Interactive Storytelling: Techniques for 21st Century Fiction (Andrew Glassner)

Interactive Storytelling: Techniques for 21st Century Fiction (Andrew Glassner, 2004). Glassner’s book is rather more effort to read than most of the other guides to interactive story I’ve covered so far: it’s hundreds of pages longer, and in a somewhat more pedantic style. It begins with two long chunks on the nature of story and the nature of games.

He begins the section on stories by introducing many standard concepts of writing from scratch: character, plot, scenes. Conflict and stakes. Three-act structures and inciting incidents. The monomyth, again (though mercifully he admits that it is not necessary to use and is not the guarantee of a good story). Maslow’s hierarchy of needs. Viewpoint and dialogue. At the end of this section — about a hundred pages, much of it consisting of example narrations from film and other sources — Glassner proposes a “Story Contract,” which he will use throughout the rest of the book to make value judgments. The contract contains the following clauses:

• The author is responsible for the psychological integrity of the main characters.
• The author is responsible for the sequencing and timing of major plot events.
• The audience must allow itself to be emotionally involved.

Glassner later uses this contract to evaluate various works and forms of interactive story (about which more below), so baking in what the author is “responsible” for gives him a way to dismiss a lot of techniques in existing work. In many other respects, the story segment is largely a not very edited overview of basic writing advice.

In the section on games, Glassner also offers quite a bit of review. Like late 90s IF theory, he distinguishes puzzles from toys (this is something that we talked about quite a lot back then). Here, again, he offers a bunch of broad background: types of games, game loops, participation vs spectating, the nature of rules; the uses of scoring; the types of resources that can be included in game design, and the ways resources are deployed. He also gets into individual vs. team sports, competition and cooperation, applications of chance, some basic game theory chestnuts like the Prisoner’s Dilemma, and a section on terminology from Go.

All of this discussion is on the more abstract end and includes examples from sports and board games as well as computer games; it’s by no means focused purely on executing a AAA first-person shooter experience, and much of his game typology is not focused on the video game industry.

As with the section on basics of story, I’d be inclined to recommend other resources in the case that the reader genuinely hasn’t encountered these ideas before (and most people seriously invested in games and story probably will be familiar with them already): many other presentations are more concise or more nuanced.

On the Prisoner’s Dilemma, for instance, Glassner spends about three pages setting up the concept, including a full one-page of narrated skit, before concluding that there’s no best solution. The rather more interesting problem of the iterated prisoner’s dilemma just gets a couple of follow-up paragraphs, which acknowledge that different strategies can emerge but don’t get into any of the details.

After the Story and Game sections, Glassner starts to compare existing interactive stories stand up to his ideals. He gives glancing mention to some interesting topics — there are a few pages on the triangle of identities concept, for instance, though he doesn’t lay it out in those terms or in much detail.

But much of his segment is about Why Everything That Exists Is Disappointing. Here, his interpretation of the Story Contract leads him to dismiss quite a lot of potential subject-matter. He doesn’t like branching narratives, feeling that they undermine the psychological integrity and break the audience’s engagement:

Branching narratives have failed to find success in the mainstream because they offer no advantages to the author or the audience in terms of storytelling, and the sheer need for the audience to consciously make decisions destroys the vital empathetic connection between the audience and the characters they are identifying with and hoping and fearing for. (249)

And later in the same chapter:

It’s hard to see what practical benefits hypertext has for an author or a reader. Theorists enjoy thinking about issues like post-modern narratives, deconstructing texts, and hypothesizing what a successful form of hypertext might look like in a possible future. These ideas are fun to think about, but they’re the wrong tools for the job of storytelling… (253)

Glassner gets to these points by noting things often already said about choice-based narrative and hypertext: that it’s possible to lose a sense of protagonist character in a very free-branching choice narrative; that player agency is attenuated if the player doesn’t know where a hypertext link is going to lead; that the author needs some way to control the size and complexity of a branching narrative tree; that it’s hard to foreshadow things if one doesn’t know where the player is going to go. All of those are genuine challenges, for which the craft of interactive storytelling has developed an assortment of solutions. But Glassner has decided that these concerns invalidate those forms entirely.

And it’s not just CYOA and hypertext that gets this treatment. He also dislikes:

• any attempt to build a protagonist based on the player’s personality choices;
• choices whose consequences are not known at the time of selection;
• difficulty settings that let the player choose (especially when the choice is presented at the beginning of the game);
• cut-scenes for important plot events;
• random elements in character behavior;
• multiple-choice conversations with NPCs, which run the risk of not including some options that the player might hypothetically want to take:

This lack of ability to act on your desires is a disaster for a story-based game: the player is ripped right out of the game’s world…

Very often, in fact, I found myself agreeing with some of the general complaints Glassner was laying out, but saying “uh, hang on…” in response to his conclusions. Yes, it disrupts player intentionality to present choices whose implications are unclear. Player intentionality is often desirable. But there are exceptions — moments in a work that derive power from the fact that the player is uninformed or even misled. And there are also different levels of information you can offer the player about what a consequence is likely to look like. I’m reminded of a great moment of participatory comedy in Portal 2, for which I’ll borrow Andrew Kauz’s write-up:

Wheatley then asks the player character if she understands what Wheatley is saying, and asks her to simply say yes. A prompt appears on-screen: a textbook example of breaking the fourth wall. The developer — not Wheatley — is asking the player to press A to speak. So, being the player and having no other option, we press the requested button.

That button is, of course, the jump button. It makes the player character jump.

Next, you’re asked to say apple. The same button appears for you to press, though this time the developer’s programming assures you that it will result in your character saying “apple.”

It doesn’t. You jump again.

Now, if you’re anything like me, you laughed harder at this moment than nearly any other in Portal 2. — Andrew Kauz at Destructoid

Glassner is not particularly into these nuances or exceptional cases, and he tends to hand out a lot of prescriptive “should” instructions for his reader. “I looked at the techniques that are popular today and found them wanting,” he remarks (page 289).

Since I’m generally more interested in reading people’s positive ideas (even if I think those ideas are weird or misguided) than in reading their negative generalizations (especially when I think they’re dismissing quality work), I started to enjoy myself more starting around page 300 or so.

In Chapter 11 (“First Steps”), Glassner starts to offer recommendations of his own. He suggests that interactive stories should offer the most possible fun for the least possible work (though he doesn’t really explain what he thinks fun is; Raph Koster might be relevant here, for anyone coming to the question fresh).

At the section “System Interface and World Interface,” Glassner starts discussing how UI design tightly unites with good world system design (or what I would call systematic mechanics) to create an experience of discoverability and agency. This is a topic we discuss in IF less than I would prefer, though I tend to yammer on about these things whenever I’m given a chance. We certainly didn’t talk about them much at all in 2004, in part because we hadn’t really broken away from the expectation that the parser was just pretty much It as far as possible interfaces went.

He also makes a concerted pitch for using constructed languages to converse with game characters. That’s something Chris Crawford has explored and which has turned up in a handful of other games both IF and otherwise. This section goes into detail about Crawford’s own Trust and Betrayal as well as a communication system called The Elephant’s Memory.

About prose and story generation, he’s intrigued, but less sanguine that computers will get anywhere interesting in the very near future. (As a view from 2004, that’s entirely fair enough: we’re not all the way there yet, and 13 years have elapsed.)

Chapter 12 talks about “Story Environments,” where Glassner introduces what I initially feared was going to be a straight pitch for a holodeck: wouldn’t it be nice to be within a world that adapted to us? But a lot of his specifics are more precise and useful than that: for instance, the idea of a two-character drama where both characters are vigorously pursuing their own aims and the player’s part in it is to provide bridges (allowing protagonists to do what they otherwise couldn’t) or blocks (preventing things that are currently possible). This lines up a bit with the technique of waypoint narrative.

Another possibility — all the more relevant in the days of multiplayer VR — is the concept of “living masks,” where the computer is translating the player’s somewhat limited enactments and natural language input into a more stylish and characteristic “in character” behavior for the imagined protagonist: playing as a riverboat gambler, the player might say “you’re stealing my chips!” and the computer might translate this into a riverboat drawl. I can see how this could be cool in theory, though I’d also venture to suggest it’s quite a lot harder than he seems to appreciate.

Games that involve the player as a director also get a look-in here, a concept that bears some connection to games of co-authorship.

So there was a lot going on here, but even in this chapter I wished Glassner had had a visit from a vigorous editor. He lumps together as “adaptivity” everything from the waypoint-style plotting above to automatic difficulty adaptation, creature simulations as in Black & White, SimCity-esque approaches, crowd behavior models, and a digression (with diagrams!) on chaotic pendulum systems. That last gave me flashbacks to my college seminar on nonlinear dynamics, but I’m not sure how much light it really shed on the problems of interactive storytelling.

Finally, in Chapter 14, Glassner takes the reader through a series of experimental works that he designed to try out his concepts — that is to say, they’re short descriptions of imaginary projects, rather than things for which he’s implemented or sketched out a detailed design doc. Several of these deploy players in groups to handle any social richness too complex for the computer to mediate: I think Glassner might have been interested by Velvet Sundown and, say, by Eve Online (which did technically exist by the time Glassner published, but might not yet have reached its famous heights of player-directed military and political content).

So, overall: there are some correct, interesting, and novel thoughts in here (especially if you’re looking for “novel in 2004”). There are a bunch of other elements that are in my view pretty misguided or needlessly dismissive, and quite a lot of content that’s way too longwinded or disorganized; and most of his most positive ideas are divorced from much discussion of how you’d actually implement anything like them. The book has higher ambitions than most of the other game writing and interactive narrative books I’ve seen, but it doesn’t quite arrive where it wants to go.