i.plot therefore i.write
Published by Mark Marino October 5th, 2005 in HCTI, generators, Poetics, Features, Text Art, Software, Education.
If you need help writing a story and are adverse to the mainstream plot-twists of Dramatica, you now have a zen-based alternative. At this past Siggraph in Los Angeles, Dr. Noako Tosa of Kyoto University presented i.plot, which she sees as part of “the future of narrative.” Enter a few words and i.plot will return a web of interlocking nodes, each filled with related words. A story is born. Users can see it in action here.
To use i.plot, users enter several (3 or less is best) search terms. i.plot searches out the associations with these terms, using Edinburgh Associative Thesaurus, and then proceeds to map out a connection. The connections involve semantic webs of antonyms and synonyms, along with a path that runs through various interrelated words. The system is not as robust as Visual Thesaurus by Thinkmap, but its large idling balls have a bit more play to them. Incidentally Visual Thesaurus, which is now for-pay, offers pronunication, multiple languages, and spellchecking.
i.plot is a combination search engine and semantic web with a bit of I Ching-like possibility space thrown in. The work builds on a previous project, at Siggraph 2004, ZEN-etic, which offered an intercultural encounter with the allegorical tropes of Eastern philosophy. The new system, however, focusing on storytelling, offers unpredictable strings of association for your input that may require a bit of a zen attitude to appreciate its relevance.
In terms of interactive storytelling, the system seems to fall along the lines of the semantic fiction generators discussed in Chris Crawford’s Interactive Storytelling. Like many sentence generators, i.plot does generate a string of terms that could be interpreted as a story, but the reader has to do much, or most, of the interpretive work.
I would offer i.plot and Kartoo as valuable prewriting or brainstorming tools, or possible parts of larger storytelling systems. Kartoo is the visual metasearch engine which groups your results around thematics. Like bubbling of your college writing class, these activities can be generative, helping you to see connections and build on ideas, when all you have is a string of words, regardless of whether or not you use the sites Kartoo suggests or the story strings produced by i.plot.
Speaking of a college writing class, some of the students of mine who have used i.plot as a brain-storming device find that they do discover new ideas an meaning when they type in search terms. Unfortunately, the vocabulary recognized by i.plot is currently quite limited, and abstract terms tend to choke the system. Eventually, the creators promise to harness Google in the aquisition of new language. This has not yet been implemented in the online version.
Mark,
not that I’m trying to evangelize you or something, but if you think that Dramatica and the Grand Argument Story theory are tools that somehow adress “mainstream plot twists”, you might underestimate their use to writers in general, and botmasters in particular.
For the writer/storyteller in general, GAS decribes how the elements of a story that falls within its analytical range relate to each other to form a logic argument. It doesn’t say anything about how this argument is symbolized to the audience. The story structures of “Star Wars, Episode IV”, “Reservoir Dogs”, and “Hamlet” make all of them GAS-analyzable - does that mean that those stories are all about “mainstream plot twists”?
As a botmaster, I find the fact that the GAS is known to be computable to be of particular interest, because it provides me with a fixed point combinator to base my bot’s logic on.
Dirk,
Point well taken. At some point, story may have to be systematized and to do that one might need to develop formal understandings of plot structure.
Can you talk some more about how you find Dramatica useful to botmasters? (Or link us to your posts regarding this).
From the (non-screenwriting) creative writing standpoint, however, workshops always seem to stress starting with the character rather than starting from the plot. To use your own terms, this is looking at characters and objectives rather than choosing from plot types. Sometimes I wonder if courses don’t emphasize this because we all have a lot of plots in us by the time we reach a creative writing workshop. But there probably is something about the kinds of stories that are generated (electronically or otherwise) when the emphasis is on specific characters achieving goals, rather than characters performing their roles in archetypical plots. On the other hand, that could have more to do with a “point of focus” that’s important for (human) creative writers (who again have already developed a more innate understanding of plot conventions) that is not as necessary for writing programs (that have no “concept” of the necessities of plot).
So in other words: For now, human using Dramatica to help write story is very different than computer using Dramatica or algorhithm to help produce story.
Mark,
you seem to regard Dramatica and the GAS as some kind of “story synthesis” tools. But that’s not how it works. Those are analysis tools. You already have to have a story to make any use of them. Then you try to fit a model (a “storyform”), and the tool tells you where the holes in the logic of your story are with reference to that model. I shouldn’t even be saying that it “tells you” anything much, since strictly speaking you “tell yourself”, by interpreting the discrepancies between your story and your model. Then you change your story, or the model, or both, and re-analyze.
There are no “archetypical plots” given which limit you (the writer, as ever, does all of the elementary generation, and I, for one, would never use something as crude as Propp functions). You can, however, fit Propp functions, or the “Hero’s Journey” model, or the Linda Segers character model, or the Syd Fields plot structure over the GAS grid, and see that they are all easily subsumed by the GAS (you still shouldn’t let your Propp functions generate anything ;-). Its documented method also encourages writers to start story development as character development, since it analyzes plot as emergent from the interactions of characters (quite the opposite of Propp functions).
How is that useful to botmasters? The most obvious answer is that botmasters have to deal with non-linearity at a variety of levels, and you need suitable models to do that. The GAS is a good story (=high-level) model for botmasters because its logic is unaffected by story-level nonlinearities (e.g. “Pulp Fiction”, a non-linear story, is GAS-analyzable), as long as you encode your story with proper reference to its “cultural bias” towards the “Knowledge” element at the plot level.
There are two constrains that any writer has to accept in order to be able to make use of Dramatica: 1) I only can produce stories that take the form of logical arguments; 2) the underlying bias of that logic will be that “Knowledge” is their default functional element of the calculus - the one that can “spawn” all other elements, or as they say in math, the fixed point combinator -, as opposed to, say, “Desire”.
If I work directly on top of GAS theory, however, I can still choose to forgo working with Dramatica entirely, and build my own model using any story element I care to as my initializing function. The mechanism I use for this is the same that the Dramatica authors use - it’s the I Ching mechanism, or more mathematically speaking, the technique of quadratic programming. I still have to give the implementing program some authorial bias - otherwise, the system would generate random output, and be unfit for the communication of any authorial intent -, but it’s just any bias I choose, which may or may not have something to do with the bias of the Dramatica program.
What I have then is a very general model of story-as-argument, where “argument” resolves to “a partially ordered list (non-linear, but not random/non-sense-ical) of all logically possible resolutions to a conflict between two problem solving methods”. That’s why we use The Story as a tool in the first place - it reduces the infinite list of logically possible conflict-resolution pairs in real life to the finite list of logically possible conflict-resolution pairs in a story context. The GAS gives a botmaster an algebraic model for that.
However, it’s an algebraic model through which the designer can interact with story structures. As a botmaster, you can go and say “Okay, I have two character classes with conflicting problem solving methods - humans and bots -, and whatever one of them does in any story I implement, the other one can find all the logical solutions to all possible problems (as biased by the character’s designer, of course - that bias actually becomes the “percieved quality” that we call “character”), either by just looking up a value in the current story structure, or by being told by that very story structure to look up a value in another story structure. But: there’s nothing informative regarding player interaction with a story in that model - we all need to try rolling our own here, and see what works :-) The GAS theory helps with designing computable story logics; it neither helps writers write stories, nor will it ever, by itself, will help machines to produce them.