Recently I’ve been looking at academic papers available at the Association of Computer Machinery portal, checking out what various scholars and conference-attenders have recently had to say about interactive fiction. (Note that while you can register with the site for limited free service, full access is by subscription, so your best bet is to access it through a registered library.)
It turns out that what these papers have to say is often some variation on “here is why conventional interactive fiction doesn’t live up to its promise: …”. They have different ideas of what that promise is.
One take on this:
Nicolas Szilas, Jason Barles, and Manolya Kavakli: “An implementation of Real-Time 3D Interactive Drama,” Computers in Entertainment, Vol. 5, January 2007, article no. 5.
Szilas, Barles, and Kavakli argue that conventional IF (like many other forms of computer game and computer literature) doesn’t go far enough in allowing the player true agency. They want the player to be able to do “anything”, and they propose a couple of example games in which the player is allowed to articulate semi-complicated things to the other characters.
In practice, the “anything” that the player can do is just as as limited by their world model and UI as it is in anyone else’s game/electronic literature; the textual part of the output, at least, reminds me a bit of Chris Crawford’s storytelling plans and in particular his mechanical Deikto language.
Some sample transcript:
You decide to start a riot against Nelson.
Jack to Malcolm: I want to start a riot against Nelson.
Malcolm to Jack: You should start a riot against Nelson.
Jack to Malcolm: I have a problem to start a riot against Nelson: the gate is closed.
Malcolm to Jack: The gate is closed because you haven’t got the hook.
You decide to get the hook.
You ask Malcolm: Do you have the hook?
Malcolm to Jack: I haven’t got the hook.
It goes on in that vein for quite a while, with entertainingly inane questions and answers between Jack (the player character) and Malcolm and the other NPCs, until the player has acquired a book of rhetoric from one of the NPCs, enhanced his abilities in spoken persuasion, and persuaded other characters to join his cause. He then succeeds in starting a riot, and wins.
There is something kind of charming about this, but I don’t quite feel that it is in all respects a step in the direction of greater literary value, or even greater agency. Most of the things the player gets to do with his newfound agency are, after all, pretty minimal in terms of plot significance — it’s basically about solving a bunch of linearly organized puzzles to progress through a predefined plot to a single predefined ending; the same format as quite old IF, but with puzzles with more social content.
Emily Short's Interactive Storytelling 
Interesting – I’ve seen the title of the article, but my university doesn’t have access to it. I’ve read other papers by N.Szilas though, which focus more on interactive and active characters with goals and values, and that did seem like an interesting prospect for IF
Oh, I think it’s all quite interesting stuff, really — and I *am* intrigued by ways of specifying social interactions and bartering that go beyond what we currently do. (I also haven’t done justice to the full content of the article, which spends a while summarizing their particular take on how to code and structure such a piece.)
On the other hand, I think this is not a situation where the player is inherently more powerful or more able to affect things than he would be in another kind of IF or video game; it’s just that he’s given a thorough set of commands for negotiating one particular kind of interaction (bargaining for objects and actions), rather than a sparse set of commands for negotiating many kinds of interaction, as we often see in IF. IF treatments of conversation tend to be especially sparse — by which I mean that an IF game might acknowledge the possibility of blackmailing the countess, seducing the dairy-maid, and telling a ribald joke to the boy who blacks the boots, but it would offer all these as menu items or similar one-shots, and not give the player the ability to specify his own variations. This is limiting, but any dense/thorough implementation of conversation has to choose a much smaller scope of things to talk about, or the design will get completely out of hand. I am not sure that the authors of the paper fully realized this fact, or (if they did) that they wanted to acknowledge it.
Unfortunately my University hasn’t got access to it either, though the Storytron is always available :)
Now, I only wish the academics (many times a dirty word, see: Intellectual Impostures) that bothered to come up with these papers in order to get an extra citation or two, were just a bit more knowledgable in game design…
Gosh, I didn’t mean this as a cheap pointing-and-laughing session — and it would be hypocritical of me to diss academics as a group. The authors of the paper are working on some legitimate projects, but I disagree with the way they frame the problem; that’s what I wanted to look at.
Actually, despite working in academia, I can quite easily say that the majority of academics, especially the post-modern ones in humanities & social sciences, aren’t the most serious people around. Obviously I didn’t think you dished anyone, really…
Indeed, I didn’t encounter any articles that adress problems of UI and user input in interactive storytelling, perhaps with exception of a discussion on GrandTextAuto.
You might also find interesting other researchers’ work on close subject
http://www-scm.tees.ac.uk/users/f.charles/
Interesting article…I’m glad I was able to access it through my University library. I had heard about this project before, although I don’t remember from where (probably GrandTextAuto). I applaud the authors for their attempt at advancing the field of interactive narrative, particularly in their use of a 3D engine, and for publishing their thoughts and experiences. As this is a field of particular interest to me, I can say we can’t get enough of this stuff, especially when approached with an academic perspective.
I agree with Emily’s comments above, however, particularly in that this does not appear to be a creation where “the player is inherently more powerful or more able to affect things than he would be in another kind of IF or video game.” But also, as Emily states, the authors argue that conventional IF doesn’t go far enough in allowing the player true agency. But is true player agency the grand target for which we should be aiming? I, for one, side with Stephen Bond on this issue, that “the effort to maximise player freedom is misguided” (http://plover.net/~bonds/playerfreedom.html).
Note that I don’t necessarily think out project (http://www.greatgamesexperiment.com/game/vespers3d/) is any more likely to be a critical or popular success than this one. Our approach is simply to add a graphical layer to what is still conventional IF (relatively speaking), so we will essentially just be propagating all of the limitations and frustrations that come with typical IF.
I’d think part of this would also be that if you’re going to be published in a Computer Science journal, the editors will be more interested in the CS side of things — self-willed intelligent agents and powerful world modeling. The phenomenon of player complicity, as opposed to boundless agency, seems like it’s more out of their field.
I would have a lot to say about your remarks on this paper I wrote with my Australian colleagues… I try to make it concise…
First, I should say that it is always a little bit unfair to judge an Interactive Fiction or Drama only by the final transcript of the experience… But I guess it is the easiest way to report an approach.
Emily, I quite agree that the piece does not have a great literate value… If it had, I would publish it elsewhere!! Interactive Drama is more focused on action (often displayed with 3D characters) so the text in itself is less interesting than in IF. Also, all text presented is computer generated (from generic text templates), in order to allow greater variation… So quality of text is sacrified to improve agency, even if I could have done a better job if I had more writing skill.
Regarding agency, well, I am a little bit disappointed that you did not see the difference. In average, there was 93 choices at each turn for the player… Ok, this is just quantity of choice, not quality but still, it means that it is different. All choices are not equally meaningful in the story, and we have improvements to make in this direction, but each choice changes the state of the storyworld so potentially changes the story.
The plot is not predefined, most puzles are not linearly organized, and the question of different ending is finally not as relevant as it seems (you can have dozens of different stories with the same “ending”). Let me try to rephrase the story to explain agency: Well, Malcolm does not have the hook, in fact Rak has it. You have the choice to steal it, but if you tell Malcom about this idea he might warn you that it is dangerous. He (or another character) might simply advice you to ask Jak for the hook… Ask, rob, it is up to you to choose. You don’t know what will succeed (I, as an author, do not know either…). If you ask, it might work, or fail. In that later case, yes, you need the book of rhetorics to convince Rak, and it leads to a another similar, but different avenue in the story (getting an object)… But it is just one solution, you can still try to rob the hook. And what about asking Malcolm to get the hook for you? Yes, you can do it, it is not a predefined action, because you can delegate any “possession goal”, as well as many other types of goals. Here we already have many different possibilities, and the story contains much more that looking for objects. You need to make allies, investigate about character’s preferences, etc. It is up to the author to define the structures leading to other types of situations in the story. Okay, you find it inane, I am just saying it is agency, and I would like to improve it to make it more artistic.
Emily, I quite agree with your distinction between thorough vs sparse set of actions. I am aware of that and in that sense Interactive Drama is more like a game than IF, because you have a limited set of “cards” in hands, this thorough set of actions, that you play freely. Note that these cards are not actions but generic actions (steal someone, anyone).
Ash, I have written two papers on UI in Interactive Drama: a conceptual one at TIDSE’04, and a practical one at IUI’06, where I propose a specific user interface. This a quite relevant topic indeed, but it happens that several researchers, including me, have started with the algorthmic side of agency, and later came to the issue of UI.
Gnome, I will not answer to your “anti-academic” statements, it would not lead us anywhere. Read the paper first.
Well, There would be much more to say. If you want to improve the demo and somehow be part of the project, feel free to contact me!
Regarding agency, well, I am a little bit disappointed that you did not see the difference. In average, there was 93 choices at each turn for the player… Ok, this is just quantity of choice, not quality but still, it means that it is different.
It’s certainly different from most 3D games I’ve tried, but no, the average IF has a similar number of implemented commands; when this is combined with the number of possible objects, the range of available actions grows very large indeed.
You need to make allies, investigate about character’s preferences, etc. It is up to the author to define the structures leading to other types of situations in the story. Okay, you find it inane, I am just saying it is agency, and I would like to improve it to make it more artistic.
For the record, I do find this quite interesting: making a model that represents social interactions in any systematic way is not easy at all. But it’s not so much a matter of giving the player more agency, as of giving him agency in a different area than before.
At the end, the difference between IF and IDtension (the name of the narrative engine discussed in the paper) might be the folowing: while both IF and IDtension have the same level of agency, IDtension puts the agency at the level of social interactions so that this agency makes the story, while IF adds other quasi-linearly strctured elements on the top of its agency to ensure the quality of the story. Am I right?
Note also that the agency is not only social, it aims at being narrative, by the specific choice of the catalog of actions on the one hand and by the user modeling on the other hand.
IDtension puts the agency at the level of social interactions so that this agency makes the story, while IF adds other quasi-linearly strctured elements on the top of its agency to ensure the quality of the story. Am I right?
Well, this is true of some IF, but not of all: there is IF that does put the agency at the level of social interaction. (Most works that are called “conversational IF” would fit this description.) Structuring features also vary.
Where I most disagree with you, I think, is in the idea that it’s possible for agency to “make story” — that is, for a good story to arise from nothing more than a sequence of actions that the player takes, with no structural features to make some of those actions more significant than others. (By significant, I mean that they result in rare or dramatic outcomes, or mark a change that the player cannot undo.)
I do not believe either that “it’s possible […] for a good story to arise from nothing more than a sequence of actions that the player takes”.
Let me explain a little more. It is one thing to say “we want agency”, it is another thing to say “agency is the solution to make a good story”.
In my opinion, guidance is needed on top of agency; Not in the form of a predefined narrative arch, but in the form of more abstract narrative features such as “promote tension, suspense, surprise, conflict, ethical value expression, rythm, etc”.