So, Game of Thrones is a thing that’s on television that a lot of people watch. I don’t, but I do know people and follow people on social media who do, and so I get exposed to some reactions to the show. A recent episode lead to this twitter thread:

//platform.twitter.com/widgets.js

The gist of it is that characters have no agency because the writer is the one in control of the whole thing. Now the argument of character agency is usually brought up when the writers are making bad decisions (see: any argument over the design of a female video game character). The truth is, no matter the in-character reason for a character’s decisions, there is always the hand of the writer behind that.

But as always, it’s more complicated than that.

Watson v Doyle

There are two views of why something happens in a story. Okay, there are more than two views, but there are two complimentary views that I’m going to talk about. There’s intradiagetic, or ‘Watsonian’ and extradiagetic, or ‘Doylist’. The terms ‘Watsonian’ and ‘Doylist’ come from Sir Arthur Conan Doyle’s Sherlock Holmes stories, where Doyle was the writer, but Dr. Watson was the in-universe chronicler of Holmes’ exploits. A Watsonian take on a piece of literature would be to explain everything that occurs within the bounds of the universe that has been created. A Doylist analysis would be examining the author’s choices as it pertains to the work as a piece of literature. Watsonian analysis usually takes place within “fan” spheres, while academic analysis tends towards a Doylist perspective, but this is by no means a hard and fast rule.

As an example let me give you a Watsonian and Doylist take on a famous character moment in literature: Hamlet not killing Claudius in Act 3 of Hamlet. The scene: Claudius has just walked out of the play that Hamlet had the players act to try and get a reaction out of Claudius (the “catching the conscience of a king” bit). Claudius has a brilliant soliloquy where he confesses the whole thing and shows his thoughts and anguish over it. He kneels, as if to pray, and Hamlet enters, having been spying on him this whole time. Hamlet raises his weapon, but hesitates, monologues, and does not kill Claudius. He leaves before Claudius reveals that his prayer was meaningless and not heard by the ones he was praying to (angels, saints, etc.).

The Watsonian analysis is stated outright: “A villain kills my father, and for that / I, his sole son, do this same villain send / To heaven.” (3.4.76-78). Hamlet believes that by killing Claudius while he is praying, he takes the chance that Claudius is being absolved of the sins of killing King Hamlet and wedding Queen Gertrude, and will be sent to Heaven because of this absolution (there’s a huge amount of Catholic text and subtext in Hamlet, if you haven’t read it in a while). Hamlet further explains that he will kill Claudius “…about some act / That has no relish of salvation in’t” (3.4.92-93), and that if he waits for that “…his soul may be as damned and black / As hell to whereto it goes” (3.4.94-95). He wants to make certain that the man who murdered his father is punished for it for eternity.

The Doylist analysis could be considered simple: Hamlet doesn’t kill Claudius here because it’s only the third act, and there’s two more acts with pirates and talking to skulls that we have to get through. But it is more than just that– there is the aforementioned Catholic theme of the play that Will is continuing here, as well as the theme of Hamlet’s indecision and obsession with the mechanics of death. Depending on the actor, Hamlet can here be a cunning strategist deciding the right moment to strike, or a scared kid who is trying to make an excuse for why he doesn’t want to kill someone. There’s also the dramatic irony that Will sets up, as the audience knows that Claudius is in no position to be absolved, and if Hamlet killed him now, it would be exactly what he wants.

So, does Hamlet have agency? Well, no, not really. Because no matter how well an actor gets into the role, or how realistic the play is being performed, or how internally logical it would be for him to do so, Hamlet will never kill Claudius in Act 3, because the text says that Claudius doesn’t die until he’s poisoned in Act 5. Or, if you’re Kenneth Branagh, when he has a chandelier dropped on him in Act 5. But that’s if we only take the Doylist approach. The Watsonian approach shows us Hamlet’s character and conflict and motivations, something Will was always really, really good about explaining, and something that is a key element of Will’s plays. The Doylist approach responds to the question “What’s my motivation?” with “Doing what the script says.” There’s more to fiction than that.

When this argument actually comes up

Suffice to say, people don’t typically pick out Watsonian and Doylist interpretations of Hamlet. Or, at least, no one I know does that. The two analyses build on one another are are typically viewed as equally important. As I’ve observed, the Watsonian v Doylist interpretations generally occur when the quality of the work is in question. In other words…

Yeah. The argument doesn’t typically go “well, let’s discuss this by taking apart the in-universe and out-of-universe reasons for this character action” but more like “the writers/actors/editors/etc. made bad decisions and made me aware of them, so now I’m going to pick this thing apart”. It’s not about nuance, it’s about quality.

Is that fair? Well, maybe. Unless the writer is intentionally making themself visible to the audience (see: the Monty Python gag where they’re saved from a dragon because the animator dies), if everyone’s doing their job right, you don’t think about how it was made over what is happening in the story. I mean, unless you’re like me and can’t turn it off.

So, some hypothetical scenarios:

  • Character A does something stupid that almost gets Character B killed (or actually gets Character B killed)– generally good writing, though it relies on the writer to follow through with character development. Some consequences are always better than none. EX: Harry and Sirius in Order of the Phoenix
  • Character A does something mean to Character B, with no motivation or ill-will between them– generally not good writing. If they already have an established relationship and are on good terms, there should be no reason for this to happen. The audience should have some clear foreshadowing, at the least. EX: Hans and Anna in Frozen
  • Event C happens, making Character A trust Character B more– generally good writing. If the characters don’t have a good relationship, there should be some time spent on learning to trust each other. EX: Katara and Zuko in Avatar: The Last Airbender
  • Character A, a novice, is shown to get better at a task than Character B, an expert, over the course of a quirky montage– it’s not really great writing, though it is rather commonplace. It’s often used to justify the main character’s special-ness or something. EX: so many, though Kung Fu PandaAnt-Man, and The Matrix

And, as always, there are exceptions to this, and there are good stories that have bad moments.

But back to agency, in good writing, the writer has clearly shown a character’s motivations. Ariel wants to visit the surface world. Dorothy wants to go somewhere over the rainbow. Everyone in Into the Woods wishes for something. Then the plot happens. With good writing and clear character agency, a character will be forced to examine these wants and make a decision– Ariel making her deal with Ursula, Dorothy immediately wanting to go back home when she gets to Oz, the whole second act of Into the Woods is about how their wishes aren’t always everything they thought they’d be. In fact, all of these being musicals, they all have these characters explaining these motivations in an “I Want” song. How to convey character motivations is a post unto itself, but regardless, there’s a particular distinction made in theater that facilitates the audience’s understanding.

So do characters have agency?

Characters have agency when characters are allowed to have agency. Ultimately, when it seems like the writer is less of an invisible hand and more like a railroading GM that needs the chess pieces they call characters to be in particular places to serve their plot, odds are the characters have less agency. And if that’s the experience you’re going for, then great, but it might mess with the audience’s suspension of disbelief. And when that happens, they will start to notice that it doesn’t make sense for these characters to be in these situations.

And maybe that doesn’t matter to you, but it matters to me, as a writer and an audience. Sure, I want a plot— characters just standing around and talking isn’t exactly my genre– but I want that plot to service the characters. One of the things I love about the much-maligned Pirates of the Carribean: At World’s End is that by the time we get to the final fight all of the motivations have been lined up and we understand why the giant ship battle is happening. Is it the greatest movie ever? No, but even in its moments of absurdity I can follow the characters, which is also why I like it better than Dead Man’s Chest.

I think it’s important to consider this stuff about agency because 1) it’s a great way to detect weak writing and because 2) it’s a great tool to use to deconstruct any writing. No, fictional characters are not making real decisions, but they are making fictional ones, and what matters is whether or not the writer can convince you that the characters are making decisions. Because you’re the one they’re trying to entertain.

I can’t write your decisions, but if I could, I might have you like, comment, or subscribe. I have e-mail subscriptions too! In the sidebar between the tags and the archives.