Understanding Video Games text-book
The dragon in the attic - on the limits of interactive fiction

Date posted: May 16, 2006
Updated: Oct 23, 2006

By Jonas Heide Smith (smith@game-research.com)
Published: July 2000

Computerized interactive fiction is at a dead end. It builds upon principles that are teeth-grindingly linear and so cannot compete artistically with other media - such as the time-honed novel. To help solve this problem, this article attempts to sketch the potentials of interactive fiction.

It really seemed quite straightforward. Slightly more than thirty years ago, programmer Will Crowther tumbled with a small program, he thought of as ‘Adventure’. It was really nothing special - most of all it was a combination of Crowther’s greatest interests: programming, cave exploration and fantasy role-playing. And mostly, it was aimed at his children.
If one combined these ingredients with the data processing powers of contemporary computers, the result became a cave exploration simulator, which only communicated with the user through text.

It was this primitive construction, which Crowther (who was actually somewhat busy laying the foundations for something as trivial as the Internet) bestowed on the public by letting it lose on the computer networks of the time. Now, this gesture was accompanied by neither trumpets nor drums but in fact Crowther had paved the way for rather large sums of money; those that would be made on so-called interactive fiction.
Obviously, Crowther should not be held responsible for the subsequent development of the business. On the other hand, there is a clear connection between his choices when coding ‘Adventure’ and later design principles. In fact, the entire business has a substantial need to break free from its history.

The story of a genre

To put something behind us we have to understand it. In this spirit, let’s take a look at what happened next.
‘Adventure’ was the cause of much rejoicing on what was generally known as Arpanet. Through this channel, it reached a group of innovative MIT students in the late 1970s. At this hotspot of technological revelations what was only dawning upon the general public had long been clear: The computer was not just a cold-blooded number cruncher. It could even be thought of as a medium. With this in mind, the group of students, who soon founded the company Infocom, constructed a larger, more literally ambitious ‘Adventure’. The named their child ‘Zork’.

As was the case in ‘Adventure’ interaction with ‘Zork’ consisted solely of text. ‘Zork’, however, was so cunningly construed that its linguistic interpretation - the so-called parser - was able to secure the flow of the narrative, even though it really surrendered in the face of even the simplest grammatical constructions.

The player might, for instance, type ‘Hit mailbox with hand’. To this the game would respond ‘I’ve known strange people, but fighting a small mailbox?”.
Even though the game does not have a clue as to what the player is trying to do, the parser perceives that the user is attempting to hit something (X) with something (Y) and responds in a way that manages to push the user back inside the framework of highly limited options. The response serves to point to the narrative logic of the genre, thus perhaps making the user behave in a more fantasy-like manner. ‘Zork’ was a huge success.

At a later time the concept was improved with complementary graphics by Sierra’s ‘Mystery House’ (1981). ‘Mystery House’ was rather displeasing to the eye and primarily used its graphics as complementary decoration to the textual interaction. A few years later, however, graphics came ‘alive’ in the sense that players could move their characters around on the screen and interact with nearby objects. A highlight of sophistication was reached at points where the graphics themselves had a narrative function that was not just repeated by the text.

But did you really need text? Creative people at LucasFilm’s game department LucasArts found the classical way to be a hindrance to the narrative experience. Flow was broken when the player was forced to guess exactly what words were known to the parser. With ‘Maniac Mansion’ (1987) the verbal guessing game was over. By use of a so-called point-and-click interface the user (using joystick or mouse) could combine a range of verbs with graphical elements of the game. For example, one could press ‘open’ and then click on a door. The reader will be able to guess what happened then, and that was exactly the point - it was logical and straight-forward.

The new interface worked so well that - with certain modifications - it is still in use. The most significant development after ‘Maniac Mansion’ occurred when, in the early 1990s, the CD-ROM storage medium allowed for the use of much greater amounts of data. This lead to an audiovisual explosion and a merciless wave of digitized film. However, as the dust settled many were left with a bitter taste in their mouths. The feeling that the niceties were a cover for a lack of narrative innovation was widespread.
Myst (Brīderbund, 1995)

More was needed, and more was achieved with the meditative ‘Myst’ (Brīderbund, 1995) that reached out to the hearts of book lovers with its literary polish and focus on exploration. Although ‘Myst’ is among the best selling games ever, the second half of the decade saw the adventure genre declining. While ‘Doom’ (ID Soft, 1994) and its brethren blasted their way past most competition, strategy games invaded most networks with their obvious multiplayer potential. The adventure genre - that is, the sort of game that values ’stories’ - was headed for an early funeral.

To be fair, at late Sierra and others have managed to make impressive use of developments within 3D graphics, but it seems as if the genre, when it comes to formal development, has hit a solid wall.
There is a very good reason for this.

The vice of linearity

Despite talk from game designers, writers, and even movie directors of the blessings of interactivity, many of them are still trapped in the past. They write books with a few options at the bottom of selected pages and they make games that are comparable to movies, which sometimes pause until the viewer has solved a crossword puzzle. If this is what one has to offer it isn’t odd that many book fans have little but scorn for the narrative potential of the computer. But it’s all due to a misunderstanding.

If someone should be blamed for the confusion, we can pounce on poor Will Crowther. ‘Adventure’ had a beginning and an end and that simply will not do. If the player is to be led towards a specific ending in a way that makes logical/causal sense it can only happen at the expense of freedom.

An example: A man enters a store and buys a newspaper. This less than epic tale can be told in a broad range of ways. But if it is to be ‘told’ interactively, the choice of protagonist (a man), a flow (the newspaper will be bought) and the ambition of the protagonist (to buy a newspaper) are predetermined.
In other words, there a not many interesting choices left to be made and the situation is likely to be of little interest to the user. On the other hand, an expert of expression, or ‘artist’, will be able to tell the story in a fascinating way in his or her medium. And this is a strong argument against interactive fiction - the user is not an expert. To be entirely blunt, the user is often a complete amateur and furthermore often does not have the time it takes (say, for a dedicated movie editor) to present the material in an optimal way.

On its own premises this argument is rock solid. But in another sense it is unimportant as it presumes that the user must be motivated to make the very same choices that the artist would have made.
And if this is our ambition the whole thing is merely a rather crude transference of old principles to a new medium.
Figure 1

The user as problem

Traditionally then, interactive fiction has been conceptualised as narratives with limited freedom of action - of the ‘Do you want to go left or right, first?” kind.
Drawing upon the writings of Danish game designer Michael Valeur this way of thinking can be illustrated in Figure 1.
I ‘chapter 1′ the user has a certain space of freedom but to proceed to ‘chapter 2′ he or she must submit to the direction of the plot and accept a significant reduction of freedom when faced with what might be called the ‘plot points’. It would, for instance, be no good if the user kills the bad guy in chapter 1 since there would then be nothing else to do. Valeur himself has compared the model to an apartment. To get to the kitchen one must go through the hall and so on.

Although this is a very precise analogy to traditional adventure games, it is a quite odd construction. Designers working within this model are often struck by terror, since consequences of user choices will multiply indefinitely - by geometric progression - unless powerful restrictions are applied. These designers must lure the player into believing that he or she is free and then waste considerable effort in maintaining this illusion.
But what would happen if one fulfilled the ambition of interactivity and set the player free?

The user as resource

That question should not be put to a writer. It should be put to an architect. The architect constructs rooms, steps back and lets people use them. The building represents the outer limits of activity while laws of nature see to it that no one starts walking on the ceiling. What is worth noticing is that few of us are continuously annoyed that we cannot defy these laws. We have no problem accepting limits to action. What may, however, be harder to deal with is that which is a mainstay of the adventure game: The door cannot be opened, because it is not meant to be used - yet.
Figure 2

Let us imagine an alternative model. Instead of demanding that the user go through a predetermined process of recognition, the setup could be something like this: The player starts in the castle vestibule and in the attic we place a dragon. The behaviour of the dragon is controlled by a number of variables, e.g. joy/sadness, aggressivity/passivity, courage/fear. If dragon and player meet, the dragon interprets any action of the user according to these variables, set by the designer. If the player comes within a certain range of the dragon, its aggression level determines what it will do and so on.

Even with a highly limited set of variables, it will be impossible for the designer to predict the process. On the other hand, fear of the tree of infinite consequences is gone - the system adjusts itself and while a large number of occurrences are possible, none are required. Because there is no goal.
We have here a sort of ‘poetics of the starting point’, loosely illustrated in Figure 2. It is a structure-less model, or to be frank: It is a frame. But the frame has a pattern, and this is important.

Deistic narration
So-called ‘deism’ is (among other things) the idea that God set up the frame and the rules and then left it to human beings to take the initiative. In all modesty, this presents a possible way for the designer to work. The frame is not only the laws that apply to the universe of the designer. It is also all the genre codes and cultural references that are incorporated into the expression and thus cue the user to assume a certain ‘role’. If, for instance, the designer constructs a cold and apocalyptically futuristic universe, film noir cues might be strategically applied to encourage the user to assume roles that are compatible with the universe (such as the role of hard-boiled private eye rather than the role of fearless knight).
Similarly, the architect’s act of building construction to some degree determines (or shapes) what the building is likely to be used for. The idea of deistic narration, then, is basically that stories don’t need to be something you have to begin with. They can just as well be something you get.

Theoretical dead ends
That last sentence, however, is the source of some controversy. While adventure games unwaveringly followed in the limiting footsteps of ‘Adventure’ several scholars - often from comparative literature - discovered the phenomenon. Apparently, their humanist background has been a problem.

Writer and game designer Michael Valeur sensibly writes: “Freedom and seduction point in fundamentally different directions”. It is hard to disagree, since seduction is usually seen as the passive surrender to fictitious premises - it is hard to be seduced with mouse in hand. But this is not exactly what Valeur means. Valeur, and I hold that this is a problem, equates narrative and seduction: “When you tell a story, when you seduce people, you need to lure them by the use of linear dynamics.” Valeur’s ‘you’ is the writer - not the user. With such statements, he rather pointedly disregards the possibility of (interesting) stories arising from appropriate narrative premises. And although his analysis is highly enlightening on many points, on this topic I believe he misses the mark.

And he is not alone. Espen Aarseth, Norwegian literary scholar, takes the subject quite seriously but still refuses to lay aside the axiom that adventure games naturally have to tell stories. He does, however, claim that there is a fundamental gap between narrative and interactivity (as does Valeur). Again, we have the fully understandable - but rather unimportant - argument that it is hard to plan a process and grant the user huge influence on this process at the same time.

It seems as if these theories build on shaky premises. They seek to define interactive fiction in accordance with a concept of narrative that is too loose and traditionally has only been used to describe linear narratives. In a way, they measure interactive fiction to Aristotle’s definition that narratives have a beginning, a middle, and an end. This is merely the act of definiting interactivity as the opposite category. The mistaken premise is that definitions, such as the ones we are talking about here, exist in pre-linguistic reality and can be tracked down. Coincidentally, Aristotle following Plato, held the very same view, believing that all things have essential qualities - that things are naturally divided into categories for us to discover.
This is wrong. Analytic definitions are useful abstractions that make complex things manageable. In the beginning of this text, I could have defined interactive fiction as pantomime. It would have been rather silly but it would not have been wrong in any absolute sense. The intense hunt for the ‘nature’ of the adventure game is misguided.

The art of world building
Perhaps it was the commercial trouble of the adventure genre that sparked creative reflections in a number of designers. The revolution, if you will, began in 1997 with Origin’s Ultima Online (which, of course, had important ancestors). The game, and here the term begins to be inappropriate, is a world filled with a range of narrative vectors. All citizens of this world - whether controlled by a person or a computer - has skills and preferences. No one can entirely predict the future of this world since every choice has consequences and the number of choices made every day by thousands of players is overwhelming.
The frame was constructed with obvious inspiration from traditional fantasy worlds: Elves, dwarfs, and orcs are commonplace. But the choice of genre is just a consequence of conservatism and designer demographics. It is not hard to imagine digital worlds build around other ages or upon genres such as crime fiction, horror or even more lyrical alternatives.
Nostalgic fans of the adventure genre may cry their salty tears. They feel that their favourite genre is passing away. If they would raise their heads, they would see that the future is a bright place full of narratives.
Further reading

* Aarseth, Espen. Cybertext - Perspectives on Ergodic Literature. The Johns Hopkins University Press. London. 1997.
* Murray, Janet H. Hamlet on the Holodeck - The Future of Narrative in Cyberspace. The MIT Press. Cambridge. 1997.
* Valeur, Michael. Blackout - erfaringer omkring arbejdet med interaktiv manuskriptskrivning. Speciale, RUC. 1998.

[A slightly different Danish version appeared in SAMSON, July 2000]

Share and Enjoy:These icons link to social bookmarking sites where readers can share and discover new web pages.
  • del.icio.us
  • digg
  • Shadows

RSS feed for this page
since June 2007

RSS feed | Trackback URI

Comments »

No comments yet.

Name (required)
E-mail (required - never shown publicly)
Your Comment (smaller size | larger size)
You may use <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong> in your comment.