Gameology 2.0

According to my reading, Laurel's book is about how we can apply tenants of dramatic criticism (specifically Aristotle and Brecht) to the design of human-computer interfaces. Laurel brings some experience to the table; she worked at Atari and Lucas Arts, and wrote her dissertation on a proposed "interactive fantasy" program. The book is essentially an answer to the question posed in the preface: "When we look toward what is known about the nature of interaction, why not turn to those who manage it best—to those from the world of drama, of the stage, of the theatre?" (xii). Obviously, readers of Espen Aarseth's Cybertext will wince at that question. Shouldn’t we be "starting from scratch" instead of force-fitting a theoretical framework intended for another subject onto "human-computer interfaces?" Yet I don't share this conviction, and found Laurel's book enlightening on many levels.

It's unfortunate (to me, at least) that Laurel is so broadly focused here on human-computer interfaces rather than zooming in on computer games. While I can readily accept that drama has something valuable to teach us about computer games, I scratch at my head at her discussion of how it could help with the design of spreadsheets and word processors. Her thesis works something like this. Like Aristotle, Laurel argues that good dramas represent a single, complete action. All other aspects of the play work towards this representation. She writes, "We expect to have a beginning, a middle, and an end" (xiii). For this reason, she criticizes games like Space Invaders that have no definite ending and praises the Lucas Arts games for providing all three. After providing a gloss on Aristotle's Poetics and his four causes, she develops some principles of drama that should apply to HCI. Perhaps the most intriguing is to "think of the computer not as a tool, but as a medium" (126). This leads to her next point: "Focus on designing the action. The design of objects, environments, and characters is all subsidiary to this central goal" (134). Thus, we should build interfaces based on the final product that the user is trying to create; the interface should not only be transparent, but invisible.

The ideal HCI would work something like this. First, it would consider the possible things that a user might be trying to accomplish. Based on what happened next, the computer would deduct what is probable, and then (and only then) what is necessary. Consider a play. At the beginning, the audience has little to no idea about what to expect—anything is possible. Gradually, as characters, settings, props, and such are revealed, the audience begins to narrow the possibilities. If the stage and costumes are from the "wild west," the audience begins to narrow these possibilities into probabilities (i.e,. it's not likely that JFK will be in the play, and probable that there will be a shootout at some point). Each "action" on stage causes something else to happen; the sum of all these causes leads to the play's final cause, or the "whole action" discussed earlier. "Catharsis" occurs when the audience puts all this together and feels satisfied; the drama, as a type of extraction of human life and events, provides the "closure" that is missing from reality and thus allows us to purge ourselves of emotion (which I think more of as a sort of existential anxiety than "emotions" like sadness, fear, and so on).

While I appreciate her discussions of HCI, I'm more interested in what Laurel writes about computer games. It's very easy to apply what she says to graphical adventure games and understand what makes bad GAGs bad. Consider: "It is key to the success of a dramatic representation that all of the materials that are formulated into action are drawn from the circumscribed potential of the particular dramatic world" (58). Whenever a drama fails to do this, the "organic unity of the work is diminished" (58). It's easy to find examples of this problem in bad GAGs. The key phrase here is "circumscribed potential of a particular dramatic world," rather than some vacuous standard like "realistic." A virtual world might have different physics, history, etc., but it's still vital that it have unity. We are able to suspend our disbelief to a point, but if there are simply too many holes in the plot or gaps in continuity, we "fall out" of the drama and experience that dreadful anxiety that something has gone terribly wrong. For instance, in Myst IV, the player is required at one point to perform an unusual mouse maneuver to "wake up" a magical snake. It's an awkward and non-intuitive maneuver, and though there are clues telling you how to execute it, it just seems wrong. Why? Because the "circumscribed potential" of the interface has taught us that the only valid mouse maneuvers are those signaled by the appropriate change in mouse cursor; i.e., a directional or "possible action" pointer. We also learned early on that you can "tap" objects at will, but doing so only has an aural (and sometimes visible) effect, but is ultimately inconsequential; it doesn't cause anything to happen that would help complete the "whole" action of the game. There are countless examples of this sort of thing in Journey to the Center of the Earth, where the interface seems so arbitrary and unpredictable that it is ultimately impossible to determine how to circumscribe the potential.

Other games clearly delineate the potential of possible actions. The Secret of Monkey Island, for instance, puts all the possible actions (look at, pick up, etc.) on the screen as part of a menu. Thus, we know exactly what actions are possible and can deduce (based on the conventions of so many stories and movies about pirates) what actions are probable. Yet we eventually find what is necessary and the plot moves forward. Thus, The Secret of Monkey Island differs from drama not in kind but in degree; the only difference is that the player has been given a script with plenty of gaps that must be filled correctly to complete the final action. The experience is even more cathartic than watching a play, because we not only see the final action completed but actually have participated in its completion.

In short, the problem is not that GAGs have been linear and we need more dynamism; in fact, bad GAGs have been too non-linear and we need to improve our handling of the necessary constraints in place to help the gamer complete the final action. Laurel has interesting things to say about such constraints. She writes, "Constraints should be applied without shrinking our perceived range of freedom of action: Constraints should limit, not what we can do, but what we are likely to think of doing" (105). This is of critical importance in the design of a good GAG. For instance—it's always bothered me that there are so many interesting objects in GAGs that can't be picked up or manipulated by the player. They look as though they might be useful; yet we can't even click on them; they are merely there for aesthetic purposes. Other GAGs tell the player directly, "You don't need that item," or something similar. This is obviously a clumsy way to handle this situation. I am frustrated because I feel my "range of freedom of action" has been unnecessary and unfairly curtailed, yet I understand that by doing so the developer is trying to make my play easier by preventing me from getting "off track." Perhaps the opposite extreme are games like Shadowgate and Uninvited, which allow players to pick up all kinds of objects that ultimately have no purpose other than to clutter up the inventory. Yet when we encounter an "aporia," we feel obligated to try each of these objects (the trial-and-error problem). My feeling is that once we've demanded trial-and-error from a player, we've already failed. Ideally, the necessary action could have been inferred from probable actions which were inferred from possible actions. When this transition isn't smooth, the player gets frustrated, and the quality of the game suffers. Furthermore, I agree with Laurel that "the standard techniques for introducing constraints—instructions, error messages, or dialogue boxes, for instance—are almost always destructive of our engagement in the activity by forcing us to "pop out" of the mimetic context into a metacontext of interface operations" (102). This problem is especially apparent in early GAGs like The Wizard and the Princess, where a horrid parser constantly draws attention to itself. Indeed, the key challenge of the game is really dealing with a poor interface—the "meta-context" becomes much more significant than the context of the game (which is banal the point of absurdity). More interestingly, Lucas Arts games are noted for their many "inside jokes" about the meta-context. I noted in my review of Sam and Max how at one point the game appears to end when the avatar looks into a wishing well and wishes that the game were over; yet this is just another joke. Obviously, one of the more interesting aspects of the Lucas Arts series is its witty handling of the meta-context; where Roberta Williams' games stumble about in zombie lurches, Lucas Arts titles dance like ballerinas.

I need to cut off this review early (lunch is ready), yet I hope to return to these themes in-depth once Mat and I actually start writing our GAG book.