People who were interested in Second Person but didn’t buy the hardcover copy may be interested to know that many of the articles are now being presented online over at electronic book review, as a continuation of the First Person thread. Since there’s a large amount of content here and it may not be immediately obvious which of these articles are IF-related, I’ve also added to the Second Person page over at ifwiki with links directly into the IF-specific stuff.
Along similar lines, Dennis Jerz has an interesting summary of Chris Crawford’s talk at Hypertext ’08.
Admittedly, a lot of this is stuff that Crawford has been saying, in various forms, for quite a long time now, and my reasons for skepticism remain unchanged. Still, there are some bits of quoted dialogue that are sort of revealing. E.g.:
Chris: “Why do I want more literary flavor?”
Mark M: “That’s what I want.”
Chris: “Then write a book!”
Chris — expecting a literary output is dismissing the primacy of ineraction. The level of interaction is where the richness. The visual and literary part aren’t expected to be rich. The medium has zero literary value, positive intearctive value. [sic]
I sort of sympathize with this, in the sense that I’m often frustrated by dismissive evaluations of interactive fiction and video games in general: critics often argue that because an interactive medium cannot provide exactly the same virtues as the uninteractive one of their choice (usually books or movies), games are categorically incapable of qualifying as art form or presenting a story “as well” as these more developed media. Repetition of content and pacing issues are often cited as particularly problematic for interactive fiction, often by people who feel that the transcript of an IF game should read like a novel if it wants to lay claim to novel-like artistic status. This essentially flattens everything that is interesting about IF — the give-and-take between the game and the player — and insists on purely formal aspects of good novel writing as being inherent in good IF writing. (I’ve argued before that good writing for IF does not necessarily have the same characteristics as good writing for a novel.)
But — and this is a pretty large but — I don’t think “interaction”, in and of itself, usually produces richness. You have to be interacting with something. You might say: fine; the player is interacting with the mathematical model that Crawford has built, which represents the feelings and attitudes of the non-player characters, and this model is so sophisticated that the feelings and attitudes can come out in many combinations and gradations.
Look at it this way, though: if we strip away the fictionality entirely, then what we have is the player putting some numbers in and getting some numbers out. Anger: 0.5! Lust: -0.37! Clearly we do need the fiction, which comes between the players and the numbers, and gives the model a meaning. (Fiction is not the same as “narrative” here. I’m using fiction in the broad sense that Jasper Juul uses in half-real, to apply to everything from the plot in Portal to the fact that some chess pieces are shaped like knights on horseback — everything that situates the abstractions of the game in some sort of relation to a real or imagined world.)
Once we admit the need for this fiction — something that goes beyond the computer’s representation of the “rules”, and gives them color and meaning to humans — then we also have to consider that the fictional aspects can be better or worse, more or less engaging. This is especially true if, like me, you’re interested in telling a specific story or set of stories with a given game; but it’s still true if, like Chris, you’re interested in giving the player a storyworld in which to come up with his own narrative. The quality of the fictional elements is not irrelevant, and it’s especially important when it comes to making the player care. The modeled character may have a wide range of emotional responses available, but if he doesn’t articulate these in a way that I identify as characteristic and at least somewhat human (with facial expressions or gestures or plausible use of language), I will always regard him as a toy or robot, not a person. I may have a fun time seeing what his response range is, but in all likelihood I will not be invested in his feelings and I may not regard the output as story-ish.
Making the player care is, in fact, an element that Crawford seems to have left out of a lot of his calculations — I’ve talked elsewhere about how his model of the ideal interactive story as just a string of choices is deficient because it doesn’t engage the player with the exposition and doesn’t give him a chance to build an investment in the story between crisis points.
And I sort of suspect that the unappealing vision of what gameplay will be like — mechanical interaction with an artificial-looking face, by means of artificial-sounding flowcharts drained of all personality — also explains some of the indifference Crawford gets from the very people he’s trying to bring on board.
Emily Short's Interactive Storytelling 
Speaking of Second Person, I keep remembering the absurd Storytron transcript in that book, written in pidgin English at about a 5-year old level. I asked myself then and ask myself now: how is anyone ever going to find this compelling?
Chris’s problem is that he has decided that the fundamental strength of the computer medium, the one thing it offers that nothing else does, is interactivity. Conversely, he claims that it does virtually everything else — meaning all the conventional literary tropes — badly. I don’t believe that either of these statements are true.
One strength of IF and other forms of interactive narrative is indeed (and obviously) interactivity. However, IF stands not on one but two legs, the other being immersion. By placing the reader in a story, rather than just expecting her to read about it, IF can make her care that much more, can make her walk a mile in the PC’s shoes. In a mystery story, I am solving the crime myself, not just reading about it; when I am being chased by a psychokiller through the graveyard, I myself am running, looking behind me in fear, etc. Interactivity increases immersion by letting me enact the story myself, but for that to work I must have all those things Crawford wants to throw away: vivid setting; fully-fleshed characters I care about; (if textual IF) strong, descriptive prose; (if a graphical game) vivid, realistic graphics, etc. How can the player ever get immersed and emotionally engaged in a story written in Daktu (or whatever), played on what looks like an Excel spreadsheet?
I’ve said it before, but what the hell: the potential of interactive narrative is not to be a sandbox allowing the player to do whatever she wants, but rather to be a way to experience a compelling plot, vivid setting, etc., in a more immersive and engaging way than can happen in conventional fiction, a form which always necessarily leaves its reader at a certain distance.
Crawford is investing a great deal of his life into something that I don’t think will ever be really satisfying — it’s always going to feel like playing an Excel spreadsheet, and is trying to answer a question that I don’t think anyone ever really asked. That doesn’t mean his work is useless, though — some of his ideas may have value in the service of IF and other more conventional interactive narratives. I’ll certainly be interested in seeing what he comes up with, in spite of my skepticism about the whole thing.
Crawford is investing a great deal of his life into something that I don’t think will ever be really satisfying — it’s always going to feel like playing an Excel spreadsheet, and is trying to answer a question that I don’t think anyone ever really asked. That doesn’t mean his work is useless, though — some of his ideas may have value in the service of IF and other more conventional interactive narratives. I’ll certainly be interested in seeing what he comes up with, in spite of my skepticism about the whole thing.
I don’t think it’s necessarily true that Crawford’s ideal exemplified in Deikto won’t at some point stand on its own. I agree that currently it just doesn’t work, but given the rise in popularity of forms like twitter stories and cell phone novels I think ‘pure’ Storytron could be harnessed in one of those mediums very effectively. Like trying to put a Zane Grey story in an illuminated manuscript, to me it looks like Chris has a new mode of expression, but without a ‘new’ platform to put it on the expression is going to be crippled.
To try to put this into different words: If we think of a game as a simulation of an imaginary world, maybe the point is not the fidelity of the simulation so much as getting the reader to understand and care about the world being evoked?
A simple sketch where each line shows something important about the imaginary world beyond will do better than a rich mathematical representation where some details are represented crudely or placed badly. Perhaps more detailed mathematical models should serve like knowledge of perspective and anatomy; they give us is a way to avoid reactions that “look wrong”.
Unfortunately I don’t know all the previous work and principles underlying Crawford’s system. But from what I gather from the linked summary, it seems that his main idea is building a system with a complex world model based on character interaction, while neglecting the literary part.
I strongly agree with some of the ideas in that summary, and strongly disagree with others. The idea I like most is the one about forgetting the plot. My idea about the perfect IF has always been a story without a predefined plot, because a predefined plot constrains the player’s freedom to act. When I read a book, I assume that I’m getting inmersed into another person’s story and must accept what is written there, that’s just the way a book works. But when I play a story which is supposed to be interactive, I expect to be able to inmerse into my own story and make my own decisions, and I’m usually disappointed by IF which only lets me guess the next “correct” action inside a prewritten static story.
However, I don’t see why this should imply that we should not aim for a rich language, as close as possible to a literary experience. He says that if we want more literary flavour, we should “write a book”. Well, in the same way I could say that if he doesn’t care about the language, he could use graphics instead. Show some glossy red hearts when one character is in love with another. But if one is going to use a tool as versatile and powerful as language, one should aim to use it well.
I don’t think it’s impossible to compatibilise a rich, non-linear world model with an advanced usage of linguistic resources. Of course, it is much more difficult than using language alone to tell a mostly static story, or than using the world model alone with bare numbers or conceptual representations. The complexity of both possibilities is not added but multiplied, since we need to generate text for every possible event, action and state arising from the world model. But I think anyone having time for an ambitious project should consider this and not be content with halfway points that won’t satisfy users. During the last years there has been extensive research about natural language generation, I was at the ACL conference some days ago and it had two or three full sessions about it. There is a lot of work being done that allows a computer to generate meaningful text which can convey feelings out of numbers. Of course, successfully applying such things to IF would be time consuming, but if someone is willing to devote years of his life to IF as Crawford seems to have done, he shouldn’t dismiss this kind of possibilities.