Excerpts from an interview with David Gaider, senior writer at BioWare, conducted for the piece "Talking, Believing, Knowing: Making Friends in Dragon Age" in Kill Screen Issue #3.
In Dragon Age: Origins, we see social skills as art to explicitly master in the way combat is mastered. How do you feel dialogue trees allow for a tactical use of social skills?
Generally it's about reaction, right? If a player takes an action, he gets a reaction from the character, as long as it's consistent with that character. The player is learning exactly what kinds of responses generate a specific reaction from that character.
So, you look at Morrigan, for instance, and she reacts very differently to things the players says than Alistair does. The player has to learn to keep that straight. Alistair sometimes will pull away if you jump too far with him, joke or talk not very seriously. For him, the conversation may suddenly push too close to home. Sometimes, he jokes right along with you. That's how it is in real life. If we're generating a kind of verisimilitude, that's when the player sheds the feeling that this is a videogame and these are scripted responses that are just dropped.
Let's talk about structure. Though dialogue trees have their limits, they also seem to be incredibly unstable. There are endless conditionals. Have you found a somewhat distilled approach to writing the primary path and branches, or do you outline the entire structure beforehand? Do you start with question hubs, or work from the outside in?
Well, it's hard. There is, first, as assumed, no "write it all" approach to dialogue trees. One would assume that it's easy, but unfortunately, this is a different type of writing altogether. I've found that people who have been fans of tabletop gaming—say, have been a game-master—have an easier time wrapping their head around the dialogue trees, just because they are used to accommodating a player. Whereas, someone who writes prose is more used to thinking, "I control the dialogue, the viewpoint, and the internal thought process." Game writers keep in mind that we shouldn't presume the player has a particular motivation or that he will respond in a certain way. It's more important that the player feels agency.
The way we break it down: Writing trees is a matter of copying the branches, allowing for reactivity, while still maintaining dialogue that fills up the "higher" parts of the story. For me, when I sit down and first start writing dialogue, I don't even write full lines. I just write down the rudimentary concepts. I'll write it all in funny phrases, like cartoon talk ("Oh my god! Are you kidding?"). Then I lay down the structure of how things will play out. Then, I go back and I fill in what the actual lines are, and expand.
You have to be careful about branching. If you branch and branch, you need to bring the conversation back eventually to a bottleneck to tie things together. Unless the branch is a significant one—say, someone has actually made a decision that changes things. Then the conversation should go off. I know some players like to think that you can just branch, then branch off some more, and that every branch should follow its own line of conversation. That is when dialogue starts to become untenable, and the whole just doesn't work.
What's surprised you most about players' reactions to your characters, and the dialogue they cite? You've noted feeling like a "behavioral scientist trying to determine cause-and-effect" when hearing of fans' reactions.
You know, it is always so surprising to me. For one, before a game goes out, all I can think about are the flaws. Back in Baldur's Gate II—right before it went out—after going through the whole development process, all you can think about are the bugs, the scenes that you cut, all the missed opportunities. You just think, "This is such a pile of crap. Oh man, what were we even thinking?" Players—they don't see that stuff. That side of development is all we writers can see, but of course players see something very different. They start reacting to a world in ways that you never really saw or thought of.
Specifically, I would say the reaction to Morrigan and Alistair both have been very positive. I love the fact that in the case of Morrigan, people fall into such extremes. They either hate her, or they love her, and that is perfect. With some other characters, you can get a reaction that is mixed, or, nobody really even cares about them and you're not sure why. During Origins pre-production, there was talk that predicted that Ohgren [a red-haired, lusty dwarf] would be the most popular character, and that really wasn't the case. People are always going to offer their opinions on characters, and the writer can't take it personally.
I began to particularly feel like a behavioral scientist when it came to the romances. When I started writing romances back in Baldur's Gate II, I was gauging the reactions. You start thinking, Okay, where can this go? How far can we take this? Where haven't we gone? I would plug one query in and trace how another character would respond.
Origins' romances came about from a number of experiences. In the last romance I'd written, I wanted to draw out if I could make a male romantic character that female players would actually love. I wanted to know what was wrong with former female-oriented romances, what was lacking there, and how they differed from actual romance. It's difficult, because you can't really go to prose for the answer, because this is an interactive romance. I went to this site called "Ladies of Neverwinter Nights," and I asked the female players there, How do you guys feel? What do you want to see in romance? What kinds of elements and moments were you missing? And they told me. They were very frank about it, and having gotten that excellent feedback, I tried to incorporate that feedback into Alistair.
Is there a type of non-romantic intimacy that Morrigan desires or is intrigued by? On my playthrough as a female human noble, I found myself developing friendship with her, or, at least, what I thought was friendship. It was enjoyable, but not entirely "necessary" in terms of the game's progress. Are there understandings, like this, that are "useless," that have no immediate benefit, in the game? Or are most understandings, developed through the conversations, useful in some way?
Useful is a hard term to define, isn't it, as it would depend on the sort of player playing. Are friendships useful in terms of direct use of the game? Well, not so much. A character has got certain abilities if you raise their approval, and there are some benefits, I guess, there. Some new abilities could open up. For the most part, building a relationship with the character is supposed to be its own reward.
From my perspective, your party followers are your cipher. They are the way you experience the rest of the gameplay. Most of the moral decisions you make in the game really evolve from things and events that the player has a very limited interaction with. There's a line of story involving Mordin in Mass Effect 2, in which he talks about how he lost focus on saving his nephew back when he was home. Thinking about saving the galaxy is difficult to wrap your brain around, but thinking about saving one person is something you can understand. The same goes with the moral decisions in Origins. There are, of course, a lot of moral decisions you're having to make. You experience your conflict through the conflict of others. You'll have Alistair saying one thing, and Morrigan saying another thing, and their sudden conflict is your subject. So, there is the experience level of the decision itself, and there is also the level at which you are experiencing everything through your followers. They are the people you spend the most time with. They are the people you are supposed to care about.
The more the player feels that he has developed a real relationship, the more the decisions in the game have relevance. That's the usefulness of the friendships to us, and in the game. I'm thinking back to Knights of the Old Republic, and [...] the problem we encountered was with same-gender players; if you were playing someone of the same gender, there was no romance. You also felt like there wasn't an actual friendship developed. You had gone down the romantic path, and then everything fell short of the actual "love" part.
So when you mention Morrigan, and how a friendship evolves with her, that's great, because that was intentional. Just because there are romances for all these characters, doesn't mean we shouldn't be able to feel like we have another kind of relationship with them, especially if they are, like Morrigan, such an important character to the overall plot. I wanted a female player to feel like Morrigan was a friend, someone whose shell she had to crack (but in a completely different way from a romantic player).
One astounding element here is how faith is enacted, and how real faith feels in this world. The heart of conversation is belief and trust, and at the core of this exchange, imagination. Take, again, Morrigan. I'm curious about how much of our response to her is rooted in our imagination of her need, of some deeply buried vulnerability. Do we rush to fill in the blanks her conversation forms the outline for?
It's a great point, because faith, as we describe it in the game, is basically about not being absolute. In order for there to be faith, there has to be doubt. Whether you're talking about the characters' interaction with the Maker, or talking about the player's interactions with another character, you can't make anyone's reactions absolute. You can't say, if you impart A, then you get B back. So you always leave room for the player to interpret. If Alistair pulls away, or Morrigan gives a little smile, you have to leave some room in the player's head to interpret what is behind that reaction. Clearly I know what it is. As long it's going there for me, it's going to be there for the player. How they interpret it may vary, but they may eventually see that your picture of the member is more open.
Like you said, with Morrigan, some people hate her and think she's cold. Some people felt like they cracked that surface and only they could have cracked that surface and saw the sort of frightened girl that was underneath, right? I know it's there, and so it's just a matter of leaving that room so that when you're interacting with her, you never quite know where you stand.
We have the approval bar. In a way, that's a bit of feedback for the player, just because we can't have the nuances of social interaction that you would in real life. You wouldn't be able to detect the variances on the level of gestures. I personally never like to leave the player feeling that they know exactly where they stand. That's sort of like real life, too. Even if someone is completely in favor of you [...] the player should feel like he's interacting with a real person here. There are degrees of uncertainty, even at that level of intimacy.
I'm curious about how much group banter is meant to influence the player's feelings about characters.
Well, when it comes to the banter—it's generally between two pre-set characters. It's a little tougher to write dialogue with the player, simply because, from a prose-writing perspective, you're not certain what the player's personality is, and you don't want to presume on their behalf. At that point, it is a bit hard to get an easy back-and-forth cinematic banter. When you have two characters, like followers, you know exactly what they think, as a writer. You know exactly what they are. It's much easier to write for their back-and-forth.
Banter allows us to have conversations in the game that would be very difficult to have with the player. I wish we could do the same with the player, but in order to get the effect we want, it would involve a little bit too much presumption about the part of the player. We don't like to do that. We like to leave the player with some agency. Maybe we'll get there eventually, to the point we can cast some kind of interchange in which the player is directing the banter. [...] Taking Morrigan as an example, her banter was designed to pick apart her feelings on faith and tease out her exact relationships with her party members. This really opens up for the player to listen. We reward a player who does listen. The player can get a feel for what kind of person she is, on top of what she is able to talk about. I imagine, for a lot of players, their first response is: "I should be able to talk to her."
On another level, you realize that not every conversation can be written for the player. "Scene" conversations develop characters according to their position, rather than having them say, I feel this. So it's the context you learn from. You can watch their interactions and see what they do, and infer from that, exactly how you feel.
What do you see or hope for in future writing for RPGs?
The cinematic element is definitely becoming bigger. I think we're trying to find a way to incorporate cinematics without making the fact that it's cinematic so expensive; you just can't have the volume of interaction that you had before, and that comes down to the art. There's this hump that we're getting over. Just adding in all the 3D animations and the cinematics was making the development so costly, that we were having to pare down the level and volume of interaction you were having. I think players were treating this as a bad thing, and it's not, because there is potential there. When we watch movies, we're seeing reactions and we're seeing emotions. You never want to separate this element from the player's imagination.
Since it's a game, you want the player to have agency, and not make everything there on the page for them to see. There is a way, I think, for writing, game interactive-ness, and cinematics to get to a point where the player feels like they're in charge. They see the writing, then see the emotional reaction on the screen in a way that's believable, and fluid, and reactive.
We're slowly getting there. The level of sophistication of our cinematics tools have been increasing leaps and bounds and I see myself, in the future, working a lot more closely with the cinematics people. Take how it was before. I remember back in Knights of the Old Republic, it was a huge deal to get the models to interact. Now, it isn't that big of a deal, and you can develop endless levels of subtlety. I'll ask the cinematics teammate, "You can do that? Really?" and he will say, "Yes! Let's do that; let's try it." This teamwork gets easier as the tools become more advanced.
Suddenly, we're opening up a different kind of storytelling. Having that new type of story while keeping the story interactive—well, the interactive element is really the goal. There is potential for something much bigger there.
Illustrations by Daniel Purvis