Thanks for responding to the call. I will now endeavor to show you that you actually agree with me in all important details.

Re: 1) I think you already agree that we agree here. Happy dance.

Re: 2) I think you’re just fastening on the wrong convention in thinking that the third season shouldn’t count as part of canon. Alternatively, you might think (although I don’t, really) that there are two properly canonical worlds, such that one of which (call it “w”) includes the third season, and the other of which (call it “Ian Proops forgot my name”) is a proper initial segment of w, which stops after the second season. Let’s name this tv show “Scotland Sucks.” Then we’ll consider w the canonical world of the official (or, say, marketed) version of “Scotland Sucks” and we’ll consider Ian Proops forgot my name the canonical world of the good version of “Scotland Sucks”. Did I really just suggest that one tv show should generate two worlds, both of which count as canonical? Yes, yes I did, mainly because all the fans (as you’ve stipulated) will consider Ian Proops forgot my name the canonical world for “Scotland Sucks,” which they consider to be a good show that ran for two seasons, and other people (like the show’s writers) will consider w the canonical world for “Scotland Sucks,” which they consider to be a good show that ran for 3 seasons but was abandoned but some uppity self-righteous “fans” after the second season. That means there are two separate conventions in play, both of which are part of the society in which both w and Ian Proops forgot my name are considered by certain disjoint groups of people to be canonical. I think the Ian Flora deep down inside of you will agree to that, so now I’ll tell you what a non-conventional criterion would be.

Suppose that your favorite atheist writes a novel that has a canonical world with no God in it and stipulates that the canonical world of said novel is possible. Then suppose that the Pope comes along and says, “Wait a minute. The canonical world of this novel has to be possible. Therefore there has to be a God in it, because God is a necessary being and all that jazz.” Then we would say, “Listen, Mr. Pope, this canonical world does not have a God in it. The work clearly implies that it lacks anything nonphysical. I mean, the title is even ‘A novel that generates a possible canonical world that doesn’t have a God in it,’ and the author gave a lot of interviews to that effect, and all of the fans conceive of the work as such, etc., etc. And so, Mr. Pope, no convention compatible with the canonical world of the novel can require there to be a God in said world.” And then suppose the Pope says, “Well, if it’s a possible world, then it’s got to have God in it.” Let’s also stipulate that all of this happens in Sweden after there’s a large earthquake and it becomes separates from all other countries, and is filled only with atheists, so that the Pope has no power to impose conventions of his own. Then I think his attempt to claim that the canonical world of the novel must, by (necessary, as it were) convention, have God in it will seem to lack the force of conventions we should be worried about. I suspect that makes the point clear enough that you can dissent if you wish.

Re: 3) I was just going to say, “OK, then let’s make them extensions of a large-enough subset of the propositions true in the canonical world,” which is just to say that they should be extensions of fanonical bases in your terminology. Also, who the hell is Uhura? Just checking.

So, where are the disagreements still lurking?

Best,
j

Those last two “canonical”s in my comment should read “fanonical.” Damn you, jargon!

1) I’m largely behind Jon’s strategy of using creative personae to get at which works build canonical worlds (provided we don’t get ourselves into trouble by secretly using works to define personas). But Sam raises an excellent issue, namely the demands that picking canon out places on the reader. Should we use “saves” to make Canon consistent? Or should we admit that the canonical world is supposed to be inconsistent? My view is that the work itself will tell us. For instance, if an author has a core of similar-lookig characters that keep showing up, the reader will be naturally curious as to whether the stories are all about the same people, i.e. whether the author is building a canonical world. I think what would tell us this, as appreciators, is whether the implicit connections between the works hold up to scrutiny. For some stories, trying to build a timeline is a very rewarding endeavor that elucidates connections and leads to some novel appreciation of the works (an example of this would be the process that Kurt Busiek went through in writing “Marvels”). In others, it’s just frustrating, because it’s clear the author wasn’t thinking about any of the things you’re trying to appreciate. This seems to be a pretty good candidate for an interpretive/evaluative component to canon. I’m not quite sure what I should think about that.

2) About the conventionality of canon: That I can’t really agree with, and here’s why. Suppose the third season of a TV show is really crappy, and no fan pays any attention to it. They write their own stories to fill in what happened, and think of the third season as an embarrassment. There could be wide-spread consensus that one just doesn’t think about the third season. I submit that, nevertheless, if a fan wanted to build an interpretation of events that occur in season 2 using events in season 3 as explanantes, he would be able to insist on doing so, overriding all of the fan embarrassment and taboo. Why? Because it happened; it’s canon, regardless of what people might like to think. Of course, you could say that there’s a convention that events derived from sanctioned works have overruling power in disputes of that sort, and so this is not a counterexample to conventionalism. But then I’d have to ask you what would count as a non-conventionalist criterion.

3) Fanon and extensions: I think the extension idea works for one kind of fanon, namely what I call “Playful exploration,” where fans speculate on things left unsaid, without offering corrections. I do, however, identify two different types of Fanon where a straight-forward extension idea won’t do the trick: corrective exploration (where events are posited that, while unsaid, were clearly far from the canonical creator’s mind, like Uhura being a strong, independent woman and leading an away team) and corrective defiance (where events are posited such that not-p, where canon expressly says p). I call all of these Fanon because they are all sets of propositions derived from a fanonical base. But you could say we’re just dealing with some homonymy here, and so all of them deserve different treatments. I’d need some considerations that there is a deep distinction between the three, though.

P.S. If you don’t know what I’m talking about when I say “fanonical base,” the short of it is that it’s a subset of canonical propositions. A proposition is fanonical, then, if it is a member of a canonical base or generated from a canonical base using appropriate principles. A lot of work will go into saying what would constitute “appropriate.”

Fair enough. Here’s a general strategy, together with a (probably contentious) claim about canon: Start with a given work. Use the sorts of criteria I suggested to determine the creative persona under which it was created, and use that to determine which other works contribute to specification of the relevant fictional world. This is enough to get us from works to worlds. Some of these worlds will be labeled canonical; some will not. Since what counts as canonical is nothing more than a social convention (there’s the probably contentious claim), how this labeling goes will be purely conventional, so we aren’t falling back on intuitions here. The important thing is that we’ve given (rather, vaguely suggested) a method for getting from works to worlds, which gives us a way to group works together and make them into one world and separate them from other works which get made into other worlds.

A possible extension of this idea for fanonical worlds is that they should be the ones that take the fictional worlds that get labeled canonical and embed them in “extensions,” where to be an extension of a world w is to fill in details left underdetermined by the specification of w.

Thoughts?

Best,
Jon

P.S. It would be nice to get some more thoughts on this from Ian now that he’s done being a thespian for a while. So, Ian, thoughts?

bracketing: i am not opposed to this. obviously, the hard part is giving a consistent account of this that works. i think one easy way to do this is use a fiction-to relation. obviously, a fiction F can refer to other works which are fictional to F. you might think what happens in family guy or the simpsons is where a narrator is telling a story about the characters. but another issue is how do you decide between consistency-saving fixes like this and an intended impossible fictional world?

creative persona: so yes, i think you might be able to ramsify your way out of it — i was just pointing out straightforwardly defining creative personas by canonical works would not work generally because we are already using creative personas to define canonical works.

“There would have to be a lot said about the identity conditions for creative personae (e.g., changes in age probably shouldn’t matter, development of psychosis probably should, etc.), but I don’t see anything wrong in principle with stipulative identification using established works of canon, and using the thus-identified creative persona to adjudicate the contentious cases.”

in fact, i think you raise some good independent criteria to define creative persona here; psychological continuity might be a good candidate. i take it your proposed procedures start with intuitively identifying canonical works. this may work for most cases, and that may be enough. but wouldn’t it be better if we could have have a definition that works generally? only if we fail should we fall back on the intuition option, because then we haven’t shed much light on the concept.

Anyway, for principle 1A and 1B (which are really two sides of the same coin), I would argue that seeking consistency in fiction is a byproduct of our logical nature in the same way that creating literature is a byproduct of our intelligence and creativity. It may well be that every new sequel to a book takes place in a separate universe.

I once started writing a series of what I called ‘Necroautobiographical microfiction.’ These were 250 word stories in which I was the main character. At the end of all of these, the main character dies in a rare and unexplained way. I consider all of them canonical.
My next story will involve the main character becoming a married bachelor and drawing a round square.
He’ll also be doing something outwardly immoral that violates the laws of physics and goes against the previous character’s history (using X-ray vision to spy on his neighbor, most likely).

In this example, I use fictionalitive resistance and bracketing for personal enjoyment. The question this example should bring up is whether we can reasonably suppose that natural properties hold across fictional worlds (including fictional worlds which seem to have no consistent properties of their own).

I probably don’t even need to say it, but in support of Ian’s resolution strategy…

They killed Kenny! You bastards!

Re: my favorite topic (i.e., me and various things I’ve said)

Also, in my last comment I said I couldn’t think of a case where there were a bunch of possibly inconsistent competing options. Shortly thereafter I thought of one. Namely, suppose an author writes a bunch of stories using what appear to be the same fictional characters (they might have the same names, physical characteristics, street addresses, birthdays, etc.), but writes each story such that the characters in it have to be different than the characters in the other stories. (E.g., he might craft the stories so that what seemed like the same character has in each story a different personality, different relationships to other characters, etc.) Then, to the extent that we think the stories really give competing options for being canonical, rather than combining to characterize one inconsistent fictional world (probably involving split personalities, lots of drugs, a statue of William Wallace inscribed with the word ‘freedom’ standing behind a protective cage, etc.), it seems to me that we would think that the author has merely established separate canonical worlds where the occupants of each world bear a fairly large amount of similarities to the occupants of all the other worlds. Further, to the extent that we think they really give competing options for being canonical, rather than combining to characterize one inconsistent fictional world, it seems to me that we would think that the author must have been operating under substantively different creative personae in crafting the stories. Why? Because, I would think, if we thought he crafted all of the stories while operating under the same creative persona, then we would think he meant to be creating one (admittedly weird, admittedly inconsistent) fictional world.

I will now shut up until somebody responds :-)

jon

I’m glad Sam (and, by extension, I) have caused some discussion. I have a quick point about the idea of comedic bracketing. I’ll also have more to say later about the idea of a “creative persona” as a means of separating the canonical from the non-canonical.

Bracketing is an important part of our appreciation of fiction, and, to recall Steve’s brief post, I would consider it a limitation of my view if it could deal only with consistent Canon, or consistent fictional worlds (I recall a portion of our discussion where it became clear that a lot of Canonical worlds are, in fact, impossible worlds). But I also don’t see a huge problem here. Remember, it’s not just expressed propositions that determine what is fictional (and, further along, canonical) in a given work. Conventions, both of genre and form, can also play a role. Someone who had only seen that episode of “Family Guy” would probably be a bit surprised to see Meg die. They might even turn to their friend and ask “is she really dead?” A more experienced viewer would probably not be as surprised that she turned up fine the next week. One might also notice from FG being a cartoon, and from one’s experience with Loony Toons and the like, that the reality is more elastic than in, say, 24. So one could say:

1) It is FG-canonical that Meg died in episode X.
2) It is FG-canonical that Meg is alive in episode X+1.

Prima facie contradiction. The easiest way to resolve it is to say “it’s a cartoon.” In fact, I think you had the right idea when you said that any sort of complex explanation in terms of the fictional world would probably seem inappropriate, much like trying to explain why Wile E. Coyote doesn’t die when he falls off the cliff.

Now we might be in trouble here, because we seem to be disrupting one of Canon’s main purposes, which is its ability to tell us about what’s already fictional in the world of the work (in this case, episode X+1). I’m not worried here, though, because we can identify certain conventions that govern the relation of presumed fictionality. It was a convention of sitcoms for a long time (and still is, maybe) that everything had to be back to normal at the end of the episode. Usually, the writers at least gestured at the return to normal within the fiction itself (a heart-felt talk, a confrontation, etc.), but maybe FG is onto something new, where you just hit the reset button and don’t worry about it. Hence, we might have a principle that anything that causes too great a deviation from the status quo is not presumed fictional.

Looking forward: This seems to be getting into some neat territory, because it seems to suggest that there is an evaluative component to attributions of Canon itself. If we work this through, we may find that it’s not as easy to draw the Canon/Fanon distinction along causal/evaluative lines as I might have made it sound in my presentation. Watch this space.

It seems a little quick to rule out using canonical works to characterize the relevant creative persona just because you’re supposed to use the creative persona to figure out which candidate works get to become canonical. In cases where there’s already some definitely established canon, we should certainly be able to pick out the relevant creative persona stipulatively (as the one that was being operated under when the established canon was produced). There would have to be a lot said about the identity conditions for creative personae (e.g., changes in age probably shouldn’t matter, development of psychosis probably should, etc.), but I don’t see anything wrong in principle with stipulative identification using established works of canon, and using the thus-identified creative persona to adjudicate the contentious cases. [There should also be a better way of saying "operating under a creative persona" but I can't decide what it is.]

It’s trickier when there isn’t already a group of canonical works, and instead there are a bunch of possibly inconsistent (possibly fanonical) competing options. Maybe then considerations about authorship and what not come in. I’m not sure, though, that I can think of a reasonable case where there is no established canon whatsoever and no way of deciding which creative persona should take precedence. Have you got a plausible one in mind?

Best,
j