The need for "real" interactivity

On another floor, a discussion is being held regarding what Interactive Storytelling really is, and if such tag could be pinned to most IF works. Well, that discussion doesn’t interest me, but it made me think of another discussion that does. I took it here so it doesn’t pollute the other thread.

Chris wrote somewhere that he wanted games to be a viable medium of artistic expression — an art form. Regarding Façade, one of the reviews stated it was the “future of video games”. Are these two things related? Is real/better interactivity necessary to make more “like art” games? My personal view is: no - the only thing for which one needs real interactivity is to produce a more interactive game. Period.

Regarding the related discussion of IF’s obligation of being more like literature, Peter (Olá, gajo!) defended it shouldn’t, but I strongly believe he missed the point, because he based his entire stand on more or less technical aspects of literature, and on “mechanical particularities of each medium”. You see, the gigantilous In Search of Lost Time, by Proust, is not considered one of the best novels ever written because of pace, well constructed sentences or an engaging story. It has “some” of those qualities, “some” of the time, of course, but it is not what it makes it stand out. Hell, regarding such, it’s not even better than your average Nik Spark most of the time! The hard truth is: it is possible to learn how to write better sentences, how to pace a narrative better, how to build better characters, how to engage the reader, etc - but you cannot learn how to write something like that (and what that infamous character was defending was that something like that exists in film and literature, but not in IF - I disagree, BTW). The mechanics of creating and consuming theater are different from those of film, which are different from those of literature, etc, but each of these mediums have many works that are viewed as artistic expression. The medium of computer games is viewed as not producing so many (if any, some will add) of these works. I don’t agree, and I’m going to talk about war to explain why.

In film, war can be used to produce action packed blockbusters, like Chuck Norris’ awesome MIA; or to produce an essay on violence, madness, absurdity and darkness, and here you have Apocalypse Now. The second is more viewed as an artistic approach to the subject than the first. But the gaming universe is seen as more prone to produce MIAs than Apocalypse Now-s, with each year seeing a handful of Modern Warfares rising, action games that make you feel like a hero for shooting cubic light-years of bad guys. Would a Modern Warfare clone be more like Apocalypse Now if one breathes true interactivity into it? Once again, I think not. But is it even possible to make a game that is similar to Modern Warfare in its mechanics, but that is received more like Apocalypse Now in any way? That, I think so.

Spec Ops: The Line, is a realistic tactical shooter as much as any other. It is even less interactive than many of its counterparts. You go around completing missions, following orders, killing moving targets with all sorts of weapons. The plot is linear, the player has little to nill agency over the actions of the PC, and there is only one moment where you can choose a different path to walk upon - and that’s the very end of the game. But The Line is considered as something different. Why? Why is it taken to be the video game equivalent of Apocalypse Now (or at least a step in the right direction)?

The authors set themselves up to produce a gaming experience that would rip the player of any ability to feel like a hero in their game, because there are no heroes in war. They wanted to address the darkness, the violence and the overall madness one experiences in real war. The trick to do this: the story, the characters, the level design, the missions, even the radio communication. It was all orchestrated to emulate a descent into the abyss, not an action-packed heroic joint. Judging from the reviews and online discussions, it worked. As a reviewer noted, the game, which manipulates the player most of the time to produce the effect it wants, has one hidden interaction ready to be used - you can stop playing at any time.

So, would this game be a better experience if it had real interactivity? Would it be more like a “serious” essay on war? Would any game be? Is real interactivity what’s missing in games so that they’re taken as a form of artistic expression?

I don’t think so.

There is also a money issue here. Most big studio films aren’t artistic essays on the human condition either - they’re entertainment pieces. There is a reason: complex films give poor boxoffice results. These usually come from independent studios/authors. With the rise of independent game designers/studios, many games are being published which are being presented and accepted more and more as a form of artistic expression. Curiously, these are coming more from visual artists than from writers.

Any game which only has one kind of interactivity to analyze is a toy, an experiment, or a drabble.

All the fun is playing the layers off against each other. If you’ve done it right, anybody who says “This game is linear” will be neither right nor wrong, but rather simplifying the case.

Well obviously interactivity is just one quality a work can have and it is neither a necessary nor sufficient condition for artistic expression. Conventional big-budget video games have become more linear; that is to say, there is less player agency in the game experience. Greater player agency is often admirable, but it isn’t sufficient for a game to be taken seriously as a work of art (whatever we’re meaning by that).

OK, I’m having a hard time getting past this passage – doesn’t Proust have really really well constructed sentences? I’d also argue that his pace is perfect, specifically his very slow pace, but it’s the part about the sentences that is getting to me.

Of course he has - but that’s not the reason the novel is taken as one of the best ever. Many books have well constructed sentences and steady paces. That doesn’t make them great books, just well written ones. A microwave oven manual can be well written.

This seems like a ‘not sufficient, therefore not necessary’ sort of fallacy. Or, even, ‘not sufficient, therefore not relevant’.

I don’t think there’s any great work that is a Great Work for one simple, neat reason. They wouldn’t be great if they were just about delivering one thing. (See what zarf said, above.)

Wasn’t the intention. My point was to address a specific critique to the possibility of IF being more literature-like. I think those attributes are necessary, but not sufficent, and actually doubt they are missing. A lot of IF authors write very well, from a purely technical/syntactical, by-the-books standpoint; and intersting pace, despite player agency, is not an unicorn.

Yes, good static literature delivers many things, but zarf was talking about delivering interactivity. My question was if games (that I’m already assumming to posess some levels of interactivity, not just one - I mentioned a linear plot, not a linear game) should pursuit more - or more real - interactivity to be closer to other channels of artistic expression.

My question didn’t came from thin air, but from some of Chris Crawford’s thoughts. Two, to be precise. First: that he wanted games to be more of an art form; second: that such could (only?) be achieved within the “promised land of interactive storytelling”.

Most of the games I think are closer to this idea of games-as-art-forms are not necessarilly the most interactive out there, which got me questioning this assumption and wondering about other opinions.

Personally I reject “how much does the player change the plot?” as the key criterion of interactivity. Many powerful interactive experiences make use of challenge or complicity or player-controlled exploration – or choice that affects something other than the plot.

Yeah. I think of the most basic form of interactive narrative (and also the most basic form of narrative, full stop) as the story told around the fire, with the audience butting in all the time to ask questions. The general plot arc is unlikely to change all that much; but changes in focus or the addition of minor elements are easily accomplished. (Not that this is a particularly workable model for IF, but the point stands.)

I think there’s a fair argument that “how much does the player change the plot?” is the key criterion for “truly interactive narrative.” The real trouble is people who claim that interactive experiences with narratives that aren’t “truly interactive” are worthless. We should be proud to say “this narrative experience is wonderful; it’s highly interactive, even though the narrative itself isn’t highly interactive.”

There’s an argument, true, but it’s only fair if you don’t use a leading phrase like ‘truly interactive narrative’, which frames things in a question-begging way.

(This is a really common failure in aesthetics. The answer to “What is art?” has very little to do with the answer to “What makes good art?” Similarly, the question “What is the essential nature of interactive narrative?” does not, when answered, give you a direct answer to “How do we make better interactive narratives?”)

Agreed with maga. It seems like it’d be less question-begging to say “this narrative experience is wonderful; it’s highly interactive, even though the plot itself isn’t highly interactive.”

All this said, I have a lot of sympathy with the aspiration that computers are great at generating things procedurally; wouldn’t it be great if they could generate narrative procedurally, in a way that goes beyond the “You can also see…” paragraph. But that’s an aspiration, and outside the Sims maybe I think procedural generation of the plot would probably be the last thing we see.

Procedural generation of plot is interesting as a tool for something, but it has one problem: it necessarily kills authorial voice.

In fact, I see this quest for better interactivity the same way I see the quest for better graphics or better AI in NPCs in shooters: tools. They won’t make the interactive experience better, but they give authors more tools to do so.

I disagree, at least with the “necessarily”; as Chris Crawford said in the Phrontisterion talks, it’s possible that “the authorial voice in procedurally generated [interactive storytelling] is expressed at a higher level of abstraction.” You can set up a world with certain assumptions, and allow the player to generate a plot within those assumptions, but the assumptions themselves will express your voice.

But the proof of the pudding is in the baking, here. My Secret Hideout demonstrates that (as does the Finnegans Wake sentence generator he did, which I can’t find, and the maze in Hunter, In Darkness). Combining procedural generation and voice with plot, let alone procedural plot, would be harder.

But isn’t the point of “interactive narrative” the, well, interactivity of the narrative? If nothing happens, story/drama/plot-wise, in response to user input, I wouldn’t call it (the narrative) interactive.

That’s not to say that a work can’t be interactive in other respects – in fact, it may be very good at being what it is! – I would just give it a different name. It’s just not the kind of thing I’m most interested in.

This argument reminds me of the people who claim that good GMs always railroad the players in tabletop RPGs. Maybe they’ve never experienced anything else, so they can’t imagine any other way a game could work. I think Ron Edward’s GM-as-bass-player style has a lot of relevance in designing a potential interactive narrative: the designer basically just sets up the initial scenario and relationships, then sits back and lets the characters (especially those of the players) interact. Every event unfolds naturally from the NPCs’ initial relationships and motivations and the actions of the players.

I picture a successful interactive narrative unfolding like a physics simulation, only in terms of interpersonal relationships: the course of events is determined by the initial starting conditions and further inputs (from the player(s)) over the course of play.

This, exactly. To extend the physics metaphor, the author’s job in such a work is to define the initial state of the world and the laws of physics governing the interactions between the actors; as opposed to a typical plot-heavy game, in which all or most of these behaviors are hard-coded, here they’re defined at a higher level of abstraction.

“Necessarily kills” is an exaggeration, I think. “Tends to handicap” might be better. If you’re particularly optimistic, you might go so far as to substitute “creates big challenges for”.

(Dwarf Fortress has a strong and distinctive authorial voice, despite having very procedural plots and characters. What actually happens is that most of the time the voice is really a lot stronger than the plots and characters; you can sometimes get an interesting emergent narrative, or more usually a fragment of one, but what you really take away from the game is the sense of an overall style that’s entirely author-determined.)

This feels ungracious while we’re agreeing about the other stuff, but why identify the narrative entirely with the story/drama/plot? There are lots of narratives that don’t primarily concern the story/drama/plot; if I can’t change what happens in the story, but I can change what the characters think about, or how they feel about what happens or has happened, then that’s still a narrative that’s interactive, isn’t it?

My claim is that there isn’t a meaningful role for an “author” to play in a drama simulation of that kind; the complexity is so great that you can’t understand the simulation well enough to get involved, except at such a superficial level that you’re hardly an “author” at all.

Imagine an AI improv night, where the audience is just one person. You can shout out the basics of a situation, and the AI actors will play it out for your entertainment. It would be fun, if the actors are smart enough, but you would hardly be an author at that point.

I see two extremes: hand-coded plots and AI plots, and a vast desert in between. It’s so expensive to develop plots in the desert that it’s more practical to fall back to one of the extremes. Which is to say that you can’t make AI plots cheaper by adding a human touch, because it’s so expensive to get the human to understand the simulation that you might as well hand-code the plot.

But the AIs come with a human touch. If I’m in the audience at AI improv night and I yell out “orange! mouse!”, some human input somewhere has decided whether the robots are going to show me a sketch about someone at tech support trying to clean orange peel out of their computer mouse, or one about someone whose pet mouse has fallen in the orange paint, based on what they’ve taught the robots about how the world works, or what will likely get a laugh out of the audience, or even just what words mean. If the AI is acting out a robo-pet character, it’s going to behave very differently if its engineer had a cat in mind than if they had a dog in mind. A love plot will be very different if the author believes that it’s possible to earnestly and selflessly love another human than if the author just got out of a messy breakup and thinks that love is a crock and we’re all lying to ourselves. When we write physics engines, we base them on universal laws that we all agree on (or some approximation thereof). When we write stories, there’s no such guarantee. Which is where the interesting stuff can come from.

In the way I’ve experienced games to date, I broadly agree with dfabulich.

The reason that changing what characters think or feel about events happening can feel unsatisfying is simply because you often can’t tell whether you did or did not cause any changes. In cases where you can’t tell if it was you, or you decide that it wasn’t, that can be as good as it not actually being interactive on the code’s side, as far as the aesthetic experience goes, no matter what the reality of the mechanical interaction was.

Common examples are where you can answer a person’s question 4 ways (multiple choice example), then the computer character says the next thing. We all know of lots of examples in games where no matter what you say, the other person says exactly the same thing afterwards that they were always going to say at that point. In isolation (in RPGs for instance, where there’s tons of other kinds of game around it that conspicuously let you know you’re doing stuff) people don’t mind or can enjoy such moments. Even in an IF game with a mixture of physical and character based stuff, enough else can be going on so you never doubt your feeling that you are an agent. When the majority of the game is about character moments, though, it gets tricky.

If the computer character actually has multiple responses depending on what you say, how can you tell in retrospect if that was so? The dialogue itself may make it clear, or it may not at all if the character is guarded or tricky. ‘By replaying the game’ is another answer to the question. But it’s the very complexity of human behaviour, that ‘interesting stuff’, that makes it hard to tell. That ‘hard to tell’ is what I see as the biggest challenge to authors. Action-actions are infinitely less ambiguous - you press the red button, something happens. You throw a rock at something, it shatters, etc. If I’m too often wondering ‘Did I change things?’ my brain loses interest and reaches for the ‘non-interactive’ assessment.

EDIT - Jon Ingold, I think, not Nick Montfort - sorry, pulling this from a Planet IF memory - also brought this issue up in his blog about Frankenstein. He rhetorically said something like ‘How do you let the player know that something has changed?’ while describing the issues around character interaction.

Anyway, these thoughts are why, to date, I’ve found the larger part of character interaction focused IF querulous.

I don’t want it to be querulous, therefore I demand everyone else solve all of these problems for me!

  • Wade