover results and/or goals?
Well, no. That certainly isn't meant to dismiss those. I think analyzing those would be a profitable point of view of role-playing. On the other hand, I think that this won't look much like the Threefold Model.
If you accepting for a moment that decision-making is the basis of the Threefold Model, it is clear that there are other ways of analyzing games. You can look at objective results rather than subjective intent -- like the transcript of events, impact on the players, and so forth. This is difficult because most games aren't recorded and the participants' accounts often differ. Still, it's useful to look for. You can also look at larger goals: i.e. what are the processes trying to accomplish?
I started on a look at larger goals in an old Forge thread, "Classifying by Social Function". However, it was pretty incomplete.
3) Does the Threefold Model only apply to in-game GM decisions? What about player decisions, or design decisions?
The Threefold Model is a taxonomy rather than a strictly defined model. Thus, wherever the definitions make sense, they can be applied. It is possible for players to make decisions on the basis of either in-game cause, story effect, or meeting challenge. Thus, the model can apply to players. The definitions can also be applied to many cases of preparation. For example, you can generate the details of a town based on either needs of the story you are telling there, or extrapolation based on it's geographic location, local history, and so forth. Now, broad early choices about the world cannot be extrapolated. For example, the choice of whether to play fantasy or science fiction, for example, isn't classifiable under the Threefold. The same is true of other classifications, though.
Now, the Threefold Model was developed to apply to actual play -- not to rules on their own or to campaign preparation on its own. However, the principle is to look at of what sort of actual play the rules and/or preparation support. This applies similarly to how other models (like Robin Law's Sixfold or GNS) are applied to rules or preparation.