Hypertext '08: Chris Crawford — Deikto: An Application of the Weak Sapir-Whorf Hypothesis

Chris Crawford has been working on Storytron for 16 years. The computer gaming industry was not intersted in or able to solve the problem of interactive storytelling. Left the gaming industry to solve that problem on his own. Has been explaining Storytron for 10 years, says that during that time he has “failed miserably.” 

About 2-3 months ago, someone said “The problem with your technology is that it’s revolutionary,” in the sense that it’s too much change coming too fast, requiring conceptual leaps that people can’t handle.

The major leaps Chris had to go through: “People, not Things!”  Entertainment or art should be about human beings, not about guns or monsters or ammo or food or spatial reasoning or puzzles. 

Chris: “Games are cold and heartless because they don’t have any people in them.”

Game people all accepted the notion that people are a good idea, but they don’t embrace it — making the same old games with characters that you didn’t interact with in a meaningful way.  Games now “don’t have any story, any really compelling content to them — they’re just things.”

2nd: (Chris says this will be harder for the hypertext audience to grasp.) The primacy of interactivity in the computer medium.  Interactivity is to computers as cinematicness is to movies — it is the essence of the medium.   You can use a computer as a slide projector, phone, movie making — but the one strength that computers have is interactivity — the prime asset of the medium. 

Mark B: Why would this crowd have a hard time grasping that?  What’s the anticipated straw man?

Resposnse to Mark M:  Interactivity is a meaningful choice for the player, achived by processing by the computer.

3rd: Forget plot. Plot does not belong in interactive applications.

Plot in interactive media is like talking about color and shadow in poetry.  Plot is a standard that simply doesn’t apply in an interactive environment… plot = the player’s actions don’t matter. Plot — pre-defined plan for the events that will take place.  (Metaplot is an advance notion of the general themes you intend to explore… that can be done interactively.)

Theoretically, an interactive environment should permit any action that’s dramatically consistent with the author’s goals.

Crawford’s First Law of Software Design: Always ask “What does the user DO?” (Not what he user sees, hears, etc.)  The choices are expressed as verbs.

List “Here are the verbs I want the user to execute.”  Demonstrated that with just a few verbs, you can identify the piece of software.  Software is defined by its verbs.

Next Leap: Linguistic User Interface (LUI)

Interacting with a computer is talking to the computer.    [Is “giving orders” really “talking”?  Is it necessary to translate an already metaphor-laden action such as dragging an item to the trash can into words? “Make friends with Betsy” — an example Chris used – is powerful language, but we communicate through gestures, proximity, eye contact… we communicate subconsciously in many ways that lose quote a bit when we translate it to language.]

Language is the protocol that people use.

Linguist — Sapir-Wharf hypothesis.  Language and the perceptual reality of the speaker are closely locked together. Language mirrors reality as perceived by the speaker.  Language is hard to put into a computer… you can’t put reality in a computer therefore you can’t put language inside a computer.

Every artist is creating a tiny universe… storyteller creates a little dramatic universe that contains only the appropriate level of detal. Create a “toy universe” and a “toy language” to go along with it.  The universe and language are a single task — you can’t create one without the other.

Next idea: Inverse Parser

The words they require are hidden.  Only show the user the options that are appropriate to the particular moment. Build a sentence, word by word, setting a goal.  You’re not the supplicant any more — “Please accept my text!”  Instead the computer does the work so that it only shows you the relevant choices.

“The computer does all the work, not the user. That’s the way it’s supposed to be!” [Hm… well, in certain cases, yes, but that would pretty much kill the whole concept of the riddle. CF the chess example in The Garden of Forking Paths.]

Storytelling is best done by storytellers, not programmers. 

Raises a gigantic problem… storytelling is algorithm-driven.  Interactivity is about thinking how to respond.  The computer is supposed to come up with an interesting reaction to user input.
Gives the example of a refrigerator door — the light goes on when you open the door.  Keeping a given reader interested requires computation, requires the language of mathematics.  And storytellers are not particularly mathematically, nor are they very technical.  We don’t need millions of storytellers to do this, we just need a few dozen to get started.

Make the tool eaiser.

Chris — getting the storytelling part working was easy… that was 4 out of the 16 years.  The rest of the time has meant making the tool easier for the non-technical person.

Demonstrated the script describing the inclination of the actor to accept a proposed deal.

Last example — “Bounded Numbers” — Bounded numbers, between +1 and -1.  You can never get a number outside that range.  All manipulation just push the variable between those ranges.  Gave the scenario of your brother has been bullied… how do you compare your confidence in your strength, and your love for your brother.  Every value gets a bell curve, based on average.  Arithmetic is a lot easier for the non-technical person if all values are based on a scale with 0 as average, 1 the most possible, -1 the least possible.

To do anything with all this you have to embrace all these concepts, which stops many people from using the technology.

The Storytron “Grand Opening” is July 15.  In 20 years, will be paving a parking lot over Electronic Arts.

Why can’t programmers tell stories?  Steeped in a style of thinking that is inimical to storytelling… it takes an exceptional programmer to cross the line. and vice-versa.

The conceptual difference is not a personality thing. The games industry has not embraced the humanistic elements of storytelling.  “They do great Potemkin Village characters.”

Chris says there has never been a character in a game that you can have a meaningful emotional relationship with. (Emily Short briefly reacted to Chris’s dismissal of the death of Floyd in Planetfall.)

When gamers are “in the zone” — they’ve cut off the frontal cortex, so that kind of gameplay is “untouched by human neurons” because the cognitive part of their brain is cut out of the cycle.

Dene noted that being “in the zone” happens in sports as well… you block some of the data from the outside, but can still be very physically active.  Being in the zone is more complex than Chris described it.  Dene offers “immersion” as the key between trivial an non-trivial activities.  Dene notes that what counts as trivial and non-trivial for kids is different from adults.

Chris says that your body has a natural correction system that kicks in when you are in the zone.  “There’s always death as a feedback mechanism.”

Chris feels the gaming culture will be seen as kiddy entertainment like Disney.

Mark M — can you make the grains of
output larger?  Invoke
d Noah Wardip Fruin’s Talespin effect — the software does robust processing, but the output is harder to appreciate as a completed work of art.

Crawford noted that cinema’s first frame of reference was the theater… static camera where the audience would sit.  It took a revolution to realize that cinema does things differently.

I suggested that the granuarity of Storytron would fill the space of a cell phone or IM.

Mark B — invoked Roger Schank‘s work on storytelling. Is Crawford reviving Schank? (Chris distanced hisself from Schank, saying he looked at his work long ago then set aside and went off on his own.)

Mark M — you can make ELIZA interesting by plaing around with ELIZA.  There are “bad examples of ELIZA.”   Mark sees the data going in and the data going out.  Once you start playing with the grains coming out, and what they look like, you start addressing its literary flavor.

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.

Mark B: Asked Chris — I’ve got this character the reader is supposed to love… how do you make the character lovable. 

Chris: Make her lovable in what she does, not what she looks like.  Make her interested in who the protagnist is.  The character should be empathetic, she will inquire, respond to the player’s statements.  Chris notes that the core point is on the mark — we don’t know how we’ll be able to come up with algorithms to define interactive stories.

Chris — the best people for the new technology are “young, angry failures.”  The established writer has no need to seek an outlet in a new medium.  “That’s how I think we’ll find our best story builders. Losers.”

In responset to my question, Chris said that he spent about a half hour with Emily Short’s Galatea and found it “interesting.” I noted that in terms of meaningful interaction with a virtural character, the interactive fiction community considers it to be the mark to beat.