Can you optimize for tragedy?

There is an awesome conversation between hellfire, Ari and myself over in the comments on Line Hollis’ post “Moral Incentives and Story Structure“. Part of the conversation, as I understand it, is in trying to understand how and even if games can be tragedies — I argued mostly against it (sort of) in “All games are comedies“. However, as is increasing becoming the case, Ari hits on exactly what I was going to be writing about today in the following comment:

“It would be interesting to define what makes a game tragic…

I like the idea of a tragic fall being the result of the player’s flaws, but all players are essentially flawed in the sense that they don’t know everything that the character knows. Despite being the connection into the virtual world of the game, the player and the character are always two different beings. At best, you can direct and control your character, but you can’t fully become him/her/it.

It might be interesting to have a game in which you do have knowledge of the consequences, yet you choose to act in a way that leads to tragedy.”

Before we can even define tragedy in games, I’d like to roll back to the source of discussion, Aristotle and Poetics, by using this quote:

“Tragedy, then, is an imitation of an action that is serious, complete, and of a certain magnitude; in language embellished with each kind of artistic ornament, the several kinds being found in separate parts of the play; in the form of action, not of narrative; through pity and fear effecting the proper purgation of these emotions.”

We already see, from this definition, that a tragedy requires activity (“imitation of an action that is serious, complete, and of a certain magnitude”), that this activity make sense dramatically (“in language embellished with each kind of artistic ornament, the several kinds being found in separate parts of the play”), that it be acted (“in the form of action, not of narrative”) and that “through pity and fear effecting the proper purgation of these emotions.”

In more modern terms, the “pity”, in its pure form, can be thought of as a “tearjerker” (intense pathos, emotion) and “fear”, in its own pure form, as “horror”. (Put, in a different translation, as “pity is aroused by unmerited misfortune, fear by the misfortune of a man like ourselves.”) The last part — “effecting the proper purgation of these emotions” — we now call catharsis. It’s easier for me to think about it as the elation of going downhill, on a roller coaster, after the intense initial climb up — it’s scary (especially for me, I hate heights) on the way up but, once past a threshold, the audience is allowed to breathe easier and enjoy the experience.

Aristotelian Tragedy also has six parts (Plot, Character, Diction, Thought, Spectacle, Song) that I am not going to go into other than to say that an extremely loose translation of them is “Interesting (Spectacle) incidents (Plot) happen to agents (Character) within a certain logical context (Thought) which they then comment on (Diction) while occasionally exposition (Song) is needed to explain certain events.”

The trajectory of a tragic hero, in another loose sense, is the following: “likable but flawed hero has, during the events of the story, some deserved downfall most likely caused by his hamartia, “tragic flaw”, resulting in discovery or self-knowledge. (Oedipus’ flaw of hubris leads to him killing his father and eventually marrying his mother. The discovery, as you might imagine, is him finding out the truth of the situation.)

The problem, as I see it and to return to Ari’s comment, is in knowledge. In order to design for tragedy, the developers or even player must know several things:

  • Mechanics of the game
  • Entirety of the plot
  • All possible endings

Here is where I return to Ari’s “all players are essentially flawed in the sense that they don’t know everything that the character knows.” A classic example of this is Metroid in which the player goes through the events of the game thinking the person in the power suit is a male but, upon completion of the game, Samus is revealed to be a woman. (Yes, I know. This is a somewhat sexist example. It should not matter to the story — it does not in Metroid — if the character is male or female.) The player starts with their assumed identity and, as the story unfolds, the player learns more about the character — this fits into performativity but will not be covered here — and must adapt to that knowledge.

For a game to be tragic, it must one in which the player, as the protagonist, comes to the moment of discovery, the audience experiences catharsis but, in the end, the charatcer who was controlled by the player is “defeated”.

A great example was supplied by Rachel:

“Portal has GLaDOS very much still alive, and Chell being dragged off. Would that qualify as a tragedy?”

As I pointed out:

“The catharsis, release of emotion, happens for the player and character upon defeat of the enemy at the end. It is a comedy; Chell wins and finds her freedom on the surface!

Chell getting dragged away, as I understand it, was added a great deal of time after the original game came out. That would, since Chell has knowledge that her defeat was short-lived in the revised version, make it a tragedy.”

In the new revised version — How strange is it that Valve can change the ending of their game? — Chell getting dragged away makes it a tragedy as her discovery that she was lied to by GLaDOS then leads to the catharsis of GLaDOS’ defeat which then, in the new video, puts her on the surface free, only to then be dragged away. Chell’s flaw, as is the player’s by reflection, is in assuming that they know better (hubris) than GLaDOS. Just because she is a crazy AI does not make her a bad manager of the facility. (This ties into the plot of Portal 2 as Wheatley makes the same choice to supplant GLaDOS  to humorous results.)

What if the player wanted to design an experience toward tragedy — “you do have knowledge of the consequences, yet you choose to act in a way that leads to tragedy”?

Here is where knowledge of the entire plot and all possible endings comes into play. In order to plan for tragedy, you must have an understanding of where the game story can go but steer it a different way. You must shape the entire experience toward this singular end. At each possible junction, the answers must be known and supplied to direct the story path to this end.

The game I came up with as an example was The Path.

You can direct each woman toward their ‘wolf’ (discovery) in the forest — flaw of hubris, “I know better” — which leads to them being reset on the path but now walking in a slow, crippled pace toward Grandmother’s house. In fact, in order to complete the game, in one sense, you must do this action with each woman. Each of them must disobey the simple edict of “stay on the path” and explore the surrounding forest.

From Wikipedia:

“If the player did go to the Wolf, then everything in the house is darker, and if the player remains still for too long, darkness clouds the screen, and something growls. Depending on the girl, doors are scratched, or furniture tipped over and broken, or strange black threads are draped across everything. Instead of ending with Grandmother, the music crescendos as the player enters a final surreal room before falling down, and things black out again. Images flash on the screen, featuring the girl being [killed] by her Wolf, before the player is relocated back in the apartment. The girl played is not there, and will remain absent.

[Note: the writer of this entry supports the theory of the girls being raped at the end of their journey. I follow the ‘death of childhood’ idea from Tale of Tales’ own post-mortem —

“Some say blindly that the game is ‘about rape.’ And while that could be one of the interpretations -and I understand it-, for me, those black-out moments after meeting her wolf are the moments of realization. Those are the times when a girl grows. And what happens in Grandmother’s House is not a murder but a shedding of childhood and an initiation to womanhood”

— and have thus switched “raped” to “killed”. I’m not completely happy with “killed” either though.]

When all of the girls have encountered their wolves, a girl in a white dress, who could be previously encountered by the sisters, becomes playable and visits Grandmother’s house. The girl will then travel through the house, now a combination of all of the end rooms of the previous girls ending with the no-wolf room. Upon reaching the grandmother, the girl appears in the apartment covered in blood, but alive. The sisters all return through the door and the game starts over.”

While The Path is an extreme case, it is actually much easier to optimize for tragedy: simply have the character die. One of the most tragic outcomes is where the hero simply does not save the day, city or galaxy. Have the character, from either a player or developer point of view, approach the apex of their journey and then simply kill them. During the final battle, the hero can defeat the villain but be trapped by their own ingenious solution to the villain’s troubles. I’m not saying this is not, in fact, “cheap” but it is an effective way to create tragedy.

One thought on “Can you optimize for tragedy?

  1. You write pretty fast. And it would appear that great minds think alike.
    The problem is indeed in knowledge: I’m still thinking over how knowledge affects tragedy. You can, of course, have tragedy through ignorance (such as when people do wrong without knowing what they’re doing), but can you truly have it through omniscience? By knowing the entire plot and all possible outcomes, the player could avoid tragedy entirely. You’d said, “What if the player wanted to design an experience toward tragedy?” If the player is omniscient, then tragedy only occurs if and when he/she wants.
    It’s certainly possible, but does it carry the same emotional impact as not knowing everything? Tragedy is all about the emotional impact, about the build-up and release, so optimizing for tragedy means maximizing this impact.

    Also consider the tension-and-release nature of tragedy: the audience know what’s coming, but they don’t know when or how. This suspenseful tension creates reader/viewer/player interest. At the moment it happens, the tension, the suspense, is released.

    For a tragedy game, the player has to be like the audience member who knows that the big fall is coming, but doesn’t know when. The player might be flawed in that they don’t know everything the character knows, but they do know where the character’s flaws will end him up: the character himself doesn’t know.
    Further, to fully make the events tragic, the character (and the player) must find out what their actions have wrought just after crossing the story’s event horizon; the horror of their realization hits them after it’s too late to stop the terrible outcome they’re responsible for. They may try to turn away, but they can’t.
    Allowing the character a way out, a final final chance to let them improve their fortune, leads away from tragedy, thus minimizing the impact.

    Would the player need in-game rewards to continue to take their character down a tragic path, (assuming that the game allowed a wide range of outcomes, positive to negative)? Perhaps just the desire to see said outcome, to complete the game, is sufficient.
    I’m on the Ishimura, and I know that the necromorphs are out there somewhere. Just walking down the hallway increases my suspense. Am I going to lead Isaac down that way, knowing he can and will be attacked? …Hell yes! I want to see what happens, I want to reach my goal.
    Comparing the tension-and-release of a good horror game to that of a tragedy, I suppose it’s fair to say that we can expect players to endure suspense so as to reach the end, with or without rewards. Of course, a devious and clever game designer would place the rewards to guide the player towards the game’s tragic ending. Or at least, I would.

Comments are closed.