User:Thijshijsijsjss/Gossamery/The New Media Reader/Nonlinearity and Literary Theory

• Read on: July - August 2024
• Read it physically in the studio, or on bootleg library, or with my annotations.

This 1994 essay by Espen J. Aarseth outlines a theory of nonlinear texts and places this in the discourse of contemporary literary theory. I stumbled upon it in The New Media Reader, after reading the preceding chapters 12 and 50. By the merit of its placement in the book, that follows a chronological timeline, the alluring title, and Aarseth's name being connected to ludology, my interest was piqued. It is a dense text. Therefore, this entry will start with an attempt of a (subjective) (digestive) summary, after which more personal reflections are explored.

Follow ups to explore:

• Cybertext: Perspectives on Ergodic Literature (Espen J. Aarseth) (monoskop link)

Summary of 'Nonlinearity and Literary Theory'

Introduction
A nonlinear text is an object of verbal communication that is not simply one fixed sequence of letters, wors and sentences, but one in which the words or sequence of words maay differ from reading to reading because of the shape, conventions or mechanisms of the text. In particular, this refers to the form of the text, not to any fictional meaning they might hold. For example:  nonlinearity and nonchronologicality are not the same.

I want to bring forward a passage from the next section, as it reflects on this definiction and introduces terms fundamental to this essay. Aarseth notes that his use of the word "text" might clash with some textual theory that regard the text as a semantic network of symbolic relations, loosely attached to the notion of the literary work. He states having no intention to challenge this, but instead wants to shed light on two perspectives on text.  These aspects follow different rules, but are interdependent.
1) Informative: the text as a technical, historical and social object. This brings rise to a fundemental characteristic of a text: in addition to the visible words and spaces (which we might call "script"), a text includes a practise, a structure or ritual of use. We can compare this to the concept of a genre, except that genre is seen prior to the text, contrary to this choreographic practise.
2) Interpretable: the text as is individually received and understood. Here, Aarseth highlights a "flaw" in this definition: it is open to interference from the interpretable level. What if the text insists on its nonlinearity? Or what if it gives instructions and cancels them later on? Or tells you to skip over some instructions, for they might be contaminated by subversive directions? The linear text can flirt with nonlinearity, but the nonlinear text cannot lie and pretend to be linear.

Behind the Lines: What is a Text, Anyway?
We must rethink the concept op textuality to allow for both linear and non-linear texts. "The text" brings about a set of metaphysics, the 3 main ones being:
1) Reading: a text is what you read, the words and phrases you see and the meanings they produce in your head
2) Writing: a text is a message, imbued with the values and intentoins of a specific writer / genre / culture
3) A text is a fixed sequence of constituents that cannot change, although its interpretations might
Aarseth opposes these notions, arguing that nonlinear literature shows us a textuality different from our readings, more fundamental than our messages and, through evolving rituals and technologies or use and distribution, subject to many types of change. The remainder of the section is dedicated to a sequence of questions, examples and ontological edge cases to provoke examination of the question of a text's metaphysics, in an effort to give a concluding definition that allows us to include nonlinear texts.

What elements belong to a text? For example: does our interpretation not start already before the first word of the script, with the author's name? Historically, author's have been aware of this, play with it (perhaps out of necessity), e.g. with pseudonymns. This suggests: the name belongs to the text, the person of the writer does not.

This begs the quesion: in what sense can there exist a script independent of text? Imagine a book from which the middle 10 pages are 'missing'. We are quick to assume the existence of some 'ideal book', deeming it more real than our copy. We do this out of disrespect for he copy, as it seems to misrepresent the 'real' text. Even though we observe a stable version, we prefer the imagined integrity of the metaphysical object. But what if the 'flaws' interfere so deeply with our sense of reception that they 'steal the show'? Aarseth makes an argument here that I do not follow, paradoxically concluding that some texts are metaphysical, some are not. Another option, he states, is to abandon the concept of the transcedental text-behind-the-text altogether.

We are presented another example (Zardos ***), one in which acts of a film are, unbeknownst to the audience, shown in an order different from the creators' intentions. Again, we can ask ourself if a new film is created. Aarseth answers: no, because they are both based on the same material potential. In this sense, a text is a limited language in which all the parts are known, but the full potential of their combination is not. There are many scales of change in a text's metamorphosis, some of which we might accept as authentic new works, others not (according to the cultural legitimacy of their method of construction, to new policital symbolism, ...). Textual integrity is a cultural construct. More importantly: so is our notion of what constitutes the text itself. A text contains within it its potential interpretability, too. In fact, a text is not what we may read out of it, nor is it what someone once wrote into it. It is a potential that can be realized only partially and only through its script. Furthermore, texts are like electrons: they can never be experienced directly, only by signs of behavior. They are processes impossible to terminate and reduce. This perspective allows us to include nonlinear texts, that often have no author or reader in the traditional sense.

A Typology of Nonlinear Textuality
Aarseth points out that this effort of formalizing nonlinearity is grounded in mathematics, in particular: topology, the study of properties of a geometric object that are perserved under continuous deformation. An attempt at a textonomical analogy is presented: the study of ways in which the various sections of text are connected, disregarding the physical properties of the channel (paper, etc). (Note that by having such a direct transposition from geometry to textonomy, the author claims to preserve formalism. I disagree.)

Textual topology describes the formal structures that govern the sequence and accessibility of the script. For a text to be described topologically, it must consist of smaller unit and connections between them. The units are called textons. An unbroken sequence of textons as they are projected by the text is a scripton. The connections are traversal functions, of which there exists an infinite set (**). The author makes an attempt to describe some variates that together define a multi-dimensional space into which the functions can be plotted. The variates are as follows:
* Topology: distinguished between linear and nonlinear texts, where nonlinear texts are those works that do not present scriptons in a fixed sequence (whether temporal or spatial). INstead, through cybernetic agency, an arbitrary sequence emerges.
* Dynamics: there are texts that are static (with constant scriptons) and dynamic (the contents of scriptons may change).
* Determinability: a text can be determinate (adjacent scriptons are always the the same) and indeterminate (otherwise).
* Transiency: a text can be transcient (the passing of a user's time causes scriptons to appear), or intransient (otherwise). If the transciency is real time, it is synchronous. If it's arbitrary, it is asynchronous.
* Maneauverability: describes the agency of traversal. Random access < links with explicit access < hidden links < conditional or complex links < completely controlled access.
* User-functionality: besides the interprative function of thte usuer, the use of nonlinear texts might be described by 4 feedback functions:
    * explorative function: in which the user decides which 'path' to take
    * role playing function: in which the user assumes strategic responsibility for a 'character' in a 'world' 
    * configurative function: in which textons and / or traversal functions are in part chosen and / or designed by the user
    * poetic function: in which the user's actions, dialogue, or design are aesthetically motivated 
(Curiously, these variates are described not as continuous, as the previous claim of infinite functions (**) would lead to believe)
 
The introduced traversal space is illustrated with an example: Agrippa: A Book of the Dead, a story project at a set pace that would encrypt itself after one showing, rendering it unreadable. If is linearity flirting with nonlinearity through the difference between its used and unused copies. It exposes the inherent instability of the metaphysical concept of 'the text itself': it becomes non-linear only if we choose to accept the 'text-behnid-the-text' as more real thatn the physical copy.

Finally, 4 categories / degrees of nonlinearity are presented:
1) the simple nonlinear text, whose textons are totally static, open and explorable by the user.
2) the discontinuous nonlinear text (hypertext) which may be traversed by 'jumps' (explicit links) between textons.
3) the determinate 'cybertext', in which the behaviour of textons is predictable but conditional and with the element of role playing.
4) the indeterminate cybertext in which textons are dynamic and unpredictable.
The following sections are a discussion of each of these categories.

The Readerless Text
(This section is concerned with the simple, non-linear text)
Perhaps the simplest way to achieve nonlinearity is with 'forking paths'. A decision between paths, entirely up to the user, produces an ambiguity different from a phrase's usual poetic (double) meaning, because there seem to be two different versions, neither of which can exist alongside the other, and both obviously different from the text itself. Forking scripts are like optical illusions, we can first imagine one, then another, but never both at the same time. When one looks at the whole of a nonlinear text, one cannot read it. And when we read it, one cannot see the whole text. Somethin has come between us and the text, and that is ourselves trying to read. This self-consioucness forces us to take responsibility for what we read and to accept that it can bever be the text itself.

The presence of the reader is further investigated with two examples of ancient nonlinear texts: wall paintings, read in accordance with the architectural layout of a temple, and Book of Changes. These seem to reject the presence of a traditional reader. Active or passive, the reader is always assumed to be a receiver, constructing meaning only. It can be argued that the same holds true for nonlinear texts: once the novelty has faded, it's back to layed back consuming. However, this counterpoint ignores that (beyond trivial) understanding of a nonlinear text can never be a consummate understanding, because the realization of its script (and not just its meaning) belongs to the individual user, who is acutely aware of their own constructive participation. As alluded to in the paragraph above, the object can only be glimpsed and guessed at. It is this frustrating attempt to harmonize contradictory scriptons from the same text that define the ritual of a nonlinear text. Importantly, a ritual might relief the user of not reading every combinatorial possibility.

Similar to the question of the reader, there is a question of authorship for combinatorial products of a nonlinear text. Queneau's Cent Mille Milliears de Poemes is compared this to the earlier example of Zardos (***). Aarseth concludes that in the latter's case, he rejected his reading as it told him he was not a real reader, since he was not reading the real text. This shock of finding yourself not to be a reader can, Aarseth argues, only (and only accidentally) happen with linear texts, because that is the only text in which the metaphysics of a real reader has any credibility and the only text in which the reader can exist as a reducible, accountable figure.

Hypertext Is Not What You (May) Think
(This section is concerned with discontinuous nonlinear text: hypertext)
Hypertext is a direct connection rom one position in a text to another. This text is only concerned with hypertext as a this general concept, not two other interpretations that make this a noisy field of study: specific implementations as hypertext systems, or texts within such systems. 

Hypertext's main feature is discontinuity. To ease the situation, they often offer additional features (overviews, index views, ...). As we would still consider 'the same' text without such add-ons a hypertext, and as they tend to fundamentally impact the reading ritual, it is hard to consider the work with and without the add-ons as 'the same'.

The remainer of the section is dedicated to discussing the example Afternoon, a story and some more frustrations the author seems to have with misconceptions in hypertext literary discourse. He makes a note, also, that in his mind paper and digital are not two axes in this model, its just that there happen to be many more electronic hypertexts.

Death and Cybernetics in the Ever-ending Text
(This section is concerned with determinate cybertext)
"A cybertext is a self-chaning text in which scriptons and traversal functions are controlledby an immanent cybernetic agent, either mechanical or human." Two historic examples are the root of computer-mediated cybertexts: Eliza and Adventure. The former was an early AI 'chatbot' playing a psychotherapist. The latter is a text adventure that went on to inspire the likes of Zork, in which the user controls a 'you', role-playing through a narrative. Aarseth notes that death in the cybertext is curious. Instead of signifying closure, 'cyberdeath' instead signifies a sort of reincarnation that implies a new beginning. This clearly distinguishes separate entities of 'user', 'main character' and 'narratee'.

There are two outcomes of a cybertext:
* Either it is successful ('the user wins'). This isn't as satisfying as a traditional text, as the 'you' is left behind and the same build-up and release of tension cannot be provided. It leads to peripety, not catharsis.
* Or it is unsuccessful (the game is not solved). Again, the you is abandoned.
Aarseth concludes that if the absent structure of narrative is the key problem in literary hypertext, in determinate cybertext the abest structure is the plot. In cybertext, the plot is not determined by the story, but instead is hidden itself, to be determined by user actions. Then found ,it does not produce anything interesting.

"The Lingo of the Cable": Travels in Cybertextuality
(This sections is concerned with indeterminate cybertext)
Indeterminate cybertext should be seen as a movement not against, but beyond genre.

We are presented with a brief summary of communication hyistory, from the telegraph to telex to mainframe computers to digitally mediated network chatting to the first MUD. Next to being a game, this was a cyberspace to enjoy complete anonymous freedom independent of your physical self. TinyMUD then allowed users to even add to the topology of the game. This is described as very anarchic.

MUDs seem to defy literary theory. Every user has a different or partial POV and the users bombard each other with temporal textons. However different this may seem from traditional novels, Aarseth assures us this is literature: 'letters, wordes and sentences are selected, arranged and dissemated to delight, impress or enrage an unknown audience.' Questions of authors and readers are also defied by MUDs. MUDs cannot be read, only (partially) experienced. Similarly, there's no knowing if a user's additions will be seen by anyone at all. 

The Limits of Fiction
Both determinate and indeterminate cybertext bring to light an important issue with ontological categories of textuality. As they are more empirical, they demand a new category -- though they rely on belief, it is not the same belief or suspension of disbelief applied to traditional novels. This new category might be called 'simulation'.

A user of fiction is responsible for constructing a mental image of the described text, but the text remains in control: the text can contradict this image at any moment. The text can make any claims that will then be true in the fiction, there's no one to contradict. From the user's POV, fiction is neither logical nor illogical. However, in some of these cybertextual examples, responsibility for coherence is shared between the text and its user(s).

Cybertext are between reality and fiction. Unlike fiction, which presents something else, cyubertext represents something beyond themselves. They a re not obliged to represent reality, but they do have an empirical logic of their own, so they shouldn't be called fictions.

The Rhetoric of Nonlinearity
We return to the question 'what kind of literary phenomenon is nonlinearity'. It is noted that it cannot be a genre (there's examples of fiction, poetry, encyclopediae, etc.). Nor can it be a rhetoric trope, as it works on the level of words, not meaning. Also, semiotically, every text is linear, as it considers a text simply a chain of signs. Another reference is brought up: Pierre Fontanier's taxonomy of tropes. We are able to place the figures of nonlinearity in his described classes. 

The corruption of the Critic
This section discusses the effect the 'ontological embaressment' has on the literary critic: if they can no longer read, how can they be a critic?

Problems of "Textual Anthropology"
The concluding section. After the death of the author, of the work and of the reading, this essay illustrates the death of the text. We are left with linear and nonlinear textualities. According to Aarseth, this change (an 'empirical revoliuton)' allows for an anthropological approach with the object of study a process (changing text) rather than a project (static text). He stresses that any theorist or critic that does study indeterminate cybtertext should do so not as an a posteriori investigator: they should be there when it happens. They must make it happen.

The next question then becomes what separates this literary anthropology from (becoming a sub-discipline of) social anthropology? Aarseth notes that these fields have always had some goals in common. For fear of their seeming convergence becoming to fuzzy, he then gives a few conflicting examples.

Finally, a summary much unlike this one is given.

(click for a pdf with my annotations)