May 26, 2003

Narrative as Virtual Reality

by Noah Wardrip-Fruin · , 9:53 pm

I’ve been working on a review of Marie-Laure Ryan’s Narrative as Virtual Reality for Computers and the Humanities, the journal of the Association for Computers and the Humanities. It needs to be short, explain why the journal’s audience might be interested in Ryan’s topic, and also give my personal take on the book. I’d be very interested to hear comments on the draft below.


Last year Andrew Stern sent out an enthusiastic email message about Marie-Laure Ryan’s Narrative as Virtual Reality. He called it one of the best books to address interactive drama.

Interactive drama is an area of investigation that attracts scholarly and popular audiences. At its broadest, it covers the wide range of computer experiences that have story content, some form of performative enactment, and a means for the audience (whether a full theatre or a single person in front of their PC) to alter some aspect of this story or enactment. The group interested in interactive drama includes English professors who see it as a future form of literature, media scholars who see it as an approach for understanding computer games, computer scientists who see it as the next major application for artificial intelligence, and entertainment executives who see it as the next stage of cinema. Interest in interactive drama has contributed to the success of past books such as Brenda Laurel’s Computers as Theatre, Janet H. Murray’s Hamlet on the Holodeck, Espen Aarseth’s Cybertext, and Mark Stephen Meadows’s Pause and Effect.

Stern is one of the leading practitioners in the area of interactive character and drama. After getting his email I decided to do a web search and see what else I could learn about Narrative as Virtual Reality. To my surprise, I found almost nothing. There were some weblog comments, but few did more than mention the book’s existence. I didn’t find a single online review, and I found few references to offline ones. At the end of this process, my interest was piqued. A book came out from a major academic publisher on a topic of current interest — and to someone as well informed as Stern it was an exciting addition to the literature, but the field as a whole had largely ignored it. It was a bit of a mystery. And so, in order to have a reason to delve into this mystery myself, I set out to review this volume.

My conclusion: the field is missing out. Ryan’s writing holds ideas that should be part of our discussion of interactive drama. I can only speculate as to why this book hasn’t garnered more attention, but I do have a theory. Upon initial perusal, it’s not clear that this is a book about interactive drama, or that it contains many original ideas. It’s frontloaded with chapters that review the ideas of others in a not-especially-gripping and frankly sometimes questionable manner. This book would, I think, be a touchstone for the field if it were a svelte monograph like Cybertext (also, as it happens, from Johns Hopkins) rather than nearly 400 pages long.

For the rest of my review I’m going to pretend that Ryan’s book is that svelte volume. I’m taking this unusual approach for two reasons. First, it would be little service to go over the less-interesting parts of the book in this review. Second, I’m hoping this review can be used as a set of suggestions for potential readers. If one reads the chapters of Ryan’s book as presented here, and reads only the chapters mentioned, I think there’s much to gain from the book. I certainly don’t agree with everything Ryan argues in these chapters, but every one of them is rewarding.

A general introduction to the concepts of virtual reality is found in chapter 2. This is useful for those who aren’t already familiar, but it also holds some attractions for those who know the field well. For example, Ryan’s well-executed breakdown of the vision of the holodeck (beginning on page 51) is one of the best introductions I’ve seen and looks quite appropriate for classroom use. Later in the chapter, the section on “Simulation as Narrative” (page 62) makes the first moves toward the ideas Ryan will explore in Chapters 8 and 10.

The book begins in earnest with chapter 7 — unfortunately titled “Hypertext.” This titling is unfortunate both because Ryan’s approach to hypertext is the volume’s greatest weakness (a topic returned to below) and because hypertext is not her primary subject in this chapter. Instead, chapter 7 is when Ryan begins to deploy her specific ideas about interactivity. She introduces the concept of “ergodic design” (building on Aarseth) in order to discuss work that includes a feedback loop in its design, causing the act of reading/viewing to produce alterations in the work “so that the reader will encounter different sequences of signs during different reading sessions” (206).

The standout feature of Chapter 8 is “The Structures of Interactive Narrativity.” This section, which describes and diagrams twelve varieties of interactive story structure, will find immediate use in my teaching. At a minimum it dispels the notion that all types of reader choice are the same. But more broadly its survey and categories will be useful for budding artists trying to think through a project, as well as for young critics seeking a way into discussing a particular work’s structures. Perhaps most importantly, these categories should be a useful bridge vocabulary for the two groups to share. Ryan also draws some controversial conclusions from her diagrams (e.g., “the potential of a network to generated well-formed stories for every traversal is inversely proportional to its degree of connectivity”) which should provide grist for heated classroom discussion.

Following a strong case study of the “interactive movie” I’m Your Man, chapter 9 is an opinionated survey of possible predecessors to interactive drama — from baroque art to children’s games to theatre. In the last of these the work of Augusto Boal is puzzlingly omitted, which leads to the unfortunate conclusion that the “vicarious interactivity” of Artaud (304) is the maximum point of interactivity before we see “the actors becoming their own audience” (305). Nonetheless, it’s a generally good survey, and notable for discussing erotic scenarios as a type of embodied immersive dramatic experience. It also lays the groundwork for the final chapter, during which Ryan deals most directly with the material about which she seems passionate. In this chapter’s strongest section, she offers the VR artwork Placeholder and the CMU Oz Project as two differing models of interactive drama. Placeholder focuses on the spatial immersion and bodily issues that are, in some sense, the baseline of VR, while the Oz project’s goals were more temporal and narrative. In concluding the chapter Ryan offers some interesting speculation as to how these strands may move into the future.

Before I offer my own conclusion, however, I must address the one shortcoming of this book that will remain glaring even if one reads only those sections I have recommended: Ryan likes to talk about hypertext, but clearly hasn’t done her homework. Hypertext comes out of the work of people such as Ted Nelson, Doug Engelbart, and Andries van Dam. But none of their names appear in Ryan’s index. Ryan writes about 1990s hypertext theorists — lumping Moulthrop, Joyce, Bolter, and Landow together as though they argued the same position — and discusses the idea that the hypertext experience blurs the lines between reader and writer. She demeans this position repeatedly, but apparently without awareness that the hypertext systems these theorists were using (e.g., Intermedia, Storyspace with the full application) made movement between reading and writing smooth, and also allowed one author to place links in the writing of another author. Neither is true of her hypertext examples (e.g., the Web, Storyspace files in stand-alone “reader” form). While her historical research in other areas seems fine, when hypertext is the subject it’s like reading a book that talks repeatedly about psychoanalysis and yet is written by someone who hasn’t read Freud. Or like reading a book that continually returns to natural selection by an author who thinks the phrase comes from Dawkins rather than Darwin. I found I could only enjoy the book if I mentally replaced each use of the word “hypertext” with a phrase like “read-only, link-and-node hypertext.”

Of course, the root of this difficulty was probably in Ryan trying to cover too much, leading her to address an area (hypertext) that wasn’t close to primary work, and so didn’t get researched as much as it should have. This doesn’t in any way lessen the contributions she’s made in other areas, and I heartily recommend this volume to anyone with an interest in what we mean by drama, performance, and narrative in connection with new technologies.

19 Responses to “Narrative as Virtual Reality”


  1. Matt K. Says:

    Every semester I have to hack away at the pernicious idea that all cybertext, whether “Adventure,” a Storyspace hypertext, or electronic lit on the Web is simply a form of “choose your own adventure” (it’s really amazing how universally acquainted all of my students are with those books). If The Structures of Interactive Narrativity material will help slay the choose your own adventure dragon then it will find immediate application in my own teaching too.

    Thanks for the review–I’ll definitely check out the book. And glad to see it’s going to CHUM.

  2. Lisa Says:

    I enjoyed this. I’ve been studying traditional narratives for a long time, but I’m still quite new to all the electronic permutations. So, it was instructive to me to hear names that were unfamiliar, grounded in relation to those I do know. (My first contact with hypertext fiction being Stuart’s asides in tutor’s meetings about his first novel, then in-progress.)

    I also particularly like that you engaged with the book so personally and so thoroughly. You don’t preclude another reader forming a different opinion overall, yet you do give a roadmap to some best bits in what could be a daunting read.

    My only editorial suggestion would be that the second mention of Andrew’s name at the beginning of the third paragraph, separated from the first mention by a paragraph dense with other personages, is a bit dislocating. Perhaps a smoother transition could be found there.

  3. noah Says:

    It looks like the appearance of the review will be quite timely, as JHUP will be bringing out a paperback this fall:

    http://www.press.jhu.edu/press/books/titles/f03/f03ryna.htm

  4. andrew Says:

    Ryan is giving a Plenary talk at the ACH/ALLC 2003 “Web X” conference at the University of Georgia, in Athens, GA, this Saturday at 10am, called “Metaleptic Machines”.

  5. Marie-Laure Ryan Says:

    Just a few comments from the author of “Narrative as Virtual Reality.” First, I’d like to thank Noah from what I perceive as an insightful and sympathetic review. My main point of disagreement is that the book starts only in chapter 7. My purpose was not to write a book exclusively on digital texts, much less on Interactive Drama, but rather, a book that would bridge the gap between literary narratology and what is increasingly getting known as “new media literature.” In other words, it’s not a book exclusively for “techies.” The first part, “The Poetics of Immersion,” deals mostly with print narrative of the realist school, because I feel they have perfected the art of immersivity. Digital texts by contrast tend to be influenced by postmodern aesthetics, whose insistence on self-reflexivity I regard as fundamentally hostile to immersion. The question I ask is whether this resistance to immersivity is a matter of ideological position, or whether it is inherent to the medium.

    About my neglect of the founding figures of Hypertext (“Ryan did no do her homework”): I believe I am reasonably familiar with this tradition, but I chose not to treat it because (1) a lot has already been written about the history of hypertext; and (2) the topic of my book is narrative, so I had to neglect the use of hypertext for informational databases. This forced a focus on Landow, Bolter, Joyce, etc. who have dealt with Hypertext’s supposed ability to “reconfigure narrative.” My question was whether or not the hypertextual format is compatible with the basically linear meaning structure of narrative. Though many texts have developed a “spatial” type of meaning (for instance, lyric texts), I believe that pure spatiality is a cognitive dead-end because the mind needs some continuous thread to hold information in memory and to keep interest alive. Even with a non-narrative essay, we are driven by the desire to know what comes next, and how it connects to what we have already read. It has been said that hypertext is closer to how the mind works than standard linear textuality because it creates analogies between seemingly disconnected elements. But if analogy is an important source of creativity it is not the only one. Look at any philosophical text, persuasive text, narrative text: the creation of meaning—what we may call thinking– is to a large extent a process of linearization. Of course it could be said that “new media literature” does not want to be ready-made meaning, but a tolkit for making meaning, what Lev Manovich calls a database. As Landow puts it, we should freely fish out of this database and make our own meaning. (But how can we, if the fragments have been selected and linked by another mind ?) Alternatively, we could make meaning by “spatializing” a linear text. Literary critics have been doing that for years. I personally find it easier (and more productive) to spatialize the linear than to linearize the spatial in a meaningful way, because the linearity of the text holds it together in my mind, and allows me to wander back and forth along this line. To me, the maintenance of some kind of narrative, and consequently linear coherence is crucial to whether new media literature will remain an academic fad, or will reach the wider audience of the educated public. Note that I am not even dreaming of an audience comparable to that of computer games, which have mastered the art of combining choice and linearity, but at the cost of narrative diversity.

    Well, enough rambling for today. I realize I have been carried far away from my original purpose—commenting on Noah’s review. But I did so in a linear way, so perhaps you did not notice.

  6. andrew Says:

    wow, how fun, reviewer and reviewee having a public dialog! cool.

  7. noah Says:

    Marie-Laure, thanks for coming by to comment!

    Of course I agree with you that the early chapters of NaVR are important to your overall project. And I quite enjoyed some of the close readings early on — the review should probably mention a couple of them. But all told (perhaps just because of my background) those chapters didn’t feel as though they were making as big a contribution to the field as the later ones. And, when I started thinking of recommending to people that they read only the latter chapters, I went through to see if I felt those chapters depended for readability on the presence of the earlier ones, and it seemed to me that they didn’t. Certainly, on their own they make a different book than they do when wedded to the earlier chapters — but they make a book that I feel I can recommend to a wider group of people.

    I’m guessing we’re going to have to agree to disagree about hypertext. But for what it’s worth, here are two points that occur to me.

    First, I believe that if you’re going to critique new media theories based on particular new media systems, you must take the specifics of those systems into account when mounting your critique. Moulthrop, Bolter, Landow, Joyce, and company were using particular hypertext systems, which blurred the distinction between author and reader in particular ways, when they did the writing you critique in NaVR. But your critique seems based on the examination of systems that don’t function in the same ways — even if they are still called hypertext systems, or even if (as with Storyspace) they bear the same name despite their very different functionality.

    Second, as you can probably tell from The New Media Reader, I’m passionate about the importance of new media’s history. I think we are a field, we have a history, and to ignore that history (as is sadly common) is intellectually irresponsible. Acting as though “hypertext” was coined by Landow is similar to acting as though “deconstruction” comes from Crews. It’s not an option one can choose within the field, in order to focus on one aspect or another — it’s simply inaccurate.

    If we look at where the term actually comes from, we get definitions very different, and more interesting for discussion of new media narrative systems, than those at work in NaVR. For example, Ted Nelson (who coined the term) wrote of hypertext and hypermedia in 1970:

    “‘Hypertext’ means forms of writing which branch or perform on request; they are best presented on computer display screen. . . . Hyper-media are branching or performing presentations which respond to user actions, systems of prearranged words and pictures (for example) which may be explored freely or queried in stylized ways. . . . Like ordinary prose and pictures, they will be media; and because they are in some sense ‘multi-dimensional,’ we may call them hyper-media, following mathematical use of the term hyper-.’”

    I understand that you see NaVR as asking “whether or not the hypertextual format is compatible with the basically linear meaning structure of narrative.” I’m saying that NaVR starts out trying to answer this question using a definition of hypertext that is the product of relying on a specific, dated set of secondary sources (which are then seen through the lens of the web and the Storyspace Reader) and by starting from this perspective it can only come to incorrect conclusions.

    Which is not to say that I think NaVR is somehow fundamentally flawed by this. I just wish the topic of hypertext had been left aside altogether, because the book’s strength lies in other areas.

    BTW, I was excited to read on the JHUP site that you’ve got another book in the works: Narrative Across Media: The Languages of Storytelling. Any chance you’d be willing to let us know a little more about it?

  8. andrew Says:

    Noah and Nick, I haven’t yet acquired a copy of The New Media Reader, and maybe the answer to this question can be found there, but I’ll go ahead and ask:

    Reading the above quote snippet of Nelson’s definition of hypertext and hypermedia, I’m not seeing how it is larger or different than what would describe the hypertext or hypermedia we see on the web, or from Storyspace… Do you imagine the difference lies in the “queried in stylized ways” part? Or that web / Storyspace hypertexts don’t typically allow the user to “explore freely” enough? (I’m thinking back to that keen observation Espen makes in Cybertext about the promise of hypertext liberating the reader – that in fact, hypertext can actually limit the reader’s freedom when compared to reading a traditional book. When reading a book, the reader is technically unconstrained to jump around to any page at any time in any order, even though of course we typically read books in a linear fashion.)

    For a moment I wondered if Nelson’s definition could encompass what we think of as cybertext; but the “prearranged words and pictures” part of the description seems to prevent that interpretation, since cybertext is capable of constructing its own arrangements of words.

    Sorry if this discussion has already occurred elsewhere in the past; if there’s a paper you can point me to that addresses this, that would be great.

  9. noah Says:

    Hi Andrew -

    Just a quick answer (after my perhaps over-long answer above).

    For some examples of hypertext/media — from that same 1970 article of Nelson’s — which are a bit different from what’s possible with HTML and Storyspace, see the series of 10 or so slides that begins here (part of a larger talk on ewriting):

    http://hyperfiction.org/talks/ewriting_hypertext/#slide07

    As for the distinction between hypertext and cybertext…. Well, one thing you see a hint of in what I excerpted above is that Nelson had a very strong view of hypertext/media as being authored. You could say that he was constructing his view in opposition to the “let’s build an AI to figure out how to present it” point of view. And cybertext definitely embraces both the hypertext kind of work and the AI kind of work.

    Of course, now there’s this Expressive AI business, which might be seen as using AI tools in an “authoring” way, and therefore perhaps bring them under the umbrella of hypertext. And to some extent I think people in the CS hypertext community think that AIish work can be hypertext, as we see in some of the “adaptive hypertext” work (which I have to admit I haven’t read that much of, but it looks interesting from afar).

  10. noah Says:

    Oh, and, Andrew, the answer to your question is yes. The New Media Reader contains the 1970 Nelson article I’ve been referencing here, as reprinted in his Computer Lib / Dream Machines. There’s also some other material from the CL/DM book, the 1965 article which introduced “hypertext,” and some of his 1981 outlining of the particular Xanadu model of hypertext. And there’s also material (both text and digitized video) from Doug Engelbart who, among many other innovations, was the first to implement the “click a link” type of hypertext. And so on. It’s a pretty thick book, but it’s designed to answer questions just like the one you asked, and to do so via the presentation of primary sources.

  11. Lisa Says:

    That was helpful. (The New Media Reader is on my Amazon wishlist, but so far no one has made my wish come true.)

    I have a question, which is not meant to be dumb or difficult (in the sense of causing a fuss), but which may be both:

    Why would an author *want* to yield the authorial control of a piece to some sort of AI engine? I find I can grasp why an author would want to make the work modular, rather than linear, but I can’t make the next leap. (Perhaps this is because I’m not a gamer?)

    And yet, and this is probably a huge inconsistency, I have been teaching others what to make of Dramatica for years now. It is, in a way, trying to replicate how the human mind solves a problem, so perhaps it qualifies as a primitive AI engine. I suspect my understanding falters here at the brink between being able to manipulate software and being able to program it.

  12. andrew Says:

    That’s a really good and interesting question. My reply is in a new top-level blog post.

  13. Lisa Says:

    Thanks. (The new post is a really good and interesting answer.)

  14. Marie-Laure Ryan Says:

    Sorry not to be able to respond responsively to all these posts, especially Noah’s, but I am spending 2 days home between trips that will take me away from e-mail and I am flooded with other things to do. So you won’t hear from me for a while.

    At any rate I have to say this:

    of all chapters of NaVR, the hypertext one was the most difficult to write, because so much has been doe about HT, and I found it very difficult to have something of my own to add. I may or may not have succeeded–but I rewrote the chapter several times and left a lot of the early drafts out. On the other hand I felt that I HAD to treat hypertext, or my book would be incomplete, because so many people associate digital narrative with hypertext.

    Second Noah asks about my forthcoming book (2004) “Narrative Across Media.” It’s nice that JHUP mentions it because they turned it down when it was still in outline form, and that’s why it ended up with Nebraska. For a table of content and short presentations you can go to my web site,

    http://lamar.colostate.edu/~pwryan, and follow some self-explanatory links (Marie-Laure, books, narrative across media.

  15. Julian Says:

    For those of you who have access to tekka.net, you can find my review of Narrative As Virtual Reality here.

    While I agree to the opinion voiced here that the book hasn’t received the attention it deserves, my review focuses much more on the first part of the book, rather than the second.

  16. vika Says:

    Julian-

    I’d love to read the review, but don’t have a subscription to Tekka. Is there another way to get it? I’m interested in the ways in which print narrative relates (and doesn’t directly relate) to electronic narrative, in historically important ways — although some would warn against over-historicizing. I haven’t seen NaVR yet, but from this discussion it seems the first part of the book might relate more to New Media Reader‘s historical emphasis than is obvious at first.

  17. miscellany is the largest category Says:
    Recent Trains of Thought
    I have become used to the various conductors’ voices on the train in the morning (although oddly, I don’t remember the ones in the evening). There are two primary conductors, it seems, that I’m likely to catch. One has a…

  18. Grand Text Auto » Satisfying Review Review on ebr Says:

    [...] s Narrative as Virtual Reality, which I’m very glad to see being further discussed. (Noah’s review on here has been one of too few texts to grapple wit [...]

  19. JoseAngel Says:

    On Marie-Laure (5):
    “I personally find it easier (and more productive) to spatialize the linear than to linearize the spatial in a meaningful way, because the linearity of the text holds it together in my mind, and allows me to wander back and forth along this line. To me, the maintenance of some kind of narrative, and consequently linear coherence is crucial to whether new media literature will remain an academic fad, or will reach the wider audience of the educated public.”

    In the case of that online literary genre known as blogs, spatial hypertext connections and a linear narrative may develop together… with the actual writing of the hypertext/blog providing the basis of the temporal linearity that Marie-Laure finds necessary as a backbone for narrative. Of course the readers of the blog may circulate through its links and texts in a number of ways, further structuring that narrativity through their reading and interpretive trajectories, and adding to the text’s hypertextual complexity as well. But the timeline of actual history, the writing of the blog, is inscribed into its structure and thereby provides a strong linear backbone, a sequential structure which must be taken into account by author and readers alike. Blogs of course challenge most vividly the notion of “narrative literature” as an already finished and pre-packaged product with a clear frame separating it from “life”, but… that is what makes them interesting.

Powered by WordPress