Avatars of Story

I received a copy of Marie-Laure Ryan’s Avatars of Story almost a year ago, thanks to the generosity of the author, in part (I assume) because the book discusses Galatea. I read part of it at the time, but other tasks claimed my attention; and though I had a good opinion of what I read, it’s only now that I’m finally able to finish.

Ryan focuses on an issue central to the theory of interactive fiction: what happens to narratology when we apply it to material other than the traditional novel, and in particular to digital media? What does interaction do to story? In the process, she has to ask even more basic questions, such as how we define narrative — or a medium, for that matter. She then turns to a number of specific kinds of new media, including interactive fiction, hypertext, web-based narratives, and computer games, and discusses their narrative potential. This is the section that is likely to be of most interest to IF authors and players, but the portion tackling theoretical questions is valuable as well.

Ryan’s approach is scholarly; her writing, though lucid and backed by solid thinking, is also dense going. Where something like Second Person is accessible enough that I would recommend it to anyone making a serious study of IF (as player or as author), Avatars of Story is consistently academic and is likely to appeal only to those members of the IF community who are interested an abstract theory. But that’s not to say that it’s not worth reading.

Narrative, Media, and Modes

Ryan begins, as I said, by taking on essential questions of terminology. What is narrative? What is a medium? How do we define these terms in order to talk about the ways in which different media affect the forms of narrative?

This is difficult territory, but Ryan deals with it well; her definitions are academically precise without violating too much one’s common-sense ideas about what these terms might mean. Media, Ryan looks at in three ways: as semiotic types (language, image, music, etc.); as technological types (digital, oral, etc.); and as cultural practices. Language, interestingly, she identifies as being bad at communicating spatial relationships or encouraging the reader to build a specific sense of the layout of the storyworld — a defect that interactive fiction arguably partly addresses, by forcing the player to navigate the storyworld explicitly. All the same, her analyses here are generally on track.

Finally Ryan offers a definition “from the perspective of transmedial narratology”:

For students of narrative, what counts as a medium is a category that truly makes a difference as to what stories can be evoked or told, how they are presented, why they are communicated, and how they are experienced.

So far this definition is not analytically very useful, but Ryan goes on to offer a breakdown of the ways in which media can differ from one another in the narrative possibilities offered: do they extend through space or through time, or both? Are they static or kinetic? How many channels of communication are at work at a given time (language only, image only, music only? or some combination of these?)? Which senses are appealed to? Of these senses, which has priority over the others?

This kind of analysis would seem to make an illustrated book into a different medium from an un-illustrated one: but this is possibly justifiable. Ryan also introduces a phrase she will use often throughout the book, “thinking with the medium”, to mean putting the strengths of the medium to use in support of one’s artistic and narrative goals.

Drawing and Transgressing Fictional Boundaries

In the second chapter, Ryan takes on the question of what “fiction” is, and whether it is a category that admits of an absolute boundary or whether there are degrees of fictionality. This is just as difficult as defining narrative and medium, but (I felt) less critical to the core of her argument.

Ryan closes the chapter with some remarks on the relative functionality of myth and science; here she reveals a weakness, her desire to generalize from the strong and precise analysis of her native material into theoretical territories in which she is relatively less experienced. She writes, e.g.,

Myth is the voice of Truth itself, the foundation of a culture, a definitive representation that refuses to acknowledge the existence of any other version.

Her footnote to this passage admits that she means to refer to oral societies only and not to literate ones like ancient Greece, which is as well, since the statement is certainly falsifiable with respect to Greece in particular. But — though she reveals her familiarity with Paul Veyne — Ryan treats myth in this section as theoretically equivalent to religious dogma, and also accepts the myth-is-primitive-science line advanced in the nineteenth century by Max Müller and now generally rejected (or at least treated with great caution) by theorists of mythology.

Narrative in Fake and Real Reality TV
Narrative in Real Time

In these chapters Ryan looks at the type of narrative that emerges from Survivor, The Truman Show, and a baseball game as narrated by commentators; these point at some principles (in particular, that lived experience requires careful mediation and editing before it can emerge as what we would call a story). If there’s a relevance to IF here, it is (I think) that IF’s narratively “wasted” passages — the times when the player types LOOK and INVENTORY a great deal, and travels through already visited rooms, and experiences nothing that is new to the plot — are more life-like than story-like, but don’t preclude the possibility of a viable narrative being remembered as emerging from the IF work after the fact.

I don’t think this is really a much-contended issue, though; I’ve seen a little discussion of it now and then, with some people taking the extreme position that IF should always be designed to minimize inactive turns (perhaps by automating important background activity ongoing regardless of what the player does). But I think most players would not subscribe to such a strong view.

Toward an Interactive Narratology

Next Ryan begins to articulate some approaches to studying the narratives of interactive works. Here she identifies some common structures of interactive narrative, and characterizes different kinds of interactive games and stories based on the perspective of the player (internal to the gameworld or external) and the nature of his participation (ontological — that is, with the ability to change what happens — or merely exploratory, with the ability only to observe). This section is solid but not entirely surprising: it mostly provides some tools of categorization, which will prove useful in later chapters.

There were a few points that I would have enjoyed seeing more fully explored. Though she talks very briefly about the fact that narratives can contrast “the actual life story of a character with the virtual paths that the character did not or could not take”, Ryan doesn’t spend much time on what this means with respect to interactive media where (as in IF) the player can play out these alternate paths, whether they lead to equally valid endings (as in the case of choice-based games) or to losing endings (as in traditional puzzle-based IF). A fair amount has been said within the IF community about multiple endings that give the player a choice; we tend to talk less about the significance of losing endings to a game/narrative, though it has been observed that the knowledge of losing endings in a horror game such as Anchorhead heightens the tension. (In fact, I remember one of the losing endings better than I remember most of the rest of the game, so traumatic did I find it.) I wouldn’t mind reading more discussion of how interactive media treat the virtual, untaken paths of a narrative; but Ryan doesn’t delve into the issue much here.

Ryan concludes with three claims about interactive narrative which she rejects: first, that digital narrative is about choice, and that the more choice is available to the user, the better the experience will be; second, that narrative can be produced out of elements selected at random, creating an infinite range of possible stories; and third, that becoming a character in a story is the ultimate narrative experience.

I agree with her on rejecting all three. The idea of the infinitely-story-productive yet infinitely-free game has been brought up with moderate regularity on the IF newsgroups since at least the mid-90s, with players speculating about how much fun it would be to be able to “do anything” and have the world “react realistically”, but I find it hard to believe that this would actually be fun.

She then elaborates on this final point by saying that participation in the plot, through avatars, is a compromise between identification and observation — a point that speaks to some of the recent discussion on rec.arts.int-fiction and on this blog about how the player relates to the player character and whether (and how) IF is handicapped in portraying truly nuanced protagonists. But Ryan then adds, I think partly contradicting herself, that people will naturally

prefer identifying with a rather flat but active character whose participation in the plot is not a matter of emotional relation to other characters but a matter of exploring a world, solving problems, performing actions, and competing against enemies… If interactive narrative wants to expand the rather limited emotional repertory of games and develop complex characters who undergo truly poignant experiences, it may have to limit user participation to a largely observatory role rather than placing the user in the of the experiencer. (125)

In claiming this, Ryan has, I think, not adequately considered the kind of pleasure that arises from exploring a disagreeable character (such as Primo Varicella or Rameses), or from role-playing someone distinct from oneself: the pleasure does partly arise, as she suggests, from a kind of compromised sympathy with the protagonist which at the same time observes his situation as an object of aesthetic enjoyment.

Experience with live role-playing games supports the idea that this is a possible or even common way for the player to relate to his protagonist: I have often played, and watched others play, characters more greedy, lustful, flirtatious, violent, ambitious, or deceitful than we would be in life, and yet at the same time felt some genuine sympathy with the character when they came to grief. For that matter, I’ve known a few people who seemed to live real life in a permanent drama-queen state of mind, always acting more for effect than out of genuine emotion. As a way of life this is unhealthy, but in games it can work well.

Writing for this kind of play is not easy. RPG authors have found various ways to encourage it; for instance, some roleplaying systems allow players to get better character statistics if they will also take on some failing or flaw. This provides hooks for the gamemaster to use in developing the story, but it also gives players something to work from when characterizing their protagonists, and part of the fun of playing is to spin that out as far as possible. (Though some flaws are more fun to develop than others. I remember one rather grim session where my protagonist was coldly pragmatic. In the scenario, this meant she was willing to torture her enemies for information, but we quickly reached a point where, even in play, I couldn’t bring myself to make the protagonist go further. Fortunately, I was able to cow our enemies with threats and didn’t have to inflict any physical damage: possibly the GM took pity on me. But where was this angst located? Whose problem was it, mine or my character’s? After all, because of my discomfort, my character manifested a bit of queasy conscience after all, which didn’t accord with the traits I originally wrote down for her; was this nuanced characterization or just an inability to play the game right?)

But I digress. My main point here, really, is that — despite everyone’s concerns to the contrary — it’s possible to present a more nuanced player character than the flat, active type Ryan describes, and that part of the potential arises from the observer-vs.-experiencer tension she identifies. I’ve talked elsewhere about techniques of characterization such as restraining the PC’s actions; hinting at attitude and backstory as the player explores the surrounding world; and providing some in-character actions just for local color. To the list, I might add (on the basis of comparison with RPGs) communicating major PC traits to the player early in the game so that the player has a chance to do a little roleplaying consciously. Perhaps the most obvious case I can think of where the player actively plays a well-characterized role (as opposed to being forced into one by the constraints of the game) is in Varicella. The actions in Varicella are tuned to encourage this — the player, adopting the role of a manipulative courtier, selects the tone of voice in which he will speak to people according to their social standing; collects secret and damning information in order to solve puzzles; and stabs people in the back, over and over. But this all works in part because Primo Varicella’s flaws and strengths — his cleverness, his conniving, his obsession with interior decorating — are all conveyed to the player before the first move the game:

You are Primo Varicella, Palace Minister at the Palazzo del Piemonte. This title is unlikely to impress anyone. Piedmont is the laughingstock of the Carolingian League, and the Palace Ministry has devolved into little more than a glorified (and not even especially glorified) butlership: your duties include organizing banquets, overseeing the servants, and greeting visitors…

Piedmont, it seems, will be requiring the services of a regent for the foreseeable future. And you can think of no better candidate than yourself.

Of course, you shall scarcely be alone in seeking the position. The King’s Cabinet is not a small body. And your fellow ministers will no doubt try all sorts of unseemly tactics in their quest for the throne. Some will try bribery. Others will employ treachery. A few may even resort to brute force. But would Primo Varicella stoop to using one of these methods? Perish the thought! You’re better than that. You shall employ all three…

There should be little to stand in the way of your ascent to power so long as you put your plan into action immediately.

Or at least as soon as this manicure is finished. One must have one’s priorities.

And despite these somewhat cartoonish outlines, Primo becomes one of the most memorable protagonists in IF. To pull this off, I think one has to convince the player that he understands his avatar, and that it’s going to be rewarding to play-act (rather than playing purely “to win”); also, to give him some opportunities in the game to perform a bit and explore the boundaries of this character; and then (what Varicella doesn’t do, particularly) play off that character nuance within the story by making it relevant to the major conflicts.

At least, that’s my current idea, subject to further experiment.

Interactive Fiction and Storyspace Hypertext

In this and in succeeding chapters, Ryan describes particular interactive media and then discusses, on the basis of her interactive narratology, the narrative possibilities she perceives in each medium.

A little to my disappointment, Ryan analyzes the narrative possibilities of IF less thoroughly than she will for several other media; she notes that world models are usually incomplete and that a subversive player can often trick IF into producing nonsensical responses, but this does not seem like a strong limitation on the type of storytelling possible.

For the most part, her description of the form won’t surprise experienced players, though she makes a few odd assertions that betray inexperience with the breadth of what IF does these days. For instance, she writes

The interactive fiction engine supports two main types of narrative. The first, and most widely represented, is the puzzle-based quest… Mindwheel anticipates the second form of IF narrative: a conversation with a system-generated character (chatterbot, in the jargon) reminiscent of ELIZA… In this type of IF there are no puzzles and no geography: the entire action takes place in the same room, and the only problem to be solved is eliciting interesting confessions from the character. (129, 131)

While there are plenty of puzzle-based quests in interactive fiction, her definition of the “second type” applies almost exclusively to Galatea: while there are other works I would describe as “conversational” and generically akin to Galatea, most of them break at least one of the criteria she offers here. Redemption and Shadows on a Mirror both have some degree of puzzle-directedness or some goal beyond eliciting information; Best of Three and The Big Mama break the single-room rule; etc. And in any case this division leaves out all sorts of other possibilities: non-puzzle-based works like Rameses and Photopia; one-room puzzle games like Enlightenment; art-show landscapes and still-lifes; pieces focusing on moral choice, like The Baron and Fate. The last two pieces weren’t yet out when Ryan was writing her book, but still: IF is rather broader than this categorization would imply.

Much, arguably too much, has been made of the creative role of the reader in digital environments. The fact that the system of IF rewrites most of the player’s input seriously dampens the claim that interactivity turns the reader into a coauthor: even though the player interacts through language, most of her contributions are treated as paratext, and she does not participate directly in the writing process. (136)

This is true. I know of only a very few works that attempt to get around this to any degree. Winchester’s Nightmare offers an unusual prompt that encourages the player to view typed commands as though they were part of the ongoing narrative structure; another of Nick Montfort’s works, Ad Verbum, pays attention to whether the player abbreviates directions or spells them out, and adjusts its output accordingly. And — largely inspired by Nick’s example — I have occasionally played with directly parroting the player’s input in specialized situations. As a rule, though, such effects are at best playful and trivial; there’s no possibility of encouraging the player to write literary prose as input, since the parser wouldn’t be able to interpret any such thing; and, in any case (and a more serious problem than it might first appear), there is no audience for what the player types, beyond the player himself. For the player to be a coauthor of a work, there has also to be a parity of readership — someone out there who will experience the player’s contributions, and perhaps even experience them interactively. Short of radical experiments like Victor Gijsbers’ ideas of a system in which the player could rewrite parts of the game as he played, I don’t see how this is to be done; nor am I certain that it’s really desirable. (Taking things to the opposite extreme is My Angel, which separates the command line from the text generated so that what appears on screen is a continuous stream of narration unbroken by the player’s instructions; all input occurs in a separate window.)

Her account of Storyspace hypertext in the same chapter suggests that it is a medium rather more limited than interactive fiction. This is perhaps surprising — hypertext appears on the surface to be closer to a novel than IF does, which would seem to imply that it shares more of the established strengths of static literature. But she writes

Thinking with the medium in hypertext also means giving meaning to the reader’s activity. In computer games, this meaning derives from the player’s identification with the avatar and from the tasks to be fulfilled in order to win… But in hypertext, as I argue in chapter 5, the reader’s involvement in the fictional world is external and observatory, and the significance of her activity cannot come from playing the role of an individuated member of the fictional world. (145)

The range of narratives she envisions being told in hypertext form is fairly restricted, to loose thematic collections of very short stories; representations of a mental experience or journey; or a mostly-linear treatment in which linked texts are entirely supporting materials to the main fictional branch. I’d like to think that the form is capable of more than that, but I haven’t ever come up with an idea that I wanted to try executing in hypertext form.

Web-based Narrative, Multimedia, and Interactive Drama

In this chapter Ryan continues the investigation of individual media by considering Flash and Director works; narratives that might emerge as in alternative-reality games from a collection of pseudo-factual websites (she calls these “archival narratives”); and finally the interactive drama Façade by Stern and Mateas, which, like Galatea, focuses on conversation between the player and the characters, but unlike Galatea functions in real time, accepts freeform natural-language input, and displays its events images and spoken language output rather than in text.

This strikes me as a somewhat odd collection of items to lump into a single chapter, Façade being so different a beast from a hypothetical short story told through the medium of fake wikipedia pages.

Still, her observations about “archival narrative” struck me as very much to the point:

Narrative texts do more than chronicle actions and events; they also provide background information that situates these events within a concrete environment. As they follow the life of the main characters, they also pick up the destinies of the secondary characters, together with their spatial surroundings… All narratives must eventually limit this accumulation of space and time, though some authors — Dostoevsky and Laurence Sterne — do it rather reluctantly. With a digital database, the decision to stop the spatial and temporal growth of a textual world no longer needs to be made by the author. It is the reader who decides who far she wants to follow trails into new narrative territories. (149)

Though this concerns something like a wikipedia-narrative (the 21st-century twist on the epistolary novel?), several of the observations seem also applicable to IF. IF also allows the player some ability to govern the extent and arrangement of exposition. It is up to the player to examine things, to look things up in books, to ask questions of characters; and while some of that material may be required for the player to finish the game, most IF also incorporates a certain amount of optional background, though of what depth varies from work to work. Floatpoint I consciously designed to contain quite a number of background stories that can be researched by the diligent participant but that are not required for the player to complete a run-through of the game; to some extent this is because I wanted the player to make his own determinations about what kind of information he felt he needed before he could make the moral choice at the end of the work. Elysium Enigma likewise invites the player to discover as much as he can of the world’s true history, though it locks away more of this information behind puzzles. (On the other hand, EE is also more explicit about how successful the player has been about finding everything there is to find, since this is tied to the score and is announced at the end of the game.)

About the archival narrative medium, Ryan writes

The reconciliation of database and narrative is facilitated whn the following conditions are met:

1. A storyline with which readers are already familiar. When the global coherence of the story is not problematic, readers can bring a magnifying glass to certain parts without losing sight of the whole plot.
2. A very modular narrative, whose individual parts are themselves more or less autonomous stories.
3. A narrative that foregrounds the setting, so that learning about the world in which the story takes place is at least as important to the reader as following the narrative events proper.
4. A database design and linking philosophy sufficiently transparent to enable readers to aim with precision at the elements of the story that they want to expand. (149)

Some of these apply to IF: certainly storytelling with a lot of optional exposition works best in a world with a deeply-developed setting, and the player’s control of pacing works best when there are identifiable ways to do in-game research. Requirements 1 and 2 are, I think, less in force in IF than they would be in an archival-type narrative because the traversal of a work of IF is a bit more structured: there is usually a goal or explicit ending, allowing the author to present a coherent narrative despite the player’s forays into optional exposition.

Much of the rest of the chapter is less relevant to IF, as it explores flash-based artworks and other forms whose interactive structure is quite foreign to interactive fiction, but at the end of the chapter Ryan turns to Façade. Her observations here are largely along the same lines that I would draw: that Façade succeeds (when it succeeds) in large part because the story (about an arguing couple) provides a good cover excuse when communication breaks down and the game doesn’t respond naturally to the player’s remarks; that miscommunication is less evident than it would be with an IF parser, but that the player also sometimes has the feeling that he is not fully in control of events. I think I ultimately — and perhaps this is just a matter of taste — come down on the side of preferring clearer agency even if it means that the illusion of the work is occasionally broken: I would rather play/read an interactive work where my input is either effective or explicitly rejected than one in which everything I type is nominally accepted but it’s never quite clear what has really made a difference.

Ryan ends the chapter with a paragraph I both accept and find unsatisfying. She argues that it is inappropriate to judge digital literature according to the standards of print literature, that different media have their different strengths, concluding:

Digital texts should not be expected to be enhanced versions of the novel, of drama, or of the cinema. Their achievements reside in other areas: freely explorable narrative archives; dynamic interplay between words and images; and active participation in fantasy worlds. (160)

My dissatisfaction arises from the sense that — after a spirited defense of differences between media — the actual achievements listed for digital narrative seem rather weak. Is “active participation in a fantasy world” as valuable as what we get from great works of print literature, movies, or drama? Is a “freely explorable narrative archive” anything but a toy? What is this giving to humanity that is important? What new avenues of understanding is it opening between people? What new forms of beauty is it revealing to us?

So in the end I think: yes, it’s a mistake to judge digital literature by the standards of print literature, but these aren’t the strengths of digital literature. I have my own, possibly crackbrained and certainly incomplete ideas about what those strengths might be, but I believe they go well beyond what Ryan has thought of listing.

Computer Games as Narrative

In this chapter Ryan takes on a debate that will be familiar to steady readers of Grand Text Auto: what is the relationship between games and narrative? Is it possible to have a good narrative in a game? Are the game components always necessarily primary?

Ryan does a fairminded job of summarizing the various takes on this question, and argues sensibly that some games have more narrative content than others; she concludes with a list of suggested questions to ask about the relationship between narrative and game rules in any given game, by way of studying how that interface works in particular cases. For the IF theorist, this list is probably the most interesting part of the chapter. Some bits will seem awfully familiar from long-standing arguments on rec.arts.int-fiction. Ryan asks, e.g., “Are the player’s actions an integral part of the plot, or merely a way to gain access to spaces where more of the story will be revealed? How are the personality and past history of the avatar presented to the player? When the player gets to know the avatar only gradually, how can he take meaningful actions for this character when he is still incompletely acquainted with him?” (201); or, a little more abstractly, “Evaluate the connection between gameplay and narrative: could the same system of rules be narrativized in many different ways, or is there an organic, necessary connection between rules and narrative?” (202).

Metaleptic Machines

It is possible that by this point in my reading I was suffering a bit of mental fatigue, but I found this chapter to be the least useful portion of the book. Ryan looks at the ways in which the boundaries between levels of discourse can be broken down: by creating code poetry, by transgressing levels of VR reality, by having aspects of computer games refer to the fact that they are games. After the glut of movies like Matrix, eXistenZ, and Thirteenth Floor a few years ago, too many Star Trek episodes about the holodeck, the “Never Again” episode of Buffy, If On A Winter’s Night A Traveler, and several IF games to remain nameless for fear of spoiling them, I feel that the idea of reality-stack-popping has been mostly played out: it’s a cool gimmick, but not much more than that. I was a little diverted by the discussion of code poetry because we’ve seen some experiments in Inform 7 poetry since the language was introduced — but again, these are charming and fun but too formally constrained to bear the burden of much significance.

At a few points in this chapter I also felt (as I did with her discussion of myth and science) that Ryan was excited by the ideas but hadn’t thought through their implications as precisely and thoroughly as she does the other ideas in the book. She does quite reasonably point out at the end, though, that no amount of metalepsis is going to cause the player any true confusion about his place with respect to the narrative.

On the whole, I found this book quite useful in organizing some of my thoughts about IF theory, even where I didn’t agree with its conclusions. Ryan sets out a good deal of terminology with sensible definitions, and offers some good categories for comparing media.

I think there’s still deeper to go: as I’ve said, I don’t think Ryan has identified what the strengths of digital literature really are. But reading Avatars of Story has helped me think about that question for myself.

8 thoughts on “Avatars of Story”

  1. If she were to open up her definition of the chatterbot style game, there’d be a lot more in it. I see nothing intrinsic to that style of game that precludes merging it with other types of gameplay, for example, by adding multiple rooms, multiple chatterbots, etc.

    If we open it up a whole lot, we could even put in some modern commercial games, like Phoenix Wright, Ace Attorney.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: