The pitfalls of narrative technique in rpg play

This is going to rpg be theory stuff, just so you know.

A thing I’ve noticed lately specifically at Story Games, but also on other gaming fora, is the increased acceptance and advocation of narration authority sharing between players in a roleplaying game. It’s a nice technique, but I also find that it is being recommended and utilized in ways that might have unexpected consequences that need to be considered in depth. I’ll write a short treatise about the topic here – I don’t particularly want to piss in anybody’s cereals if they find that unrestrained sharing of narrative authority brings them happiness, but it’s not correct to call it the universal panacea of roleplaying, either – there are solid reasons for refusing to introduce this technique into every single game you might ever wish to play.

The basic phenomenon of narration

The layman does well to notice that “narration” has basically nothing in particular to do with narrativism, an important theory term on its own right. This is often a superbly confusing matter for the majority of participants in rpg theory discussions, as they either assume narrativism where mere narrative is concerned, or assume narrative technique where only narrativism is indicated.

When we discuss “narration sharing”, we mean the type of roleplaying technique where first one player narrates things, then another. “Narration” here basically means adding to the fiction developed during the game with little overview, or none at all. An author basically “narrates” when he writes or tells things about a story: first this thing happens, then this other thing, and so on. Much of the communication in roleplaying is not “narration” in this sense of the term: for example, when you declare the actions of your character, this is not narration in most traditional games, as those declarations will still have to be processed through the game’s system or GM’s consideration before they will be properly considered part of the fiction. “Narration” is when there is no such constraint; you open your mouth, and what you say goes straight into the fiction because it’s been pre-vetted in some manner – you’ve been given the right to speak as to the fiction at this moment.

Narration is a crucial technique for many roleplaying games. One of the most spectacular games in this regard is Matt Snyder’s Dust Devils. One of the ideas introduced in the game was that while there was relatively little narration in other parts of the play procedure (stuff is introduced into the fiction not as narration, but by consentual negotiation dominated by the GM), a single player would be randomly chosen to narrate the outcome and immediate effective consequences of all conflicts. This powerful technique of suddenly giving plot authority to a single player in resolution of a conflict has since then been duplicated in dozens of games, and with reason: a single voice with authority may ensure that the conflict outcome is resolved in a consequential and suitably dramatic manner in a way that would not happen with nearly as much impact if the details were negotiated.

Narration is also used in traditional games, of course. I would even say that wide GM narration rights are a cornerstone feature of traditional rpg design.  Typically the GM would have you roll the dice and then tell you what happened as a consequence in the fiction; in this the GM is utilizing his narration rights. Some games and gamers would even extend the GM narration rights into full control over all aspects of the fiction.

(Although I don’t describe it, there are other places where narration rights routinely come up apart from conflict resolution. Scene framing is typically an example of narration rights as well: conditional on player veto, the GM narrates directly into fiction the set-piece from which play resumes after a scene change. The difference is mostly in which authorities are executed via pure narration as the concrete activity.)

Conch-passing

“Narration sharing” is when the narration rights are distributed dynamically, such as in Dust Devils. It’s an exciting technique, but one that I see being used in pretty futile ways now and then. Specifically, there is a certain game that is the logical outcome of fuzzy narration-sharing in game design. It’s usually called “conch-passing” in these discussions, and it goes thusly:

At the beginning of a session of conch-passing a player has the conch, and that signifies that he has the narration rights. Therefore he starts telling the story. Certain rules procedures tell us when the conch is given away, and to whom; at its simplest the player just continues as long as he likes and then passes the conch to his left. The player who gets the conch now has all narration rights, so he continues telling the story where the first player left off.

Obviously enough game designers don’t usually follow narration sharing into its logical conclusion in conch-passing; we just say that a game is “essentially conch-passing” if it undermines other, more valuable, concerns inadvertedly by encouraging attitudes and playstyles related to conch-passing. Specifically, conch-passing lacks many features that are mandatory for different types of roleplaying games, so if your game reduces into conch-passing, you might find that the system no longer adequately supports roleplaying of the sort you wanted. There are games that are intentionally conch-passing as well: Once Upon A Time is a well-liked conch-passing game that complicates the conch-passing by introducing player goals, allowing players to steal the conch and by requiring various stock elements to make an appearance in the story.

Backstory authority and character advocacy

I’ll focus on one particular problem that seems to be becoming endemic in Internet discourse. This problem is when narration sharing is introduced uncritically into games that involve strong backstory authority and character advocacy, two very common roleplaying game elements. Let’s define these things first:

Backstory authority

Backstory is the part of a roleplaying game scenario that “has happened before the game began”. The concept only makes sense when somebody has done preparatory work for the game or is using specific heuristics to simulate such preparation in real-time. For example, if the GM has decided in advance that the butler did it, then that is part of the backstory – it happened before the player characters came to the scene, and the GM will do his job with the assumption that this is an unchanging part of the game, even if the players might not know about it. Similarly a player character’s personal history is part of the backstory in a game that requires such. Backstory is specifically separate from what might happen during play itself. We say that somebody has “backstory authority” if he is allowed to determine something about the backstory, simply enough.

Character advocacy

Players can have different roles in a roleplaying game. Typical overarching categories are “player roles” and “GM roles”, which are fuzzy and historically determined expressions of natural language. One type of player role is when the game requires a player to be an advocate for a single player character – this advocacy thing is an exact theory term, unlike the fuzzy concept of “player role”. When a player is an advocate for a character in a roleplaying game, this means that his task in playing the game is to express his character’s personality, interests and agenda for the benefit of himself and other players. This means that the player tells the others what his character does, thinks and feels, and he’s doing his job well if the picture he paints of the character is clear and powerful, easy to relate to.

Both of the above techniques are immensely popular in roleplaying games. For example, Dungeons & Dragons involves clear and strong backstory authority as a crucial part of the game: the GM not only should prepare a dungeon ahead of time for the game, but he is also allowed to amend and expand on his preparatory work during play on the premise that his task is to present the game world as fully as necessary for the players: there is no ambiguity about who gets to decide what is inside a treasure chest: unless somebody changed its contents during play, the GM refers to his notes or imagination and decides what should be in the chest. Only when player character history is involved might the GM ask a player’s opinion on a backstory issue.

Character advocacy is also a common ideal in D&D, although I do admit that there are readings of the game text where advocacy is not present. Still, it is common to claim that the ideal of the game is that each player invents a player character who is a full personality, and then represents this character’s choices and actions in the game for the benefit of the group. (An alternative reading might be that the player’s job is to create a mechanically powerful character that he then uses to win challenges set up by the GM. Thus this is not as entirely obvious as the above.)

There are also roleplaying games that do not involve the above concepts. For example, my game Zombie Cinema doesn’t have unified backstory or a backstory authority as such: the game is explicitly No Myth, as it presumes that the setting and characters and everything else in the game is only thinly present as story elements introduced by the players on the spot. Any backstory presented to the group is handled as an extension of other authorities the players wield, such as the right to describe a scene and thus imply things about the location’s past. To preserve story coherence and provide twists into the story the game doesn’t use a backstory, but rather evaluation of reasonability (if somebody starts spouting some wacky shit, you stop the game and talk about it) and pure imaginative inspiration (assuming your wacky shit is not too wacky, it gets into the game and thus changes the story on the spot, even casting doubt on what has gone on before).

Games without character advocacy can be tricky because traditionally game design has operated from the faulty assumption that all games involve an identical, overarching player role that only requires the player to “play the character”. As I mentioned above, it’s not even entirely clear what you’re supposed to be doing in D&D. My favourite example of explicitly non-advocating game design is the Turku School Manifesto, which is pretty clear about elevating subjective experience as the primary task of the player; the player’s task is not to display his character, but to experience it.

The problem

When we bring the above terminology together, I can finally express my issue: I think that mixing narration sharing uncritically with backstory-heavy games and advocacy-model narrativistic games sucks ass. Examples:

• My brother Markku likes narration-sharing a lot, narrating stuff is one of his big loves in roleplaying. Now and then he gets proactive about introducing various methodologies into his gaming, which often ends up with him asking his D&D players what sort of monsters they would like to meet in the next encounter. Of course it’s fine if he likes this (no intent to call Markku out here specifically), but to me it seems completely awry and awkward to break the GM backstory authority and allow the players to narrate whatever they want. There’s no excitement and discovery in finding orcs in the next room if I decided myself that there would be orcs there. This fundamentally changes my relationship to my character.
• Somebody at Story Games suggested in relation to 3:16 (don’t remember who, it’s not really important) that a great GM technique would be to leave the greater purpose and nature of the high command of the space army undefined so the players could make this decision when and if their characters find it out. So maybe they find out that the great space war is a hoax or whatever. I find that this is completely ass-backwards for this sort of game: the players cannot be put into a position of advocacy for their characters if those same players are required to make the crucial backstory choices: am I supposed to myself decide that the space war is a cruel lie, and then in the next moment determine how my character is going to react to this knowledge? Doesn’t that look at all artificial?
• In another thread a similar claim was made about Trail of Cthulhu – that is, somebody described how he’d played the game with the players having the right to invent backstory by paying points for it. I’m not that vehemently against this in this case, as I don’t know ToC that well. Still, I’m almost certain that this is not the intended reading of the game text, and it definitely deviates quite a bit from how the game works if you assume an objective, GM-controlled backstory. My first instinct would be that I wouldn’t be that interested in playing the game if there weren’t a carefully considered, atmospheric backstory to uncover; it’s an investigation game after all.

The “Czege principle” is a proposition by Paul Czege that it’s not exciting to play a roleplaying game if the rules require one player to both introduce and resolve a conflict. It’s not a theorem but rather an observation; where and how and why it holds true is an ongoing question of some particular interest.

The problem we have here, specifically, is that when you apply narration sharing to backstory authority, you require the player to both establish and resolve a conflict, which runs counter to the Czege principle. You also require the player to take on additional responsibilities in addition to his tasks in character advocacy; this is a crucial change to the nature of the game, as it shapes a core activity into a completely new form. Now, instead of only having to worry about expressing his character and making decisions for him, the player is thrust into a position of authorship: he has to make decisions that are not predicated on the best interests of his character, but on the best interests of the story itself.

This is pretty much just my own opinion, call it an observation – I think that a logical division of tasks is important for a roleplaying game to such a degree that it actually prescribes and explains much of what we find interesting in the game in the first place. Specifically, I find that the riddle of roleplaying is answered thusly: it is more fun to play a roleplaying game than write a novel because the game by the virtue of its system allows you to take on a variety of roles that are inherently more entertaining than that of pure authorship. This is why many people find conch-passing games to pale next to a proper roleplaying game; the advocacy/referee/antagonism division of responsibilities is simply a more dynamic, interactive, emergent and fun way of crafting stories than undiluted and complete dramatic control for many of us. Authorship is work, advocacy is game.

Some games explicitly choose to place players in an authorship mode, conch-passing style. A pretty recent example of a major game in this vein is Houses of the Blooded: I find it a remarkably clear exposition on a game where the author truly believes that a roleplaying game is improved if you require the players to engage the game at all times with the best interests of the shared story in mind. As an example of this thrust in the game, consider this: the conflict resolution mechanic of HotB does not actually resolve a conflict, it only determines which player will have the authority to narrate a resolution. Specifically and explicitly, if you win a conflict for your character, you may determine that the character will actually lose because you find it better for the story. Could the authorship agenda even be formulated more clearly in the rules? Those sorts of rules are completely counter to the idea of character advocacy, which requires a player to think of and express the needs and wants of his character; thus games that take care to formulate a strong advocacy position for the player do not require him to make choices that are contrary to the character’s interests.

As a point of comparison we might consider Dust Devils and Zombie Cinema, both of which make use of narration rights while preserving character advocacy. In both of those games the player’s task in character advocacy is separated by a procedural firewall from his tasks in dramatic coordination. This is drastically evident in Zombie Cinema, in which you might during one turn first be required to frame a scene with dramatic coordination in mind, then advocate for your character in the scene, then take on narrative authorship to describe how a conflict went down. The game makes use of three different “modes” a player needs to engage in rapid succession, but this does not dilute any of them (I think; I’ve met people who disagree) due to how the player needs to only consider each of them separately in different phases of the game.

Naive narration sharing

I’ll segue towards my point here (about time, I’m sure). What I described above as the problem is naive narration sharing, which is the belief that you can just take any old roleplaying game and execute its gamemastering role with an extra dash of mutual cooperative narration. It’s such a nice idea, isn’t it? We don’t want to be evil authoritative GMs, after all, so doesn’t that mean that we should be inclusive and bring everybody in on the authorities available? This is simply wrong as an universal claim: while there are games where narration-sharing is central, and there are games where you can add great dollops without breaking the game, there are also games that do not withstand it. D&D and 3:16 are just arbitrary examples here, there are many others, much more than games where you can make narration-sharing the central element. Conch-passing is not very popular in roleplaying, generally speaking.

In case the reader doesn’t know this, I’ll explain next how narrativistic roleplaying games usually work. This explanation should amply illustrate the problems of naive narration sharing. In case you already know about advocacy + bangs = fun, feel free to skip the next bit.

The standard narrativistic model

I won’t explain what narrativism (Story Now) is here; if you don’t know, find out.

The above discussion on broken advocacy is very relevant to the whole field of narrativistic games for one simple reason: all but the most experimental narrativistic games run on a very simple and rewarding role distribution that relies heavily on both absolute backstory authority and character advocacy. If you introduce shared narration into this model, the game breaks up into consensual conch-passing (a particularly annoying form of conch-passing, that) almost instantly.

Here’s how games like Sorcerer, Dogs in the Vineyard, some varieties of Heroquest, The Shadow of Yesterday, Mountain Witch, Primetime Adventures and more games than I care to name all work:

1. One of the players is a gamemaster whose job it is to keep track of the backstory, frame scenes according to dramatic needs (that is, go where the action is) and provoke thematic moments (defined in narrativistic theory as moments of in-character action that carry weight as commentary on the game’s premise) by introducing complications.
2. The rest of the players each have their own characters to play. They play their characters according to the advocacy role: the important part is that they naturally allow the character’s interests to come through based on what they imagine of the character’s nature and background. Then they let the other players know in certain terms what the character thinks and wants.
3. The actual procedure of play is very simple: once the players have established concrete characters, situations and backstory in whatever manner a given game ascribes, the GM starts framing scenes for the player characters. Each scene is an interesting situation in relation to the premise of the setting or the character (or wherever the premise comes from, depends on the game). The GM describes a situation that provokes choices on the part of the character. The player is ready for this, as he knows his character and the character’s needs, so he makes choices on the part of the character. This in turn leads to consequences as determined by the game’s rules. Story is an outcome of the process as choices lead to consequences which lead to further choices, until all outstanding issues have been resolved and the story naturally reaches an end.
4. The player’s task in these games is simple advocacy, which is not difficult once you have a firm character. (Chargen is a key consideration in these games, compare them to see how different approaches work.) The GM might have more difficulty, as he needs to be able to reference the backstory, determine complications to introduce into the game, and figure out consequences. Much of the rules systems in these games address these challenges, and in addition the GM might have methodical tools outside the rules, such as pre-prepared relationship maps (helps with backstory), bangs (helps with provoking thematic choice) and pure experience (helps with determining consequences).

These games are tremendously fun, and they form a very discrete family of games wherein many techniques are interchangeable between the games. The most important common trait these games share is the GM authority over backstory and dramatic coordination (I talk of these two extensively in Solar System, which is also a game of this ilk), which powers the GM uses to put the player characters into pertinent choice situations. Can you see how this underlying fundamental structure is undermined by undiscretionary use of narrative sharing? The fun in these games from the player’s viewpoint comes from the fact that he can create an amazing story with nothing but choices made in playing his character; this is the holy grail of rpg design, this is exactly the thing that was promised to me in 1992 in the MERP rulebook. And it works, but only as long as you do not require the player to take part in determining the backstory and moments of choice. If the player character is engaged in a deadly duel with the evil villain of the story, you do not ask the player to determine whether it would be “cool” if the villain were revealed to be the player character’s father. The correct heuristic is to throw out the claim of fatherhood if it seems like a challenging revelation for the character, not ask the player whether he’s OK with it – asking him is the same as telling him to stop considering the scene in terms of what his character wants and requiring him to take an objective stance on what is “best for the story”. Consensus is a poor tool in driving excitement, a roleplaying game does not have teeth if you stop to ask the other players if it’s OK to actually challenge their characters.

Conclusion

I hope I’ve outlined my argument in sufficient detail. What I’m getting to here is that I don’t find it convincing how lightly many GMs seem to give away their backstory authority even when playing games that absolutely rely on the GM’s ability to drive home hard choices by using these same powers. There are other types of game that have similar problems (D&D for instance has nothing to do with the standard narrativist model, but it still sucks for slightly different reasons if you make playing the setting a matter of group consensus), but that issue with narrativist games seems exceptionally clear to me on account of how very clearly these games are written and how well-known the theory of their function is – the only reason to introduce extra dashes of shared narration in these games is well-intentioned foolishness, it seems to me.

Towards a theory of player roles

Old hands at rpg theory might compare and contrast this treatise on player roles with the concept of “Stance” in early Forge theory. A player’s stance towards his character describes the decision-making interface he uses in controlling the character; for instance, “actor stance” means that the player only uses in-character knowledge in determining the character’s actions. I do not consider player roles and stance theory to overlap to any considerable degree.

I probably should write a couple of words about player roles, as the above exposition goes pretty deeply into that. This is something that Ron Edwards explicited in 2006 or so, I think; I’ll rephrase the pertinent bits from memory.

The tasks that need to be performed for a Shared Imagined Space to form and play to be executed successfully are various and can be categorized in different ways. There are something like 20 tasks that I could list right now and probably give examples of games that distribute these tasks differently among the players, too, so I think I’ll skip on the list. These are things like “knowing the rules”, “framing scenes”, “playing protagonist”, “playing support character”, “narrating consequences” and so on and so forth.

(There’s also the important concept of authority, and particularly the different types thereby. This post is getting a bit long, though, so I’ll leave that part implicit for now. The backstory authority thing up top should be enough for now.)

The important bit is that these tasks are usually not explicitly dealt with as a list in rpg design. Rather, what you usually do is that you consult your intuitive understanding of how roleplaying games have usually gone about things, you pick one of the models in your experience and then modify it until you get what you want. This is the way you get to something like Mountain Witch, which involves a pretty normal GM role (“normal” is, of course, not the same as standard – there’s wide leeway even within traditional design parameters on what the GM does or doesn’t do), except he doesn’t know the backstory.

Most games tend to go with a blurry game text that references “players” and “gamemaster”, but mostly rely in determining what these roles actually do on tradition and a close reading of what the game mechanics require the players to actually do. For example, almost no game actually says what it means for a player to play his character in teleological terms – what the player is trying to achieve in the short term by playing the character.

For these purposes it is useful to example games in close reading and find out what it is, actually, that the game requires of a player. This whole post has actually been an overview of how certain types of game require players to be engaged in the role of advocacy (“I play my character to express him into the story”) as opposed to authorship (“I play my character to fill the narrative role allotted to him”). Both are called “playing your character” in different game texts, but psychologically and practically they are rather different processes.

What other types of player roles do roleplaying games use and make rewarding for their audience? There are definitely different types of gamemaster roles, but the role of the common player-participant is particularly interesting in this regard: I understand very well how advocacy works psychologically and why it is fun to play the advocate while the GM rains down all seven sorts of shit on the poor character, but what other options are there? Aside from character advocation, I know of at least character authorship (as in The Houses of the Blooded), character substitution (as in a game of D&D played without advocation) and character identification (as in Nordic immersionism). Then there are games where everybody is some brand flavour of GM, like Universalis or the various common conch-passing games.

Sources

As usual, I don’t think that what I write here is particularly controversial in the context of Big Model rpg theory. I’m just focusing on certain aspects of the matter and restating things a bit, not saying anything that wasn’t said by 2006 at the Forge.