Two types of “non-reality” appear in online role-playing games: the virtual and the fictional.  I put the term in scare quotes because representations and ideas, as such, are themselves real, even if what they represent or refer to don’t exist.  What exactly is the relationship between the virtual and the fictional — what is the nature of their “non-realities,” and how different are they?  Clearly it isn’t that one represents a reality whereas the other doesn’t: in a virtual world, one can have a human avatar, but one’s av can just as easily be a dragon or a pulsating blob, neither of which represent the natural form of the person behind the avatar.  Also, especially in a virtual world like SL, most people don’t engage in role-playing, so their avs are virtual but non-fictional whereas the RPers have virtual forms but also portray fictions.  It’s notable that in SL, where one has near-total control over an avatar’s look, people almost never create avs that look like themselves; but there is no pretense that they do either, so on the whole their avs aren’t exactly fictions.  One could say that within the context of the virtual world, the av is what the person really looks like.  (One major complication is when the av and the person don’t have the same gender: for one reason or another, this riles a good many people, some of whom call it “pretending” or even “deception” whether or not it’s intended as such.  Also, a few individuals have no stable appearance — from one day to the next, they use a completely different av, often not even humanoid; it’s a form of play.  But I have to leave those complexities aside for now.)  But that context includes the element of pretense or “as if” thinking, which would seem to blur the difference between a virtual world and a fictional world.

It is tempting to say that “virtual” is an ontological term, “fictional” an epistemological one.  It might also be correct, but it’ll take a fair bit of analysis to decide.  It would make sense of the fact that in a virtual world like SL, one soon treats the environment as though it were real; or rather, one realizes it is real, but as a digital realm, not a physical one.  Correspondingly, some people object to contrasting Second Life with “Real Life,” as SL is real; they contrast the virtual world not with the “real world,” but rather with the “physical world,” “atomic world,” or terms of that sort.  Online RP’s fictions thus supervene upon virtual entities, not fictional ones: avatars in RP don’t sit on fictional chairs, they sit on virtual ones.  This is the same as in physical-world theater: when actors play a character, their character is fictional but they sit on real chairs, not fictional ones (unless they’re pretending that there’s a chair, which is a different matter).  In contrast, a character in a novel sits on a fictional chair.  Following this reasoning, the stage character supervenes upon the actor: this is connected to the way everything on stage becomes semioticized as signs of signs.  That’s even clearer in the virtual world, since virtuality is fundamentally semiotic: in Peirce’s sense, it consists of iconic signs.  It’s easy to slip into the positivist idea that signs aren’t “real,” but under critical realism, they certainly are.  Understanding virtuality as an ontological category would make some sense of the fact that in child play, a mud pie can for all intents and purposes (except for eating) be a real pie.  One doesn’t cut a sign of a pie, one cuts a pie.

I’ve argued that dramatic characters are virtual agents; now I have to reconsider that idea.  “Fictional agents” might be better.  But maybe not: it depends at least in part on how one understands the actor’s body.  Again, a chair in the performance is an actual chair (not a fictional one), except that it has been semioticized as the sign of a sign, which perhaps means it has been virtualized.  Is the only part of a play performance that’s truly fictional the characters and their activities — i.e., the scriptive level of performance?  That seems correct.  However, that still doesn’t provide an answer to the virtual agent vs fictional agent problem.

Or, perhaps that distinction between the virtual and the fictional is incorrect.