“Weird stuff can happen in folk tales”: Ed Key talks meaning, morals and evil bears in Forest of Sleep

When the designer of 2013’s ambient exploration game Proteus announced his next project would be an experimental, generative storytelling adventure based on Eastern European fairy tales, a light turned on in my head. Literature, and its sinewy flex beneath the surface of video games, is of interest to me. Proteus was a brilliant exploration of making myths in computer-generated landscapes, and Forest of Sleep looked like a gorgeous fiction adventure – part-cartoon, part-structuralist romp.

Also, the promo material has a picture of a bear holding a balalaika.

Ed Key’s Forest of Sleep will see the player guide three children through a forest, encountering characters and being given quests. Whether or not the player decides to follow these story-pointers will dictate how the overall folktale plays out. This will all be done without text, with the game instead relying on visual language to hint at events and relationships.

Still in development, the game sees Key collaborating with animator Nicolai Troshinsky and a handful of other artists and producers. I spoke to Key about his work on the game, and how the computer-player folklore collaborations might play out.

TM: How’s progress been going on Forest of Sleep?

EK: We just got a bit of funding from the UK Games Fund. Our current plan is to make a playable Act 1 of the story, although we’re not releasing the final game episodically. Our current aim – which is really the hinge of the project, because if we can get this working we can get the whole thing working – is to make sure you can play through this 10- to 15-minute prototype segment and come away from it understanding that a story has happened.

TM: So how are you going about structuring these generative folk tales? Are you looking at folklore through a structuralist lens – taking the approach that they’re built up of common movements and characters?

EK: Yes, but there’s also the link to modern storytelling here, like episodic cartoons, which all follow this fairly limited set of dramatic structures. Because of the incidents within them, they feel different and surprising, and they have a measure of anticipation.

I should really say that thinking in terms of these structures is quite new to me. Nicolai and Hannah [Nicolai Troshinsky and Hannah Nicklin, who are also working on Forest of Sleep] both come from much more of a story-making background. Between us we’re getting into this structural idea of narratives. Vladimir Propp is the big figure when you talk about folk tales and structuralism. I haven’t read his book [Morphology of the Folktale], but I’ve read the Wikipedia page. That’s my homework! I’ll borrow it off someone.

(Above: Vladimir Propp)

TM: How do you straddle that balance between building structures of a story, and then blending in the structures of a game, where players want a sense of agency? How do you negotiate the tension between those two things?

EK: It’s more interesting for things to be subverted – for expectations to be set up then subverted by the game in interesting ways that feel like they acknowledge the original expectation. Like turning up to a quest and the guy is dead, or in his final breath he says, “Go get that evil bear that did this to me,” or something.

You can allow the player to do the subverting, or the game can do it. An obvious example is that if the player is told to go somewhere, they could decide to go somewhere completely different. Rather than that being handled in the way of RPGs like the Elder Scrolls series – where if you get a quest and don’t do it, it just waits for you forever – if you go the wrong way in Forest of Sleep, that act will be about someone who ignored instructions and went the wrong way.

You’re set off in motion, to do something, but you’re not railroaded down that path. You’re given a grain, which you can go with or against. There are also choices within those. Subplot choices, or flavour choices, of how you interact with supporting characters.

TM: That’s a bit like something from [Cardboard Computer’s southern gothic/magic-realist adventure game] Kentucky Route Zero – the idea that making a decision in some way, even if has no real impact on the action, adds a sense of ownership to the story. It’s also interesting what you mention about having a narrative to follow, but being given a chance to rebel against that. It sounds a bit like [Davey Wreden and William Pugh’s interactive postmodern short] The Stanley Parable, where you have that conflict between the narrator and player.  

EK: Both of those are really good references. Although they’re static authored games, the way they treat player choices is interesting. [Inkle’s branching adventure] 80 Days is another big reference point.

TM: It also reminds me a bit of a bot, like Magic Realism Bot, where the story is randomly generated but people read into it – bring meaning by imagining it. I guess some of the things that come out of Forest’s tales might be similarly chaotic… I wonder how that aspect balances with the morals you get from folk tales. If these stories are created through a mishmash of player decisions, how can you have an overarching moral to the story?

EK: I think the predominant moral of most folk tales – or at least a major moral – is: ‘Do what society expects of you. Don’t get big ideas. Don’t go off the path in the forest.’ I was recently reading some of Aesop’s Fables, and they’re the worst kind of tedious do-what-you’re told moralising. The more fun ones are where anarchic things happen, or where unfair things happen. We definitely don’t have a plan to make stories with neat morals. It’s more about leaving things open and ambiguous.

One of my big obsessions on this project – I’m such a fake philosopher – is how meaning arises from meaningless stuff. More generally, how does that work within generative projects? Obviously within a built system there is an authorial intent, but the outputs of the system can be quite meaningless, as they don’t have a conscious context. The object of inquiry here is, can you have something where you build a context up over time, with the player, where something becomes meaningful because it’s consistently reinforced?

TM: Like creating meaning through repetition?

EK: Sure, yeah. That is a big part of how something becomes meaningful. It’s the same with places. If you walk down a road once, it doesn’t mean much to you necessarily. But the way home is meaningful because it’s a route you go through again and again.

TM: Can we talk a bit about the aesthetic style? I’ve seen before that you’ve mentioned Yuri Norstein’s animation Fox and Rabbit as a touchpoint. Could you say a bit more about what you’re pulling on there?

EK: I should say that’s really Nicolai’s area, but yeah, Yuri Norstein is a Russian animator who did most of his work in the Soviet era. Fox and Rabbit has a lot of similar visual elements to those we’re using. It has a decorative border, and these inserts that are pictures within picture bubbles, which remind you of things characters are thinking about. We’re using them as our prime method of describing interactions between things.

https://youtube.com/watch?v=Tr1UPxkZsVA

There’s also the whole world of book illustrators – mostly Eastern European bookmakers and people that fit into that tradition. Janusz Stanny and Józef Wilkoń, for example. Their style is a mixture of semi-representational and abstract images, along with decorative borders that mirror what’s in the scene. Ivan Bilibin is another artist working within that area of textures and borders, where you can do alternate representations of things in the scene.

TM: There seems to have been quite a few games recently that deal with text, and making decisions within the text. I know there isn’t actually text within Forest, but it still seems to be based on directing a text of sorts. Do you see this as an emerging area of gaming? One that taps explicitly into more traditional ideas of literature and storytelling?

EK: I find it really hard to identify trends – I have a natural resistance to them – like when Proteus was grouped in with others as a genre. That said, The Yawhg is another game along those lines. It’s very small and best played with four people on a sofa. It’s a sort of multiplayer game where you make choices and other people laugh at what happens to your character. I’d really recommend it.

Another cool game that we’ve referred to a lot is a board game called Tales of the Arabian Nights, based on One Thousand and One Nights and from the 1980s. It has a big Risk-style map where you move your character around and have encounters in cities. There isn’t much skill involved, but like Kentucky Route Zero, it makes you feel like you own your choices. The other nice story game I’d like to mention is The Quiet Year, which is a map-drawing storytelling game. It’s a peripheral influence to Forest of Sleep. It’s not that we took the mechanics and brought them into our game, but it comes from this background of storytelling games.

(Above: Ed Key)

TM: What was the impulse for doing a game like this? Hearing you talk about it, there are definite overlaps with the generative and voiceless Proteus, but from the outset it does look like a very different game.

EK: Yes, so there’s a comparison to be made between the two games not having text. Not that I’ve got anything against text, but I guess I’m interested in the idea of presenting someone with images and letting their mind wander and connect things.

Where Forest came from originally was, halfway through making Proteus I took a break and started making a game about an expedition – going up a mountain and coming back down again, and how you plan your food and so on. That morphed into a more fixed folk tale story about being in the forest when your parent falls ill, and your group needs to go into the next valley and find medicine. Then I started talking to Nicholai about generative narratives, and he suggested making a game about folk tales. His reason for this was based on the sense that weird stuff can happen in folk tales, and you don’t question it so much.

Also, there’s a thing fairly specific to Russian folk tales, in that you have characters that recur across several stories, like Prince Ivan or Baba Yaga, who are kind of archetypes. The way these characters recur felt like it lent itself to a generative system.

Over the course of the last few years,

Forest has swung between being a game with resource systems comparable to [procedurally generated space adventure] FTL, to being a pure narrative game. Now it’s found a balance between those two things. 80 Days is a good example I like to quote on this, because that has Fog’s health and your money. That latter number in particular affects the mood of your journey. It’s more of a resource system that’s interwoven with the narrative, rather than the resource system being the thing that’s generating the narrative, like in FTL. That game is a story of either you succeeding or failing, whereas we’re trying to make something where there isn’t a fail condition… where it’s more something that colours the mood of a story.

TM: Something where it isn’t possible to lose. Where the game doesn’t ‘fail’, but just becomes a sad tale.

EK: Yes, exactly.