A Home in the Uncanny Valley

The Unreal Geography of Video Games

I don’t entirely remember the first video game I ever played; it was either Bump ’n’ Jump or Moon Patrol, sometime back in ’83 or thereabouts. What I do remember is that my hand-eye coordination was woefully underdeveloped, and I wasn’t even remotely competent at the then-high-concept idea of driving a vehicle capable of leaping under its own power. But I kept spending quarters to keep crashing them — or to be mauled by muscle-bound karate fighters, or have my plane shot down repeatedly, or to get flung out of a Ferrari Testarossa to the understandable dismay of the driver’s pixellated girlfriend — because it was escapism, not competition or high scores, that compelled me to hit the arcades. By the time I was in 6th grade in the late ’80s, video games had reached that point where the atmosphere had become just as crucial as the gameplay mechanics, and while there wasn’t any real artfulness or strategy in throwing the same set of kung-fu punches for a half-hour or making sure the superbike you were controlling didn’t hit a billboard, you did it anyway because you wanted to see what lay beyond the next boss fight or finish line.

"I hear you get to the hidden temple and there’s spears that come out of the wall and knock you into a pit with spikes in it."

"No way! I can’t even get past the part where you have to fight the green dude outside the cave entrance. Sucks that I’ll never get to see it."

"Just learn how to get the hurricane kick right. And get some more quarters. I’m serious, the last boss’ hideout is totally cool, it’s like Master of the Flying Guillotine."

The compulsion to keep going and see the next new arena was something relatively new in the mid-/late ’80s, especially in a genre of entertainment that previously banked on getting people to play strictly to achieve a high score. It was a feeling you couldn’t quite get from the simple, repetitive layouts of games like Elevator Action or Pac-Man — an sort of anticipatory excitement, like watching an action movie of which you had no guarantee you’d see the ending, unless you were really good or had a lot of tokens. Most of these evocative places were technically limited, two-dimensional playing fields, often reduced to simplified caricatures of urban streets or the fields of war or the latest blockbuster movie — which was fine. Just so long as the enemies looked cool and the level designers kept throwing new things at you, graphical evolution could take its sweet time. Meanwhile, I spent some of the periods I should’ve used studying during class planning out elaborate side-scrolling levels for some nonexistent installation of Double Dragon, where the fights would take place in parking garages or subway trains or rock clubs, and it would be awesome.

Ten years later, everything had changed: Wolfenstein 3D and Doom had popularized the idea of moving a video game character in a three-dimensional space, while Tomb Raider on the PlayStation and Super Mario 64 on the Nintendo 64 took the leap into a completely polygonal game world. These were the titles that brought the concept of game space to a whole other level of detail, and slowly but surely the entire landscape of video game design shifted radically. Forget about just running around with a static, one-angle background scrolling behind you — now you could go wherever three dimensions would take you. The dungeon you were running through had a ceiling and three other walls now, and you could move laterally in a space that actually had a sense of depth to it.

With each successive generation of technological advances, something new kept raising the stakes: in 1998, Half-Life featured the sprawling, breathing Black Mesa research complex, complete with working vending machines; 2000’s Dreamcast cult hit Shenmue featured an ambitious replication of Japanese harbor town Yokosuka; 2001’s Grand Theft Auto III starred the free-for-all go-anywhere/do-anything urban environs that ushered in the “sandbox” era; and, with the perilous, wildlife-filled organic jungles of Metal Gear Solid 3, the surroundings weren’t just becoming as memorable as the gameplay; they were actually starting to dictate it. While the button-pushing mechanics, magically-appearing floating icons and visible life bars were a reminder that you were still playing a video game, the immersion started to become more enveloping — it’s one thing to still see falling Tetris blocks after you close your eyes, but it’s another phenomenon entirely to recognize a specific room in your favorite first-person shooter and know its many lifelike details just as well as you’d know your own desk at work.

With the current generation of video game graphic technology as expansive as it is, the big focus right now seems to be building a better human. And with each advance in technology, each attempt to capture a more profound and distinct sense of realism, there’s a strange, perilous point that computerized innovation has to traverse. Roboticists, CGI artists and others involved in the fictionalized replication of the human form refer to it as the “uncanny valley”: a hypothetical point where a robot or 3D computer-rendered character becomes detailed enough to closely resemble a humanoid figure, but is just “off” enough that it causes strong negative reactions. Theoretically, characters that don’t quite resemble photorealistic humans — say, the cartoonish cast of a recent Super Mario game or the caricatures in The Incredibles — are easier to empathize with because we recognize them as not significantly realistic and therefore it’s easier for their humanlike characteristics to stand out. But a character that skews closer to a lifelike depiction of the human form while getting a minor detail or even an intangible “feel” wrong (the motion-captured almost-humans in the movie The Polar Express, for instance) come across as unnerving, since the closer they get to resembling actual humans the more their non-human characteristics and artificiality come to the forefront — at which point viewers get creeped the hell out.

But aside from a sense of humanity, one of our strongest connections — psychologically and emotionally — is to our sense of place, and when something in an otherwise-familiar environment doesn’t quite click, it can be almost as discomforting. It’s a feeling hinted at in the 1964 Twilight Zone episode “Stopover in a Quiet Town, ” where a couple wake up in an unfamiliar house, wander outside and gradually realize that everything around them is a realistic but artificial prop, and that they’ve become toys in a giant child’s playset. Fire up your PS3, Xbox 360 or Wii, and you run the chance of getting that same feeling. I can’t count the number of recent games that whip up elaborate physics engines and promise to let you do anything within the bounds of the game’s reality — but when you try to smash an ultra-realistic table with an ultra-realistic sword, the thing doesn’t so much as budge. It’s a Catch-22 of sorts: the more a game visually resembles real life, the more we expect it to resemble real life in every other way, and when it doesn’t, our fun’s jostled a bit and we end up confused or frustrated or unsettled. And since video games require you to be more acutely aware of every aspect of your immediate surroundings than your day-to-day life might (assuming your day-to-day life does not involve constant fighting, stealth espionage or maneuvering a vehicle at insanely high rates of speed), the flaws stick out all that more prominently.

Take 2003’s Project Gotham Racing 2, a game that prided itself on being one of the most graphically-advanced titles of its time, where shiny detailed polygonal sports cars raced through photorealistic renderings of over ten major cities, spanning the globe from Chicago to Hong Kong to Sydney. As detailed as these cities were, with the assorted routes in the game based on actual streets and landmarks and with painstaking detail reproducing the actual signage of the area, there was a widespread consensus that there was something clinical and sterile about the tracks. While the racing circuits were situated in the midst of largely-populated metropolitan areas, there was no sign whatsoever of any existence of human life, as though somebody had dropped a neutron bomb. For the sequel, Gotham developer Bizarre Creations made the decision to park a few stands filled with spectators at various points in the new game’s racetracks, a small touch that, though difficult to see at 180 MPH and hard to hear above the roar of the digitized engines. helped avoid the ghost-town atmosphere of the previous game.

Ghost Recon: Advanced Warfighter has a similar problem, though it’s more disquieting in the context of its genre (the first-person tactical enemies-of-America-killing niche of games with Tom Clancy’s name on the box). Set in the nearish future in what appears to be an immaculate recreation of a terrorist-infested Mexico, each hotel, apartment, rail yard, warehouse and street-level furnishing looks exactly how it should, but aside from the scattered enemies in the game, there’s no sign of any civilian life anywhere. There could be some theoretical curfew in place in the story, and in practical gameplay terms it would probably be an unrealistic bringdown to have your badass commando fantasy interrupted by the unsanctioned ’collateral’ death of a wayward grocery shopper or a student on his way home from the laundromat, but that doesn’t make the proceedings feel like anything less of a high-tech, movie-set game of live-ammo paintball.

A missing populace isn’t the only way to stir up feelings of displaced uneasiness in a game environment. It’s safe to say that Dead Rising doesn’t give off the aforementioned feeling of being alone and abandoned, since the vast majority of the game’s got zombies packed to the rafters. In an unofficial, legal-disclaimer-strewn nod to George Romero’s Dawn of the Dead, Dead Rising takes place inside a shopping mall in the fictional town of Willamette, Colorado — and, in theory, it’s one of the most faithful fictionalizations of a single realistic modern-day location to date in a video game. Every conceivable architectural detail you could think of in a modern shopping center is there, from the obligatory movie theater and food court to an indoor roller coaster and sprawling park plaza to all sorts of maintenance tunnels, generator rooms, restrooms and security areas. The designers even made the decision to have one section of the mall in the midst of being remodeled to give it that extra feeling of a living work-in-progress. But while the layout and the architecture are remarkably realistic, the one thing that gives the whole proceedings a sensation of dislocation is the fact that every single store in the mall is fictionalized. Most of the Willamette Mall’s stores seem like fairly credible imitations of the typical retail outlets you’d find in a shopping mall, played straight without any real visual comedy — there’s nothing along the cartoonish lines of "Ammu-Nation" or the innumerable sex-pun-reliant store names that run rampant in the Grand Theft Auto titles — and there’s surprisingly little of the cross-cultural distortion that arises when a Japanese developer attempts to depict an American setting. But there’s a couple of places that feel just a bit off: I don’t recall ever going to a shopping mall that featured an antique store, much less one without any major department stores attached. The closest the Willamette Mall comes to a sizeable anchor are a grocery store and hardware store that don’t seem entirely large enough to justify drawing so many shoppers (undead or otherwise). As anyone who’s lived near a dead mall has witnessed, the lack of a major anchor is enough to bring down an entire shopping center, and after going through the game’s intro — where it’s revealed that Willamette is an isolated, small-town type of place with a population of just over 50,000 — it seems inexplicable that it managed to support a mall that size. Then again, considering that you’re spending most of the game fending off zombies that want to tear chunks out of your neck flesh, the subtle incongruities in the game’s setting are like little flashes of disconcerting unrealism in a place where horror and weirdness are the order of the day; in that way, it actually works to the game’s benefit.

Speaking of Grand Theft Auto, there’s always been a few things about its locales that rang blatantly false: the all-too-convenient natural disasters or construction snafus that just so happen to impede your progress until the game decides you’re ready for the next level (or, in this case, neighborhood); the condensation of what should be a major city into a space roughly a quarter of the size of the city it’s emulating; and, crucially, in San Andreas, the illusion of time and space that separates three real-world-analogous metropolitan areas. As impressive a feat as it is that Rockstar somehow managed to fit a microcosm of the American Southwest on one map, it feels a bit weird being able to drive from pseudo-Los Angeles to pseudo-San Francisco in about eight real-world minutes. Still, there’s a neat workaround to that: eight real-world minutes is equal to eight in-game hours, so as the light changes and the sun rises or sets, it gives off the illusion of a longer journey, and in a greater sense a bigger space — pulling back and looking at it reveals that the whole of San Andreas is actually a lot smaller than the real-world metro area of, for instance, Minneapolis/St. Paul. Given its comparatively simple graphics and its absurdist, caricaturized version of urban life (the GTA series has been read as a parody of the United States as perpetrated by a bunch of smart-ass Scots), there’s not much point in worrying about the game’s environmental and civic logistical unrealism (though I have to wonder what they were thinking when they designed that sprawling Vice City mansion that’s nothing but two rooms, a roof and a whole lot of halls and stairways).

Eventually, as graphic advancements continue apace, the conflicts between realistic immersion and practical game mechanics are just going to get more tangled and hard to solve, which is why it’s worth noting that many of the best games from the last generation of consoles — whether graphically complex or relatively basic — opted to shoot for a unique stylization that supercedes realism and puts computing horsepower behind the realization of an artistic world. The bright, simplistic and deliberately low-polygon world of Katamari Damacy, the cel-shaded, neon-riddled pop-art/graffiti setting of Jet Set Radio Future, and the cartoonish whimsy of the last several years’ worth of Mario titles all benefited from a unified aesthetic vision that pulled you in and kept you there — if there was something that felt off or wrong about an area in the game, it was meant to. (Not that everything has to be brightly-colored and cheerful to pull that off; Shadow of the Colossus and Gears of War have managed to throw a ton of graphic horsepower behind some remarkably striking setpieces that feel completely logical and real as well as artistically unique.)

Once I’m done with this article, I’m probably going to wind up playing the recently-released demo for Crackdown, a GTA-style game where you’re some sort of bionic super-cop with Six Million Dollar Man powers running amok in a highly-stylized comic-book metropolis. It looks a lot like a real city, only with all the colors turned up and the lights brightened and the overall atmosphere infused with a couple hundred cc’s of outlandish near-futurism. It all makes sense in the same ridiculous, over-the-top way that the faux-slums of Final Fight did. And, incidentally, you can drive an SUV that, in keeping with a quarter-century of video game tradition, can jump under its own power. Visual realism has its place, but sometimes it’s good to be reminded that you’re really just pushing the same old buttons.