©1997 Josephine Berry & Micz Flor

After seeing two scenes between soap opera lovers, neighbours or market place hustlers, you can easily anticipate the last five minutes of any episode. Why is it that it only ever takes two episodes of a soap opera to become initiated?

Looking through the TV Times one day you come across the matching diagram - a complicated but nevertheless illuminating two-dimensional set of love and hate relationships between the characters of a soap. It reveals if they are together, if she wants to but he doesn't, if he hates so and so because he knows that...., it is an engine for generating continual problems and potential conflicts, for then, now and in the future.

At this point each episode of a soap opera, produced for 25 minutes of TV time, becomes an almost predictable effect of the underlying structure. The soap has two axes: one is the linear daily unfolding of events in real TV time, and the other is the lateral static network of possible interactions, which looks like - you guessed it - a hypertext structure. This structure, which also resembles a flow chart, is one of the reasons for the existence (of the myth?) of computer generated soap opera dialogues or narratives. Hitting further below the belt, you might even argue that this could also be assumed of many Hollywood mainstream productions.

*snowfields* is a soap opera that has found its natural home on the internet, where the chronotope of each episode can open to reveal the spidery narrative network that lurks beneath. As the use and focus of the internet rapidly moves away from the elite network of academic thinking and into the abyss of commercial and mainstream trash culture it hosts an increasing number of soaps - and rightly so, where else should they find the most appropriate habitat?

What is *snowfields*

*snowfields* at first seems to be a static image of a map with a familiar ring to it - a somewhat truncated image of a city which seems to have grown over centuries and is not based on a grid pattern as many modern and/or American cities do - it looks European. It is a city almost without information, without names, only offering the 5 kilometers as a unit of measurment in the bottom left hand corner. We chose East Berlin because it's both real (a grown city with rivers, trains, living areas), and ghostly, for it has ceased to exist in this former guise.

The map is active. By clicking on it you enter the text based part of *snowfields*. What follows is a small image of the area you clicked on with an accompanying text. This text tells you a story about something happening, people interacting, or a situation into which you're thrown. It might only indicate the time of day, or it could deliver a fully fledged dialogue. This is the soap opera, this is what's happening in *snowfields*.

The text is fully editable. It comes in a textbox (like many internet subscribe or e-mail pages) and has a 'send' button underneath. Having edited and/or read the text you can send it out - to us - and we will archive it as a part of the reality of what has happened in *snowfields*.

Selecting the same area of the map again only moments later will produce a similar but nonetheless altered version of events there. The characters might be the same but their interaction will have changed. In a more abstract sense, what you see is not what you get, but rather only one permutation of all possible combinations of narrative 'bites'. When you click on the map your computer loads information from the web, and your browser composes an episode on your very desktop.

Each episode of *snowfields* is a text written in HTML and JavaScript, containing not only one but many possible stories. These stories are numerous but not unlimited. The characters' interactions are random, but nevertheless determined by the code. Those determinations themselves are nothing but a proposition, entertaining, possibly thought provoking, but yet fully editable and ephemeral - depending on whether or not the delete button is employed. This situation offers three different texts: the one read by the browser, the one read by the reader and finally the one which the reader writes themselves.

How do automation and text come together?

One of the best known examples (or mind games) of text, meaning and automation is Jorge Luis Borges short story 'the library of Babel', written early in the 20th century. In this story a group of librarians are maintaining a huge library which consists of a seemingly unlimited number of books which are filled with all possible combinations of the Latin letters on 400 pages, 40 lines per page and 40 characters each line. The astounding phenomenon arising from the possibility of such a library is the idea that on the one hand each book is unique, whist on the other hand most of them are meaningless heaps of letters, numerous versions of the same book with tiny spelling mistakes, translations of the same text in many different languages and so on.

The seemingly infinite number of books is actually a predictable number. Following the very basic mathematics of probability (by merely multiplying the factors) describes the size of the library of Babel. Foucault used this story as a vivid example for his contention that the task of literature is to hold up a mirror to language from within, and reveal the fact that it is a property of language itself to spin out endless permutations from its rudimentary elements. However, these are not endless and that is the fascinating fact.

In order to play with this probabilistic mind game you need a limited number of units and a constant number of locations in which to place them to start from. Borges decided to use the Latin alphabet (subsequently barring any Arabic or Chinese texts from his library - however, numerous translations of such texts still exist within the equation). You might narrow this down further and decide to only use the vocabulary of a certain language. That way you can cut down the amount of noise and chaos and dramatically cut down the number of possible combinations. The process of using already meaningful clusters of smaller units (i.e. letters) to create bigger units (i.e. words) is to create order out of chaos. *snowfields* works with such meaningful clusters on a much higher levels and uses clusters of English words which are grammatically correct and inherently comprise meaning.

The JavaScript in *snowfields* seems to randomly assemble text from these clusters and put them on the screen. However, this randomising effect is contingent on data such as the time of reading, month, year, and the age of the computer with which it's being read.

Authors and Automation

Selecting and composing elements from the multitude of possibilities that surround us and feeding these clusters into the programme is what we - the authors - do. Assembling (or composing?) a story from these possibilities is the work of automation within *snowfields*. Authorial control is limited.

Composition is replaced by automation, providing a tension between elements which is sometimes far more surprising and interesting than the predetermined clusters (i.e.sentences) of the authors. These frictions between two elements, randomly thrown together by the computer demand the explanation of the reader. Given the expectation of a coherent and well composed text on the web page, at this point the process of reading becomes more active and imaginative - the break, the in-between requires explanation. This even works for us, the authors of *snowfields* - it can be surprising, funny, strange to see what is merely one of many possibilities that we have input ourselves. Running the programme turns us into readers and we know as little what to expect as anybody else.

Automated Scripts

But automation of narrative clusters does not just make the familiar strange, it also highlights an inherent quality within narrative itself. The surprising ease with which a 'dumb' computer programme can create sense out of these elements speaks volumes about the already automated nature of narrative. Like in the example of the tragedy in which the audience follows the fluctuations between harmony and conflict almost by heart, narratives adhere to a number of formulae developed and honed over generations of cultural activity. Following this line of argumentation it is possible to say that many systems and codes within narrative are automatic - a word which is not intended to be negative here. In this sense, the soap opera in its decentralised, non-linear structure (as described above) could even be perceived as a consequent and contemporary development of the dramatic tradition as we know it in the context of a Eurocentric tradition.

Even though the attributes of the characters within a narrative provide points of identification and opposition for the reader, it is mainly the relationship between the characters which carry the plot and provide the ongoing momentum. So the attributes and the intention of the individual (described by Max Weber as the minima socialia) set the agenda and build the potential, while the friction between characters, the clash of intentions, the incisions set it into motion.

The decentralised and multifaceted structure of relationships and possible interactions between characters makes the soap opera an appropriate test field for introducing the apparatus, the machine into the narrative field. Furthermore, the inclusion of an automised device does not only mimic the narrative scripts (as described above), but it also provides clashes and connections which have the potential to spin the narrative outside of the field of expectations.

Criticism of Hypertext

The much vaunted inclusion of the reader as writer within hypertext is more often than not an unfulfilled promise. The distinction made by Barthes in 'S/Z' between the 'readerly' and the 'writerly' text has become the key note of this discourse. When Barthes speaks of the pleasures of writing, this is meant as the process of creating a text and not merely rearranging fragments within it. To this extent the internet and its inherent language, HTML, are in fact very poor tools with which to create such a writerly text. There is not much to write in the www apart from filling in your e-mail and VISA card number when ordering software. 'Creating your own text' has become a euphemism for plotting your own route through a set of limited (albeit potentially unlimited) components.

The inclusion of the automised device within *snowfields* side-steps the painful issue of authorial coherence. With automaton not author acting as composer of text, a situation is able to occur in which the text can develop in a state of aesthetic and rational anaesthesia. (The text having reached such a point is displayed in a fully editable text area on the screen and can be submitted by pressing a button.) The reader/writer submits their text into an intellectually unprotected environment. This pervasive ambivalence towards the ever changing outcome of the narrative releases the reader/writer from the responsibility for its final meaning. On the other hand, the realisation that their addition will play a significant, if impermanent role in the soap's development creates a meaningful context for participation. The individual imagination has a consequence and result which affects the text's development in an unerasable sense. Using this combined method of automation and composition, the reader not only experiences the pleasure of creation but also in reading their own creative output in a constantly defamiliarised manifestation.

The *snowfields* project is always editable. Following the notion of computer-based writing as working on one of many texts ( displayed on the screen and infrequently stored on the disk) which can easily be altered, changed, reverted and moved around, even the archive is editable and by returning an altered version of a previously submitted text the reader/writer creates a new layer. Such layers produce a multidirectional network of narratives.

*to be continued*

Josephine Berry & Micz Flor

Berlin, 05.Sep.97

[x] top of page