At NarraScope last weekend, Mark Bernstein (“Those Trojan Girls”, previous observations on hypertext narrative) was passing out a booklet entitled “The Fellow Who Caught Fire.” On the left-side pages are sections of a story; on the right-side pages, commentary about how stories are presented. Some of the areas for discussion are familiar from non-interactive literature, such as framing, tense, and person. Others dig into topics like explicit choice and link placement in hypertext narrative.
Since not everyone will have access to these, I thought I’d talk a little about what it contains, and some thoughts I had in response to Bernstein’s questions and provocations.
The passage on framing doesn’t really dig much at all about how it’s different to frame an interactive story, although there’s a little hint at this: “We might, for instance, allow the protagonist to recall the events long after they took place: this gives us access to that character’s thinking but places some limits on both plot — we can’t describe things that the character doesn’t know — and story.” Framing an interactive narrative might mean committing to a particular ending of that narrative, shifting the range of diegetic agency from “what will happen at the end of this story?” to “how will we reach this outcome?” or “what will we understand about the outcome when we get there?”
Another possibility, even wilder, is a narrative that randomly or not-so-randomly offers different framings to the reader. In her talk for the GDC AI Summit this year, Nicole Lazzaro discussed the idea of AI selectively reframing the same story with different titling or art (not even necessarily a frame story, but rather a set of packaging choices) in order to shift how the player is likely to experience it.
Elsewhere Bernstein talks about exposition, and how we choose what information the reader needs, given that an interactive narrative could focus on many different themes: “Interactivity makes these questions harder. In some readings, our story might concern adolescent rebellion while in others we focus on the reaction of mankind and the effect of that reaction on the hero. We might then need to write this passage in different ways — and we might also need to ensure it appears only after we have committed to one focus or another. Yet that point of commitment might not be obvious…”
My interactive translation Endure, though tiny, explicitly makes thematic focus part of the mechanic: the user’s choices about which text to translate first will inform what thematic material is considered important for later portions of the translation, so that if you focus first on the Cyclops, you’ll get a different final text than if you focus first on Odysseus’ cunning.
By far the most common technique I encounter, though, is to let the reader/player decide what they’re curious about, and explore non-critical points of exposition deeply or shallowly, with whatever focus they prefer. For me, this is one of the key purposes of providing exploratory passages in interactive story, whether the exploration mechanic is research (as in Analogue: A Hate Story or Her Story), physical movement (as in Gone Home et al), or conversation (as in Galatea).
On morality, Bernstein writes: “When our stories contain plausible computational agents that enact humanity, we should treat them humanely. So should the audience. The consent of agents is particularly perplexing: can we marry Ophelia on the Holodeck?”
The “so should the audience” certainly expresses a wish rather than a guarantee of reality — there’s extensive evidence to show that players will treat NPCs terribly. I have received some irate email from players who wished that they could sexually assault characters I wrote and were annoyed that I had not created verbs in my parser-based games to enable them to do so.
More recent games do interesting things with how they model character desires and undercut the assumption that the player’s wishes are paramount — I’m thinking of Meg Jayanth’s work in 80 Days, often centering someone other than the protagonist. In my own Choice of Games project, I have some characters who are always available for romance; one character who is interested in the protagonist at the outset; and some characters who have strong personal preferences that dictate whether they will reciprocate interest. So my solution here has been, not to thwart the player in all cases, but to suggest that the pliability of the NPC is itself an expression of that character’s personality.
Bernstein’s section on Links is the most technical, and assumes that the reader is familiar with his own terminology in this space, though many hypertext and Twine authors speak in partly or completely different terms:
“We typically encounter four kinds of links: time shift, recursus, renewal, and annotation. Even the humble annotative link might be epistemic, argumentative, or ironic…”
…though later he briefly raises the topic of stretchtext and cycling links, terms that may be more familiar. He concludes by lamenting that the poetics of links are not well understood at all.
Remedying that in any kind of academic way would doubtless require close study of a number of works, but this provocation made me think about what my own readerly assumptions tend to be when I encounter links (though a work may then defy those assumptions). Among these:
- A cycling link is offering me an in-line option among a few different items. Because it’s in-line with the other text and because it does not advance the story to a new passage, I tend to assume that it represents a choice with less diegetic impact than other choices, and that my selection in the cycle will either be completely reflective (how do I feel about this situation?) or be something that is making a minor and partial difference to later outcomes (e.g. by tweaking a stat)
- A stopping link — one where I can click one or more times to get to a point where the text becomes static — offers me an opportunity to interrogate and correct the narrative. (This is my term, not Bernstein’s, and it’s possible he would call that something else.) I tend to associate this type of link with situations where the narrator is meant to be unreliable or the protagonist is engaged in self-deception which we have the opportunity, as readers, to correct.
- Stretchtext is offering me more information about something, but that information is best understood in the context of the information at the surface level of a passage. I should read the whole passage and then click the stretchtext to understand this aspect more deeply. Conceptually, that’s similar to what the marginalia links do in Harmonia: though they are not stretchtext and explicitly appear separate from the main body, the idea is that we are looking at follow-up information that is best understood after the main text.
Bernstein also speculates about the placement of links on different parts of speech or in different areas of the page, and here I don’t see as much continuity from one piece of hypertext to the next. Perhaps the main thing that catches my attention there is link density vs sparsity, and whether there’s one, several, or a large number of links available at a time. As with choice menus, I tend to find large numbers of links overwhelming and single link situations constraining. Both of those effects can be useful in the right context, and howling dogs of course contains an artful application of hiding an important link in a thicket of ineffectual ones.