Pretty sure this was the first piece I ever successfully pitched. It ran in the Back to School issue of Kill Screen in…several yeas ago.
Thanks to Chris Dahlen for taking the chance on a recovering academic and unproven blogger, and also to him and Ryan Kuo for their editing. They did their best to make this readable.
Also ugh so many first-person present-tense “scene setting” sentences.
I’ve just left the Smithsonian Museum of American History in Washington, D.C. The air is stagnant and humid, and my friend and I are looking for a metro station. “There’s one by the National Archives,” I say. “I think it’s up here on our left.”
The National Archives are situated north of the Mall, about halfway between the Washington Monument and the Capitol Building. I know this because I once had to go there to steal the Declaration of Independence from a robot who believed he was one of its original signers.
Then, D.C. was a post-nuclear wasteland, fairly unrecognizable except for a few surviving landmarks. It was also virtual, housed in Bethesda Softworks’ Fallout 3. Now, I’m in the actual city for the annual conference of the American Library Association—but my friend and I are still using the Fallout 3 map to find this metro entrance.
Why would I rely on the videogame map rather than an atlas of the street? The visuals are not the same. The distances are not the same. Walking down the street in D.C. in mid-June is nothing like sitting on my couch in my Pittsburgh apartment in mid-November.
So I started researching the relationship between virtual space and physical space—and how the space inside your head figures into it all.
And it starts, as most quests do, with rats.
In the July 1948 issue of The Psychological Review, Edward C. Tolman published “Cognitive Maps in Rats and Men.” The paper detailed a series of experiments designed to study how rats learn their way through mazes.
The basis was simple: Take a hungry rat, put it in a maze with food at the end, and with enough repetition, it will become more and more efficient at running through that maze to the food. What interested Tolman was why. One school of thought suggested that paths were learned passively, as a result of moment-to-moment stimuli. Think of it as the rodent equivalent of pattern recognition: “Turn right, turn right, turn left, turn right.” Another school claimed that the relationship between turns was crucial, and that this relationship was actively constructed in the mind of the rat.
Tolman and his team placed rats in a maze where one path led off of a circular area and made a ninety-degree turn to the food. After a number of runs, the setup was changed. The circular area and the location of the food remained the same, but there were now several paths branching out. The original path was blocked, while one of the other spokes led straight to the food.
A statistically significant number of rats, after finding the original path blocked, selected the new one and found their goal. The rats, Tolman surmised, were forming a cognitive map of their environment: a set of mental relationships that allowed them to conceive of the food’s location relative to their own, even though it was out of sight. He described these maps as “broad”; that is, they were not limited to the “narrow” path the rat had traveled, but took into account the area as a whole, coherent space.
This seems pretty straightforward, but it was a big deal for understanding spatial cognition. Tolman was the first person to show that rats had a sense of “over there” which could be demonstrated as long as the environment was built to allow it.
This is the most interesting part of Tolman’s findings—that environmental factors can aid in the creation of these cognitive maps.
“Imageability” is a term used by Kevin Lynch in his 1967 The Image of the City to describe “that quality in a physical object which gives it a high probability of evoking a strong image in any given observer,” or how easily a space can be constructed in the mind. In Tolman’s terms, high imageability results in a broad cognitive map.
Lynch proposed that a strong mental map was significant to feeling comfortable in a space. People in general do not like to be lost. I’m talking really lost—not “Hey, let’s go for a drive and see where we end up,” but the kind where you can’t begin to orient yourself. Being lost isn’t not knowing where you are. It’s not knowing where (or what) anything else is in relation to you.
In Lynch’s work, a person’s cognitive map of their city is made up of five mental elements. Paths are linear spaces through which one can move. Edges are linear spaces through which one cannot. Districts are larger areas that have a consistent character. Nodes are points that a person enters, such as intersections and plazas. And landmarks are points not entered but used for orientation—anything from a tall building to a mailbox to a pothole.
But in a person’s cognitive map, the distinctions are fuzzy and often shift depending on the individual’s perception. A landmark like a tall building, for example, becomes a node if you cut through its lobby to reach a street on the other side. And the cognitive map is simpler than reality.
We select parts of our observations, simplify them to these elements, and relate them to one another in our minds. If Lynch’s model is right (and plenty of spatial cognition research suggests that, while it may be simplified, it is not inaccurate), then a virtual space designed around it should be easier to understand as a consistent whole.
I found, when navigating Tallon IV several months after completing Metroid Prime, that I was still able to find certain locations with little effort. The Tallon Overworld, Magmoor Caverns, and Phazon Mines acted as districts, unified by coherent art styles. But because they were designed to be separate spaces, their boundaries were stronger than those of districts in the real world. I didn’t just think of the edges as impenetrable: They actually were. Color-coded doors and glowing items out of reach, like an ice beam or missile upgrade, served as landmarks. And the elevators connecting the different areas became key nodes in my mental map; if I couldn’t remember what the path to an elevator looked like, I still had a pretty solid idea of where it was located—even without using the in-game map.
This kind of design is incredibly useful in a game like Metroid Prime that requires a lot of travel through the same spaces. By reducing the amount of thinking (and map-checking) required to orient yourself, it adds a sense of easy travel and makes backtracking feel like less of a chore.
Games aren’t reality. They’re not tied to an actual physical space, and they don’t have to follow the laws of physics. Like any other art, their relationship to reality is representative. A game can only provide suggestions to the entire space—but the suggestions it provides and how it provides them is vital to making you fill in the gaps. A game can evoke a real space, one that appears to follow the laws of the real world, but we as players are always going to have to meet it halfway. And that’s important to how the game world feels.
When Harry Beck designed the iconic London tube map in the 1950s, he transformed the complicated London Underground into an easily understood diagram. Rather than recreating every turn of the lines and the actual distance between each station, Beck simplified and abstracted them, drawing straight lines for the tracks and equidistant marks for the stations.
The typical tube rider experiences the space more or less as a straight line between landmarks. Barring some emergency, they won’t leave the train between those stations. They don’t need to know where they are in relation to the London topography—they just need to know their destination.
Videogame spaces fall somewhere between Beck’s map and the transit system it depicts. They’re a map with no physical analogue; “Ceci n’est pas une pipe” indeed. But because it’s a map through which we move, it’s easy to forget that it is still a map. And so we look for the bathrooms, and wonder where people sleep. But maybe instead of thinking of the game map as a recreation of a coherent physical space that follows the same rules as the real world, we can think of it as an interpretation—one that has been simplified by the designers, like our mental maps.
As I walk down a city street in real life, the edges of the path can be transgressed; landmarks can become nodes. There is an inside to every building. I don’t know what is inside, but I know that it exists. I could construct it in my mind, and imagine the stories occurring within, but I don’t. If I did, I would go insane. So I simplify, like the designers of Grand Theft Auto.
Plenty of games give you landmarks that suggest the limited space through which you move is related to a larger one. You can see out the windows of Rapture in BioShock. In Mass Effect and Mass Effect 2, you often move along a limited path with views of distant landmarks that you cannot reach. Looking up in the middle of the Citadel, you can see the extent of the space station; the sense of scale is palpable. But you can’t go there.
When games let you go there—when they let landmarks shift to nodes, as they so often do in our cognitive maps—they can create a powerful sense of expansive, connected space. This is true even in the most linear games.
Consider another Citadel, from Valve’s Half-Life 2. During your journey through City 17, the Citadel looms over the landscape. You always know where you are in relation to it. When you finally enter it late in the game, you have a sense of what surrounds it. The relationship between “here” and “there” provides an easy framework of progression and leverages spatial relationships to build narrative tension.
All roads lead to the Citadel, literally. The only roads that actually exist in City 17 are the ones you travel along. But your brain is used to constructing spaces out of just that information, and so it’s less work for you to imagine the spaces in between the paths and the landmarks.
Once you’ve broadened your map of City 17, there’s enough empty space for your imagination to run wild. The best game spaces don’t just tell you a story: They tell a story with you. They create a frame on which you can drape your own meanings; they are a conversation between you and the environment. It’s a story that creates a world that is more than the actions you’ve taken in it. It’s about the spaces you’ve traveled, but also the spaces you haven’t.
The blank space on a map is a powerful trigger for the imagination. When a game space understands this, you become more involved in that space. The map is essential to the stories that occur within it. And as Lynch observed when he cautioned against over-designing environments: “A landscape whose every rock tells a different story may make difficult the creation of fresh stories.”