Intermittent Mechanism

The saga continues. This one’s dedicated to the Siren franchise, which means it’s a more in-depth version of some ideas I first poked around in in the tail end of this lesson plan.

I wanted to finish up this ep because it caps off a four-episode sequence that begain with ep 4. But my hiatus from this series is beginning now. Next up: catching up on interesting games from 2018.

Script below the jump.

This is an appreciation of the games in the Siren franchise. It’s the seventh part in an ongoing series I’m doing on horror games, and I would really recommend against jumping in right here. Instead, I’d recommend watching episode 4 and episode 6, at the very least, in preparation.

Before jumping right into the Siren games, I want to home in on one particular visual trick that you find in horror cinema:

You’re watching a horror movie. It’s shot an edited in a fairly standard way. You got your shots, you got your reverse shots. The camera is on a tripod. It pans, it tilts, but for the most part it remains steady from shot to shot. Then something happens. You get a shot that moves more than normal. Maybe it’s handheld. Maybe it’s steadicam. But whatever the case, it moves more freely. It seems embodied. It peers around corners furtively. As you’re seeing the scene, you get the impression that the view you’re watching belongs to someone who doesn’t want to be seen. And when you see a character within this view, you get the impression that they’re in danger. Someone’s watching them, without their knowledge—and that person’s probably up to no good.

This is the Stalker Cam. It is an effect that’s extremely simple to produce. But it’s surprisingly complex in its operation, because it hinges so heavily on the viewer’s imagination, and their assumptions about visual language, to do the actual work of making the scene scary.

Sometimes, we’ll get explicit indications that the stalker cam is the POV of an actual character in the movie, with clear and undeniable violent intent. But this is relatively rare. Much more often, the film will equivocate, not pin things down so concretely. Maybe this is the point of view of a killer, maybe it isn’t. Fake-outs are common. Sometimes, we’ve just been seeing the POV of a friendly character. Other times, the fake-out is even more extreme. We assume, because of how the camera is moving, that we’re seeing the POV of someone … but then a character will look directly into the camera, and not react. Turns out, this view doesn’t belong to anyone. Sometimes impressions of embodiment and malevolent intent that we have, as viewers, are just that: impressions, and nothing more. Sometimes a camera is just a camera. As I described in the previous episode, this fosters a paranoid visual system. As viewers, we become primed to suspect that the certain types of shots belong to someone with violent intent—regardless of whether this is actually true for each of these shots. Steve Neale describes this at work in Halloween, and you can find it at work in many other horror films, as well.     

As I laid out in my identification episode, the film theorist Christian Metz distinguished between “secondary identification”—our identification with on-screen characters, and “primary identification”—our identification with the cinematic camera itself. The stalker cam collapses these two types of identification, because it insinuates that the view we’re seeing belongs to a character within the fiction. In decades past, a fair number of feminist film critics rejected horror films outright, claiming that they forced the viewer to identify with the killers, promoting a misogynist worldview. And I don’t think they were entirely wrong. We definitely do identify with the killers in these stalker cam moments. But it would be a mistake to think that we only identify with them. There are other, simultaneous layers. The film’s visual narration is forcing us to identify with the killer, but it’s an uncomfortable culpability that’s being involuntarily forced on us. Meanwhile, the film’s story encourages us to sympathize with the women who are being stalked. As I’ve said before, the term “identification” has a tendency to to smoosh together several concepts, in a way that sometimes limits its usefulness. This is one of those times.

Despite being a simple effect that is utterly ubiquitous in horror cinema, particularly slashers, the stalker cam is extremely difficult to effectively employ in horror video games. You see a few stabs at it in the fixed-frame era of horror games. There’s one moment early on in Alone in the Dark when all of a sudden the camera switches, seemingly unprompted, to a view from outside a window, looking into the room where our character is. And then a giant rat smashes through the window, and suddenly the camera’s back in its usual place, to better facilitate combat. And Resident Evil would occasionally throw pre-rendered movies on the screen, showing the POV of a zombie or other monster as they make their way to the location of our player-character. But these moments weren’t interactive.

And it’s pretty easy to see why they weren’t interactive. Cinema provides lots of opportunities for for camerawork to be expressive. In fact, as viewers we’re so keyed in to the expressive possibility of camerawork that even something as simple as handheld camera shots become weighted with meaning, either consciously or subconsciously. Although the camerawork in games can be expressive—see, for example, all of the high and low and tilted angles in Alone in the Dark and Resident Evil that I explored in the first video in this series—the realities of gameplay mean that it has to be brutally functional, moreso than in cinema. And, as players, the way we react to what we see onscreen is conditioned by the awareness that games are essentially software, with basic usability requirements. When we see a camera trailing behind our character, we usually don’t assume it’s the point-of-view of another character within the fiction. (Although I suppose it could be.) Instead, we usually assume that that’s just a convenient place to put the camera, so that we can get the best view of the action.

All of this brings us to the Siren franchise.

After the release of the first Silent Hill game, its director, Toyama Keiichiro, left Team Silent and joined Sony Computer Entertainment’s JAPAN Studio, where he formed a team called project SIREN. The first game created by Project Siren was a PS2 survival horror title: Siren, released in 2003, also known as Forbidden Siren. This was followed by the sequel Siren 2 in 2006, and then Siren: Blood Curse, a semi-remake of the first game released for the PS3 in 2008. I’m going to talk about the entire series in this video … although I’ve never actually played the second one, because it wasn’t released in North America, for some reason. So I won’t actually be talking about that one. Let’s just forget about that one.

The Siren games play out on a relatively small number of maps—locations such as houses, hospitals, mines, and small rural villages that will be re-visited several times over the course of the game, as players return while playing as new characters, given new objectives. Populating these maps are the game’s main antagonists, known as shibito. These used to be regular human villagers, but now they’re zombie-like creatures. But they’re not like the zombies in Romero’s Night of the Living Dead. They’re more like the zombies in the fourth Dead movie, Land of the Dead, where the zombies have started exhibiting humanlike behaviors again, engaging in repetitive actions that are sort of wind-up parodies of the labor they used to perform when alive. The villagers that got turned into shibito were primarily working-class, and this is reflected in the repetitive actions they undertake, such as hammering and weed-pulling. As long as you’re quiet and don’t disturb them, many shibito you encounter will just keep attending to their work and ignore you.

Getting shibito to successfully remain unaware of you is a central goal of these games. The Siren games are very much stealth games, in a way that Silent Hill never was. There are three main design decisions that coalesce to inform the game’s focus on stealth. Taken together, they form something utterly unique. Say what you will about these games—and there are a whole range of valid opinions as to whether they work or not—but they offer a play experience you really can’t find anywhere else.

First, you generally don’t have a weapon in the games’ earliest levels. Later on, you do get weapons, both melee weapons and firearms. But the fact that you don’t have them early on forces you to understand the maps from a stealth standpoint first, before later becoming empowered.

Secondly, there are no health packs. Your health regenerates. If you get hurt, you can just find a quiet spot, wait it out, and be right as rain again. But on the flip side, the shibitos’ health regenerates, as well. The shibito you take down never truly “die.” They just fall into a dormant state. Wait long enough, and they’ll get up again. This serves as a two-way disincentive. First, it dis-incentivizes blasting your way through a level when you have a firearm, attempting to completely empty it out of enemies. Since a levels’ objectives will often force you to traverse it back and forth, and by the time you get back to your previous position the shibito you shot will have woken up again, there’s really no point to that. It wastes ammo for nothing but some temporary solace. But secondly it simultaneously dis-incentivizes going slowly. Many stealth games reward patience. Siren doesn’t, really. If you spend too long waiting for one enemy’s patrol route to line up, the enemy you previously took out is gonna get up again, and back into the action. So you want to plan carefully, but then act as quickly and efficiently as possible.

Thirdly and finally, the shibito all share a hive mind—and it is one that your player-character can tap into. The game features a mechanic called “sightjacking,” where your player-character closes their eyes and, through a process similar to tuning a radio, can see the world through a given shibito’s eyes. This is how you learn the vision cones and patrol routes of the shibito. Shibito are creatures of habit, so they move like clockwork, and, as in most stealth games, the point is to figure out the gaps in their patrols, and exploit them. The twist here is that you do this by taking the time to literally see what they see, and then figure out how to move outside of the bounds of that sight.

This is just straight-up wild. It does the seemingly impossible: it finds a way to implement the stalker cam as a mechanically-useful feature. Large swaths of playing a Siren game—I would recon hours upon hours of play—are spent looking through an enemy’s eyes, in first-person, watching them stalk the map, hoping that they’re not able to spot you in your hiding place. It’s not always clear what they’ll notice and what they won’t. And sometimes slipping up just a little bit, moving just a bit too loudly for instance, will alter them enough for them to wander off their usual route, to explore a little further, and to see you. And then you actually see yourself being seen, see your character being stalked in true stalker-cam fashion.

So, that’s the good news. These games implement a type of visual address that seemed nigh-impossible in games, charting something bold and new in the visual depiction of danger in games. But I can’t in good conscience make a video of pure praise for the Siren games … so now for the bad news.

The first Siren is not a welcoming game. The individual chapters are accessed through a maddeningly opaque menu, which supposedly illustrates causal links between past and future events. The idea seems to be that defeat for our characters is extraordinarily likely, and the only way we can reach a victorious ending is through a chain of vanishingly-unlikely events. So the game makes you play through the same chapters repeatedly, with additional objectives and clues tacked on, which you have to piece together and complete in order to unlock alternate versions of future events, in which the protagonists actually succeed. It’s not a terrible idea, but the execution is just mind-bogglingly awful. The objectives are unclear to the point of being unfair. And the causal link they supposedly set up is just often absurd. For instance, in one chapter, the secondary objective is simply to “find the nurse’s shoe.” And then you have to tromp through the level, avoiding shibito in complicated and time-consuming ways, and then on top of that you have to press “X” in front of every bush you see, looking for a shoe hidden in some invisible spot on the map the game does nothing to indicate. And then you find the shoe, and … apparently this means that the character Risa can now confront her sister Mina, a nurse who has transformed into a shibito? But she … couldn’t do that unless the shoe was found? And can we just rewind a bit? Because … dude, why are you so perturbed that there’s a shoe left in a bush? There are literal monsters everywhere. Any one of them could have carted off Mina, leaving her shoe behind in the bush. Just … what?

But that doesn’t even hold a candle to the most ludicrous causal chain in the game. The game’s most laborious puzzle requires noticing  a four-digit code on the side on an out-of-the-way wall, realizing that this code is not for the padlock right behind you, but instead is in fact a timecode for a specific moment in an otherwise blank tape recording, which you have to fast forward to, in order to hear the code for that padlock (“six … seven … two … three”). Then, upon opening the padlock, you get a hand towel, which you then have to soak in water, then place in a freezer. And then … nothing happens, until much later in the game, when you return to the same area later as a different character, and have to retrieve the now-frozen towel, suspend it across a small gap, and place a piggy bank on it, to create a sort of makeship timer: so that when the towel thaws and loses its stiffness, the piggy bank will fall and smash, distracting an eating shibito who was previously guarding an ID lanyard. Picking up this lanyard is actually the goal of this entire chain of events! It unlocks … something. For some reason. The idea that a normal human being, without recourse to a walkthrough, could actually put this sequence of actions together is preposterous. This goes down in the annals of terrible moon-logic adventure game puzzles.

Needless to say, I wouldn’t recommend playing the first Siren without a walkthrough. I recommend Shunichiro’s, from 2004, still hosted on IGN. It’s quite comprehensive, and it makes the game MUCH more playable.

Siren: Blood Curse is a different beast. It drops all of the cause-and-effect pretense of Siren. You navigate through chapters in a linear way, through an easy-to-understand menu. When you’re given a task, it is explained clearly, to the point where the next place you need to go is marked on your map. And although it loses some of the character of the original game, I overall think these are good choices. Blood Curse is much more approachable, and I would definitely recommend it as someone’s entrance into the series.

That said, the Siren games never reached the popularity of Silent Hill, and Blood Curse presents a good case study in why. Even while being massively streamlined in comparison to the first Siren, Blood Curse is still kinda hard to play. And a lot of this comes down to how much players are expected to sightjack hostile NPCs to learn the ins and outs of each map. As I said before, the game expects players to plan movements extensively, before executing them quickly. But usually the best way to learn how to move through a map is to actually, well … move through the map. It’s hard to understand the layout of a virtual space if all you’re doing is passively watching the viewpoints of NPCs as they wander through it.

This is not surprising. There’s actually a famous experiment in perceptual psychology where the researchers put two kittens on a carousel. One was able to walk around of its own volition, while the other was in a cart thing, a sort of kitty-chariot. It was just dragged around, on the basis of the other cat’s movements. The kitten that was allowed to walk around grew up to be a healthy cat. But the other one was messed up. It didn’t have good depth perception or “paw-eye coordination.” So they theorized that in order to be good at visually-guided behavior, it’s not enough to just have a visual experience. You need to experience that sensori-motor loop, to know the visual consequences of your locomotive actions. The learning curve for Siren is killer. In the beginning, it’s hard to make sense of the maps strictly by way of the shibito’s POVs, which in turn makes the views themselves close to useless. After you stick with it, a positive feedback loop kicks in: the more you move around on the map, the better you come to know it, which in turns makes the information you get from sightjacking shibitos much more useful. So the game gets geometrically easier. I also think it gets scarier, because you have a better understanding of the danger that’s portrayed in the stalker-cam moments. But there’s no getting around the fact that the learning curve is rough. You really have to get off that cat-carousel for the game to start clicking.

Overall, the visuals of Siren are among the most radically non-egocentric of any game I’ve ever played. And I don’t mean “egocentric” in terms of “vain” or “conceited.” I mean it in its perceptual sense. The game expects us to comprehend a version of visual experience that isn’t tethered to a single body. In the first game, even the map reflects this philosophy. It shows the ins and outs of the space, but it doesn’t have any “you are here” dot. It doesn’t mark the surroundings relative to your player-character’s body. This contributes to the game’s steep learning curve, but I think it’s consistent with the game’s overall philosophy. We’re not supposed to be thinking egocentrically about the one particular avatar we’re controlling. We’re supposed to be thinking holistically about all of the moving parts, all of the pieces moving around on this map, and find a way to integrate seamlessly into this system, so seamlessly that we remain unseen. We’re supposed to adapt to the hivemind—in fact, one of the characters we play as in the first Siren transforms into a shibito over the course of the game, a fact that is only revealed as a twist as we’ve played as her in her final level. (“Tomoko!”) This is a fantastic little story beat that meshes mechanics and themes really well.

This episode concludes a mini-sequence that I began with episode four. The next episode is not going to be a direct continuation of any themes from this one. In fact, I’m still not entirely sure what it will be, yet. But, rest assured, I do have plans to continue the series. So stay tuned! And, as always, thanks for watching.