In a response to my recent comments on his work, Mark Bernstein writes:
I wasn’t actually talking about ate and hamartia, or not only about them: there’s a simple logistical contradiction that lies at the heart of IF. You’re the hero. You’re in a tight spot. Things seem hopeless.
>What do you do?
>Well, what would you do? What would James Bond do, or wily Ulysses? They’d do something brilliant, totally unexpected, something nobody would have thought of. They’d do the one perfect thing that only they could do to get out of this tight spot.
So, you rack your brains. And you come up with something incredibly clever, unexpected, and far-fetched. Something perfect! But I’m just a writer, not a hero: have I thought of your incredibly clever strategem? If I have, you’re deflated: it’s not heroic after all, it was just a puzzle and you’ve supplied the correct answer. A tough puzzle, maybe, but (obviously) the author was here before you.
And if I have not been here before, the game’s going to say, “I don’t understand.” So, heads you lose, and tailsyou lose.
I have several responses to this:
1) Bernstein’s characterization of “what a hero does” does not describe the role of the hero in tragedy. A tragic hero is rarely like a folkloric trickster hero. There are roles for such characters in ancient drama, but they appear more often (a bit transmuted) as slaves in New Comedy, or in satyr plays. Tragic heroes are more usually confronted with a terrible situation which they must resolve through determination and decision (Antigone making up her mind to refuse her uncle; Iphigenia resolving herself to death; Prometheus, Ajax, and Philoctetes refusing to be coerced or persuaded or shamed by people worse than themselves; Medea, Electra, and Orestes screwing up their courage to take fantastic revenges). Oedipus arguably is characterized as intelligent, but in the course of the play this is put to work mostly to find out a (long past) story, and it doesn’t resolve his problem; it only makes things worse. And, at that, I’m not sure that we’d say he invents a “really clever scheme”: he more falls into the truth by luck. Come to later tragedies: yes, okay, Hamlet thinks up a sly bit of business with that Mousetrap thing, but he is not essentially a character of wit; he in no way resembles Bond or Ulysses.
If we want to present tragic characters in IF, and make the player take the role of a tragic character, then we need to offer tough and painful choices. And indeed this is a direction taken by a good bit of recent experimental IF, though I think we have more to learn about how to do it well.
2) Bernstein’s description of the player’s emotional response doesn’t correspond to observed reality. This is a subjective matter, and if Bernstein says he himself feels disappointment when he pulls off a really clever solution to a problem in IF, well, okay: I believe him. Quite possibly some other people do too. But this isn’t my reaction, or the reaction of quite a few other people I know. Finding that the author has anticipated a crazy action doesn’t reduce my sense of my own cleverness; it makes me feel that I and the author have been clever together. A few such moments have given me quite a surprising flash of fellow-feeling or admiration for the author, while at the same time making me feel that I pulled off something pretty good in the role of my character.
3) The “I don’t understand” problem is a serious one, as I admitted in my original post, but much of IF design theory has to do with solving it. Part of the job is managing expectation: teach the player what kinds of things his character is likely to think of doing; teach the player what kinds of actions are possible and probable in the social world of the story. After all, world-view and culture are much, much more constraining than physical reality: maybe it would be possible for the hero to steal or beg to resolve his poverty, but we’re writing about a character whose sense of honor does not permit such actions. It’s perfectly fair not letting the player do something that is fundamentally out of character or beyond the conception of the player character. In other words, use the story to determine — and explain — the ways in which you constrain the player’s freedom.
The other part is thoroughly and deeply implementing those things that do fall into the game’s domain of action. If we achieve a deep implementation through simulation, then it’s entirely possible that the player will be able to find solutions that aren’t anticipated by the author. The chief challenge here is that it’s really hard to hand over social actions to the control of a simulator. It’s comparatively easy to write a physics simulation that will account for whatever the player does (once we’ve defined the problem domain to be small enough). It’s harder to write a conversation simulation that’s equally effective, which means that for those kinds of interactions the player really does have to depend on the author “getting there first” — but, again, if we define the problem domain to be specific enough, it’s then possible to fill it in pretty well. It just takes a hell of a lot of writing.
Summary: no, I don’t think we’re going to get Hamlet on the Holodeck, soon or in the long term. I don’t think that the Holodeck can be done, technically; I doubt whether it would be a medium of meaningful artistic expression if it could. But what contemporary IF attempts is smaller and at the same time (I think) more narratively promising.
Emily Short's Interactive Storytelling 
I also was confused by the ‘Oh no, the author got here first!’ complaint. Certainly, the author ‘gets there first’ in static fiction as a matter of definition. And this is equally true in hypertext fiction, unless I’m missing the mark somewhere.
There are two endeavors I can think of where the person experiencing the work ‘gets there first’: improv theater and roleplaying games. And indeed, being unexpectedly clever is a great joy and treat in both of those mediums, and one of their great strengths. We’re here crossing the line into direct creative creation of a work of art, though, and if you really crave this sort of experience, the best way to do so is to write or otherwise create your own stuff.
As another random data point, I, as Emily, am usually pleased to have reached a solution the author has anticipated, and the fact that the author got there first is (for me) irrelevant.
It could perhaps be further noted that in highly rule-based simulations of worlds, solutions to problems can indeed be arrived by the player but unanticipated by the author. This is sometimes known as ’emergent behavior’, but it can more often be classified as ‘a bug’. I can recall one story of a computer RPG where some goal was on the other side of a gorge: the gorge could not be jumped, but items could be thrown from one side to the other. An unanticipated solution found by some enterprising players was to kill one of the party members, throw them across the gorge, then cast ‘resurrect’ on them.
If emergent behavior/bugs are indeed what Bernstein is after, this is indeed achievable in IF through managed expectations, high points of contact between the player character and the world, and rule-based simulations. Research into this field has been sparse, partially because initial forays have not actually been as rewarding as one might expect. The right person could probably make good headway in this field, though.
The feeling of pleasure at reaching a solution planned-for by an author (my memory of solving the royal museum while sitting in a computer-free high school classroom is a delicious one) seems to me likely to be that of solving any other puzzle in any other (non-computer) game, or, to bring Nick Montfort’s favorite perspective in, in solving a riddle.
Perhaps Bernstein’s complaint is that this sort of “positive” experience is contradictory to the “positive” experience of immersion, and that the one undercuts the other. Or perhaps he is simply claiming that only one of those goals is worth pursuing, or at least only one at a time. (Or perhaps he merely doesn’t find the idea of solving a riddle valuable; I don’t know.)
As someone who loves immersion as well, though, I am, of course, on the same side of the fence as both the previous commenter and, presumably, Ms. Short.
(Lucian: As to emergence, you generally just need deeper, richer simulation. And in the article Emily indeed already said: ‘If we achieve a deep implementation through simulation, then it’s entirely possible that the player will be able to find solutions that aren’t anticipated by the author.’ In fact, Metamorphoses is one of the few steps in that direction in the IF world, and though I imagine there is little anyone would classify as emergent, at least one can suspect that there are stories people could tell of their playthroughs where Emily would say “oh, I never thought of that”.)
Those “aha” moments are the addictive delight of playing IF. And, from the other side of the wall, one of the small pleasures of authoring it.
It seems to me that IF is fundamentally cooperative, not competitive. The author might invent some puzzle, but the point isn’t to defeat the player (which is trivially done), but rather to pose an interesting problem that can be solved using what you already know.
Of course, if you have to explain the solution, it’s not a fun puzzle. The problem is very much like writing humor where if you have to explain the joke, it’s not funny, but if it’s too obvious, it’s also not funny. (They say comedy is hard…)
If this analogy works, then it should also be true that it’s much easier to say genuinely new things in either regular fiction or in the non-interactive portion of a game, while the interaction has to be fundamentally based on the familiar, though hopefully in an unexpected way. Even relying too much on new information explicitly mentioned in the game can be a problem, since the reader might have been skimming. I find that the most common reason I get stuck in an otherwise well-written game is due to some lapse in attention.
So it seems that IF relies a great deal on the writer’s familiarity with readers’ abilities and interests and on the reader’s diligence in paying very close attention all the time. Maybe this explains why IF is such a small community.
If this analogy works, then it should also be true that it’s much easier to say genuinely new things in either regular fiction or in the non-interactive portion of a game, while the interaction has to be fundamentally based on the familiar, though hopefully in an unexpected way.
I’d say: the interaction has to be fundamentally based either on the familiar *or on what can be taught*. The author can teach the player to be quite proficient at a new magic system or the like (preferably not just by mentioning things, as in a lecture, but by setting puzzles of ramping difficulty); so long as the author’s system is self-consistent and the syllabus, as it were, well-designed.
Just to replug: The Hobbit (1982 Beam game) has genuine emergent gameplay. Whether or not this added much to the interactive experience is another matter.
Bernstein has a response to your response to his response, so just to add to the madness I wrote a response in my blog.
One important thing to remember about IF is that there is a level of suspension of disbelief involved in convincing the player that they genuinely control what’s going on. This is the reason why a bad “I don’t understand that” response can be so devastating to a game; it shatters the suspension of disbelief necessary for the game to feel interactive.
On the other hand… it is true, certainly, that any solution the player can come up with while playing as Bond must, generally, be something the author foresaw; but then again, any solution Bond himself uses in a movie is just something written into the book or script, after all. In both cases, the effectiveness is in convincing the viewer to believe otherwise, at least for a moment.
Case in point: The Phoenix Wright games are completely, totally linear in their courtroom scenes. At each point where the player makes a decision, they either choose correctly and are allowed to proceed, or choose wrong, get penalized, and are forced to choose again. Described like that, it sounds ghastly. And yet, despite this, the games work. They feel more interactive–and feel like there’s more interaction with their characters, as adverseries–than many games with objectively more ‘interactive’ systems.
Why is this? Simple. The game’s high-quality writing, careful use of a constrained situation, and carefully-planned scenerios manage to (most of the time) subtly put the player’s deductions and decisions in the place where the game’s developers want them to be. The individual decision-points are worked into a larger whole, and the player (in order to proceed) is directly encouraged to piece them together, putting themselves in the main character’s position and generally suspending their disbelief. Assuming the game is doing its job (and it manages it admirably, most of the time), when the player presents a crucial piece of evidence, they’re not thinking “Aha, this is the solution they wanted”, but “Aha! I’ve tripped this witness up in their own lies at last!”
This is the sort of thing that interactive fiction needs to strive for. It is the difference between, say, watching a bunch of people walking around and talking on stage, and watching a play.
True, complete interaction is not only likely impossible, but also of debatable worth; if you were really just going to set the player loose to do absolutely anything they want, with no restrictions, then what is the purpose of the “fiction” part of interactive fiction? Naturally the player needs to be constrained, naturally the solutions they discover must be written by an author… if they weren’t, there wouldn’t be a _story,_ just aimless interaction. The trick is to guide the player into playing their part in that story without their realizing it.
Aquillion: yes, I’d agree with most of this. The trick (I’m increasingly convinced) is to teach the player in what specific sphere of action they are allowed control (and what vocabulary they should use to exercise that control).
The trick is to guide the player into playing their part in that story without their realizing it.
This opens up the question of choice, which we’ve occasionally discussed here before: we might give the player some options about which way to go with a story. But here again the design will need to be such as to direct the player towards the reasonable options and make him ignore the ones that we haven’t coded — that is, he should feel as though the options he’s allowed are completely natural ones for the situation the character is in.
>But I’m just a writer, not a hero: have I thought of your incredibly clever strategem? If I have, you’re deflated: it’s not heroic after all, it was just a puzzle and you’ve supplied the correct answer. A tough puzzle, maybe, but (obviously) the author was here before you.
Great minds think alike. The end. Q.E.D.
P.S. I had a whole article on thought statistics and the criteria on which an idea is deemed clever, yet nothing beats the above social meme for concision and common ground for all target groups. Maybe next time.
there is a difference between James Bond and Hamlet or Oedipus, in terms of acting, not-acting, and self discovery, but also in terms of life and death. These differences may suggest a difference between heros and players, and thus different IF experiences. Bond lives on by figuring things our correctly, quickly, and with elan. Hamlet hesitates exquisitely then dies as a result of ethical and oedipal decision to act, and Oedipus solves one puzzle only to discover he is the cause and solution of the next. If Tragedy marks the end of myth, throwing being’s ends open to chance, an unreadable god, daemon agon, or partially known physical laws, then IF fiction can recuperate a plot, a myth, and produce master spies who live to die another day, or it can produce players who, like beings everywhere, perform, decide, and act, within the already given plots, systems, and genres. The question is be a hero or become a player — that is, figure out the moves or perform the acts, step-by-step, or step by step with threshold shifts, for perception of objects, of experience emotions, of orientations within worlds. If how we solve — use objects, think in metaphor, ask for help, or how we play — which directions do we go, what orients us in the worlds, what do we turn to for authority, what we repeat — paths, conversations, items we choose to keep, etc., if, there were a phenomenological element that added perceptual threshold shifts to curiously wonderful puzzles, then I, olayer, could become spymaster, melancholic prince, warrior lost at sea, blind seer, by moving through perceptual shifts which orient me differently within the world, its objects, and its meaning. This is an argument for a phenomenological base for IF game theory, not the tool of learning, or fun of figuring, which would require a name change, for exercise seems better than fiction. fiction plays you, until you see how it is, to the extent “is” applies here, you. The punctuated equilibrium of the flOw game is a way to go, though maybe something shows how we figure and what we figure figuring is, really. did I win or did I play.
You played to win.