Hello! This is part of a series on analysing RPGs in light of broader theories about games. In part 1 I introduced the history of attempts to describe games, and in part 2 I focused in on the specific question of whether RPGs conform to the model of games devised by philosopher Bernard Suits, weighed against arguments by Vi Huntsman and C. Thi Nguyen that they aren’t.
Towards the end of that article, we zoomed through some of the ways RPG mechanics can vary, and zeroed in on the specific class of mechanic unique to RPGs (and certain types of wargame): what Vincent Baker described as ‘fiction->tokens’ and ‘fiction->fiction’ arrows, i.e. rules that deal directly with the imagined scenario.
To understand such rules, we need to dive into the question of unwritten rules. This will involve going over the history of RPGs a bit—albeit mostly the bits that I was either around for, or can read about in The Elusive Shift.
- unwritten rules, rulings, and interpretation
- the procedures of RPGs
- so how does the DM learn to make these decisions, then?
- ignoring people
- the nature of RPG rules
unwritten rules, rulings, and interpretation
Last time, I mentioned Stephen Sniderman’s article on unwritten rules in passing—now it’s time to really get into it.
Writing in 1999, Sniderman insightfully observed that every single game has numerous ‘unwritten’ rules along with its explicit, formal rules. For example, in tic-tac-toe, if you procrastinate and do not take your turn in a ‘reasonable’ amount of time, the other player will feel you are playing the game wrong. Nowhere is it written that players must take a ‘reasonable’ amount of time, and the procedure for deciding what’s reasonable takes into account an infinite amount of context that could not be explicitly described.
For roleplayers, it is interesting to see Sniderman bring up the question of ‘rulings vs rules’, since ‘rulings over rules’ has become a key axiom for the OSR movement. Rulings here are in-the-moment decisions about how to resolve a situation not explicitly addressed by the rules. For Sniderman, this mainly involves decisions about how to deal with outside context or improbable events intruding on the game, like a fire alarm going off during a chess tournament, a piece of equipment malfunctioning, or players swarming a baseball pitch.
The unwritten rules also cover the subject of ‘sportsmanship’. There are endless actions one could take which would obey the letter of the rules but would be rejected as absolutely against the ludic attitude… at least if they’re done too much, because there’s usually an equally nebulous leeway for accidents. (By contrast, many sports have explicit rules for handling fouls with scaling punishments like temporarily sending a player off the field, and the result is that fouling is sometimes the correct strategic play if the benefit is better than the potential penalty!)
The unwritten rules also cover the rituals for entering the game and exiting it. Sniderman describes what that would look like as an explicit procedure.
When my buddies and I play tennis, we meet each other at the court at a prearranged time, take out our tennis racquets and some balls, warm up for 15-20 minutes (hitting ground strokes, volleys, overheads, and serves), and eventually someone asks, “Ready?” or perhaps “Ready to play?” If anyone says no, we continue to warm up. If everybody says Yes (or nobody says No), we toss away all but three balls. At this point, I (and presumably the others) understand that the actual game is going to begin with the next serve. There is never a formal announcement that play is about to begin. At most, the server will hold up a ball and the others will nod or wave.
Such procedures can be formalised; Meguey Baker’s 2021 article ritual in game design discusses an approach to doing so. In general, RPG groups will develop their own habits: for example, perhaps the DM says ‘So, last time…’ and starts talking about characters and at that point the game is on.
Similar informal procedures cover situations where a server in a tennis game might be invited to ‘take two’ serves because their concentration was disrupted for ‘unfair’ reasons. Sniderman observes the various unspoken social rules about which of the official rules should or should not be followed, what subjects are fair to talk about during a game, how to judge whether a ball is in or out, and so on. To fully specify every aspect of the game would lead to an infinite regression of rules… and also ‘meta-rules’ for resolving rules disputes.
What makes this work, Sniderman says, and avoids players bogging the game down in interminable arguments about rules, is that the players act as if there are various ‘meta-rules’ to do with the overall purpose of the activity even if they have never been formalised. Broadly, they are there to play the game, not argue about rules; so they must suspend disbelief and act as if the rules are tractable, working to maintain the magic circle even as they play tactically within it.
In other words, we can operate on (at least) two distinct levels of cognition at once. We can play any game as if it had an autonomous existence, even though we know perfectly well that the players create the game each time they agree to play and that any player at any time can destroy the game by quitting, by arguing, by stalling, or by any number of other spoilsport tactics.
I find Sniderman’s article really interesting… in general! but also for this subject, because so much of what happens at an RPG table is governed not by explicit rules printed in a book or worked out beforehand by the players, but by the norms of a play culture that has been learned by observation and imitation…
the procedures of RPGs
An RPG is played through a conversation. It’s not unique in this—board games and card games are also played at least in part through conversations—but we’ll need to take a look at the dynamics of that conversation to understand how RPGs work.
The procedures that make up a game—who speaks when, who has narrative authority over what, what sort of utterances count as game moves—became a particular focus for game designers in the 2010s, but of course the question goes back right to the beginning. So let’s zoom out from most of the rules that designers tend to focus on, and take a moment observe the overall structure of RPGs. We’ll start, once again, in the 1970s.
the procedure of oldschool D&D
In Chapter 2 of The Elusive Shift, Peterson observes that exactly how you sit down and play D&D was not actually spelled out by the original rules. In lieu of an explicit procedure, the book provided an example dialogue between a caller and a referee:
Both speakers phrase their statements as contributions to a common story, as if they are taking turns adding sentences to a fictional work in progress—in the course of the transcript, neither challenges the other’s authority to make any utterance.
But the referee speaks in the second person, and the caller in first-person plural. At certain times, the referee rolls dice which lead them to say ‘you hear nothing’ or ‘you hear shuffling’. Before long, the players explore a room…
By way of ascertaining what the room looks like, the caller sets the party to work with statements such as “We’re examining the walls, ceiling, floor, and contents of the room itself.” The referee then provides a detailed description of the gnolls’ lair, and through a process of steady inquiry into the details of the space, all phrased as specific actions such as “Each trunk will be opened by one of us” and then “Check the trunk for secret drawers or a false bottom,” the intrepid caller uncovers hidden plunder.
This defamiliarising description of D&D highlights just what an odd game this must have seemed in the 70s. Peterson pulls out another example from the 1880 wargame Strategos as a precursor, which explicitly says that players must give a ‘positive statement of intention’. For D&D players in the 70s, however, this principle appears to have been possible to intuitively grasp and imitate.
It took a couple of years for game texts to address explicitly how the caller should consult the other players to determine the party’s action (though of course, by this point, the caller role was on its way out). Also in question was how long the party should be allowed to deliberate before stating their action. As with everything, this had many different schools of thought, from ‘basically unlimited’ to attempts to limit discussion to the brief time passing in the fiction, 10 seconds or so.
All of this stuff was therefore the province of ‘unwritten rules’ and ‘play culture’. But this was fine: the game was learned by watching other people play and then joining in, so in those early days there was no need to lay it all out explicitly. Still, in this sense and many others, Peterson writes that the original D&D was more of a ‘build a game kit’ than a game, and hence it resulted in many different interpretations.
the procedures of modern RPGs
Modern RPGs have evolved a bit from this template, but most of them more or less work the same way. The game is still a conversation. The ‘caller’ roll has largely been eliminated, and players generally speak for their own character, and have more or less explicit limits over what they can narrate.
As well as declaring their actions, they might speak ‘in-character’; this may or may not be prefaced with narration like ‘(Character name) says…’. If the game involves various languages being spoken, implicit translation conventions will usually be in effect, sometimes addressed explicitly (“I say ‘I surrender! Don’t shoot me!’” “Is that in Elvish?” “Um, yeah.”)
Besides narrating their character’s actions in a prose-like way, players will now also frequently invoke game mechanics. In D&D: “I will use my bonus action to activate ‘Flurry of Blows’, that costs a ki point. I got 14 and 17 to hit.” In Apocalypse World: “I think that counts as Going Aggro?” This class of utterance is similar to a board game; declaring it is a speech act that activates that game effect. How these game-mechanical declarations relate to the shared fiction varies a lot depending on the game.
At most tables, the players are constantly slipping between ‘game speech’ and ‘non-game speech’ and the players are adept at understanding which class of utterance is being made. A player might crack a joke, not intending that their character says it; as well as prefacing speech with narration, they might adopt a slightly different ‘character voice’. Occasionally this boundary will not be communicated, and a player might respond to an ‘out of character’ utterance ‘in character’, but usually this is easy to resolve.
If they are playing in text chat rather than verbally, different types of statement might be indicated by formatting: out-of-character statements or game-mechanic invocations might be written [in brackets] ((of various types)), in italics, in different channels, or whatever else the game supports. There might be a norm that, for example, character actions be narrated in the third person. But if a player slips up and forgets the formatting, or a novice isn’t yet familiar with the customs, it is usually easy enough to figure out their intent and correct for it.
the exhaustive ‘trad’ designs
Let’s continue our history in more detail. I’m going to pass over AD&D, and the designs of the 80s and 90s in general, because I wasn’t around to play them, and I don’t have any Elusive Shift-style book on them history. However, as far as I understand, it was a much more mechanically complex game than original D&D, attempting to cover all sorts of different situations that might come up in a game, along with many more options for modelling player characters. It was still, however, a hodgepodge of different systems, something which other RPG designs attempted to correct. It seems like the 80s was when the idea of RPG rules as a ‘physics’ of the game world began to emerge, resulting in ‘crunchy’ designs like GURPS (1986) which also attempted to put everything on a unified mechanic.
By the 2000s, TSR went bust and D&D had been handed over to Wizards of the Coast, a company which had built itself selling Richard Garfield’s card game Magic: The Gathering. Magic and similar CCGs define a kind of interface through turn order and rules ‘keywords’, allowing the cards to be self-contained rules modules which can interact robustly with other cards. (More or less robustly, anyway! Tournament Magic has judges on hand to resolve tricky rules questions.)
This design proved incredibly lucrative: WotC could continually print new cards and cycle out old cards to evolve the game, meaning the players would perpetually need to spend money.
So perhaps it’s not a surprise that D&D 3e (2000) and 3.5e (2003) evolved into a similar generic framework, into which vast arrays of modular rules elements like feats, spells, classes etc. could be inserted. These were printed in dozens of books to be used along with the main D&D rulebook, known to players as ‘splatbooks’. On top of that, WotC came up with a clever device to cement their monopoly: the ‘Open Gaming License’ invited other publishers to design D&D-compatible rules and advertise them as such. And while D&D was, as ever, by far the dominant RPG, other RPGs by major publishers such as White Wolf also adopted extensible paradigms like this.
These modular mechanics were explicitly open to the player, so it was natural for players to make ‘rules declarations’, like ‘I power attack for -3’.
Despite this much more expansive formal system, playing such RPGs still involved a huge amount of ad-hoc adjudication to relate the fiction to the formal procedures of the game. Even so, it also attempted to be fairly exhaustive. The Player’s Handbook and Dungeon Master’s Guide attempt to cover all the different situations that are likely come up in a game, and recommend mechanics, such as difficulty classes for rolls, in each situation. ‘Here’s how you handle drowning’.
I’m not sure exactly how early this started, but certainly by 3.5e D&D, the play culture strongly favoured a precise, legalistic interpretation of these rules as a standalone formal system, and D&D players would approach the game in a similar way to Magic, searching for obscure rule combinations to exploit—and they found them in droves. As such, many players at this time particularly loved to talk about edge-case interpretations of the formal system, even when these would never take place in a plausible game.
For example, in D&D 3.5e, characters in combat would take turns, with the option to delay their turn until another character had taken theirs. Also, a character could freely pass an object to another character in combat. So, someone came up with a concept of the ‘peasant railgun’. A large group of characters would stand in line, and each would delay their turn until the previous person had taken theirs, then when their turn came up, pass an object—a stick, for example—to the next player in the line. With this technique, an item could be moved an arbitrary distance within a six-second combat round. After this, the argument would go that the stick would be moving an enormous speed and therefore it could be released like a railgun projectile… somehow slipping from unphysical D&D rules to real-world physics.
Nobody really thought they would do this in an actual game of D&D (I mean, maybe as a joke), but it reflects an attitude towards the game rules as the ‘physics’ of the fantasy world, similar to a computer game simulation. So, with the theoretical toolkit-katamari we’ve been building up, what’s gone wrong here? Well, the norm or ‘meta-rule’ of the game is that the rules should be used to create a compelling fantasy story. This is clearly out of the domain of applicability of the D&D combat rules. Even if you could argue it’s a correct interpretation of the rules as printed, it would be a violation of the spirit of the game.
If someone actually tried to create a ‘peasant railgun’ type situation in a D&D game, what would happen? Well, if I was DMing it, I’d come up with an ad-hoc ruling that better represents the fiction of the game. I would decide how quickly the item can move along the line, perhaps taking into account factors like the dramatic tension of a time limit, or perhaps looking up information on real-world bucket chains, depending on the situation.
In short, I would be doing a small bit of on-the-spot game design.
In the late 2000s, 3.5e gave way to 4e, which presented an even more self-referential formal system, focusing the game on tactical combat—something which led to a whole lot of arguments. It was also around this time that two major splits happened to create what would become the two major poles of indie games: the Forge/story games tradition, and the OSR.
the OSR and their rulings
The OSR (Old School Revival/Renaissance depending on who you ask) was a movement that began around 2008 aiming to return to a certain ideal of ‘old school’ D&D (and roleplaying in general), dispensing with all the unnecessary complexity of modern editions. There are many different claims of what, exactly, the OSR entailed, and many different games that belong to the tradition in some fashion. Here’s a historical overview written in 2021 which seems reasonably thorough and demonstrates how incredibly nebulous and fractured it became. Much like the demoscene, the OSR would be pronounced dead with some regularity.
Conventionally, OSR games revolve around weak characters trying to explore a dungeon full of dangerous enemies and hazards. The characters are vulnerable and character death is likely, so rather than aim to kill everything that moves as a 3.5e party might, they are expected to sneak, run away from strong enemies, and exploit their environment for every advantage.
One of the core fixations of the OSR was playstyle. This article from 2022 by Travis Miller lays out two of the main contentions reasonably succintly: ‘rulings over rules’, and stating your actions in accordance with the fiction rather than invoking mechanics. (The other major facet is to use the original D&D rules, or something closely related to them.)
‘Rulings over rules’ appears to date back to at least around 2008, when A Quick Primer for Old School Gaming was published; mind you, it soon attracted criticism as well, so it can’t be taken as the only ideology of the movement here!
Still, essentially the principle is this: the rulebook should not attempt to cover every single scenario that might come up over the course of a game, but instead give a general framework which the referee/DM can use to scaffold an ad-hoc, situational approach based on what they think is reasonable. (‘New-school’ games also provide such a framework, indeed nearly all traditional RPGs provide a list of skills which can be rolled as situationally relevant, but the OSR tradition tends to be more simplified.)
This sits in an interesting tension with the fact that the OSR also favoured highly lethal, ‘striving’ play in Nguyen’s dichotomy. Just like ‘Free Kriegsspiel’, the players would be trying to predict how the referee would interpret their requested action and present their action in the most favourable light. Players might learn that the referee in their game expects them to describe checking every nook and cranny of the dungeon in explicit terms, as in the Quick Primer.
But does an OSR referee really have absolute license to decide what happens and what rules to adjudicate it with? No, they have ‘unwritten’ expectations, of the sort discussed by Sniderman. A referee who hated a certain player and always made them fail, no matter what, would likely be seen as refereeing the game wrong. Similarly, a referee who undermined too much the illusion of a consistent fantasy world without good reason would be failing at their task, and probably wouldn’t stay the referee for long.
so how does the DM learn to make these decisions, then?
In certain story games like Apocalypse World, the text of the rules specifies a fairly explicit procedure that the MC is expected to follow. Most of their actions are supposed to be made through ‘moves’; they further have a list of ‘principles’ to guide what they say. Some are quite specific, like addressing characters rather than players; others are more abstract suggestions like ‘barf forth apocalyptica’. Under certain circumstances, like a failed player move, the GM should make a ‘hard move’ that harms a character. There is also an explicit procedure for setting up the first session and a set of tools for modelling the game world behind the scenes.
With such scaffolding, Apocalypse World is considered a relatively easy game to play in an improvisational way with minimal prep. Other story games will eschew the GM role entirely.
On the other hand, trad games such as D&D hand off a large amount of latitude and creative decisions to the DM. The DM is left to create the scenario, figure out an opening scene, decide how best to model the world and so on. As such, D&D games have a large space of variation, depending on the tastes and methods of any given DM.
The books do give extensive advice, but it’s no surprise that a player who has not seen the game in action would still have little idea of what to do ‘in the moment’. To help give the players an on-ramp, both TSR and WotC published ‘modules’ or ‘adventures’; many other games do the same. These define a scenario and provide maps, text to read aloud, and whatever else is thought to be needed to run a game. They also act as a demonstration of what the DM might ultimately create for themselves, if they end up ‘graduating’ from modules to creating their own game.
The community at large would also create huge amounts of discussions, play reports, ‘actual play’ podcasts, tools, and so on. Game-agnostic methods, like structuring a game as a ‘hexcrawl’, would be spread not by officially published material but community word-of-mouth. Growing up in a small town in the UK without any local gatherings of roleplayers, it could reasonably be said I learned how RPGs work from forums such as Giant in the Playground and took those received ideas to run my first game of D&D with school friends.
It is fair to say that a large part of what a GM is doing is game design. A common criticism of games in recent years is that they are incomplete designs, leaving too much work to the GM that should have been the designer’s responsibility. That’s a huge question, though, so we’ll come back to it.
how free is a DM?
In trad and OSR games, and even some story games, the GM’s heavy burden of work is balanced by their absolute authority. By both explicit rules text and widespread convention, they get to be the final say on what happens in the game. They can freely change the rules and whatever they say is the ‘truth’ of what happens in the fiction.
However, they are not completely free. Should they fail to engage the players, the game will probably not last. The GM’s position is in this respect like a performer before an audience. Their authority exists only within the magic circle, and it is broken when the circle dissolves.
Since the players know the GM is not completely free, and that they are familiar with the game’s broader play culture and perhaps the GM’s personal style, they can make predictions about what sort of rulings the GM is likely to make. So, even in the contexts that the rules do not address, there are some constraints created by the dynamic system and room for strategically pursuing goals. Indeed, this could be viewed as a ‘hidden information’ game: the players are trying to discover the ‘rules’ implicit in a GM’s methods as they play the game.
ignoring people
There are many types of roleplaying game which do not involve a GM-equivalent and have all players in the same type of role. As well as story games like Fiasco or Dream Askew, roleplaying in MMOs or forums will often involve a shared responsibility for maintaining the integrity of the fiction.
How can this work, when at any point, any player could show up and narrate something that is not in the spirit of the game? Quite simply: the other players, either implicitly or by explicit agreement, ignore that person’s contribution and carry on their own narration as if it did not exist.
In an MMO RP venue, players may be carrying out conversations in-character in small groups. All the narration and dialogue is inserted into the same chat log, rather chaotically, but players soon become adept at filtering out the parts that are relevant to them, or perhaps spin up a private group chat for some mechanic assistance.
In this case, the situation is very much like improv theatre: a character will make an offer by declaring something, and another player might respond to that offer by acknowledging it with their next action, creating a new offer. In this case, Bernard Suits is not entirely wrong to observe a goal of continuing the RP indefinitely: a good RP statement gives the other player something concrete to work with.
A player will sometimes complain that nobody wants to engage with them, and often this is because they haven’t fully understood the norms of the RP space: breaking with its aesthetic and narrative assumptions, violating some of the few explicit rules like ‘don’t narrate another character’s action’, or interfering with the pacing of another player group’s scene. In this case, the other players might issue a correction, or just shut them out.
So players in a ‘freeform’ RP are not completely free either: there are nebulous, unwritten rules which must be followed or they will be pushed out of the game.
the nature of RPG rules
This survey of procedures has hopefully hinted at the murky depths of the ‘system’ shaping what sort of narrations people are likely to make in RPGs. Even equipped with game books with hundreds of different rules covering all sorts of different situation, because RPGs involve the ‘fiction’ as a core element, players of RPGs will be applying nebulous, imprecise judgements of what it is appropriate to narrate when their moment comes up and when to invoke game rules. They may be motivated by an idea of the ‘reality’ of a fictional situation, by genre convention, by respect for the other players’ boundaries; they may be influenced by the last book they read or other games they’ve heard about. In a certain sense, the players are designing the game as they play it.
And yet, there are many parts of RPGs which are constrained by a rigid, formal-ish system. RPG rules can forbid certain classes of narration or utterance, and call others into being by conjuring up game constructs or elements of the fiction. This creates the space for the type of ‘play’ described by Salen and Zimmerman, the ‘free movement within a rigid structure’. Like Sandra Snan’s ‘bones of steel in a cloud of description’, the rules give the players a scaffolding to support their narration.
So, in the next article, let’s look at how some RPG rules work, and how they compare with old Bernard’s concept of a ‘constitutive rule’.
Comments