From Sherlock and Buffy to Klingon and Norrathian Platinum Pieces: Pretense, Contextalism, and the Myth of Fiction1 Peter Ludlow
University of Michigan
Departments of Philosophy and Linguistics
firstname.lastname@example.org “If Peter Ludlow is a journalist, then I’m a railroad tycoon whenever I play Monopoly.”
Vice President for Corporate Communications
Electronic Arts Corporation
1.0 On the two way flow between popular fiction and the real world There are those who would dispute the point, but a good case can be made that real world individuals are often characters in works of fiction. So, for example, London (and Bakers Street) are characters in Conan Doyle’s Sherlock Holmes books, and New York City is often a character in Woody Allen movies. When, in Annie Hall, Albie Singer pulls Marshall McLuhan from the line at a theater and gets him to debunk some bad McLuhan exegesis, I believe that it is really Marshall McLuhan who is a character in that movie and who is doing the debunking. Cases like this can be multiplied (nation states, famous persons, historical landmarks, familiar food, drink, modes of transportation, and indeed most of the furnishings of works of fiction are real). There are puzzle cases (for example, are Bill and Hillary Clinton characters in Primary Colors? Is Albie Singer supposed to be Woody Allen?) but these don’t undermine the idea that real individuals (and substances) somehow inhabit fictional worlds.2 Less often remarked are the cases where objects begin their existence in works of fiction – typically popular fiction -- but somehow manage to spill out into the real world. Characters don’t step off the screen as in Woody Allen’s Purple Rose of Cairo, but less dramatic cases can be found. A classic example of this is the Klingon Language of Star Trek fame.3 Klingon is a language that began as a reference to a fictional language spoken by a fictional extraterrestrial race in a fictional universe. Yet, somehow, today there is a quasi-official Klingon language institute (http://www.kli.org/), people who speak Klingon almost exclusively, and alleged cases of persons who want to raise their children in Klingon-speaking households. Is it really Klingon? Well, that might depend on whether the makers of movies and television shows set in the Trekkie universe (or perhaps trekkie fandom) take the Klingon experts to be authoritative. I understand that they often do.4 Even better cases are available if we expand our examples of popular fiction to include Massively Multiplayer Online Roleplaying Games (MMORPGs). These are online games (accessible via the internet) where players assume a character in a fictional (i.e. computer generated) environment and move about that environment, interact with other players, and change the environment in what might be thought of as a giant work of collaborative fiction (some games like Lineage,5 in Korea, boast millions of users).
In a typical MMORPG one constructs an avatar – a cartoon-like character – that one uses to navigate the virtual world and through which one interacts with other users. One can build virtual homes, make virtual objects (e.g. weapons), virtual clothing, and modes of transportation. One can also organize with other users to create social institutions and objects ranging including discussion clubs, virtual theater troupes, virtual mafias, and in-world newspapers which report on the activities of these virtual groups -- my own newspaper, The Second Life Herald,6 being a case in point.
The case of game currencies is particularly interesting. MMORPGs typically have an official game currency in which nonbarter transactions take place. In Second Life, for example, the official unit of currency is the Linden Dollar. In EverQuest, the official unit of currency is the Norrathian Platinum Piece.7 These currencies begin their life as a kind of monopoly money having no real value outside of their value in roleplay within the game, but this quickly changes (and often against the efforts or the game owners). The currencies soon come to have a real world value, and an exchange rate that can be tracked on eBay and other exchanges such as Gaming Open Market that are established for precisely these purposes.8 It is important to understand that the activities within these MMORPGs can generate significant real world wealth. Persons in these games invest their time building objects, homes, and institutions, and these objects become prized. I may wish to buy a virtual home from an experienced virtual home designer so I will pay for it. If I don’t have sufficient in-game currency to buy it I may well go outside of the game and buy the currency on eBay with US Dollars. The designer has, in effect, created some real world wealth.
Depending upon the competence of the game designers, the currencies of these games can be more or less as stable as real world economies – more stable in many cases (over the past four years you would have been better off keeping your money in Norrathian Platinum Pieces than Argentinean Pesos and many other currencies). Particularly competent game managers such as Linden Lab9 watch the game money supply at least as closely as the US Federal Reserve tracks the US money supply.
The effect of all this is that the economies generated within these games are substantial. In a widely discussed study, the economist Edward Castronova (2001) has calculated that EverQuest, which has around 400,000 users (fewer than many Asian games) has a per capita gross domestic product that would make it the 77th largest in the world (just behind Russia but ahead of Bulgaria). In total, the economy of EverQuest is about the size of Namibia’s. Norrathian Platinum Pieces are as real as many other real world currencies, it seems.
The mystery in this, of course, is how it happens. It was already difficult to see how real world objects could make it into a work of fiction; we were struggling with trying to understand how real individuals like Marshal McLuhan and New York City could get into fictional worlds. Now the matter is compounded: How on earth did Klingon and those Platinum pieces become real?
I want to reemphasize that this is not just about money; certain kinds of fictional social groups and institutions also begin their lives as fictions within MMORPGs and spill out into other MMORPGs. For example several of the “mafias” and user-created quasi-governmental organizations that were spawned in The Sims Online have moved on to other MMORPGs such as Star Wars Galaxies, There, and Second Life.10 If it hasn’t already happened it is easy enough to imagine these organizations also taking up operations in the “real world.”
My own experience with the Alphaville Herald is a good illustration of this general point about in-game social objects acquiring a kind of uptake outside of the game. When I first entered a MMORPG called The Sims Online I named one of my characters Urizenus (after Urizen in William Blake’s Book of Urizen), declared him to be the editor of a newspaper called TheAlphaville Herald, and set up a blog to chronicle virtual events within the game. After blogging a series of articles that discussed unsavory aspects of the gameplay, and further articles critical of the game owner Electronic Arts (EA), the Urizenus account was terminated by EA. Yet when I blogged the termination story along with other stories of online events, I found that many of the stories were picked up by “real world” media outlets including Salon.com, Wired News, The New York Times, CNN, NPR, BBC Online, The Boston Globe, The Detroit Free Press, Italy’s Corriere della Sera and La Stampa, France’s Liberation, Spain’s El Pais and Moscow’s Izvestia.11 Many readers and media outlets (in particular the reporters) took the Herald to be a genuine newspaper. The Alphaville Herald somehow became a real newspaper (or at least as close to a real newspaper as blogs get). Again, the pressing question is how such a thing can be possible.
In this paper I want to advance a thesis that is highly contentious and no doubt quite impossible to believe on first hearing. However I think that if the position is put clearly enough the position is not absurd at all and that it has some very real merits. How do we explain the two-way flow between fictional and real worlds? How do fictions become real? Answer: they don’t become real; they always were real. There is no such thing as fiction, and there are no such things as fictional objects. There are, however, certain predicates that are only satisfied in limited contexts of use, and this gives the illusion of different kinds of entities (fictional objects), and different modes of existence (fictional existence).
More specifically, the idea is this: In the case where we have props or actors involved, certain predicates (‘is a vampire’, ‘is a stake’, ‘are fangs’, ‘is a slayer’) may be true of those props and actors in limited contexts of usage. For example, consider Buffy The Vampire Slayer star Sara Michelle Geller. The predicate ‘is a vampire slayer’ may be true of Sara in certain limited contexts (e.g. when she acts or when we watch the show and are caught up in it). In a case where there is no actor involved (as when we read a book that has not been adapted for theater or screen) we can say that certain general claims (e.g. ‘there is a slayer having certain properties’) are true in a limited context (as when we read the book).
This proposal will bear certain important similarities to pretense theory – the idea that one is engaged in a pretense that such-and-such is the case (for example, the pretense that Sara Michelle Geller is a slayer) – but differs in one important respect: The core notion of pretending is dropped altogether. Rather than saying that in certain contexts we are pretending that such and such is the case, I will argue that in those contexts it is simply true that such and such is the case (albeit true only in those contexts). Once the relevant contexts are identified, the notion of pretending and/or the introduction of a PRETEND operator in the semantics become redundant exercises at best. From the point of view of the semantics of “fictional” discourse, the PRETEND operator plays no interesting role and is arguably harmful in that forces semantic theory to abandon the principle of semantic innocence and leads to a number of difficult semantic puzzles.
Although the conclusion is reasonably straightforward (if provocative) the path to the conclusion is not short. Getting there will require that we first take up several important topics. I’ll begin (section 2) with the introduction of pretense theory and then turn (section 3) to some of the puzzles that have arisen with trying to deploy a PRETEND-that operator. In section 4 I’ll draw a comparison to difficulties that have arisen in the defense of presentism in the philosophy of time, and will suggest that at root the difficulties are cut of the same cloth. In section 5 I’ll suggest that we might take a leaf from alternatives to presentism in the philosophy of time, and I will show how neatly the “new leaf” works for the treatment of cases like Norrathian Platinum Pieces and Klingon and the more traditional puzzles (in effect, I’ll rely on the new, apparently harder cases, to show us the path back to the proper treatment of more traditional cases like Sherlock Holmes and Buffy). In section 6, I’ll begin to lay out the (off the shelf) tools that we will need to deploy in the positive proposal and in section 7 I’ll put together the pieces in a kind of rough draft of the thesis. I’ll then take up an apparent problem about identity (section 8), and will conclude (section 9) with some general thoughts about the project and the general doctrine of the myth of fiction.
2.0 Pretense Theory To understand my proposal, it will be useful to consider the theory that I am reacting against – namely pretense theory. Perhaps it would be better to say that I am building on pretense theory, for I think that pretense theory gets quite a bit right, and I’m not sure I can explain the positive proposal without having the pretense theory on the table. According to pretense theory (associated most closely with the work of Kendall Walton) when we engage with a fictional work we are involved in a kind of pretend play or pretense.12 Accordingly, when an actress like Sara Michelle Geller plays Buffy the Vampire Slayer she is doing just that: playing as or pretending to be a slayer. When we watch the program and are engaged by it, we too may be pretending that there are vampires, and that there is a slayer named Buffy. We may be pretending that Sara Michelle Geller is Buffy. Notice that this view is distinct from the usual supposition that Sara Michelle is representing Buffy or engaged in mimesis.
In my view pretense theory is a clear advance; it skirts a number of intractable difficulties about representation, and is easily extended to handle a broad class of cases. Indeed, this basic approach to fiction is so popular and so extensible that it has been taken up across the board in philosophy as a foil against all sorts of suspect entities. Don’t like an ontology with abstracta like numbers? Well, we could adopt a pretense theory about mathematical objects – we merely engage in a useful pretense that they exist (a pretense that is especially useful when we are engaged in science or business).13 Don’t like odd entities like flaws in arguments, holes in cheese, or the average man? Pretense theory can be invoked again: We are merely pretending that there are such things. Do you like possible world semantics but not the ontology of all those worlds and counterparts? You can be a pretense theorist about them.14 Worried about the reality of moral claims? Pretense theories are available here too.15 One of the more helpful aspects of pretense theory is the idea that certain of our activities of pretense serve as props which “generate” the fictional object within the pretend circumstance. For example if we are playing vampires and slayers and I use a drinking straw in lieu of a real wooden stake, then the straw is a prop and we can say that “in the fiction” or “in the pretense” the straw is (generates) a stake. This is handy, because it gives us something that might serve as a hook to hang our talk of fictions on. That is, fictions are often based on *something* real – at least in cases where we act out the fiction.
It’s less clear what we are to say in the case of a novel or forms of fiction that are not performed with physical props. Walton (1990; chapter 2) suggests that perhaps spoken words, linguistic forms, or, in the case of imagination, a kind of mental data structure might serve as a prop. An alternative would be to say that in these cases we are not engaged in a pretense about something, but rather that we are pretending that such and such is the case. When we read, for example, we are pretending that the world is such that a certain state of affairs holds in it. There need be no props except perhaps for the actual world itself.
The basic idea behind pretense theory can be incorporated into the semantics of natural language in the following way (I’m not saying all the practitioners would put it exactly this way). There is a pretense operator which we can call PRETEND. We can quantify over anything we choose so long as the quantification takes place safely within the scope of the PRETEND operator. For example I can quantify over unicorns so long as the quantification is within the scope of PRETEND, because then it merely follows that I am pretending that there are x’s, such that those x’s are unicorns. Typically we won’t bother making PRETEND explicit. If I am watching Buffy and say to you “Buffy just impaled a vampire”, I won’t bother with PRETEND because you and I know that we are engaged in pretense – it goes without saying. The operator is almost always implicit and unvoiced, but it is always there. Or at least that is the theory.
Others, including Lewis (1983), have proposed operators of this form, so we need to be careful in contrasting the Lewisian proposal from the pretense theory proposal. For Lewis, the ‘it is a fiction that’ operator worked just like modal and tense operators worked within his program. In the modal case one is quantifying over possible worlds and in the tense case one is quantifying over future and past times and events. The fiction operator is like this in that one is quantifying over other worlds – fictional worlds – which are inhabited by individuals like Sherlock Holmes (or his counterparts) etc.
Whatever may be said about the merits of modal realism, it is hard to see that thinking of fictions along the lines of possible but unactual individuals preserves our intuitions about fictions. There is a great deal of discussion of this elsewhere, including Kripke (1973), so I set it aside for now, noting only that the central problem will have to do with our intuitions about cases like Superman (i.e. even if some individual showed up from a planet named ‘Krypton’ with an ‘S’ on his chest, a red cape, and superhuman powers, we would not want to say that it was Superman). In a nutshell, given that Superman doesn’t exist and is not based on a real life individual, there *couldn’t* be Superman. He can’t exist in *any* possible world. Possibilia just aren’t reasonable candidates for fictional objects.
One advantage to the Lewsian theory, however, is that it does do a good job of handling the commingling of fictional and real objects. In effect, both kinds of objects are on a par. Pretense theory has a way of doing the same thing by relying upon props (at least in the cases where props are available), but taking advantage of props in this way requires some deft footwork. Several examples from the semantics literature illustrate just how carefully we must step and just how crucial props (or at least something like them) are to the whole enterprise.
3. Troubles with PRETEND-ing Commingling fictional and real objects doesn’t seem like too big of a problem when real life objects and individuals are introduced into the fiction – in that case we have what amounts to quantifying in.16 There is an individual, Marshall McLuhan, such that we are pretending that Albie Singer pulled him out of line at the theater. There are, of course, standard problems about quantifying in that have been familiar for half a century now, but at least they have the virtue of being familiar if not entirely tractable. For example, if I am pretending that we are flying to the Morning Star am I thereby pretending that we are flying to the Evening Star? In some cases that will be part of our pretense and in some cases not. Consider a case were it isn’t (for example, suppose we are pretending to be pre-Babylonian space travelers). Now it looks like we want to give ‘the morning star’ narrow scope with respect to our pretend operator to reflect that we have the intention to go to, say, the Morning Star and not the Evening Star. But wait! We may still want to say that it is the real Morning Star that is a character in our fiction and that we are flying to. So what do we do? If we keep the descriptions outside the scope of the PRETEND operator we lose the distinction between the thing we are pretending to do (fly to the morning star) and the thing we are not pretending to do (fly to the evening star). If we tuck the descriptions within the scope of PRETEND, then we are pretending that there is a morning star and, in effect, we are losing the insight that we wanted the actual planet Venus to be a character in our little fiction. No surprise that these puzzles should arise here, and at least we have some idea what the solution strategies are and where they break.17 More intricate are the cases where we have what might be called “quantifying out”. Cases like this are familiar in the aesthetics literature as well, though perhaps not exactly thought of in these terms. The most familiar case is (1)
(1) Sherlock Holmes is smarter than any living detective.
To see the potential problems here, first notice that here we can’t stick the whole sentence in the scope of the PRETEND operator as in (1a).
(1a) PRETEND (Sherlock Holmes is smarter than any living detective)
The first problem with (1a) is that its truth conditions are consistent with a case in which it is part of our pretense that all living detectives have been replaced with inept detectives. Accordingly, one might go for the fix of pulling the quantification over living detectives outside of the scope of PRETEND.
(1b) [any x: living detective x] PRETEND (Sherlock Holmes is smarter than x)
But this doesn’t solve the problem, either, for (1b) is consistent with a case in which it is part of our pretense that all living detectives have been dumbed down.
A further possible concern – a more subtle one to be sure -- is that it may not be part of our pretense that Sherlock Holmes is smarter than any living detective. This may just be a fact that we deduce or observe (for example, on reading the Conan Doyle novels I may remark that “no living detective could be that smart.”)
One might try exploding the comparative into a kind of bi-clausal analysis, giving PRETEND scope over only one of the clauses, as in (1c).
(1c) there is a degree d, such that PRETEND (Sherlock Holmes was smart to degree d), and No living detective is smart to degree d.
This strategy has several weaknesses, however. In the first place, it does not seem necessary that my assent to the truth of (1c) should require that there be some specific degree of smartness (some numerical quantity) such that Holmes is smart to that degree. If you asked me “precisely how smart is Holmes?” I would probably want to answer by saying “I have no idea; I’m just saying he’s really smart – smarter than any living detectives.” The problem of course, is that we are groping for something that we can safely quantify over across these contexts – and even the de re quantification over degrees is highly suspect.
The other tripwire here is that PRETEND is a hyperintentional operator; like ‘believes-that’ anything within its scope may be sensitive to substitution down to the lexical level at least. Accordingly, any analysis we introduce within the scope of PRETEND is not guaranteed to preserve truth value, even if the analysan is otherwise logically equivalent to the analysandum.
If we are prepared to bite this fusillade of bullets, the general strategy still fails when we consider cases that are only slightly different. Consider (2)
(2) Bertrand Russell resembled the Mad Hatter
Ignoring tense for the moment, the issue is where to stick the PRETEND operator. We obviously don’t want to stick it over the whole thing, because we aren’t pretending that Russell resembled the hatter. He did resemble the Hatter – or at least what the Hatter is supposed to look like. Accordingly, we want Russell (and the relational predicate ‘resembled’) to be outside the scope of PRETEND and the Hatter to be inside the scope of PRETEND. One idea would be to try to find something that degrees of smartness did in the previous case, but this time things don’t work quite as smoothly. What is the missing ingredient – the thing that we can get away with quantifying over de re – this time? One possibility would be to introduce images:
(2a) There is an image x, such that Bertrand Russell resembled x, and PRETEND (the Mad Hatter looked like x)
Are we really pretending that the Mad Hatter resembles a particular image? (Notice here that the hyperintensionality of PRETEND rears its head again.) And what does it mean to resemble an image anyway? It seems odd, after all, to say that I resemble a picture of myself.18 Of course the picture may resemble me, so we might think that reversing the order of the resembler and resemblee will help here:
(2) There is an image x, such that x resembled Bertrand Russell, and PRETEND (x resembled the Mad Hatter)
But now we have two problems. We have the original problem that when we truly utter (2), it doesn’t seem like we are pretending that some image resembled the Hatter, nor is it even clear what that would mean to engage in such a pretense. If the Mad Hatter is the target (i.e. the resemblee) then what is it to pretend that a picture resembles him? But the second problem is that Russell has now gone from being the resembler to being a resemblee. That certainly doesn’t seem right.
Nor does it help to dispense with the resemblance relation altogether in favor of something like isomorphism as in (2c).
(2c) There is an image x, such that Bertrand Russell is isomorphic to x, and PRETEND (the Mad Hatter is isomorphic to x)
Quite apart from what it would mean for a fictional object to be isomorphic to an image (or anything, for that matter), and quite apart from the issue of what it means to quantify over pictorial images (assuming there are any19), there is the issue of whether it is really a pretense of ours that The Hatter is isomorphic to that or any other image. This suggests that what The Hatter looks like is under our control to some extent, but that certainly doesn’t seem right either. Do we need to introduce a higher order quantification over pretenses now, so that we say there is a pretense according to which Russell resembles The Hatter? Or do we need to specify a specific pretense? But whose? And how?
The general problem of quantifying out gets even worse if we consider Peter Geach’s (1967) Hob, Nob, and Cob case and convert it into a case where they are pretending there are witches rather than just deluded into thinking that there are witches.
(3) Hob is pretending that a witch blighted his mare and Nob is pretending that she didn’t do that but that she put a pox on his sow and Cob is pretending that she considered doing those things but baked a cake instead
Now the issue is that there are three different pretendings at work. No single pretense operator covers all of these cases. So when in the successive pretenses we find the anaphor ‘she’, what exactly is that supposed to be picking up? If you think telling a story about this will be easy, first read the last 40 years of literature addressing Geach’s puzzle.