What is fictional?

I promised hard-core theory, and I hereby deliver on my promise. What I'm going to say is partly inspired by a short discussion I had with Vincent Baker on whether all truths about a character are in the minds of the players; it is partly inspired by and a continuation of a discussion on the anatomy of roleplaying I started a while back on The Forge; and also in no small part inspired by the book Mimesis as Make-Believe by Kendall L. Walton, which I have just started rereading.

There is a set of related issues I want to talk about, but for now I'll concentrate on a very simple question. When we are roleplaying, we are creating a piece of fiction. (Albeit a piece of fiction that is different from normal written fiction in that it is dynamic, constantly changing, instead of static. This makes all the difference in the world.) This piece of fiction comes with a fictional world (a concept which I will also say more about below). In this fictional world, some things are true and others aren't; instead of using the unwieldy "A is true in the fictional world", we say "A is fictional". Here is the question: when we are playing a roleplaying game, what things are fictional, and what makes them so?

This is a small but crucial question in the bigger project of developing an anatomy of roleplaying. (That's right, the activity, not the games. For the games, go look at John Kirk's design patterns.) If it sounds too abstract to you, that's fine, you don't have to read it - but I think understanding what we are doing at a fundamental level is important, and in the end I will talk about applications.

A preliminary point: I will be talking about roleplaying games while they are being played. One might wonder what the product of a session of play is, and what is made fictional by that product, but I won't. What interests me is the activity of roleplaying.

Roleplaying is an activity, and it is done by concrete people is a concrete social situation. It is this social situation and the actions of the people we are trying to understand, and fictionality should somehow be based in that. So let's try out a simple proposal:

Something is fictional in the fictional world of the game if and only if all the players agree that it has been established - explicitly or implicitly - about the fictional world.
At first sight, this seems to work. If player A has said "Harry kisses Mary, and they fall in love!", and all the other players agreed to accept this into the fiction, then it has become fictional that Harry kissed Mary and they fell in love. On the other hand, if nobody has said or implied this; or if it is just the case that somebody intends to say it; or if somebody has said it, but people are still disagreeing whether or not it should be accepted into the fiction - in all those cases, it is surely not (yet) fictional that Harry kissed Mare and they fell in love.

But there are also problems with this definition. Let me first discuss a scenario that Walton talks about in his book (in which the words 'roleplaying game' do not appear, unfortunately). Suppose two guys, let's call them Bob and Jim, are walking through the woods and decide to play a game of make-believe: all the stumps in the forest are bears! Whenever they encounter a stump, they shout "watch out, a bear!", and run away, or whatever.

What is fictional in the world of this game? Well, if there is a stump somewhere in the real world, it is fictional that in the fictional world, there is a bear in that place. By virtue of the convention Bob and Jim have set up - which boils down to: whereever there is a stump, you should imagine a bear - stumps have become props in their game that generate fictional truths. A stump being somewhere makes it fictional that there is a bear in that place. And here is the rub: Bob and Jim might not know that there is a stump hidden somweher in the thicket they are walking past. Yet it is fictional in their game that there is a bear in the thicket, and it is even fictional that they pass him by unawares.

This point easily generalises to the kind of roleplaying games we play. Suppose that a group of players have established, beyond doubt, the convention that they'll play My Life with Master strictly by the rules. They've started playing, and look up the rules as they need them. None of them has yet read the part of the book that is about the endgame - so nobody knows that in the end, the Master must die.* And yet their convention makes it fictional, given the rules, that the Master will die.

Another example. A group of four people + GM have just started playing The Mountain Witch. The GM deals everyone a Fate Card, face down, and puts away the others. Nobody has yet looked at their Fate Card. And yet the rules of the game and the (implicit) social convention that they'll play the game by the rules makes it fiction that the party of ronin has the four Fates that have just been dealt.

In the board game Cluedo (I think it may be called Clue in English) - in a sense also a kind of roleplaying game, though extremely limited in the importance of roles - it is already fictional who has committed the murder, even though all the players are still trying to find out who did it.

What all these examples make clear is that something can be fictional even though none of the players know about it. So what is fictional cannot simply be those things which all players believe to have been established. On the positive side, we have seen that every time, it was the prior establishment of some kind of convention as to what things have authority to make other things fictional that lies at the basis of what is fictional.

So here is the new proposal:

Whenever a group is roleplaying, there is a set of implicit or explicit rules - some cultural, some specific to the group, some specific to the game being played - that establish which things have the authority to make statements fictional, and how they are to do this. Something is fictional if and only if it has been made fictional by something that has thus been imbued with the authority to make it so.
Sounds trivial? In a sense it is, but thinking about it can give us some interesting insights. Here's one: the things that are fictional need not be the contents of a Shared Imagined Space - if we interpret that term literally. They need by neither Shared, nor Imagined.

We've already seen examples of what is fictional not being Imagined: if nobody knows about it, they can't imagine it. Here's an example of a different kind: somebody has made an authoritative statement at some point of the game, but all the players have forgotten about it. Perhaps they are even telling something that directly contradicts it. Does this statement still make something fictional? That depends on the conventions the group uses. In my group, there's an implicit convention: "If nobody cared enough about it to remember it in time, it's probably not important and we should forget it." But in a heavy Sim group, I can totally see the following convention existing: "Once we've established that something is fictional, it is and remains fictional. If we forget about it and start telling things that contradict it, we are fucking up the world." In such a group, statements that everyone has forgotten could still make things fictional.

Here is a design consideration: if you want the group to adopt a rule like the above, you should give them the tools to remember what has been established! Think about how important the internal consistency - the integrity - of the fiction is to you, and build your game accordingly.

I also said that what is fictional does not have to be shared. How is that possible? These rules that establish what is fictional and what is not are shared, surely? They are. But imagine a heavy immersionist game. This group might well use the rule: "Whatever you think and feel when you are immersed is (fictionally) what your character thinks and feels." Here, parts of the fiction are only accessible to certain players - yet these parts are validated as truly part of the fiction by the social rules that govern the game. When I play Polaris and my Mistaken tells me that my character feels hate rising within him**, I certainly cannot counter by exclaiming that that is not what I am experiencing at the moment - being immersed and experiencing something does not make it fictional that it is. But in a heavily Immersionist game, this might be a perfectly valid claim that makes the other retracts his statement.

Design consideration: make sure whether you want some parts of the fiction to be private, perhaps in the interest of immersion, and if so - protect them! This is all perfectly valid.
In a sense, The Mountain Witch also has private parts of the fiction - those things concerning their own Fate that the players have already decided upon without revealing it. Since they have the authority to makes statements about this, their deciding that something is the case could already be said to make it fictional. However, in this game one should reveal one's Fate, whereas the Immersionist is under no obligation to reveal all the details of her character's inner life to the other players.

Here's another consequence of my theory of fictionality: roleplaying is not simply a process of establishing new facts about the fiction, that are added one at a time and accumulate to tell a big story. Here's why: first, narrating something, even when it is accepted as a valid part of the narration (what I elsewhere called the 'shared text'), does not mean it makes anything fictional; second, which things do and do not have authority can change during the game; third, changing the basic rules can change the fiction.

Let's take them one at a time. If you are allowed to narrate something, that doesn't necessarily mean you have been given the authority to add things to the fiction. Shades is an example of this: oftentimes, you will be narrating things that actually turn out to be not true in the fictional world. You are allowed to tell them, yes, but you cannot imbue them with fictionality. That may happen later in the game, or it may never happen.

Second, something which at one time had authority may not have so at another; and vice versa. The familiar phenomenon of 'retconning' makes this clear: statements that were once accepted as authoritive are later relieved of this authority. (Sometimes, retconning may have to do with a mistaken belief about what statements had authority; I'll not dwell on this subtlety.) Here's a more substantial example: suppose we're playing D&D3E, and we have the social convention that we stick to the alignment rules. That gives the rules "All paladins must be lawful good" authority to establish the sentence "all paladins in the world are lawful good" as fictional. Now, at some point the players may say: "let's forget about this alignment stuff - chaotic evil druids and true neutral paladins are cool!" Effectively, they negate the authority of the rule, and it is no longer fictional that there are only lawful good paladins in the world. (However, is isn't yet fictional that there are paladins in the world who are not lawful good. Until some non-LG paladins come up during play, this may simply be undecided.)

Third, the very social rules that give out authority may change during play. "Let's ditch this D&D shit and continue our campaign with The Shadow of Yesterday!", some guy may cry out - and if all the others agree, the social rules of the game have now changed and authority has been transposed from the D&D book to the TSOY book. This also makes many things fictional that did not use to be so, and it makes many things non-fictional that used to be so. Moral: you can majorly change the fiction simply by fooling around with things surrounding the narration and without narrating anything concerned with the fictional world at all.

Ok, but what does all that mean for game design? If nothing else, it serves to illustrate that there is way more to roleplaying than people saying things related to what is happening in the fiction. Yes, it is not enough to look at the processes that govern this stating of things, there's a whole lot of roleplaying going on around that. Material props can generate fictional truths; part of the roleplaying can be private; part of the roleplaying is not concerned with establishing in-game events at all, and yet influence the fiction directly. And we, as roleplaying designers, should look at all of these processes and shape them in ways we think will help to create the gaming experience we would like to bring to people.

Does the above give the designer any hard and fast rules that he can take to his design table and work with? No. But it does map out a (small part of a) space of things he might want to think about.

* Actually, My Life with Master does not necessarily end with the death of the Master. The formulae are such that it could take an infinity of time to kill him (even though chances are that it will happen pretty soon), and we can safely assume that no group will be so invested in the game that they'll spend that kind of time on it.

** This may not be quite kosher, by the rules, but we tend to allow this kind of thing.


  1. Well done--this is a phenomenal post and it touches on a lot of my issues with the "shared imaginary space" term in common usage.

    In many games things will be fictional that have not been "established and validated" and I think this is an important and often overlooked context.

  2. Very insightful, Victor. I think I'm going to have to consider these points before I can really discuss them in full, but I am especially intrigued by (the implication that I read into this) the proposition that fictional facts have more than two states. It's not 'fictional' and 'not fictional', there seems to be a full range of 'private', 'proposed', 'under consideration', 'ratified', and the like.

  3. Marco: thanks. I intend to write more on the 'SIS;' and related concepts in the future.

    Joshua: Yep, there are many possibilities there. Fictional proposition can be (at least) shared, private, temporarily forgotten, unknown (for one of several reasons).

    Between fictionality and non-fictionality, there is a whole range relating to different strengths of authority. If, in Shades, I narrate how my character walks through a red door, it is probably fictional that the door is red as it seems unlikely that the colour of the door would be distorted by the character's self-righteous perception of the past; but who can really tell? A somewhat different example could be a game with lots of setting information and a social rule in force that "the setting information is true unless it is very advantegeous to the story to change it when it becomes important". The propositions implied by the setting information are probably fictional, if it makes sense to say so, but one cannot be quite sure. There is a lot of room for subtleties and borderline cases here.

    Then there is the 'process of giving authority'-axis: propositions can be proposed, ratified, under consideration, rejected, 'on hold', and so forth.

    So there are at least three axes on which propositions can be assigned a position, all of them allowing of more than two positions.

  4. Excellent post Victor.

    It occurs to me that something like this happens whenever my groups front-load our game by using an established setting. Everything in the setting book becomes "proposed" fiction. Sometimes we take it at face value, sometimes we don't. However, we always consider it seriously and so it has more weight than just something sitting there giving no weight or influence to the game.

    Whereas I know lots of groups that once they front load the setting is the fiction, and the events in game must conform to it and not the other way around; and on the opposite side lots of groups that make everything up as they go along with no preloaded proposed fiction at all.

    Very interesting, I will have to think upon this further.

  5. An interesting post, wanted to comment on this bit

    "Here is a design consideration: if you want the group to adopt a rule like the above, you should give them the tools to remember what has been established! Think about how important the internal consistency - the integrity - of the fiction is to you, and build your game accordingly."

    in my post on Metaplay I talk about alot of the basic processes of establishing the game world, one interesting mechanic, which I quite liberaly stole from Universalis (which is interesting comparing to your theory since the conduct of the game is about establishing a fiction, and has very explicit mechanics on how credibility is apportioned in establishign elements of the fiction/fictional world)

    Anyway, from my rules draft


    Canons are the facts and details established about the narrative environment and the narrative. These details include the important events, facts & details, and action outcomes within the game, as well as the traits of game entities. Canons are established via production scripts, through entity design, narration and other scripts, and as the result of the actions of personae within the game world. Collectively, canons comprise the “reality” and continuity of the shared game world. Canons are given mechanical treatment so that the shared game world remains consistent and knowable by all of the players. Players cannot casually create facts or details that ‘go against canon’. "

    Indeed, there is a lot of attention to nailing down and making sancrosanct elements of the fiction, the construction of the production bible, careful keeping of records and establishment of entities.

    For instance, all players are assumed to design game entities that they hope will be introduced to the narrative environment (the ficitonal world). So they are fictional creations by one player, that have at least been vetted as mechanically sound by the Setting Guide (more interesting bits regarding the apportionment of credbility/validation among group), but aren't part of the narrative environment. So they are proposed, and approved fiction, but not "legitimate" parts of the fictional world. One more interesting bit, is that players are awarded more crediblity pull in the form of Nomenar when another member introduces an entity they designed to the narrative via "Royalties".

    So like I said, all kinds of explicit mechanics. Additonally, any player can object to the introduction of elements to the Narrative Environment, either on a personal level, by using the Rewrite scripts. At the same time, the setting guide is given authority to automatically reject introductions that go against the Production Bible. So my game has lots of checks and balances built in against changin the fictional world that don't fit the pre-established vision of the fictional world. all interesting bits. Also to reiterate, prior established elements or guidelines regarding the fictional world are given extra pull, via the additional costs to overturn Canon, or automatically guarded via calls for Executive Rewrites by the Setting Guide (players would have to call a vote spending their nomenar and getting other players to agree to change.)

    So anyway, good stuff there, and it intersects with what I am doing in my game, and how Universalis is setup.

  6. I finally got around to reading this, Victor. I found myself nodding agreement throughout the article, but I found your terminology difficult--specifically, your use of "fictional" to mean "true in the game world". And "non-fictional" seems ambiguous as it could mean "not true in the game world" or it could have the weaker implication of "not part of the canon of the game world", i.e., "up for grabs".

    On this general topic, I always recommend the Forge thread you referenced, as well as Markus Montola's two articles on diegesis construction, mentioned here (a Google search on the titles will locate them easily). I also have some discussion and related links on my LJ, in the two posts regarding the Lumpley Principle here. An important discussion (for me at least) was at Thomas Robertson's LJ, in this entry.

  7. This makes me wonder if you've read Wittgenstein's _Blue and Brown Books_.

    There's a lot you can do when you start looking at role-playing at this level. The most obvious is that you can simply lie to the players about what's happening.

    I had a player who through some plot mechanic was supposed to become paranoid. So I kept having him make perception rolls for no reason. Then I began writing him notes telling him that the others in the party weren't being completely honest about what they were doing...

    But, you could simply run players through inconsistent stories and then put them in a room to see what happens.


    ps - Sometimes, when it had been a while since our last session, we didn't quite remember where we'd left off, and pick the game up too early. Which lead to some odd effects...


Post a Comment

Popular posts from this blog

Keeping the narrative pressure on

Thoughts on a Trollbabe session