Wow, a lot of blogposts in the last week. I guess I've just been in the right mood for it...
Anyway, Raph Koster just made a post defending the idea that fun comes from learning new skills. This reminded me that I've been contemplating the issue a bit recently, and have some ideas to toss into the discussion. I'll begin by saying that I agree with the last point in Koster's post - the theories that have been presented don't necessarily conflict with each other. If you had a direct argument going on between an advocate of 8 kinds of fun and the learning theory of fun, I would say that the two camps are talking past each other.
First, I'll say something about "8 kinds of fun." They are sensation, fellowship, fantasy, discovery, narrative, expression, challenge, and submission. My immediate reaction the very first time was "aren't some of these more important to games than others?" At the time I thought that some seemed more "game-like," by which I meant that they were more particular to games. In particular, "sensation" doesn't seem very game-like, because books and movies can offer spectacles at least as well as games can, and clearly "narrative" is the kind of fun where literature is king. It's true that the narrative and the presentation of games can be important though, so this left me confused for some time.
I think I understand now the source of that confusion: the above list, though a good list, is mis-titled. They aren't "kinds" of fun...they're aspects of fun, or components. That is, they are the various parts that make up a fun experiences. While someone might look for games that have good stories, they're probably not playing the game strictly for the story (again, if all they wanted was to see a good story, they would just read a book). It may be true that they want to see a good story in the game, but that's different from just wanting to see a good story - the game setting changes the meaning of the desire. What I'm getting at generally is that the items on the list above can all greatly contribute to the funness of a game, but they just don't seem core enough to the game experience to properly be called "kinds" of fun. It seems to me like we skipped a step when we call them such.
The fun-as-learning approach is at the opposite extreme - it's so fundamental that it becomes limited in how much in can tell us. It's true, given a sufficiently broad definition of "learning," but it's far from being the complete story about what makes games fun. If we recognize that the 8 kinds of fun are really components of fun, and then combine it with the fun-as-learning approach we get a much better picture of the whole process - these elements combine in different ways in order to create interesting learning environments which we call games.
I still think we're missing an intermediary, however, which is the real "kinds" of fun. Like I said, I think that only something very core to the gameplaying experience can count as a kind of fun. I doubt I'll get this right on my first pass, but I have in mind 3 different core types of experience, all of which are built out of the 8 components of fun and all of which entail some kind of learning. What I currently understand to be the 3 kinds of fun are: creation, submission, and competition. These are actually each in the list of 8 components, so if you prefer, I'm just making the list more concise by eliminating redundancy.
If we recognize that games are sets of rules, then the relationship between these kinds of fun should be fairly intuitive: creative gameplay comes from control of the rules, submissive gameplay from exploring the space within those rules, and competitive gameplay from comparing performance within the rules. In reality, the boundary lines aren't so clear, and a game can have appeal as providing more than one kind of fun, but I think these can be separated out as three core kinds of gameplay. I'll discuss them in the opposite order I introduced them in:
Competitive gameplay is the kind that we're most familiar with, and is easy to understand. In its purest form, competitive play depends on the barest set of rules, which exist only to establish a sense of fairness between the competitors. In fact, the rules here are defining what is fair, and by choosing what is and isn't allowed the rulemaker is deciding what kind of player will fare best in the game, but at the least the rules should assure the players that they each have a fair shot of proving their superiority in the game. Probably the purest example of competitive games are simple feats of physical ability: races, weight-lifting contests, and the like. In these contests, the rules are not very interesting, so the interest is entirely about the competitors, their training, and what they're able to achieve. In digital games, competitive behavior emerges in just about anything that is multiplayer, but it can also arise in single-player games if the player believes that they and the computer are in a "fair" contest. This is largely illusory, since computers can be programmed to perform better or worse at a task, but it's interesting that it's nonetheless relevant to the player when they're in a competitive gameplay mode - they become angry when they perceive the computer "cheating."
Submissive gameplay is a kind of gameplay we've seen grow substantially since the advent of digital games, but exists in plenty of non-digital examples. Submissive gameplay involves accepting the rules of the game as law, and then exploring the space of possibilities left within those restrictions. This is common in digital single-player games, where the player accepts the rules of the world, the limits of their controller, etc., as essential parts of the game. He makes a silent pact with the designer, that in exchange for accepting these limitations, he will be compensated by a rewarding game experience. This happens in person in the case of tabletop games, where the players submit to the authority of the gamemaster (I joke to the GM of my most recent campaign that his title implies that the players are "slaves", but in retrospect, it's not a joke: playing a table-top game IS play-as-submission). The purest incarnation of submissive play, however, is puzzles. In a puzzle you have a set objective, and often only one solution, but if the path leading to that solution is interesting, then players will submit to spend their time trying to do exactly what the designer wants, and fun may happen as a result.
Finally, there is creative gameplay. The most obvious example is the game designer's career - really it's just extended gameplay that happens to produce something people will pay for. Another clear example is the role of the gamemaster in a table top game, as mentioned above. Sure, the gamemaster generally abides by the rules set out in the rulebook for whatever game he is running, but he is free to invent "house rules," and expected to flesh out the environment, decide what kind of encounters the players can reasonably be expected to overcome, etc. In short, the experience of being a gamemaster is fundamentally different from the experience of being a table top RPG player. Your feedback doesn't come in the form of experience and treasure, but in the reaction of the players to your game - when they have fun, you feel accomplished.
I may have missed a fundamentally different kind of gameplay, but every example game I can think of makes use of these three varieties of gameplay. An RTS game has submissive play and competitive play (it's a good example of a genre in which players are wary of the computer "cheating"), Oblivion has submissive and creative gameplay, etc. Also note that each kind of gameplay can appear in non-obvious places. If you are given a tool within a game that allows you to control an aspect of the environment, and it's not designed to be used only for one specific purpose, you are experiencing creative play - by altering the environment, you're getting to test out different rules to see what kind of effects they have.
There's plenty more to say on the subject, but this post is too long as it is. Let me know if you can think of a core kind of gameplay that is not one of these three.