Janet Murray's response
Nick Montfort is a committed advocate of the genre that he calls Interactive Fiction (IF). He is helpful in clarifying its enduring charm and distinguishing attributes, and also in raising our awareness of the non-orthogonal nature of the many categories currently being brought to bear on storytelling and gaming in digital environments. Identifying genres and distinguishing interpretative categories is a crucial part of extending the coherence and expressiveness of the new medium, and Montfort makes a persuasive case for IF as its own digital genre with clearly identifiable attributes, with clear formal roots in the ancient story/game known as the riddle.
His identification of characteristics reinforces my own confidence in the helpfulness of looking at the digital medium as a single entity with many genres but with four chief and defining characteristics: the procedural, participatory, spatial, and encyclopedic affordances of the medium.
Montfort points out that it is the world-making quality of IF that distinguishes it from text in other media. World-creation is something that novelists and other storytellers do, and also something that some gamemakers do. Games derived from stories, such as Dungeons and Dragons, which is loosely derived from Tolkien's fantasy universe, combine the instantiation of detailed, richly imagined verbal description with the pleasure of enacting behaviors that actively create belief in the world. The world becomes more present when we can act in it. On the computer, the world gains an even stronger presence because it becomes instantiated in an artifact that has behavior. We can think of this artifact as a machine, as Nick suggests, but we experience it as interactors as a performative world, a world that reveals itself not only in words but in actions. Of course, in IF the actions are embodied in words, but they are actions nonetheless. The behavior of these objects comes from the procedural power of the computer; their responsiveness from the participatory affordance; the detailing of the worlds that reinforces our belief builds upon the encyclopedic quality of the digital medium.
To the extent that the world of the interactive fiction is described as a geographical place or a physical space, we do not merely read about it - we navigate it. The power of the original IF games was explicitly centered in the navigation of caves and dungeons, and the pleasure of such navigation was extended into the shared networked spaces on multi-user domains (MUDs). As I point out in chapter three of Hamlet on the Holodeck, the spatial property of the digital medium is quite independent of its multimedia affordances, because it is derived from the procedural and participatory properties of the medium. The space of Zork is real to us even though there are no pictures, because it is consistently scripted and therefore navigable by command. This power of the computer to create navigable space is one of its most expressive affordances, and text-based environments have provided some of the most magical experiences of this representational power.
IF was one the first and most energetic genres to be developed on the computer because of the fit between the enjoyment of a fantasy world and the development of the procedural, participatory, encyclopedic, and spatial qualities of a new medium. In one sense, the magical domain that IF explores is the inner space of the computer itself - the hackers that explored and elaborated upon the spell-rich dungeons also referred to themselves as "wizards" in their "real life" work of programming.
But today's magic is tomorrow's old technology. World-creation, especially spatially concrete world creation, poses a serious design problem as IF moves into increasingly sophisticated computational environments. In the 1970s when text-based adventure games were invented, and well into the 1980s which saw the flourishing of the form in the Infocom games, the standard computer interface was the command line. Hypertext was still a gleam in Ted Nelson's eye, word processing was in its infancy, and the graphical interface driven by Doug Engelbart's mouse was languishing in the invention orphanage of Xerox PARC waiting for Steve Jobs and his design team to come find it and adopt it.
Navigation through digital space has changed a great deal since 1980. Interactors are used to surfing the Web by mouse-clicks, arriving at Web "sites" and departing from them. The change from file transfer protocol, which users thought of as the "uploading" and "downloading" - or shipping and receiving - of containers of bits, to hypertext transfer protocol, which users experience as personally traveling from one place to another is, among other things, a change in dramatic expectations. Montfort's IF is different from hypertext, but in the online MOOs it often co-exists with hypertext. The navigation systems are then in conflict: do we go from room to room by typing "N" "S" "W" "E" as in Zork, or do we click on doorways in a picture of a room or on a compass rose? Does the screen change visually in answer to our commands or is place instantiated solely by text? Do we set up multiple channels of interaction, one for typing or clicking through directions, one for typing commands other than directions, one for conversing with virtual characters? Do we use the same line text box to type out these three very different modalities of interaction?
Ben Shneiderman, one of the founders of the discipline of human computer interaction, has articulated the core design principle of "direct manipulation," which is widely accepted as a desirable goal for structuring interaction. Typing "N" in order to move through a virtual space is far less direct and less manipulative than clicking on an image of the space or even an image of a compass. Is there a reason why the design criteria of other digital environments should not apply to IF? Is there a pleasure in indirect manipulation, in refusing to concretize the space? After all, the existence of an affordance should not dictate design. We choose between the panorama and the close-up in photography, designing by exclusion as well as inclusion. Is there a valuable trade-off in not allowing the interactor to move by mouse-click, or to click and drag a picture of coal into a picture of a diamond-making machine? Why not structure the IF world as a resource allocation simulation with pointing and clicking and dragging of objects? Why would we choose to keep the world in words and to keep the interaction verbal? Why privilege typing words over pointing to things, the keyboard over the mouse? Is the category of "Interactive Fiction," like the category of "literature," intrinsically verbal?
Montfort appropriately speaks of a "natural language" interface as one of the hallmarks of the form, but the use of commands drawn from natural language runs the risk of raising the interactor's expectations of what can be expressed, while severely limiting their actual expressiveness. The command-line, programming code interface can conflict with the literary aspirations of the author. In online MOOs it is common to see verbose descriptions of spaces, whose tone and length evoke bookishness if not literary merit, combined with the restricted code of the command line. These two very different modalities create a discord, which is further heightened if the interactor is engaged in conversation with a character within the story.
So IF has certain intrinsic design difficulties, a built-in awkwardness in the way it represents spatial navigation and the inconsistency with which it handles language. And yet it continues to draw devoted practitioners and interactors. It is, in Montfort's view, a still vibrant tradition.
Why does IF work despite these design difficulties? Perhaps the answer lies in its structure as a riddle. Riddles, unlike puzzles, are always verbal and are based on a conversational exchange. They are intrinsically interactive, and have a formal syntax, a variant of call-and-response structure. A riddle is a word-puzzle, framed as a conversation. The riddler poses a question that has to be reasoned out. But usually there is a surprise or misdirection involved. "What's black and white and red all over?" we asked in my childhood and children still ask. The answer: a newspaper. This riddle, like many others, only works in speech because we distinguish between the homonyms "red" and "read" in writing. In fact, it is a way of calling a child's attention to the category of homonym, or of helping a child to express her discomfort with such anomalies. Riddles are most popular with children who are just learning to read, or just learning how to reason things out, to hold more than one idea in mind at the same time. Riddles are like puns in the pleasure they take in the expressiveness of language. They are popular in Shakespeare, as are puns, perhaps because he was writing at a moment when an entire society was adjusting to the spread of literacy, the result of the maturity of the printing press and the form of the book. A riddle is a kind of debugging of our cognitive apparatus, calling attention to possible mistakes in verbal or logical decoding. Sometimes, as in the riddle of the sphinx or the graveyard riddles in Hamlet, they explicitly point to a larger human conundrum: the riddle of our consciousness and our mortality that underlies all gaming and all storytelling. Riddles are about the power and the limits of representation and communication in themselves.
IF is a riddle most of all because it is a conversation. It is not a conversation with an imaginary character, a chatterbot like Eliza, though it may include characters. It is a conversation with the author of the imaginary world, who is challenging the interactor to solve the puzzle, to figure out what the author has in mind, to debug their own interactive processes, repeating the sequences until the desired ending is reached. In the early online games there was no way of saving one's position or undoing moves. The space could be traversed at will, assuming there were not locked doors, but time was relentless and irreversible. As in a conversation with another person, you could not unring a bell; as in an obsessive or superstitious ritual, the only way to get it right was to do it in exactly the acceptable order, no matter how many repetitions it might take to get it right. An interactor learning an IF environment had to memorize the sequences (or record them on paper) and say them back in the right order to please the god of this magical world. Meanwhile, the author is taunting or encouraging the interactor, and in either case making clear his or her own cleverness. Like the poser of the riddle, the author of an interactive fiction exists only as a conversational partner. Like the person to whom a riddle is posed, the interactor is in a contest, drawn in by a desire to "match wits," with the riddle-poser, to test the operation of their own cognitive processes against the trickery of the master.
For my own taste, I prefer the riddling of a murder mystery to a dungeons and dragons plot, and so the only interactive fiction that I have found appealing was Deadline, an old Infocom game.
Would I play it again in command line form? Probably not: I'd rather play a game like The Last Express, which sets up its puzzle in sound and images. I find the voice of the narrator somewhat irritating, as I now find the once delightful voice of Henry Fielding in Tom Jones. Like that eighteenth century novel, IF stands at the beginning of a narrative genre, and its emphasis on the narrator is perhaps in part a sign of its self-consciousness.
But would I read anything that Nick Montfort has to say about the qualities and pleasures of IF? Absolutely.