Analysis: The Romance Problem

[In this analysis, writer and designer Emily Short considers the problem of romance in gameplay, examining several examples to highlight common pitfalls -- and rare successes.] A recent NowGamer article quoted Peter Molyneux, talking about how videogames could grow stronger in the romance department: "For a truly emergent (and yet well-written) romance simulation to be possible, there needs to be a way to not only generate “romantic content” based on player input – and how this would actually mix with pre-recorded dialogue is unknown – but the characters involved also need to be able to build on organically derived behaviours and motivations. The midpoint between The Sims and Baldur’s Gate, in other words." I agree that's one way it could happen, and definitely the way that Peter Molyneux would be most likely to endorse, all things considered. What I worry about is that, if one frames the problem simply as a problem of building a "romance simulation", the result will be a slightly slicker but even less narrative variation on the dating sim: Statistics reflecting how much time you've devoted to each of your possible mates, but no connection between that simulation and the other kinds of gameplay available in the game, and no solid structural ties between the sim events and the rest of the story. Fable 2 already goes a bit in that direction, rewarding the protagonist with romantic relationships when s/he's spent enough time making flirtatious faces at the lover of his or her choice. Interactive Romance Or Interactive Romance Novel? Interactive fiction romances have a history back to the '80s, and they've often taken a different route to handling romance from any that Molyneux describes: there are scripted and pre-determined interactions with The One, but at the end the player gets to pick an ending, either going off with the romantic hero, settling for "let's just be friends", or abandoning him entirely. This is the tack taken by Amy Briggs' wonderful and underappreciated Plundered Hearts, as well as by subsequent amateur IF including Kathleen Fischer's Masquerade and my own (I admit, rather Plundered Hearts-influenced) Pytho's Mask, where it's possible to go so far as to betray the obvious romantic lead and hook up with his enemy. There are a couple of advantages to this approach. I like it -- and used it -- in part because it somewhat subverts the aching cultural conservatism of genre romance. Your average bookstore romance may be aimed at women and may feature female characters who are in some respect powerful or independent; it's no longer the 1970s, and the plot line of the gorgeous secretary and her hunky boss has mercifully passed away. And depending on the specific imprint, a romance may fall anywhere on the sensuality scale from a few chaste kisses to torrid erotica on most pages. But some essentials remain, including the absolute guarantee of marriage and children for the hero and heroine, and secondarily often an emphasis on the overpowering strength of family ties in general. There's an ideological reason, beyond mere convenience, why so many romance novel series follow a set of siblings, who may have disagreements and (of course) will have different tastes and traumas but ultimately will wind up in harmony, reaffirming the bonds of family. Letting the player decide whether to get married or not -- and whether or not to define her protagonist by her relationship at all -- feels pleasingly freeing to me. Of the games I noted above, Masquerade probably makes the most significant use of this choice: thematically the game is only partly a romance, and more largely about female self-determination in a male society. Forcing the player to have the heroine marry at the end would have betrayed the meaning of the piece: yes, it should be an option, but it cannot be the only option. Using multiple endings also removes some of the design burden: it lets the author write a game that is intended to make one of the characters seem attractive, but provides a design out if that intention fails. One of the tropes of the romance novel is that the hero and heroine are incredibly attracted to one another from the first instant they meet, that every touch sets off fireworks of lust, and that sex, whenever they have it, is just flawlessly thrilling and soul-mergingly significant. But it's easy to write stuff like "His oddly calloused fingers brushed hers as he accepted the wounded puppy, and Lucinda's core thrilled in awareness of that searing touch, even as she recognized that before her stood the fulfillment of her secret dreams: a manly, red-headed veterinarian who also played the banjo." It's much harder to tell or show or imply to the player that she's attracted to someone and have it stick. What if the body type of the hero doesn't do it for her? What if his dialogue comes off as cocky or his voice acting reminds her of her annoying uncle? What if banjo fetishism is just not her thing? Plundered Hearts manages by inserting faintly parodic cut scenes between the protagonist and her hero, in which the player is powerless to resist -- a neat trick that equates the non-interactivity of these events with the traditionally irresistible force of sexuality. At a gameplay level it works, but it works at the cost of slightly distancing player from protagonist. Letting the player decide whether to end with a romantic relationship, or (along similar lines) choose which of several romances to pursue, lets the author off the hook a little on the job of generating a specific emotional reaction in the player. But there's a risk here. Both Molyneux's suggested strategy (let the player completely control the romantic shape of the story by relying heavily on simulation rather than structure narrative) and Infocom's (give the player a structured experience but let her decide for herself how the protagonist is motivated to react) use player choice as a way to avoid some authorial responsibility for the player's experience. Sometimes that's completely appropriate, as Masquerade demonstrates. Often it isn't -- and indeed I suspect it is likely to be less, rather than more, acceptable in games that do not specifically bill themselves as interactive romance novels. If the romance is not the whole point of the venture but rather a subplot, then it probably has functions within the larger plot -- like making the player care about someone or something, as demonstrated in the Grand Theft Auto IV example Molyneux analyzes -- that may founder if the player chooses door A instead of door B. Moreover, adding a "romance simulation" to a game isn't always going to be compatible with the rest of the gameplay. What if we want the romantic plot to emerge from and be consonant with the gameplay, to be knit into the procedural level of storytelling? Hang On, How Does Romance Work? Your romantic interest is supposed to be on your side, your best friend and confidant; a source of support in bad times and an inspiration in good; compatible in mores and lifestyle and the all-important sense of humor; an intellectual and emotional companion whose need for you healthily balances your need for him. Right? Wrong. I mean, in life, yes, that's wonderful if you can get it. In a romance plot, however, it doesn't work at all. In a romance plot the conflict has to arise from the romance itself. Sometimes that can be due to hostile external circumstances (he's a Montague!), but often the best romance plots arise from character conflicts between the very people who are in love: Harriet Vane, afraid that marriage with Lord Peter will lead to her personal and intellectual subordination. Lucy Bart and Lawrence Selden, in sympathy with one another at a fundamental level but driven apart by Lucy's social ambition and Lawrence's wavering opinion of her character. Beatrice and Benedick, both happy in singledom and much more confident and comfortable in battles of wit than in romantic self-revelation. Darcy and Elizabeth, whose flaws are so important that they make the title. Romance novels and romantic comedies often fall back on a fairly small and contrived set of standard conflicts: he doesn't want to commit or fears falling truly in love; she suffered a trauma and can't trust people; for some reason one of them told the other a whopping lie that they must farcically continue until Act III. But even in the more canned and generic examples, a romance is a push-and-pull affair, in which the Other is sometimes desirable, sometimes exasperating, sometimes both at once, and may make demands that are very difficult to meet. The Gameplay Of Romance In the examples Molyneux outlines, gameplay designed to support a romantic arc takes one of two forms: 1. Conversation/social options in which the player gets to choose whether to be affectionate/romantic or not, and in which the affection score gradually mounts up to provoke warmer feelings in the Other, perhaps earning romantic or erotic cut scenes; and 2. Missions for the sake of the Other, either because the Other keeps getting into dangerous scrapes and needing to be rescued, or because she sets these missions as a test of affection. (Typically it is a she, though Ciao Bella is an exception in the other direction.) In the worst of these cases, especially in games aimed at a heterosexual male player controlling a heterosexual male protagonist, the player is effectively encouraged by the gameplay to regard the Other as a cold-blooded manipulator who is withholding gratification in exchange for goods and services: a problematic model of male-female relations and one that reinforces some fairly corrosive stereotypes, though I don't think most designers intend the effect. Where there are romantic conflicts in games, they're far more likely to arise from external conflicts (someone has kidnapped the princess, again) than from internal ones. Just occasionally there are more interesting cases where the potential romantic partner may or may not be trustworthy, a staple of genre noir and anything involving James Bond. But it is rare for game conflicts with a potential romantic partner to arise from differences in belief or personality; for the romantic partner sometimes to help you and sometimes to get in your way; for the negative conversation/social options to be part of the expression of a developing relationship rather than a way to drop your stats so low that The Other doesn't like you any more. What's needed, from a gameplay perspective, is a romantic partner who is sometimes also functionally the villain. There's a reason people write buckets of fanfic about the secret love of Draco Malfoy and Harry Potter: passionately clashing with someone is a form of intimacy. It raises the emotional stakes between those two characters far more reliably than attempts to portray attraction in interactive form. One of my favorite IF romances of all time takes just this format: in Graham Nelson's Jigsaw the elusive character Black is alternately friend and foe, someone whose agenda you're actively working to undermine but to whom you're strongly attracted and whose help you occasionally need. I'm absolutely not arguing that romantic relationships in games should be abusive or violent, only that the protagonist's own goals should sometimes conflict with the romantic interest's goals in ways that unavoidably affect play. Maybe we want to let the player choose to sacrifice his goals in order to keep his romantic interest happy. Maybe we want to force a showdown and then let the player try to pick up the emotional pieces afterwards if he wants to. Different solutions for different stories. But the romantic conflict has to be reflected by the gameplay itself, and it will be most effective if the driving forces are personal rather than environmental. At that point you give the player the agency to make the romantic interest happy or sad or hurt, far more deeply and far more integrally than by consistently picking one flavor of conversation option. You create the possibility of anger, shame, forgiveness, compromise, resolution; of world views and character traits that either gradually converge or force the pair to split permanently. Note that none of that recipe says anything about how procedural the implementation can be. I'm not against Molyneux's "romance simulation" as long as it is well-integrated with the rest of the gameplay and the narrative. If we want AI, if we want a simulation that shapes itself to some degree to the player's choice, if we want to approximate an aspect of The Sims 3, then perhaps we might automatically create a love interest for the player whose moral traits and personal history are partly compatible, partly in conflict with the ones the player chose for the protagonist. (I always found my Sims to be most fun when they had a few incompatible traits.) What matters -- however it's implemented -- is giving the player mixed motivations and the means to express his choices through the core gameplay. This is also, incidentally, why you can't expect a writer to write you a strong romantic lead if you bring him in after the design is already set in stone. But you knew that. [Emily Short is an interactive fiction author and part of the team behind Inform 7, a language for IF creation. She also maintains a blog on interactive fiction and related topics. She can be reached at emshort AT mindspring DOT com.]