IF in the ACM literature, Part Four

I’ve been saving these for last, because they’re really the juiciest: a couple of articles authored or co-authored by Mark Bernstein.

Bernstein is the founder of Eastgate Systems, a company promoting serious hypertext. They sell a small but — within the hypertext community — highly respected collection of hypertext fiction and nonfiction, at serious book prices: much of it runs from $25 to $45. And they produce and sell Storyspace, a tool for hypertext creation. This is a niche market: the major works are self-consciously literary or pedagogical, and I think it would be fair to say that IF in general is a more populist form. At the same time, hypertext is a more successful niche market than IF: how many of us are selling IF game files at $45 a pop? how many would feel ballsy enough to try? And, leaving aside the commercial, hypertext also gets studied more extensively by academia, taught in more new media courses, and generally considered more serious.

Some of this has to do with accidents of community — with the sorts of people who happened to be drawn into creating each kind of thing, and with the ways in which they framed and presented their finished products. But I also think the media favor different kinds of content, and it’s interesting to look at why.

Hypertext is a medium which occasionally seems IF-like — after all, a Choose Your Own Adventure could easily be rendered as hypertext — but it offers fundamentally different possibilities to the author. Marie-Laure Ryan has compared the genres — I thought fairly even-handedly — in Avatars of Story. Jimmy Maher has written about a few hypertext works in SPAG, and his comments reflect the bewilderment and disenchantment that IF players sometimes feel when they encounter the storyless meandering and opaque interaction style of these:

Then we have the inevitable Michael Joyce hypertext fiction. The selection here, Twelve Blue, has come into my life several times in the past. After literally years of study, I still can’t figure out just how the interface is actually supposed to work. Perhaps Joyce is the anti-Apple of the computer world, or maybe my feverish clicking around the screen in search of something that will lead to more lines of overwrought prose is some sort of commentary on the state of postmodern life. Lines like this sound like extracts from one of the annual tortured emo-kid Competition entries: “Follow me? What choice do we have but love, what season after?”

I must admit that I also haven’t ever experienced a work of literary hypertext that I found as immersive and compelling as a good IF game. Some of this is due to the trend — which Ryan remarks on in Avatars of Story — that hypertext tends to describe internal states and mental wandering, or to involve a patchwork of very short anecdotes each of which can be told on a single page. The work copes with the nonlinear experience of the player by avoiding any kind of linear plot or long-term development that could be disrupted by our mode of access.

It is, I think, possible to write good fiction — or good writing of some kind (fiction might be the wrong word) — that explores a set of ideas fluidly, through anecdotes and images disconnected from time. Annie Dillard’s American Childhood is a memoir, but one in which time holds the reins loosely; I could imagine it being given hypertext form. But it’s also easy for this kind of work to seem formless, self-obsessed, and even dull, especially to people who are looking for interactive storytelling where the player gets the starring role.

Paradoxically, there also is some tendency for IF authors — I hold myself occasionally guilty of this — to feel that hypertext authors have it easy. We can all write HTML, right? How hard is that? The point is driven home by the fact that most IF languages can fairly trivially be used to write hypertext-alikes, without exercising more than a fraction of their programming power: see “Space Under the Window”, or “Threading the Labyrinth”. This is a bit unfair, since hypertext authors face some serious challenges of content design, as I just said, and there is a point where managing a large amount of data or large number of pages becomes as difficult as designing new procedural code. Storyspace is mostly devoted to offering the author lots of ways to keep track of his design-in-progress, and it’s clear that that’s a non-trivial task.

The fact remains that I know fairly few hypertext devotees in the IF community.

Bernstein has a similarly unflattering perspective on IF, as it turns out, and he expresses it at some length in several articles. For example:

Interactive fiction tools that attempt to construct detailed world models have thus far proven disappointing, in part because the combination of world models and the widespread assumption that the reader should enact the protagonist’s role is fundamentally incoherent.

(This from “On Writing Sculptural Hypertext”, Mark Bernstein, David Millard, and Mark Weal, Proceedings of the thirteenth ACM Conference on Hypertext and Hypermedia, 2002. 65-66.)

In other words, he thinks the reader/user should not have any real agency within the work: the reader may do things that affect the order in which texts are presented or the selection of texts used out of a larger body of possible texts, but this does not in any way mean being allowed to act the role of a protagonist and make game-world-embedded decisions. The reader sits outside the work, beside the author, moving bits around.

This gets clearer in another article, where he explains what’s good about Thespis, a system for hypertext-like development that does keep track of characters and do some moderate world modeling: (“Card Shark and Thespis: exotic tools for hypertext narrative,” Mark Bernstein, Proceedings of the twelfth ACM conference on Hypertext and Hypermedia, 2001. 41-50.)

The unimportance of the point-of-view character may be essential to Thespian hypertext. If the reader’s point of view is the hero protagonist, for example, the reader is led naturally to test the limits of the possible. That’s what heroes do. The drama rapidly devolves into a negotiation between reader and the world model; the reader asks to do unexpected [sic], the system typically responds with incomprehension. Ironic detachment makes things worse, not better, the reader-protagonist still wants to test the rules, and detachment invites the frigidity and sophomoric contempt that so often mar computer entertainments. (45)

…and again railing against the parser:

The game is rigged, and constantly calls our attention to the deception. Whenever we struggle against the bonds of fate (and the boundaries of the system), we’re told, “I don’t understand.” The more we struggle — and the more conviction and intelligence be [sic] bring to the action — the greater the likelihood that the system will find no appropriate response. (Ibid., 47)

This is good stuff to pick apart, because it’s so annoying to be on the receiving end of (“frigidity and sophomoric contempt,” eh?) and because in so many ways he’s right. IF does constantly have this challenge: to teach the player how to interact. Every new work, in telling a new story, must restrict its player’s actions in a new way. It must focus the player on actions that are implemented and important to the story; it must give some kind of answer (other than “I don’t understand”) when he tries something that should be story-relevant but which is not allowable. But it need not allow the hero-protagonist to do literally anything, or even “anything that the world physics would allow”, or “anything that I the player would do if I were there”. Bernstein writes:

Even if we could experience Hamlet on the holodeck [referring to Janet Murray’s image of fully interactive literature], it wouldn’t work. Tragedy requires that characters be blind (as we ourselves, at times, are blind); if you let a sane and sensible reader into the room, everything is bound to collapse. (47)

…and that’s true, but it doesn’t observe the player/protagonist separation that we have often talked about in the IF community. Nick Montfort has written about how the protagonist’s character often emerges from what the game doesn’t allow the player to do (see his article in Second Person). When we keep in mind that a sane reader/player can manipulate and guide a blind/misguided protagonist, the problem seems less paradoxical.

I don’t quite accept the Aristotelian formulation that tragic characters must have a tragic flaw — at least, Aristotle’s description of tragedy is as often wrong as right, when we hold actual Greek plays up to it, and I don’t find that it works any more consistently as a guide to more recent literature — but we might amend the tragic flaw to “a tragic limitation”: some inability to see the truth, some lack of self-control, some fatal inclination, which makes the hero less than a god, and traps him into the story as it happens. This might be juster to the Greek than the traditional translation anyway: hamartia is a missing of the mark in archery, the slipping of a foot on a dangerous slope. Ethically, it is an unintentional crime, a manslaughter rather than a murder, as well as (in the New Testament) a sin. An interesting related word is ate, common in Homer. Ate is what made the suitors foolishly pursue Odysseus’ wife; it is what made Odysseus’ men eat the cattle of the sun, and die for their wrongdoing; it is what made Aegisthus ignore Zeus’ warnings and seduce Clytemnestra. Ate is frequently translated “madness”, but it connotes as well not quite being able to comprehend what is going on, not being able to read the plainest omens in the world, being too manic to even try for self-control.

I present this little vocabulary lesson because I think a tragic character need only display hamartia — mistake, limitation or error — and not ate, a more specific kind of failing that involves, as Bernstein says, (sometimes willful) blindness to what is truly going on. It is hard to inflict ate on a player character and yet still let the player understand the story. There have been a few attempts, games with severely unreliable narrators in which what the game describes is not what is really happening, but most of them don’t work very well. The tension of stories about ate-driven protagonists always arises from the audience knowing what the protagonist does not, from the audience watching while the protagonist becomes more deeply mired and trapped. Bernstein is right: it’s awfully hard to get the player/character interface right in that situation. You either wind up fooling the player until the very end (in which case there is none of the rising distress that comes from watching someone screw up on-stage), or you clue the player in but then force him to keep playing the protagonist as though he were blind (in which case the player gets more and more irritated with the artificiality of the whole set-up). Perhaps the best job of it I’ve seen is “Shade”, which walks a delicate line between concealment and revelation. “Delusions” might be a good runner up, though it blew some of the effect through lack of focus. But this is hard to pull off.

Games where the protagonist just has a serious problem of some kind, though? Those can be great. Look at “Rameses”. Or look at “Narcolepsy”: I didn’t find that entirely successful, but whatever else didn’t work, the protagonist’s habit of falling asleep at weird intervals *did* work just fine. The player can get to play the rational part, the mind and the will, of a character, and still leave a lot of room for the IF work to dictate the irrational fears and desires, the physical handicaps, the social maladjustments.

After all, we don’t get to dictate those parts of our real-life personalities, either.

12 thoughts on “IF in the ACM literature, Part Four”

  1. Man, with all of the great work being done lately (the ACM literature and your responses, Nick Montfort’s thesis and other published works, Dennis Jerz’s archaeological success, and so on and so forth), it would seem like an ideal time for an academic conference on IF theory, design, and authorship…

  2. Certainly, it seems like the momentum and pressure are building for a conference of some kind. I think the people are right who advocate for piggy-backing on some established convention with its infrastructure in place. But then we all show up and talk very, very seriously about IF for a weekend, and have a good dinner. Then do it again the next year.

    I’m conflating two ideas here, I think, the academic conference, and the fan convention with panels. But maybe it’s time to make a presence at both.

  3. Thanks so much Emily for these articles – your comments (especially in these posts relating to ACM papers) have been entertaining and thought-provoking.

  4. In 2001 I wrote a brief response to Bernstein and Greco’s paper on Card Shark and Thespis, a system that I believe had some commonality with our then-in-development architecture for the interactive drama Facade. It was published in the edited volume “First Person”, alongside their paper.

    Click to access SternResponseToBernsteinGreco.pdf

    I should note that my perspective/preference for protagonists in interactive stories is that they are ideally 100% controlled by the player, without any narration or avatar semi-autonomy telling me (the player-protagonist) what I am supposed to be feeling/thinking. I realize that’s a departure from typical IF and many games.

    An excerpt from my response:

    “‘Even if we could experience Hamlet on the holodeck, it wouldn’t work. Tragedy requires that the characters be blind…’ I agree, it seems likely that certain types of stories such as traditional tragedy may not work as an interactive story, for the reasons Bernstein and Greco describe. Instead authors will need to tell the kinds of stories that do work interactively. Façade is a more open-ended, explorative, psychological situation. Is this drama anymore? We hope to understand this better once we get a chance to play with the finished work.”

    Now in 2007, after playing the finished work, my feeling is that yes, a coherent, satisfying dramatic story with enough flexibility to progress and end in a variety ways, created in real-time as a collaboration between player and system, is possible. Facade doesn’t achieve that, but it has informed me enough to believe in interactive drama as a viable form.

    Hamlet, and other tragedies requiring _ate_, are not well-suited to interactivity. Instead, work to find narrative forms that do work.

    Regarding the player testing-the-system’s-limits, and play devolving into player-system negotiation: this is of course a major challenge for player-as-protagonist approaches; Facade had this issue in spades.

    But what we found is, many (most?) players want to have a good time playing an interactive story, and assuming the story itself is enjoyable and the infrastructure supports satisfying agency, those players will often cooperate, even collaborate, with the system, playing within the sensible boundaries of the domain of the narrative. For example, those players treating Facade as a joint performance fared well and enjoyed themselves. (Cooperation doesn’t mean “you must follow a linear track”; there still must be a wide range of interesting avenues supported.)

    But even for players amenable to cooperation/collaboration, they will need to experiment to understand where the boundaries are. We want to maintain believability and suspension of disbelief, and need to gracefully handle testing-of-limits. A solution is for the NPCs to communicate to the player the limited conversation topics and limited narrative directions they _can support_, what Murray called “scripting the interactor”, avoiding “I don’t understand” responses. The developer needs to put the required effort into supporting a broad and deep range of play on those focused topics and directions. As I mentioned in my 2001 response, that is very difficult to achieve, but should not be abandoned for that reason.

  5. p.s. Emily, I just read your “What Would James Bond Do?” post, and see that in your third point, you said similar things about managing expecations and deeply supporting the focused domain of the narrative. :-)

  6. Hamlet, and other tragedies requiring _ate_, are not well-suited to interactivity. Instead, work to find narrative forms that do work.

    Not with the player as the protagonist, I think; that’s true. But I could easily see the player as a side character, one of those who is observing the train wreck in progress and trying (of course with no success) to stop it.

    Otherwise, as you say, I think we largely agree. (I have still not finished reading my copy of First Person from cover to cover: I’ve skimmed through, but it needs more thorough attention sometime…)

  7. It is hard to inflict ate on a player character and yet still let the player understand the story.

    Actually, an interesting way to do this is to play on a player’s sense of completionism; hide an ending (or some other kind of content) behind the “bad” route (or choice, or what-have-you), then make the player aware of this fact, tempting them with the promise of sweet, delicious content (whether story or otherwise). Taghairm, for instance, had this effect on me, and I think the new game Undertale does something similar with its bad route. Of course, if you do this, you always run the risk of the player refusing to take the bad path (like what happened with you and Taghairm), so it requires careful writing to make sure the player gets a complete experience whether or not they take it. It’s certainly not a perfect solution.

    Of course, the Emily Short of 2007 may not have known this, but the Emily Short of 2015 can probably understand it just fine. (Hi, I’m just searching through the archives for mentions of “Hamlet on the Holodeck”, pay me no mind.)

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: