The Last of Us and Pittsburgh: Real Game Worlds

Games Features The Last of Us

I learned to take photographs in Pittsburgh. Its topography taught me how to compose a complex image: houses clinging to hillsides, bridges across rivers and valleys. Uneven and winding terrain meant that no matter where I looked, there were always multiple layers and levels of things going on.

It’s where I learned to see.

Pittsburgh is not a place that sees itself often on a screen (a local film office and state-wide tax credits have created certain eras where that is not entirely the case). Living here for 13 years I’ve seen its obsession with its image, a cyclical repetition of the same themes (we’re not smoky! most livable city!) every decade or so since the urban renewal projects of the 1950s: the “Renaissance”. There was a second renewal plan that started in the late 70s called “Renaissance II”, and I think now we’re in the middle of…IV?

That civic obsession trickles down to an individual level, leading the more media-obsessed of us to look for our city in films like Perks of Being a Wallflower and The Dark Knight Rises. Though in the latter it was playing Gotham City, the point is that both films were shot here. There were physical cameras here, capturing images of the city. We play a game, see whether the city that is constructed through cinematography and editing lines up with the city as it exists.

It usually doesn’t. Still, there’s a concrete relationship to the place of filming that often takes precedence over a different, more abstract feeling of the place, of the city as lived versus the city as seen on-screen.

I suspect this kind of disconnect is less experienced by people who live in New York City or Los Angeles, large urban areas who appear on so many screens and whose size is so great that every use of the city does not need to feel “right” to any particular viewer. It’s easier to believe that something set in New York City or Los Angeles could take place somewhere else in that city, maybe.


A certain type of videogame recreates cinematic experiences. No, wait, that’s too loaded a term in this context, where it takes on this idea of scripted cause-and-effect with controlled camera angles that frame what you see (whether or not they frame them well, well…not everyone is Gregg Toland).

What I mean is there’s a lot in videogames that a lot of players only know of through other media, through movies and television. When we say a game lets us feel like we’re storming the beaches of Normandy, we know what that is like not from actually storming the beaches of Normandy, but maybe because we’ve seen Saving Private Ryan.

Audiences know that you can’t make a film without a camera. For a long time, Hollywood-style editing and cinematography developed to minimize the audience’s awareness of the camera. What would have been unacceptable in these films, visual acknowledgement of the camera’s presence (lens flares, water or fog on the lens), was permissible in documentary films. Eventually it became part of the cinematic language of narrative films. In a weird way, acknowledging the camera is “there” capturing an image became a way to draw people into that image, rather than pushing them away from it.

When you move the “camera” into the sunlight in The Last of Us, light reveals a dirty lens. During a cut-scene the “camera” shakes a bit. These artificial cinematic “imperfections” are animated in the service of “realism”. But visual cues that say “hey, there’s a camera here!” when there is a camera become a lie when there is no camera, no lens, no visual “imperfection”.

And maybe a certain quest for “realism” dooms these games to be copies of films. They can’t invent their own imagery because they make meaning out of the familiar. Other games, like those made in Twine or other text-focused styles, or more abstract games not seeking to replicate the audiovisual cinematic experience, are freer for it. Their author-creators are not bound by some attempt to create an “objective” reality within their electronic world. They don’t have to recycle cinematic images to make their spaces legible.

Realism isn’t about the real world; it’s about how much something reminds us of the real world, how much it DOESN’T challenge our understanding of the world, how well its digital and our mental models mesh.


I’ve lived in Pittsburgh for 13 years. I have a very specific-to-me notion of what the city is. I have an abstraction in my head (the only way to grasp something the size of a city is through an abstraction), and I want to see how well it lines up with the abstraction of the city in The Last of Us. Things I expect to see in Pittsburgh: bridges (the game has one!), a slightly disorienting tiering of buildings on hillsides, rivers. Funiculars.

When you look down a street or over a building in Pittsburgh (in my mind’s Pittsburgh), your view terminates in hills. Sometimes they’re a hazy blue. Pittsburgh’s topography was created by rivers cutting through the Allegheny Plateau, and the result is that no matter where you are, you’re usually either looking down or up at some kind of view.


Your initial encounter in the Pittsburgh of The Last of Us occurs in and around a convenience store under an overpass, steel girders forming a cage around you and your opponents.

It’s a bit on the nose, but it’s also a place I know. Well, sort of.

That kind of overpass architecture is located in one part of the city, and there are no convenience stores under it. Nor are there sandwich shops advertising “heroes” (we call them hoagies, thankyouverymuch).

You make your way to a flooded area. A modified version of the wayfinder signs that are all over the city hangs from a broken lamp post—its colors and design are different, but still recognizable. It’s like when a TV show doesn’t want to license a junk food brand, but instead of turning the bag around they modify the bag.

Lack of licensing makes the city weird. There are posters for “Duquesne” cigarettes—Pittsburgh is built on the site of the British Fort Pitt which was built on the site of the French Fort Duquesne—yet no sign of Duquesne Pilsener. And anyone who’s ever been to Pittsburgh knows there’s black and gold everywhere. These are the colors of all three major league sports teams: baseball’s Pirates and hockey’s Penguins and football’s Steelers. This Pittsburgh, though, has nothing related to sports anywhere. In-game, maybe that’s because of the martial law that locked down the city, though one would expect the local pride that apparently rebelled against said martial law would have looked a little bit like football fandom (college students flipped a car and threw a bus shelter through a library window the last time the Steelers won the Super Bowl). It’s funny, isn’t it, how something so crucial to a place’s identity can be locked down with intellectual property law?

Downtown Pittsburgh is a muddle in my brain, so I don’t expect much of the layout of the city as presented in the game to feel wrong. My mental map is more a general set of relationships between places, their relationship to landmarks. And some of those landmarks are in The Last of Us’s Pittsburgh.

There’s Fifth Avenue Place (minus its real-world Highmark branding), the US Steel Tower (its rust-colored facade sporting an “NA” sign instead of its real-world “UPMC” logo), and your goal, the Fort Duquesne Bridge (signs for its real-world twin, the Fort Pitt Bridge, are visible in the early parts of the city, but the bridge itself is not there).

tlou pittsburgh bridge.jpg

tlou pgh bridge.jpg

Those missing corporate logos, like the missing sports teams, prime me to expect an off-brand Pittsburgh. Movie posters advertise a joke about the Twilight films. It ties Boston, where the first part of the game takes place, and Pittsburgh together, this bit of mass pop culture. More of that off-brand “realism” that ends up being a joke about trademark law.

Like I said, I don’t really remember downtown well, though I can navigate it. Because it’s a fairly small area with fairly diverse architecture, there aren’t enough buildings for the facades to blend into an indistinguishable mental mass. It was jarring to come around a corner and see the 900 block of Liberty Avenue modeled in-game, but not because most of its storefronts were different and its mural changed to advertise a different company. It was jarring because the sign for Broadway Army Navy Surplus was identical to the sign that hangs above the entrance to that very store on that exact block in the real Pittsburgh.


One of Pittsburgh’s most famous buildings is the Allegheny County Courthouse. Completed in the late 1800s, it was designed by architect Henry Hobson Richardson, the man whose work is categorized as “Richardsonian Romanesque”.

It’s an imposing building (almost ridiculously so): gigantic stone blocks, towers and arches everywhere. A bridge over a street connects the courthouse to the old jail, also designed by Richardson. The effect of the two buildings is that of some kind of massive fortress, smack in the middle of downtown. It’s fortified, visually imposing; it’s the kind of building that, when you think about zombie attacks (we get that a lot in Pittsburgh thanks to George Romero), you think it would make for an excellent home for survivors.

It’s not in the game.

Well, that’s not entirely fair. There is a building that resembles it in style, but it’s much smaller and has a sign on the side that indicates its available for rent. You see it from one of the rooftops of the hotel (modeled on, I believe, the Omni William Penn) in which you spend a large chunk of your Pittsburgh time.

tlou courthouse.jpg

Staying in a hotel in downtown Pittsburgh is a bad idea in real life because there’s very little that goes on downtown in the evenings (except in the Cultural District, which is home to the ballet and the symphony and a few theaters and is, well, very specific in the kind of culture it considers “Cultural”). It turns the idea that Joel and Ellie have to fight their way through a downtown hotel, dodging hordes of people trying to kill these two “tourists”, into a dark little joke.

Maybe the courthouse would have been too videogamey, an obviously “special” space unlike the luxury hotel’s backdrop of decadent banality. Hotels are great for videogames because we expect them to have repetitive architecture and little past. Scores of identical rooms with identical contents are relatively easy to duplicate computationally, and the impermanence of living in them means they can be character-free. The mandates of “environmental storytelling”, the conventions (or clichés) that signal “something happened here”, aren’t as stringent. The history of events in a space are antithetical to the hotel. Running a hotel involves exorcising every guest before the next arrives, so a digital hotel has little need to be populated with the detritus of lives “lived” in them before the player shows up.

“Environmental storytelling” makes a game a detective story: The player finds a variety of physical objects in the game and pieces together what’s happened before their arrival. The problem is that a lot of the time, picking up and piecing together this evidence is at odds with the shooting and looting you’re doing to get from point A to point B as quickly as possible.

This makes narrative games’ relationships between space and time very interesting. In most games time passes spatially. Where film compresses or extends narrative time through editing and film speed, games essentially lock time to a specific area. It may take a player several hours to get through a particularly difficult section, but only once they’re through does time begin to move forward again. Only the successful navigation of space is real.

Sometimes this can be caused by resource limitations: Sprawling open world games that allow a player to return to a place they’ve already been will often only change dialogue after the player triggers forward motion of time by reaching certain areas in the game. Time waits for no man, maybe, but game time usually waits for the player.

Environmental storytelling tries to sidestep the weirdness of narrative time by having unchanging physical evidence: Everything has already happened to it. But it still depends on that approach to time in order to give the player the chance to move at their own pace and take in information as they see fit.

The approach is often (though not always) about looking at “real” objects in a “real” space because of the idea of cinematic realism, that even though these environments are constructed, things shouldn’t be abstract.

This sets it apart from approaches to, say, set design, where colors and shapes don’t have to have actual structural properties. They can be abstract, evocative rather than didactic. This brings them closer to the abstract meanings of spaces and images and colors we experience in our day-to-day lives.

But a cinematically real game space doesn’t just have to remind the player of a “real” space. It also needs to be navigable for both the player and the computer-controlled avatars that move through it. There is a difference in many games between what you can see and where you can “go”.

The repetition of a hotel makes it easier to carve out this play space. It takes advantage of theme and variation in the architecture: The player has a basic expectation of repetition in the architecture they can see, so the developers can vary the paths through the space without having to make dramatic changes to what’s outside that path. Any duplication within the game space can be written off as a “realistic” hotel space.


Maybe that’s how you make a game place out of a real place—the specific spaces you move through don’t need to be accurate to their real world counterpart as long as some key traits are preserved. If you place landmarks in the right places, then seeing those landmarks relates the game-place to the real place.

In the game’s Boston, it was the golden dome of the Massachusetts State House. In its Pittsburgh, it’s the Fort Duquesne bridge, as bright-yellow as it is in real life, spanning the Allegheny River north of the city and connecting downtown Pittsburgh to the North Side. The North Side was once the city of Allegheny, a majority of whose citizens voted against its annexation by Pittsburgh in the early 1900s. The law at the time required a majority of pro-merger votes from the combined population of both districts, making it easy for the nearly-three-times-more-populated Pittsburgh to swallow up the smaller municipality.

The bridge itself was finished in 1963, but it took six years before ramps on the northern shore of the river were completed. Though it wasn’t used until 1969, a college student drove a station wagon off the end of it in 1964. He survived.

Joel and Ellie follow in that chemistry student’s tire tracks, sort of: They leap into the Allegheny River to get away from Pittsburgh and eventually wash up on the shore downriver. They meet back up with two survivors who had been in the city with them and the group heads north toward a radio tower.

And the geography starts getting really weird.

The game has you constantly traveling westward. When you’re in the city and heading toward the bridge, it’s on your right—in-game north (its twin, the Fort Pitt Bridge, doesn’t head out of the city to the south in the game).

The group moves north, away from the river, to get to the radio tower. The problem is that if you’re west of the city at this point, and the river is south of you, it’s flowing the wrong way. Facing the river puts the city to your right, with the water flowing right to left. For this to be the case, you’d need to be on the south bank, not the north bank, of the Ohio River. But looking back at the city, the bridge you jumped off of does stretch to the bank you are on. But based on the direction you head to get to the radio tower as demonstrated on the map you looked at before, the city lies to the west.

But quick, let’s shift gears. The river widens dramatically near where you wash up, and the group finds a boat washed up on what appears to be the edge of a lake or ocean. There’s a journal entry in the boat, a bit of the old environmental storytelling about how the ship’s captain (call him “Ish”) spent time at sea during the early stages of the infection, and how he had returned now to see what was left.

Now, most environmental storytelling in games is expository. It’s supposed to be part of an objective reality (that is actually a cinematic reality). Unless something is obviously flagged as “unreliable”, covered in stereotypical signs of mental illness, years of games have trained players to take things literally. Look at Pittsburgh on a map. Barring an as-yet-unmentioned catastrophic climate event that resulted in the return of the North American Inland Sea, taking a boat out to the ocean from Pittsburgh means you have to go down the Ohio and Mississippi Rivers to the Gulf of Mexico.

You can see it as a flaw, a spatial bug, where the intended space doesn’t match the space as constructed.

There’s another reading, one that’s outside the game itself. If you look at the relationship of the landmarks: the city to the west, the river that flows eastward to a body of water that looks suspiciously like an ocean, you could see how this section of the game could have originally taken place outside of Boston, or somewhere else on the East Coast. The lobster traps in the ship’s hold make a lot more sense. But then a few minutes later when you enter a sewer system, there’s a sign on the wall for the District of Columbia water authority. Maybe Ish’s sea was the Chesapeake Bay.

It doesn’t really matter if your interpretation here is right or wrong. This is a different kind of environmental storytelling: software forensics or maybe archaeology, trying to reconstruct the game’s development from the physical-digital evidence left behind.

tlou army navy sign.jpg

tlou pgh army navy.jpg

And that’s pretty much all of the city of Pittsburgh in The Last of Us. A section or so later pretty much nails a certain kind of Pittsburgh residential neighborhood, all winding streets and retaining walls and houses with ground-level exits on multiple floors. But there’s something other than architecture that is important to creating place. What determines the character of a place is what you do there, and in this game that means what (and who) you kill.

Pittsburgh is where Ellie gets her first gun, but she’s still regularly taken aback by the violence of Joel’s fighting (Ellie, like the on-screen audience of a film musical, reacts to what she sees in a way that is supposed to cue how we, the off-screen audience, react). It’s where she makes her first kill.

I expected to fight lots of zombies in Pittsburgh, the city of George Romero. But that’s not the case—most of your opponents in the city are the other survivors. A Pittsburgh so self-reliant that it’s become pathologically destructive toward outsiders outweighs its cinematic heritage as land of the dead.

Because of its local setting, there was buzz outside of videogame circles when The Last of Us came out. My dad, who quit videogames around the time he started regularly losing to my brothers and I due to “cheating” NES controllers, asked me about it, and about why there are so many zombies.

I could have gone into some of the socioeconomic reasons (zombies are the ultimate in personal property paranoia, where your neighbors and friends unambiguously turn on you because they’re brainless and trying to eat you, or how late-capitalist consumer society is yadda yadda yadda), but I think that, at least for games, it comes back to computation and cinematic realism.

We know how zombies are supposed to act (thanks, again, to film) so when they mindlessly attack (whether shambling or running) it doesn’t trigger something in our mind that says, “wait, that’s not realistic!” This is crucial, because as game images get more detailed and seemingly (SEEMINGLY!) less abstract, our expectations of what happens on screen becomes more specific (if a two dimensional, eight pixel enemy doesn’t take cover and flank us, well, that’s ok). The complexity of visual representation requires, for a certain (noncritical) approach, a complexity of behavior to match (“behavior” being physics, artificial intelligence, general causes-and-effects).

Videogames are software. They’re built on systems, and systems are supposed to repeat with consistency—when the images that system generates are filled with signs of “realism”, then the system must be consistent with the reality outside of itself as well as the reality it generates. Remember, software bugs are where the intended output of code does not match the actual output. The uncanny valley is what happens when the abstraction does so much of our imagining that we’re free to pick out its shortcomings.

Zombies work better as visual than as text because the horde and the movement are more efficiently conveyed in image than in text. It would take pages and pages and pages to describe the zombies in the threatening detail they need to be intimidating from within the work—as opposed to the threats that you create in your own head from the suggestion of ‘zombie’.

So if you increase the visual complexity of the zombie model (make it grosser, basically) and increase the number of them the computer can show on screen at once, you can increase these threats. It just so happens that these particular things are something certain kinds of technological innovation is designed to enable.

And, again, because films have taught us how zombies act (how “zombie” is shorthand for a mindless being), we don’t expect anything complex in the systems that govern their behavior. They’re the perfect metaphor for a myopic march toward a surface-obsessed technology.

Zombies are hotels. They’re easily computationally repeatable and designed for specific kinds of (terrifying, soothingly anodyne) emptiness.


We get the game spaces that technology can provide. No matter how detailed these simulations become, they’re still constructions. They’re a space filtered through people filtered through code filtered through computers rebuilding those spaces on a two-dimensional plane that uses the Renaissance technique of perspective to give them a third dimension.

The desire to see that screen as a window into a world is a powerful one, that it’s a representational image. But it’s not a clear window. It’s a stained glass window, everything abstracted to polygons.

The Last of Us manages to abstract Pittsburgh in ways that are at times familiar and other times bizarre. It reduces the complexity of the city, fits it into the space bounded by technology and licensing and ideas about game design and visual media. The things it opts to keep don’t attempt a one-to-one recreation of the city as it is, but they do a pretty good job of not being unrecognizable. Just enough of the right dots for someone familiar with the city to connect.

Brian Taylor has spent too much time in Pittsburgh.

Share Tweet Submit Pin