CR: There's still a scale, though. I mean, even in this industry, you've got guys like Tim Schafer, who wrote every line of dialog in most of his games, has his name on the box, and Fumito Ueda, who does Ico, and Shadow of the Colossus. But on the other side you've got Valve, who goes so far as to list everyone on the credits alphabetically and few titles. And then you've got everyone in between, and I think most people are probably somewhere between. That does seem to... and all those people have made great games.
PR: Yeah, and here's the thing. Obviously I'm being super...
PR: Yeah, and I'm disregarding the efforts of people who've done incredible work in independent game development, which is largely auteur-driven. I mean, debatably, a game like flOw isn't going to happen without Jenova Chen sitting there making that game happen. Right?
It doesn't matter who's working with them on it or not, ultimately, as good as the contributions may have been of other people on that, somebody needed to sit there and say, "Listen, guys. This is a game about flow." That's like an abstract idea, and to make that work, someone's got to hold that in their head. Yeah, I'm kind of being obnoxiously one-dimensional in my definitions of these things...
But in the specific context of writing, I think what I'm getting at is that authorial control, and the desire to cling to it, is a dangerous tendency in games. Because we're in an interactive medium, and our job is not to make the player feel anything, not to give the player an experience. It's to enable the player to do the things that make sense in the kind of metaphor of the game.
CR: That's kind of what I was getting at with that idea of kind of, novel to film to game, where you've got someone like Stephen King who, speaking of Kubrick, hates Kubrick's reading of The Shining - but at the end of the day, in film, the screenplay, or the original author's intent, becomes subservient to the film. A lot of people worked on it, but sometimes you've got the crazy auteur. And then that's what I kind of meant with games going sort of even further than that.
PR: Well, okay, you know where I'm going to be proven dead wrong, and so let me try to approach this with as much humility as possible. Where I'm going to be proven dead wrong on this, is when we finally support tools for the generation of systemic AI behavior that are so intuitive, like, as Chris Hecker was saying, essentially as easy to use as Photoshop, that are so intuitive that, at that stage, it really does become about generating a set of dialogs. And the AI has the intuition to do what an actor does, and do what a stage director does.
And at that point, maybe a writer can step up and say, "Okay, you know what? Let me just tackle this one on my own." You know? Which happens. It happens in film, and it certainly happens in other traditional media. Maybe we will hit that point where I can sit there and hold rehearsals with my AI until I get something that feels like the game that I want. I mean, I don't know.
CR: You've probably got a while before you're proven wrong on that one.
PR: Yeah, hopefully I'll have retired by then, or been committed to an insane asylum or something.
BS: How do you record dialog, when you don't know exactly when, or in what context it will come up, if that is indeed the case?
PR: Yeah, that's a really hard problem. In fact, the way we approach it is, it becomes a minimization effort, where we know that, to a certain extent, we are going to have to say, "Well, there are a dozen characters who could possibly deliver this information to the player. Let's see if we can build that scene as much as possible out of pieces that are reusable."
And that just involves an enormous amount of up-front planning. And luckily we have some pipeline... like we do have some tools for managing that. And then at that stage you know that maybe 10% or 20% of the content you record is throw-away. You know? It will never be heard. But hopefully the 80% will be usable elsewhere.
BS: Forgive me for not knowing this, but does the player-character speak?
PR: No. The player-character doesn't speak, and in fact this was one of the motivations we had for letting the player pick his avatar from the buddies. Because what that gave us was sort of the idea that the buddies represent kind of a mirror that we hold up to the player.
So as the player begins to understand that these are guys just like me - they're here, they have their own agendas, maybe they're even here to do the same stuff that I'm here to do, they're taking mission for the factions, I'm taking missions for the factions... The idea of that is that when I hear them talk, and I hear their back stories, and I hear the rationales that they give for doing what they do, I kind of see a moral spectrum there, and I know that I fit somewhere in there.
And as the player, hopefully I'm trying to fill that avatar up, as an empty vessel, with my own beliefs, and my own worries, and my own doubts. And that way I can kind of see how I fit in. And these guys represent a kind of a buffet of moral positions on that. Right? And that's the reason why. Because we didn't want to give the player-character a voice. Because, again, we didn't want to presume to know what was in the player's head.
BS: Writing to a character that doesn't speak, it's basically like writing a very long series of monologues, which is... it's strange.
PR: Well, yeah. It's super weird. Because everybody speaks in an oddly expository way. Although it has...
CR: Like Gordon Freeman.
PR: Yeah, yeah, exactly. Oh yeah, well, and that's the good thing about it. It's not like we were the first ones to do it. We can look at the way other things work. And again, because we were working this chunky modular approach, it gives us a little bit more freedom to be slightly more mechanical about how we deliver information to the player. So yeah, is the player going to feel like some of the dialogs are a little... odd? Probably.
As much as possible, we tried to offset that by having fairly naturalistic deliveries. Like, trying to get the actors to be very conversational. I mean, okay, is it David Mamet conversational? No, probably not. But we're trying as much as possible to... "Yeah, okay. I'm going to listen to this conversation. I'm going to get some essential information out of it. I'm going to get a sense as to who these people are, and whether I like them or not. And then the information is added to my objectives screen." Right? At the end of the day, we have that as our safety net.
BS: If enemies aren't really spawning in very much, do you have the ability to cleanse the game world of humanity?
PR: Well... no. We're not spawning enemies into active areas that the player is in currently. But there's enough fluidity to the world that... I think we justify the idea that if I leave a location, and I go across the map and do something else, and then I come back to that location a couple of hours later, that there are different guys there. I think we would not feel quite so justified in saying, "Oh, hey, he's not looking behind that tree. Drop a guy there." We won't go that far with it. We don't do low-level spawning. But we're willing, at a high level, to say, "Let's repopulate this location."
BS: It's seems difficult, just from a high-level perspective, to write, or to design a story for a game in which - or any scenario, really, although we do it all the time - in which there's a character in a world. Everyone is that character's enemy, and that character is still alive and going. It's very... it's an odd place to be, and taking a step back from it, it's like, how do you really write for that convincingly?
PR: Well, one of the central conceits that we indulge in is this idea that the two factions are kind of in a state of impasse. They're in a kind of uneasy détente. Neither one really wants to win outright, because neither faction wants to govern the country. So there's the sense of there being kind of like a low-level, low-intensity conflict, kind of on the boundaries between their territories. But the rest of the time, they're kind of guided largely by necessity.
So what that means is that both factions are willing to work with the player, in spite of the fact that he's occasionally doing jobs for the other side. They view him as a tool. He's effective. He's powerful.
As his infamy builds, they're even a little scared of him; they respect him. And I think the idea is really that, yeah, when you go into their turf, they're not warning their guys that you're coming, and telling them to leave you alone. You're still going to take a bullet from their side as well. But there is that sense that they're not necessarily in a position where they can just rally all their efforts to kill you, because they're also kind of holding each other off.
Like, if the player gets into a chase, and he leads - I don't even know if I'm answering your question, exactly - but if the player gets into a vehicle chase, and leads guys from one territory into the other, there are consequences to that. Like suddenly fights will be breaking out between AI, and all this kind of... So I think the feeling ought to be to the player that, you know, "I'm just one guy who's kind of weaseling my way around in this world." Right? And then later on, as the game progresses, he does become more of a target.