I’m not one of those designers who gets hung up on terminology. Jargon has its place, but if we spend all our time defining or debating terms, we’ll never get to the real ideas. On the other hand, if we don’t ensure that we agree on the meanings of terms we’re using, we’re very likely miscommunicating.
Sometimes the term “player agency” is used for something that personally I would call “narrative control.” I see these as different from each other, and explaining why will allow us to move into discussing both concepts in terms of a game’s mechanical design, as well as in worldbuilding.
I refer to player agency as the freedom of choice a player has in regard to their character’s actions. Say a GM provides a description like, “The starship’s bridge has a small side compartment. When you open that compartment, the terrifying alien leaps out and attacks you.” If the players didn’t say they were opening that compartment, the GM is infringing on player agency in a big way. When a GM “railroads” the players—which is to say, gives them only one option at any given time, so they stick with the planned story—this also challenges player agency, perhaps just not as blatantly. In these cases, these are GM issues (or potentially adventure design issues), not game design issues.
Game design violates player agency most often when it limits a character’s in-game options. Prohibiting a wizard from wearing armor, even with—or perhaps especially with—an in-world justification, restricts player agency, for example. And for some games, or some gamers, that might be okay. It’s a very simple, low-impact way to enforce a genre convention (that being that wizards wear robes and knights wear armor). Trying to enforce that genre convention by saying that anyone can wear armor, but wizards have a big penalty when they do, is much friendlier to the players’ agency, but it’s less simple. Is it worth the extra mechanics? That’s up to the designer.
I designed Numenera (and the Cypher System that lies at its core) to enhance player agency with the concept of Effort. Effort is a limited resource that a player can choose to use when an action really matters to them (or, perhaps more appropriately, to their character). In other words, it distinguishes a player’s feelings about an action. Let's take a character who is good at climbing. From the PC's point of view, climbing a tree to get a look around and climbing over a wall to get away from a pack of horrific beasts chasing them are two very different actions. And yet, many games treat these two situations the same mechanically when the GM asks for a roll to determine success at climbing. If player agency could be defined as a player’s control over their own character, the mechanic of Effort increases that control.
Some people also call a player’s contributions to the setting player agency. For example, if a player gives the mentor who trained them in their backstory a name, they would say that’s player agency.
I, however, would not call that player agency. I would call it narrative control. And the only reason that I quibble about the terminology is that I think both player agency and narrative control are interesting but very different aspects to game design. You can have a game with lots of one and absolutely zero of the other.
For example, a game might never impinge on player choice but might never give them the ability to narrate aspects of the game world beyond the characters. Or conversely, a game might grant a great deal of narrative control to the players, such as dictating the actions of NPCs, inserting elements into the setting, and so on, but then only allow a set of prescribed actions or moves when it’s time for their character to act in order to manage the flow of gameplay.
When I’m running a game and a player says their character wants to find the local pub, I might ask the player, “Okay, when you find one, what’s the pub’s name?” That’s narrative control. Typically, these types of details are entirely the purview of the GM, but here, in a tiny dose, I’m ceding a bit of narrative control to a player. Now, I would do this only if it fits the game and if I know the players would appreciate it. If the player appreciates it, it might invest them just a bit more in the setting and a bit more in the story (because they’ve contributed to both). However, I know full well that for some players doing this would break their immersion. Much like we covered long ago, they are playing the game because they want to discover and explore a fully realized fictional world, not one that they contributed to.
Narrative control describes when anyone (or anything) dictates something that happens in the world outside the player characters themselves. Because the default for most games is that the GM has narrative control, it’s not worth even considering. That’s just the GM’s role in the game. However, once any level of control of anything outside the purview of a player character is granted to a player, that’s worth noting. These are the situations I’m likely talking about if I talk about narrative control. In other words, someone is usually controlling the narrative, but it’s only worth discussing if it’s a player.
Occasionally, neither GM nor player have narrative control. In such cases, the game system (usually in the form of the dice) has control. Rolling for “wandering monsters,” for example, is a form of narrative control that the game system might impose. The GM didn’t decide that a band of bugbears showed up, the dice did. Player death can also be a way the game controls the narrative, sometimes creating a situation that neither GM nor player really desires.
If a character is running after someone and there’s a dice roll involved, the outcome of that roll can involve the issues of narrative control and player agency. Should the player roll really well, and that roll is interpreted as their character leaping over an obstacle in order to catch up to an NPC they’re chasing, that’s granting player agency, assuming that the leap was the player’s idea. However, if the roll is interpreted as the NPC hesitating too long as they approached the obstacle, thus allowing the character to catch up to them, that’s narrative control being taken by the player (if the hesitation was the player’s idea).
This all goes back to the idea of the shared imaginary space and who’s dictating what happens in that space: the GM, the players, or the dice. It’s really all about whose ideas are being implemented and what they are changing in the imaginary space. If it’s something to do with the PCs, it’s probably a matter of player agency, and if it’s not, it’s probably a matter of narrative control.
As a designer, I am an absolutist when it comes to player agency. That is to say, players control everything about their character intentions and actions by design. I never want the game to tell a player what their character does, and I never want to empower or encourage a GM to do that either. When it comes to narrative control, I feel like it’s a knob that I can turn up or down depending on the game I’m working on. For example, in the game that I’m working on right now, Stealing Stories for the Devil, I’ve turned up the dial on the players’ narrative control considerably. Most sessions involve a heist of some kind, and when the characters are “briefed” on the mission ahead of time and make a plan, the players are actually granted a huge amount of narrative control. The players are literally designing the obstacles and challenges they know they will face (they know it because it’s a part of the briefing the characters are given).
I chose to do it this way so that, like in many heist stories, the characters are experts at what they do—even if the players are not. If a player knows their character is really good at climbing, they can use their narrative control at this stage of the session to say that there are air ducts through which they can climb and enter the building they need to infiltrate. By playing to their own strengths, they are not only going to showcase how excellent they are, but they are making a plan of action all at the same time. (Of course, the GM can add their own input into this process, always ensuring that there will be plenty of challenge—and there are specific mechanics for the GM to add in their own twists and turns once the mission has started to introduce the unexpected.)
But then I went a step further by designing the characters in Stealing Stories for the Devil with the ability to make small changes to reality. Using this supernatural power, if they need a thumb drive to copy some data onto, they can open a desk drawer and find one there. If they need to slip past a security guard, they can make it so he hasn’t gotten much sleep lately, so he’s nodding off as they go by. And so on.
This is, in a way, an in-world way to grant players narrative control. You could even say that by making it a function of the character, the game makes this (limited) narrative control into an issue of player agency. Reality changes the way the player wants it to, and the GM only decides how much stress this puts on a character. (Characters overreaching can knock themselves unconscious or worse, so most of the time, players reign themselves in.)
Aside from risking a player’s immersion, the main potential drawback to granting narrative control is one of game balance. Although in a game like Stealing Stories for the Devil, there are plenty of checks on the players’ control, it’s still a game in which the combination of a willful player and a lackadaisical GM could probably result in a less than optimal experience. But it’s a game designed for at least moderately experienced GMs and players, and more importantly, a game that tries to educate both GM and players regarding the intent of the game’s mechanics.
In the end, I think the terms player agency and narrative control are only important because the concepts, I believe, are distinct. Or, in some games, they certainly can be. Although I feel like a broken record, this isn’t really an issue of right or wrong choices, but rather things to be aware of. Both are concepts that will exist in every game, so the designer should determine their approach to both deliberately.
What is the difference between "narrative control" and "shared world building?" Your example of the name of the pub I would normally consider the latter. However I'm now thinking that the "shared world building", at least in this case, is the affect of "the narrative control." I like this, but, where would you get "shared world building" without "narrative control?" I would like to think that it is possible without it being an out of game activity. Also can you have "narrative control" without "world building?" -what would that look like?
Another question: Would you consider your example of the sleepy guard or USB key to be in the same vain as Blades in the Dark's heist mechanic, only for the present? Either way I like it - YOINK!
I take issue with some people deciding if it is not a sandbox where they can go anywhere and do anything, then they are being railroaded. Its never happened to me back in the day, but I get the impression its more common now.
If the DM spent two weeks writing material for his village, the forest and a dungeon, it is in poor taste if the players decide to choose a far away city on the map and just go there, forcing the DM to adlib and make up stuff with no preparation, and it is not railroading if he or she redirects them to their content.
Obviously experienced DMs, or module designers will be much more flexible with making it feel like a sandbox. But lets face it, preparing professional grade sandbox content allows dozens of options, each with different outcomes increases the work of the DM tenfold.