Original Publication Hyper/Text/Weory, 51-86. Ed. George Landow. Baltimore: Johns Hopkins Universib Press. 1994.

Nonlinearity and Literary Theory Espen J. Aarseth

Electronic miting d require a simpler, more pusitive literary theo'y. -J. David Bolter

The future can only be anticipated in the form of an absolute danger. -Jacques Derrida

In this essay I oudbe a themy of n&ear texts and investigate some of its possible implications for the practice

, of literary theory and criticism A norhear text is an object i of verbal communication that is not simply one fixed:

1 sequence of letters, words, and s r n t m c e r but one in which the words or sequmce of W O Y ~ may differ reading m : readmgbecause of the shape, conventions, or mechanisms of the text Nonlinear texts can be very different from each

' other, at least as different as they are from the linear texts. In the conceptual framework presented here, the linear text

m y be seen as a specla1 case of the nonlinear in which the cmmn6on is to read word by word from beginning to end Recently, because of the computer, certain types of nodinear t a t s have received attention kern educational, technologic^ and heoretical circles. Now may be the time to broaden the scope of interest and to examine textual nonlinearity from a

general point of view. Over the past two decades, the spread and radical

development of the computer as a means of cultural and aestheric expression has created a challenge to the paradigms of cultural theory that has not yet been

I systen~atically answered Studies of specific cornputer- mediated phenomena often suffer from alack of insight into neighbring phenomena, again caused by a missing

b kame of reference, a general theoretical overview based on a broad comparative study, and a dialectic h e n neighboring fields. This is not least the case in literaly theory, in which technological issues traditionally haw been

met with verylittle interest. During the past decade. however. such issues have seen a marked increase of ateention, perhaps not totally independent of the success induction of electronic word processing as an academi tool. The wwcl processor has served to familiarize the literary scholar with some aspects of the new text

technologies: but. due ta its mflaboratiw arid emulative nature (the way elecmnic word processing assumes the goals of the earlier technologies), the more radical potential of textual computing is easily ignored and the computer is gratefully perceived as less threatening than it actually is. This essay, unlike the others in this book [Hpr / rex t /

I TJleoty], is not concerned with hypertext Instead, i E shalI t q ~ to take a step back, to investigate the larger ( repertoire of textual forms of which h m can be said to

be one, Hypertext, when regarded as a type of text sham 1 with a variety of other textual types a fundamend trait I I which we defined as nohcarity It must immediately be pointed out that this concept refers only to the physicc-

I logical farm (or arrangement, appearance) of the texts, and ' not to any fictional meaning or external reference they might have. Thus, it is not the plot, or the narrative, or any other d - k n o w n poetic unit that will be our definitive agency but

i the shape or structure of the text itself. A narrative may be 1 ~ e r f e d y nodnear (for example describing a sequence of : events in a repetitive or nonsequential way) and pt be

represented in a to tdy h e a r text. The advent of computer-mediated tertuality seems to have

left many ofthose theorists and critics who noticed it in a terminological vacuum. In their eagerness to describe the b r m new reality. they let a few words like electronic and hypertext c m r many different phenomena. Behind the elecmnic text there is a large and heteveneow variety of phenomena. and as we shall see. a compu~r-mediated text may have more in common with a paper-bas& one than with one of its electronic brethren

After considering some fundamental problems with the concept of textuality, I shall propose a typo!ogy of nonlinear texts based on principles extracted from various sarnpIes, and then I shall outline the main foms of nonlinearity Smce the paradigms and practice of literary theory cannot remain unaffected by its encounter with nonlinear literature. except by pretending it never happened, I both discuss new applications of literary theory and suggest some possible new departures.

Behind the Lines: What Is a Text, Anyway? 1

The text as a whole and as a singular whole may be compared to an object, which may be viewed From s d sides, but never from all sides at once. -Paul R i c m

TO present nonlinear texhlality as a phenomenon relevant to

t e d theory, one must rethink the concept of textuality to

comprise linear as we] as nonlinear texts. 'The text," as it is

commonly perceived entails a set of powerful metaphysics

that I haw no hope of dispersing here. The three most important ones are those of rmding, writing. and stability. Regardless of m u d contradiction, these three work together to control ow notion of what a text is. For our

purpose, they can be summed up as blows: (I) A text is what you read, the words and that you see before your eyes and the meanings they produce in your head. (2) A text is a message, imbued with the values and intentions of a specific writer/genre/culture. (3) A text is a fixed sequence of

constituents beginning. middle, end) that cannot change, although its interpretations might In opposition to these notions, I argue that the lessons of nonlinear literature show us a textuality Mewent from our readings (and our readings of 'reading"), more fundamental than our messages, and through the evolving rituals and technologies of use and dis.tributrm, subject to many types of change. I do not for a

moment believe that my constructed bitlarism of the nonlinear M and the linear text or any of the other perspectives in this essay are any more free of a metaphysics than any previous textual theory, but I hope they are better flzited to identiCyingsome of the relevant issues of t exhd communication. My use ofhe word text is seemingly at odds with that of

certain schools of textual theory that regard the text as a semantic network 05symbolic relations. Imsely attached to

the notion of the iiterary work I do not intend to chdlenge that idea: I believe that it belongs to a different aspect or level of the same object. We then have two pqxcfis: the text as a technical historic$. and social object and the text as it is

individually received and u n d e d These aspeEcs. which m might call the infomati~w and the interpretable, are governd by different rules, but they are interdependent and influence (and sometimes intrude on) each other in many ways.

The informative aspect of the text is usuaUy the had^ to see, because it is the most obvious. In addition to its visible

attacks of illegitimate nonlinearities, thm weds to order us to go at o n c ~ to page SO for further insWms and skip the intervening pages that, we are told have been contaminated by subwrsiw directions? These hypotheticaI cases, whch are far from impossible, illustrate a pecdar M c power of the linear text orrer the nonlinear the

hear can flirt with nohear i ty , but the nonlinear cannot: lie and pretend to be linear.

But let us return to our metaphysical question, which really is a serious one What is a text? Or, to rephrase it. Which elemen'ts and effects belong to the text and which do not? The p o s t s t r u c t u h are fond of discussiw this question in (and in relation to) the preface or the foreword but since I do not have such places at my disposal in this Book but only a chapter. I shall not argue with them Instead consider this: dws the author's name belong to n text? It is usually only found outside the text-on the cover, m the

catalogue, in the book review, and in some cases in the top or bottom margins of the page: but it can be argued that, along with the text's title, which is also found outside the text

proper (not "enclosed" in it), the wmh that make up the author's m e are the gingle most r n e g f u l phrase of the text. Ofthe text, but not in the text Imagine the difference between a tmt by E? G. Wodehouse and a text by Aptha Christie: no problem there. It does not wen haw ta be any spedfic books; we h o w the difference anyway, The fact that

we may know something about the authors behind these

names k not anywhere near as important as what we know about a text. once we h o w it is by one of h Once I pi& up a book by Ken Follett, I have already started the interpretation of i t long before I haw started on the first page. Even if the name itself is unknown to us, its hints of gender and cultural background are meaningful

Authors have always known these dungs, In antiquity and the emIy Middle Ages, some writers would use the name of a famous author to get their ideas read and spread-not as a

dainous forgery kth the god of short-term benefit but as a

way M enhance the endurance and pmitim of heir wmk. Think of it as a land of benevolent computer virus. In more recent times, female writm used male pseudonyms: the Action was even better if a fictitious author could be constructed Still, "serious" authors use pseudonyms for thex less serious work; that way the weigh of their "me" name win not misIead their readers' m o l l s and intqiretations. This shift works well even if the connection

s to s h a nu& a p~ [textas;

between the two names is known; it is he name, nathe person behind if that is impartant. The name bdmg text, the writer (as in ghostwiter) does not. Our distinction bemeen the text and he shpt

of mass-produced and -distributed copies leads to h, fundamental question of in what sense the shpt.

independent text (the so-&d real tm &d copies) mn be said t o This distinction my seem so much quaint and unnecessary contentiousness, ht, pw of the textual ontology-r, to coin a name for our fidd texmomy-presented in &is e s q , it helps u - v that the stability of paper-based dcmanents is as n mduct of wr metaphysical belief in a transcendmta! ~n

inherent qualiq of the physical object I~II@W a book in which some of the appear to be

missing. or the print is unreadable every 16 pages. or some of the pages are repted while an equal number omittd Em if this copy is the only one we ever see, we autamaMY assume that it is not supposed to be this way and rhat a mom correct version eists I t may never have been printed; but to us. who can imagine it perfectllr (except for the missing wor& of course), it is sd more real than the one we are holding Far instance, in Terry Eagleton's Litenmy 7%wry, there are two rhapters bearing the number one; the first

titled "Ineoduction: What is Literature" (p.l), and the other "The Rise of English" l(p.17): Since my copy is from the eighth printing (1990) and the book was first: published in 1983, it is &ely that there is a version with only one first chapter, but we nevertheless assume that this is what the text meant. a$ that the introduction got numbered by mistake. We do this out of lack of respect for the ~ o p ~ ; it appears to r n i s l e p m t the "real* text, wen if such a thing may never have 6 ~n short. we prefer the imagined integrity of a metaphysical objea to the stable -ion that we obsem. Which one is more real than the other? As 10%

as we are abIe to imagine and reconsttuct ~ K I ideal ~ r s i o every&ng appears to be fine. and our rnetaphpia mz intact But what if the flawed version inkrfm deep1

with our sense d rewon that it, in more thm a ma of sp&g, steals the show? Following our rnetaphysidl@' we would haw to say that a new text had been a@tedshEe & 4ternat.i~ would be a script wi thu t a text. But because o f i ~ unintentiond origin, this new text cannot be metaphysidy equal t o the text it replaces and SO we af*e left with a paradox: some texts are metaphysd m e noL

and If we do not know their origins. we have no way ro tell the dif€erence.

The ahernative, of course, is to ~bardon the concept of a real text-behind-thetext altogether. On S a d a y , Februaly the 7th, 1987. I saw John Boorman's Zurdm (1974) at the Bergen F h Club. Or ddI? As it happened somehow the reels got mixed up md were projected in the sequence 1,2,4, 3.5. The film is a weird allegorical adventure, from a bar&= future in which technology has h o m e inexplicable and supernatural to everyone but a secluded p u p of very bored immortals. The title is an anapamatic Asion to The Wizard of OZ and the story contains many surreal and fantastic elemen&-not least. it seemed to me, the s~dden jump in the narrative, followed after a while by a just as strange flash- back. When the fifth reel came on, however, I slowly started to suspect that this rather crude montage technique was neither Booman nor his film company's doing. but most

likely a mistake in "reeI t h e . * By then the damage was done, and I had had the confusing privilege of bang lost in the materiality of a &-a strangely appropriate experience, somewhat p d e l to that of she main character, played by Sean Connery, a barbarian who manages to get into the secret place, the Vortex of the immortals, to see their strange

customs and technology (and their eventual desmction) from the inside, By virtue of the altered sequence, an unintended

cinematic experience, a new expression, was mated But was it: a new film? I am tempted to answer, no. Not because I feel that a film (or any other artistic "worr) has to be the intended and consecutive design of a conscious, neatiw operator, but because both the origind and the heretical sequenms are based on the same m a t e d potential. In this sense, a text M. a film is like a k t e d language in which all the parts are known, but the full of their combinations is not The mutation of Zarda was mated by a hidden possibility in its channel, not by the i n d u c t i ~ n of a new code or principle

There are many scales of change in a text's metamorphosis: unintentional (the blunders of a typesetter or projectionist in the dark), usurpatmy (a remix of sampIes from a musicd reco&g, a hacked version of a computer game). plagiary (one composer's unacknowledged variations on a theme a€

another), and subversive or estranging (the "cut up" t a d

experiments of W h Burroughs and John Cage). to suggest a few. Some of the results of some of these

User-f unctionality. Besides the in te~re tn t ive function of the user, which of course is present in the use of both linear and n o h e a r t ex td ty , the use of n o h e a r texts may be described in t e r n of four active feedback functions: the explorative function, in which the user decides which "path" to take: the r o ~ e - ~ l ~ n f i n c t i m , in which the user assumes strategic responsibility for a 'character" in a "world" described by the tm the conturative function, in which t w o n s andlor traversdLhctions are in part chosen andlor designed by the user; and the poetic function, in which the user's actions, dialogue, or design are aesthetically motivated.

Any type of text can be discussed according to these categories; I avoid the primitive and theoretically

uninteresting division between electronic and hard copy texts as well as the nebulous concept of interactive fiction. The model is equally applicable to a child's interrogation of a

storyteller and a researcher's conversation with an artificial intelligence program, or a radio broadcast of The Wind in the

Willows. The best way to test a model is to see how we1 it stands up

to new data. Since I developed mine in 1991, a new text type has appeared invented by the science fiction author William Gibson. Hk Apppa:A Book of the Dead (1992) displays its script at a h e d scrolling pace on the screen and then encrypts it by a technique cryptically known as RSA, rendering it effeceively unreadable after that one projection."

Leaving the more obvious jokes aside (better make reservations down at the libv, quick!), this is clearly one more of those one-of-a-kind texts for which "the medium is the message" seems to have been intended But that should not stop the empirical literary critic. I must admit to a

curious feeling of unease here. Agrippa perversely obeys h e logic of cultural capitalism beyond the wddest dreams of publishers: it is the non-reusable book At the same time it obviously subverts the metaphysics of textual mass production. How? By being a copy that destroys its text+ or a text w M destroys its ropy?Agrippa is a unique lesson in textud ontology, a linear text that seems to f h t with -.

nohear i ty , not through its convention or mechanism but through the Merence between its used and unused copies. The individual copy-as-text is linear, because there i s only one sequence: first, the decrypted scripton once, then the re- encrypted one fox ever after; but the text-as-copy may turn out to be either of the scriptons and is therefore nonlinear. Rather than accept that this paradoxical resdt undermines

my l inearnohear distinction. I contend that by deshoylng its maversal function it exposes the inherent instability ,f the metaphysical concept of 'the text itself'' Thus, A @ ~ ~ ~ becomes n o h e a r only if we choose to accept the *text-

behind-thetext" as more red than the physical object hat can refuse to be read. As for the rest of our categories, Agnpp is a rather unusual combination of a static, determinate, and transient text with compledy controlled

access to scriptons.

1 As a simplified synthesis of this model I now propose lou,

; pragmatic categories, or degrees, of nonlinearity: (I) the simple n o h e a r text, whose textons are totally static, open

I and explorabIe by the user; (2) the discontinuous nonlinear

text, or hypertext, which may be traversed by "jumpsM (explicit links) between textons; (3) the determinate *cybertext," in which the behavior of textons is predictable

but conditional and with the element of role-playing and (4) the indeterminate cybertext in which textons are dynamic

land unpredictable. The weakness of this simplified model is Ithat some nonlinear texts, such as those that are both static I and indeterminate, fall between the gener&ed categories.

However, it is not uncommon in cultural theory that generahation means loss of precision, and it shodd always

be weighed against the usefulness and convenience of the simplification and the fact that a more rigorous and

unmitigated model exists. The rest of this essay &cusses each of these four

categories, some of the texts that can be said to belong to them, their attributes and pec&wities, and their importance to literary theories and to the practice of literary criticism.

The Readerless Text Nonlinearity can be achieved in many ways, the simplest of which is a script forlung out in two duections on a surface, forcing its witness (the user) to choose one path in

preference to another. In such a case (for example, the "dream maps" in Kathy Acker's Blood and Guts in High School), the user can immediately afterwards take the other path and thus eventually view aIl parts of the script sirnultaneouslyP The verbal oscillation created by two equally possible combinations, the choice of which is entirely up to the user, produces an ambiguity different from the usual poetic double meaning of a word or phrase, because there seem to be two

different versions, neither of which can exist alongside the other, and both obviously different from the text itself. Lie

optical illusions, we can imagine first one, then the od not both at the same time. When we look at the wholl such a nonlinear text, we cannot read it; and when we

we cannot see the whole text. Something has come be us and the text, and that is ourselves, trying to read. T

f consciousness forces us to take responsibility for whal read and to accept that it can never be the i-M itself. 1 text, far from yielding its riches to our critical gaze, ap to seduce us, but it remains immaculate, recedes, and r left with our partial and impure thoughts, like unwot-t &rims beseeching an absent deity.

However, if a text cannot be conquered it is all the I suited for worship. The wal-inscriptions of the templt ancient Egypt were often connected two-dimensional one wall) or three-dimensionally (from w d to wall an

room to room), and this layout allowed a nonlinear arrangement of the rehgious text in accordance with t symbolic architectural layout of the temple?'

Without doubt, the most prominent and popular no text in history must be the famous C h e s e work of or, wisdom, I Ching or Book of Changes, one of the great cla antiquity, which was used for thousands of p a s for meditation and as an oracle. It is not, as is sometimes s

the oldest text in Chinese and world literary history, b~ well over three thousand years old and or iba tes h m symbol system said to have been invented over five thc years ago by the legendary Fu Hsi" Other notables, an: them King Wen, the Duke of Chou, and Confucius, havl developed and annotated the text down through the a8

the text is s d being rewritten and mutating, adapting modem society and its paradigms."

, I Ching is made up of swty-four symbols or h e x a p w h i h are the binary combinations of six whole or bro

1 rchangingJ lines (64 = 26). A hexagram (such as nr. 4 Ko/Revolution) contains a main texton and six small or ! one for each h e . By manipulating three coins or forty

! / yarrow stalks accordmg to a randomizing principle, tv / from two hexagrams are combined producing one out

f 4096 possible sniptons. This scripton contains the an: I a question the user wrote down in advance. The extrer I dwer openness of the formulations, the sense of ritua i imlved in throwing the coins or s&, and the stran) i ; personal communication between the user and the bod i almost always make an answer extracted from I Ching I relevant and sometimes even divinely inspired

linear-nonlinear distinction. I contend that by destr% ravcrsal filnetion it exposes the inherent instability of g

concept of "the text itsew Thus, nonlinear only if w choose to accept the 'text-

:nd-the-text" as more real than the physical objem ht refuse to be read As for the mt of our categories,

i ppa is a rather unusual combination of a static

rminate. and mnsient text with completely conhued ss to scriptons. 9 a simplified synthesis of this model I now propose four ~ a t i c categories, or degrees of nonlinearity: (I) the jle nonlinear text. whose texlons are t o d y static,

explorable by the user: (2) the discontinuous n o h e a r or hypertext, which may be traversed by "jumps"

licit links) between textons: (3) the determinate text%" in which the behavior of textons is predictable :onditional and with the element of role-playing and (4) ndeterminate cybertext in which textons are dynamic ~tnpredictable. 'She weakness of this simplified model is some nonlinear texts, such as those that are both s u t k ndeterminate, fall between the generalized categories ever, it is not uncommon in cdtural theory that rahzation means loss of precision, and it should alway

eighed against the usefulness and convmience of the lification and the fact that a more rigomus and ltigated model exists. e rest of this essay &cusses each of these four ones, some of the texts that can be said to belong to , their attributes and peculiarities, and their importar mry theories and to the practice of literary criticism.

Readerless Text m e k t ) l can be a c h i e d in many ways, the simplest of is a script forking wt in two directions on a surface,

kg i ts wifnw- (the user) to choose one path in m c e another. In such a Ease (for example, the 'dream " . " Kathy Ackerk Bloodand Guts in High School). the an immediately afterwar& take the other path and ' m d y view d parts of the script simultaneousl~" erba oscillation created by two equdy possible inations the choice of which is entirely up to the useL

ces an ambiguity &&rent hm poetic double lng of 1 word or phrase. because there seetn to be two 'nt versions, neither of which can exist alongside the

both obviously different horn the t a t itself Lhe

optical illusions, we can imagine first one, then the other, but not both at the same time. When we look at the whole of such a nonlinear text, we cannot read it; and when we read it, we cannot see the whole text. Something has wme between us and the text, and that is ourselves, tqmg to read. This self- consciousness forces us to take responsibility for what we

read and to accept that it can never be the text itself. The text, far from yielding its riches to our critical gaze, appears to seduce us, but it remains immaculate, recedes, and we are left with our partial and impure thoughts, hke unworthy pilgrims beseeching an absent deity.

However, if a text cannot be conquered, it i s all the better suited for worship, The wall-inscriptions of the temples in ancient Egypt were often connected two-dimensionally (on

one wall) or three-dimensionally ( h m d to wall and from room to room), and this layout allowed a nonlinear m g e m e n t of the religious text in accordance with the symbolic architectural layout of the temple.'"

Without doubt, the most prominent and popular no rhea r text in history must be the famous Chinese work of oracular wisdom I C h i 3 or Book of Changes. one of the great dassics of antiquity, which was used for thousands of years for meditation and as an oracle. It is not, as is sometimes stated

the oldest text in Chinese and world literary history. but it is well over three thousand years old and originates from the symbol system said to have been invented owr fiw thousand yearsago by the legendary Fu Hsi." Other notables, among them King Wen. the Duke of Chou, and Confucius, have dewloped and annotated the text down through the ages; and he text is sti l l being rewritten and mutating, adapting to modern society and its paradigms.'*

- I Ching is made up of sixty-four symbols or hexagrams, which aze the binary combinations of six whole or broken ! C h g i r ~ g " ) Lines (64 = 26). A hexagram (wrh as ny, 49.3 I Ko/Revolution) contains a main texton and six small ones, I one for each line. By rnanipdahg three coins or fortynine t I yarrow stalks according to a randomizing principle. textons horn two heragrams are combined producing one out of 4096 possible scriptons. This scripton contains the answer to a question the user wrote down in advance. The extremely

i clever openness of the formulations, the sense of ritual involvd in throwing the coins or stalks, and the strangely

I personal communication between the user and the book

) almost always make an answer e c t d From 1 Ching seem I relevant and sometimes wen divinely inspired

Unlike historic texts with a fixed expressiot~, such as ( I ; Beowulfi I Ching seems to speak uniquely to us across the i l

millennia, not as a distant mirror that can be understood in a I

phiiological or romantic sense but as an entity that somehow 1 1 1 understands us and exists for us. This almost religious effect

1 ;

can be partly explained by the repeated updates and the fact that the text was intended to be useful and &realy relevant to events in ~eopk's bvcs, but it seems to me that it is the explicit and elabomte ritual, largely unchangd through the ages, that creates the textual pwsemce that allows us to be ! , 1

naive users-not readers but agents of the text closely related to the lzsers of three thousand years ago, despite the epistemoPogical interventions of time and culture. The Book

; 1, ofChanges may not b the world's first text, but it is certainly

1 the first expert system based on the principles of binary I I

computing that vmy much later became automated by elect~icity and the vacwum tube. Bath typesof text &cussdl so far seem to reject the

I presence o f t h e mclitional reader figure, as it is implied and

I : applied in the theories of literature. As an individual, this i 1 ,' pale and uncontroversial character never mattered much to us critics anyway and then only as a construct on which to hang the baser pleasures of the text; he is our poor and predictabIe cousin, slave to the rhythm Iost in the textud pleasure dome like the ball in a pinball macliine. Later, for the i reader-response theorists, he became a thumbtack with which to pin down the variable of literary meaning when it could no longer be located in the text. Active or passive, the reader is always portrayed as a receiver of the text, going

I quietly about the business of consuming, constructing meaning only, a fixed but evolving character at the end of the text's production line, defined by the conventions and strategies of reading Of murse, it can be argued that y elation ship is no Merent for nonlinear texts, once the shock of an alien form is gone and the ~art.i& convention 1 is understood and mastered. This wunterpahk which may

I

be called the Verfremdung-a~gument has much merit. but it ignores the fact that the understandmg fieyondaividl of a nonlinear text can nwer be a consummate understanb~,

1 , ' because the reahition of its script (and not just its meanind belongs to the individual user, who is acutely aware of his or her awn cmstructiw participation. Since the object 1 1 i s unstable both in a syntactic and semantic sense, it cannot be read, only glimpsed and guessed at. Much of the initial discomfort Felt by the usw of a nonlinear text is caused by i ts

not behaving as a real text shod& once the strangeness is gone, the user knows what to expect which i s not to expect everything. The users learn to accept their position as agents: of the text, sometimes hppily, as in the case of the Book of Changes. and sometimes unhappily, as with the forking directions texts. The difference between these two types of experience can be explained by the presence or absence of an established (meaningful) ritual, which must absolve the user 1 from rhe burden of reading, which in the case of nonlinearity

may be defined as the frustrating attempt to harmonize contradictory scriptons fmm the same text. The user of I Ching &s the scripton directly to his or her individual situation, and the interpretation, following the ritual of producing the hexagram can only be done by theindividual.

Tlus fall from readership should not be confused with the dewr destabilization effects of so-called rnetafictions, in which the opposite point-readership confirmed-is made. Even {and e spe~dy) the famously "unreadable" tern subversively

observe the metaphysics of the general reader: the dmr would not be locked if the owner did not beliew in thiwes.

Few texts drive home the point of the readerIess text more abmdantly than Raymond Queneau's Cent Mine Milliards de

PoPrnes (2961)." In this short book, ten pages are cut into fourteen one-line s ~ p s , and the user i s invited to flip the strips inaividudy to Sonn 100,000,000,000.000 Werent combinations. As it turns out each of the 140 strips (or textons) is a sonnet line, and the result of any combination is a scripton in the form of a formally perfect sonnet. Here is sonnet number 65 957 658 052 33 6:

Quand I'un aveque I'autre aussitbt sympathise que convoitait c'est stir une horde d'escrocs des dtres indkcis vow psien t sans franchise il ne trourle aussi sec qu'un sac de v i m fayots

L'un et Taube a rajson nan la foule insoumise qui dochaddwenant jetait ses oriptaux aller a la grand ville est bien me entreprise lknfant pur aux yewc bleus aime les bedingots

Du pble d Rosario fait une belle mt te on pNe Ic marmot qui plong sa rnenotte Iorsquon rwient au port en essuyant u n p i n

Ne fallait pas si loin agiter ses breloques on transports et Ie rnarbre e t debris et d&fraques

This may not be the most exciting ofiyrical pee% but i s unique in a very speaal sense: 1 have never read B before, and chances are that neither has anybody el- who Was it me, or Queneau (and if so, in 1961 or 1g92r), or perhaps the text itself? Will anybody ever read numbcr 65 957 658 052 317? For one person to read all the somets is

clearly impossible, and even a very small frac t i ~ n - ~ ~ ten million-would take at least one hundred years. Cent Milliard' de P&es effectiveIy mocks the theoretical notions of writer and reader, while the power of the text is clew+ demonstrated. (What it does to our notion of the sonnet is perhaps better left. unsaidJ "ObviousIy the possibilities of he book as format are being strained ta the limit," comments William padson. who goes on to propose h m e s as 'Bn ideal candidate for a cmp~tterized ver~ion."'~ Contraary to PauIsan,

I suggest that the fact that it is a book is just as significant. and if it seems easy to implement as a computer p r o p m . that is because of the simple and unstrained elegance of its

idea. The difference between these experiences and my

experience with Boorman's Znrdoz is that in the latter case I could, based on my cultural competence, deduce the actual

existence of a version that was independent of me and the possibiity of a proper reading &at could be conducted by at1

easily imagined proper reader, but not by me. In other words, I rejected my reading because it told me that 1 was

not a real reader, since what I was reading was not the real text. The shock of discovering that one is not a reader can only happen (and only accidentally) with a linear text, because that is the nnly text in which the metaphysics of a real reader has any credibility and the only text in which the reader canlevist as a reducible, accountable figure. In addition, reader has-until now-always been defined by literary theorists with only the linear text in mind If want to know what is going on between nonlinear texts and their usem, we must come up with a concept that imp both more and less than reading and redefines literarl satisfaction as well as hemeneutic behavior.

Hypertext Is Hot What You (May) Think l-lypertext, for a11 its packaging and theories, is an arnazind~ simple concept. It is merely a direct connection from one position in a text to another. However, when we speak of

hypertext, it can signify at least three M e r e n t things: (1) h e general concept, as outlined above; (2) an implementat

the concept, usudy a computer application called a hypertext system, with idiosyncrasies and enhancements that make it different from other systems; and (3) a text- embedded in (and defined by) such a system As an unfortunate result, many assumptior~s made abut the general concept of hypertext are really about a specific implementation. Added to that are the political conjemres about the benevolent effects on h e structures of power between writers and readers, teachers and students, government and the public, in which the g o d guys seem to be winning, at Ieast in theory. Only the first of these reIation- ships will be discussed here. and only because of the assumptions about the effects oh hypertext upon the figures of author and reader.I5 (Of course, impIicit in the term hypertext is a sphere of meanings bcyond the operational. Those who luould play on this potential cannot completely escape its dark side: the excessive, the abnormal, the sickly.)

Although the term hyperte~t was first used by Theodor H. Nelson in 1965 (compare Nelson 1987). the modern origin of the idea is generally accepted to stem from Vannevar Bush, whose article "As We May Think" (1945) described a possible solution to the scientist's problem of keeping up with the ' m n g mountain of research." in the form of a "sort of mechanized private fie and library," a machine for storing. annotating retrieving, and linking information: the rnernex.'" Although Bush emphasizes the "&-the linear orderitig of interesting items from the 'maze of materials availablev-he aflom his user to go off on little side excursions. Bush was no techno-pessimist (at the end of the article he wen envisions the neural jack of the 1980s ~ b e r p u n k science fiction!), and we can hardly blame him for not coming up with a complete "web view" on hyprtmhality in 1945. But it should be

pointed out that in his fascinating vision-his p o e t i c e nonlinearity is much a ~roblem (the "maze) as a soIution

(the "trail"). Where he clearly concurs with his a p o s t l ~ is in his fms on user-createdlinks and annotations. This may seem more radical than it actually is , with subvelsive tical

consequences for the world of literature and art: but R U S ~ ~

user is clearly modeled on the traditional academic author, who can carry out his critical comparisons and annotations of sources with the same serene &tame as before, only much more efficiently.

The principle of hypertext should not be linkdl to a particular ideology or poetics because it can be used (and of

the text "seem to haw a mind of its own."lR Thus Afternoon, a r p b l y the first literary hypertext, turns out to be something more: a cybertext disguised in hypertexfs dothing

It is hard to classify Afternoon as a narrative (or *a story," as the text paradoxicdly tides itself). Although within most of the individual scriptons the voice of a first person narrator relates events to a namtee in a t~aditional manner, the unpre&ctable h g i n g a f scenes (as one trail of related scriptons abruptly stops and another begins) constarltly undermines the wrmld-be reader's attempt to identify with the narratee, as well as the identification of the narrator and the ( implid author or exo-narrator, as it were. In Afternoon there seems to !w an anti+nanator a t w r k , giving the narrator (and me) a hard time. In linear experimental texts

the subversive effect is sometimes achiwed by a "distanw btween narrator and narratee" and sometimes by the "loss of namtee"-the narrator as ~olipsist '~ In Afternom+ however, the relation between narrator and nmatee appears relatively normal; while the distance between the user and narratee on one side and narrator and atlthoron the other is stretched to the limit by the unreliable lfnks Far from feeling like Landow's 'reader-author" (117). who has no problem constructing "meaning and narrative from fragments provided by someone eke," I felt constantly sidetracked turning and turning in the dating text, dead sure that important things were beingwhispered just beyond my hearing I cannot deny that it was a very fascinating hterar y experience.

It can be argued h a t the text I encountered w a s (in more than one sense) not the same as the one discussed by Stuart Moulthrop, J. David Bolter, and Landow. From their accounts it appears that they usd a different and more advanced version of Afternoon b hypertext system, the guthor version" of Storyspace. which allows writing and adding h k s , and most significantly contains a global view, a graphid representa tion of the topological relations between all textons and hks, My version was in

Readingspace, the stand-alone reader program that Afternoons publisher distributes. Consequently, my encounter, "one saipton at a time," with Afttmoon was very different from theirs; for the global view, ewn if they did not use it, gave them a safety net that I lacked While 1 was lost

1 in the labyrinth, they could be "up there" with its creator- - but only up to a point. Whatever changw they might impose, it would onlybe on their own copies; Joyce; text

wodd stand unchanged In this. hypertext is not &fierent fxom paper-bascd linear texts. The balance of bemeen readers and writers is not changed by hypertext alone, nor by its enhancements, but by the pofitid and economic logic of society (to use some slightly inaccurate &ch+.s). % may change, under the influence of technological change and other things: but until it does,

hypert& i s just one more "instrument in some yepresentationaI en terpr ig to barrow a phrase from SamueE Delany. To expand the notion of hypertext by subsuming other

computer-mediated textual communication phenomena such as Usenet (see Bolter, 29) or intertextual allusion (see Landow, 10) will only render the smcept useless for critical &course. Landow's term -implicit hypertext" implies that an dusion anda link are essentially the same, but we only need a hypertext urith both links and allusions to see that thev work Merently and must be considered twu separate literary i n s m e n t s . Boltw, eager to proclaim the end of 'the printed book," plays along with h e metaphysics of logocmtrisrn and reduccs print on paper to barely a comer of its multiform nature: 'h printed book speaks with a

single voice and assumes a consistent character, a persona, before its a~dience:~ For "the electronic text," however, h s

no longer applies, because "it is not a physical artifac~" To go against Bolterk rhetoric, I would say that instead of having two sets of opposed attributes, one connected to the 'printed' and one to the "electronic" text, we have a number of different text types, some paper based and some digital. wid1 the greater variety among the digital ones, and the paper based most centrally placed. Thus, there may be more difference between two digital texts than between either of - those and a paper text.Allusion, reference, quotation, and linkingare d different functions of i n t e d i t y , just as

Usenet newsgoups, eIectronic mailing lists, hypertext systems, paperback bestsders, and flysheets represent different modes of t d t y ,

As the analysis of Afternoon indicates, literary hyperte& seem t o pose interesting perspectives for students of literature. The question of nonlinear narrative versus anti. narrative should not be decided by the evidence from only one text (even if it exists in ma versions), and perhaps We

need a new tern~inology that lets us name the representation and composition principle that relates to nonlinearity as

narrative relats to linearity.

Howwer. one traditional term I;eems almost perfect to describe literary hypertexts Afternoon does not represent, break with the n d On the contrary, it finds its place in a long tradition of expenmend literature in which one of the main strategies is to subvert and resist narrative- The nmd ('the new"), Srom Cervantes to the Roman N o m u , has always been an anti-re, and A f t e m n is but its l a k t confirmation.

Death and Cybernetics i n the 'Ever-ending Text

I'm not sure that I have a story. And if I do. Trn not sure that everything isn't my stay -Michael Joyce, A f i m n . a sto y

If literary hypertext i s a new form of computer-mediated textuahty. rjbertext is a fairly old one, going back to he 1960s if not longer. Cybw is derived from cybemetia the name of Norbert Wiener's science of "control and communication in the animal and the machine," again derived from the Greek kybernet*~, steersman (compare governor). A +ext is a self-changing text, in which scriptons and mavenal functions are controlled byan immanent cybernetic agent. either mechanical or human There are many species of cybertext. and my distincbh between determinate and indeterminate tries to set up ar! important division between two main p p s : those that Q,

be predicted (for example, one set of user actions will 4% yield the same set of sniptons) and those that cannot%

second p p will be discused in the next section. The histmy of computer-mediated cybertexts can be

traced to two %rent sources, both originating SromklielC; of computer science, and both with their memorable n-ter-: The first, fro, created by Joseph Weizenbaurn in 1966m an early success in the field called artificial i n t e b g ~ % mother of all ddague programs (Pav SHRDLLl Ram&+ countless others), Eliza played the part of a & o t h W ~ asking the user questions and wnstructing further- using information from the answers, Usudy, clialop turned mher Pinteresque as soon as the users d i s c 4 Elizni very rnechmical nature; kt Weizenbaum's ~h effectively demonsmted man's needs for comrnunicatjcrlr matter with whom (or indeed what), andm important literary genre-the artificial conversationah-wa~

textual communication evolved, from e-mail via mailing lists and newsgroups to s o - d e d on-he chat, such as the intmting phenomenon Internet Relay Chatm

At the end of the 1970s. with the spread of the highly popular Adventure over the networks, it was to be expected that someone should combine instant textual

communication and adventure gaming In the falE of 1979 at Essex University, Roy Tmbshaw started the development of the Multi-User Dungeon (MUD) on a DEC System-10 m a i n h e , a task taken over by Rcbad Bade in the summer of 1980" The first MUD was a successful game, with users scoring points by W n g each others* characters or finding hidden treasures and eventually reaching the powerful status of wizard ht it was also much more than a game; it was a cyberplace where people could enjoy complete anonymity and freedom from their social and physical selves md take on any persona they cwld think of, doing things with words that they wouId normally never do. Thus a new mode of t d expression was initiated different even from the telegraph: the user had to be very quick and formulate short, unrehaable sentences in seconds, or die. Dorothy Parker and Ernest Hemingway would have loved it. Like Adventure before it. MUD spread out globally on the

academic computer networks, w a s soon copied and changed into other types of multi-usa hts In the summer of 1989 at Carnegie Melon University. James hpnes pmprnmed a MUD with a significant new feature in addition to creating their own characters, the users were allowed to expand the MUD'S textual descriptions, adding their own landscapes to the topography of the MUD. This MUD, known as ? " M U D and reachable from any computer linked to the global Internet, emphasized socjd interaction and building There was no merit system; if your character was killed, it simply got an insurance fee of 50 pennies. The co-creativity of the users was a very anarchic step from the first MUDS. %yMm lasted from August 19,1989, to April 28,1990, when its data base of descriptions became too big to handle, fiUed up by more than 132,000 user-defined objects, each of which could contain several textons.

When regarded as literary objects, MUDS seem to defy every concept of hterary theory. Every user has a different (or several different) and +a1 pmpective(s), and the users bombard each other with tcxtons meant only to last as long as they are not scrolled off the s-n. MUDS are like constantly meandering rivers, developing new -es that

cross and *cross each other and are filled w i h all sons of

peculiar flo~arn and jetsam. And suddenly, in he mid& of chaos, a group of characters may start singing in unison fie Yoyodyne song from Thomas Pynchonb The C y ' n g o f ~ o t qg, 'High above the LA freeways, / And the ttaffic's whine, / Stands the well-known Galactonics / Branch of Y C ~ ~ ? P Strange things happen at sea.

Compared to a nineteenth-century navd TmyMm appears t o d y different: transient, dynamic, i n d m i n a r e , with explorative, role-playing configurative, and pwrjc wr. functionahty, h d yet this k !itmature: letters, words, and sentences are selechd arranged and disseminated to delight; impress, or enrage an unknown audience. The scriptons,

which can be funny, poignant, sleazy, silly. obnmious, or noisy, u s d y come in a heterogeneous mix With more than a q characten in the same room, it takes a hardened 'MUDder" to

keep mck of what is going on. Special+purpose MUD-dimt programs that have been developed to run on the user's l ~ d machine and ease communication provide functionaltty that

is not part of the MUQ itself, such as filtering out noisy characters and automating offen-used wmmands. Not all characters one meets on a MUD have realpmms behind them. and several characters might be played by the same person. An earIy automatic character Iso-mlled bod on TmyMUD was called Terminator, had its own office, and was, like its cinematic namesake, programmed to kill If you paid it

200 pennies it would go and pester any character you specified Bots were simply external programs built using various artifidal mtdhgen~ techniques and logged on by t he i r creators to TinyMUI3 just &e human players, but wudy recognized by their sonlewhat poor communication s U .

A discussion of MUDS in terms of authors and readers is devant: a M W cannot be read, only experienced fmm the very narrow perspective of one or more of the ugerj characters, with a lot of simultaneous scriptons h n g b e ~ o n d reach: and the user cannot be sure that a particular contribution wiU ever be experienced by more than a few people, or, since the other characters might all be d d a l persons or conholled by the same real person. by myone at an.

The Limits of Fiction An important issue raised by both determina~ and indeterminate cybertexts is their relation to the ontofogjcal categories of t e x d t y : fiction, nonfiction, p w , &+ama- etc. In the case of cybertexts such as Adventure and

TinyMUD, the most obvious choice, fiction, is not obvious enough.Advtntu~ invites- a belief from the user, but thii is not the same belief or q n s i o n of disbelief that must be mstained by the user of realistic or fantstic novels. Cybertextuality has an empirical element that is not found in fiction and that necessitates an ontologid category of its own, which might as well be called simulation In fiction the user must construct m e n d images that

somehow correspond to the world described in the text The user is responsible for the images, but the text is in control

and can dictate changes without any deference to external logic, From the user's ~"pective. fiEtions are neither logical nor illogical If the fiction claims that elephants are pink, then in the fiction they are, because nobody is *there" to contradict it, A fiction, then, is not about something that d m not exist but about something that it is meaningless to contradict.

In AmEenhrre, the responsibility €or coherence is shared

between the user and the text. If the you+character drops a sword in one place. lea- and comes back, the sword is still there. In other words, there is a systematic contract between text and user, hke the causal one that aists in the real world and which, &e fictions, can be empirically tested. In T i M U D the simulation of reality is wen closer to the real thin% since the conversations the us& character conducts with other characters often have the s i g s of real c o m t i o n s .

Simulations are somewhere in between reality and fiction: they are not obliged to represent rcaliy, but they do have an empirical logic of their own, and therefore they should not be called fictions. Unlike fictions, which simply present somehng else, cybertexts represent something beyond themselves.

The 'Rhetoric of Nonlinearity As we have seen, the profound challenge of n o h e a r texts to the basic concepts of literaq theory makes it difficdt to

discuss &m in common literary terms. Even to the extent it is s t d possible it should be done with caution; and if we can be sure of nothing else. we may be certain that contradiction will be the uninvited master trope of our discourse. But stdl-what kind of (literary? semiotic?) phenomenon is nonlinear textualirj? Is there a name or recognized class for tho device (or better. set of devices) of nonlinearity7 Do some domains of literary theory knd their vocabularies more easily to its description than others? (If so, those are the ones most worrhy of spino on.)

computation may be determinate or contain a random function that makes it indeterminate. Polygenesis can be determinate (for example, when the user types a sentence to Elizu. its response can be predicted) or indeterminate (as in the MUDS). A further classi6cation of the figures of nonlinearity such as distinguishing between different types of forks, h. random functions, plygenetic modes, and so o n d not be undertaken here.

The Corruption of the Critic: How can literary th"y attack the &ties of

nonlinearity? How can we cut them up, read into them. de- scribe them so they fit in our narratives? How can we Iink them to our totems and mntrol their hidden inechanisms? Hypertext seems already we1 on its way into the m o n . Fs this a good sign? Conquests, unlike discoveries, are seldom accidental On the other hand there is no such thing as litwary theory there are only theories and theorists. A d texts. Literarythmry, more than most academic &ciphnes. has always been uncentexed and fragmented a widening we of readings and interests linked to countless philosophies, like a true Barrhesian texle scriptible So if hypertext should find a home, why not here?

This essay will not answer any of the big questions: What win hypertext do to the ways we think about rexts? How wiIl it resist the ways we are going to think about it, and be remembered as something other than an in-house pet, a dead kadtion of literary experiment, explained and pacbgcd from the start? How will the powerful but extremely primitive logic of the link affect our discursive methods?

If hypertext has connected well with literary studies, cybertext. a much older textual phenomenon, has gone by largely unnoticed An article or two, a few doctoral dissertations; the lack of interest is signifiant, and may have several muses. One is obvious: adventure games are games, and that is not our department Neither is the similarity between I Chng, Queneaui PoPmes, and Adwnturr too &gat. first sight. Perhaps, also. the adventure game. for adl its trivia and popular appeal, is too radical to be recognized because it disfigures not only the reabgprocess but also the reader. Literary critics have generally scorn4 prosaic texts that too openly captured their users-in which the relationship between reader and narratee became too

[ intimate, ladung ironic distance or Verfrrmdung Like the

L telegraph, such texts fa11 between accepted categories. in this

case between lyrical poetry and prose. Afternoon on the other hand, with its subwrsiw anti-namtor, has seemingly no problem with this, and can be welcomed and configured into literature and the hteraly.

The key difference between Afternoon and cybertexts such as Adwnture and TiyMUD is what the virtual reality researchers call immersion: the user's convinced sense h e artificial environment is not just a main agent with whom they can identify but sumunds the user? In cybertextd terms we could say that the user assumes he strategic and emotional responsibility af the charackr, or that the distances between the positions of main charaaer narratee. and user have collapsed

To the critical institution, this ontological embamwment becomes an ethical one. How can we be critics if we can no longer read? How can reviewers of cybertexts face the kt they pbab ly missed large numbers of saiptons? And worse, not onIy will we haw to admit that we barely made it to first base, but in the exploration of indeterminate cybertms we will be reviewing the results of our own strategic and creative jnvestment~.~

Problems of "Textual Anthropology" This crisis in criticism might not amount to anything terrible, but it. could be wed as a new departure for literary hermeneutics. After &e celebrated deaths of the authm, the work, and reading, the text is now giving up he spirit, betrayed by its most trusted cmpanion. the sifier. What is left is linear and nonlinear t d q , or better, h e a r and nonlinear rextualities. This empirid evoIution makes possible a shift in method from a phlological so an anthropological approach in which the object of study is a process (the changing text) rather than a project (the static t e ) . On-he phenomena and particularly the MUDS, with their fluid exchanges of textual praxis, offer unique opportunities for the study of rhetoric, semiotics, and d t u r a l communication in general.

MUDS and s~milar nonlocal f o m of instant textual

communication can be studied from many perspectives in the human sciences; psychological, sociohgid anthropd~gica'h

h p k t i ~ p!&-ophical historical, etc. Shades of these dl inevitabIy h d their way into the literary and t d

p e m that we might expect from our own discipline literary theorists and critics do engage in the study of indeterminate cyber- it should be with an awareness

the old role of a posterioH investigator no longer suffices. Like the user. the critic mrrst be there when it happens. Not on]y &at but, like the particjpant observer of social anthropology, he or she must make it happen-improvise. mingle with the d i v e s , play roles, provoke response.

What, may we ask, will then be the difference between this literary an&ropo1ogy and a rea! anthroplogjst; investigation of on-line phenomena? In other words what keeps criticism from changing into a sub-&tipline of trarljtiond social anthropology? First. it must be noted that social an thmpology and literary theoty already have several perspectives and goals in common, and a recent history of mutual influence In cultural anthropology, cultures are treated as texts to be interpreted and subjected to critique? and even the problem of anthropological method as a literar) process has become a concern."

In the transient social textualities, the ontologies of the two traditions might seem to converge, and the boundaries between cultural anthropology and literary dieory may appear fuzzier than ewr. It could therefore be useful to explore some problems and conflicts of perspective &at

might await eventual partnerships of the two fields. Since MUDS and other indeterminate cybertexts are closed signification systems, that is, textual types, they should not be analyzed as traditional cultures or subcultures. The postorganic anthropology solicited in a recent essay on the phenomenon known as cyberspace is perhaps just another tenn for what literary critics haw been doing since Plato? 7: be analyze$ and defined a culture must be shown to &t independently of any one signification system. When a science starts to confuse its metaphors with its empirical substratum (for example when 'texts" become teuts), it is dangerously dose to becoming a mythology. An anthroPOIogy of MUDS. i. forstance, should not see as its primary object the rituals and interactjms between the characters inside; but rather the relation betwm the outsid participants (the users) and their inside symbolic actions. Literary theory, on dte other hand should not focus on the social behavior made by s d ~ b o l s , but on ho'

the sign system is used to consmvt and explore the possibility of a text-based representation of idenhty If a

cooperation between anthropology and textual criticism is t

be achieved, the two dishphes should not tr)r to do each other's work ox rnistake the other's ontology for its

ween lyrical poetry and prose.Afternoon on the othW +& its subversive anti-narrator, has seemingly no I with h i s , and can be welcomed and c~nfigund into .e and the literary :ey difference between Afternoon and cyber- such ?&re and EnyMUD is what the virtual realty lers call immersion: the us& comrinced sense that

environment k not just a main agent with hey c a n identify but sumunds the user." In

terms we codd say that the user assumes he : and emotional responsibility of the character, or distances h e n the positions of main character,

?. and user have collapsed critical institution, this ontological embarr;l-ent ; an ethical one. How can we be critics if we can no tad? How can reviewers of cybertexts face the fact ibably missed large numbers of scriptons? And ot only d we have to admit that we barely made it w e , but in the exploration of indeterminate :ts we will be reviewing the results of our own

: and creative inwsments?

Ems of "Textual Anthropology" ;is in critids~n might not amount to anything but it could be used as a new departure for literary wtics. After the celebrated deaths of the author, h e td reading. the text is now giving up the spirit, I by its most t~usted companion, the signifier. What hear and nonlinear textualiv, or better, linear and

textualities. This empirical evolution makes a shift in method from a to an llogical approach in whch the object of study is a

Ithe changing text) rather &a a project (the static 1-line phenomena and ~artfcularl~ the MUDS, with id exchanges of texcua praxis, offer unique nities for h e study of rhetoric, semiotics, and -0mmunication in generd.

similar nonlod forms of instant textual imion can be studied from many perspectives in the 5ences: psycholo~cal. socioIogid, anthropol+~, , p h ~ o s o p h d histnrical, etc Shades of these Y find their way in to the 1item-y and t d Ves that we might expea from -own discipline If 1mrists a d critics da engage in the study of inate cybertexts, it should be with an *at

the old role of a posteriori investigator no longer suffices. Like the user. the cr i t ic must be them when it happens. Not only that but, Eike dle participant observer of social anthropology, he or she must make it happen-improvise, mingle with the natives, play roles, provoke response.

What, may we ask, d l then be the difference b e m e n this litera y anthropology and a rear anthropologfst's investigation of on-Ime phenomena? In other words, what

keeps criticism from changing into a sub-discipline of traditional social anthropology? First, i t must be noted that smid anthrupology and literary theory already haw several perspectives and gods in common, and a recent history of mutual influence In cultural anthropology, mlmres are treated as texts to be interpreted and subjected to critique,36 and even the of anthropological method as a literary process has become a concern?

In the transient mial &ties, the ontologies of the two traditions might seem to converge, and the boundaries between cultural anthropology arid literary theory may appear b i e r than ever. It codd therefore be useful to explore some problems and conflicts of perspective that might await eventual pam~rships of the two fields. Since MUDS and other indeterminate cybertexts are dosed signification systems, that is, textual types, they should not be analyzed as traditional cultures or subcultures. The postorganic anthropalogy solicited in a recent essay on the phenomenon known as cyberspace is just another

term for what literary critics have been doing since plat^.^ To be analyzed and defined a culture must be shown to exist independently of any one signification system. W h e n a science s ~ s to confuse its metaphors with its empirical substratum (for example when 'texts" become texts], it is Gangeroilsly dose to becoming a mythology. An anthropology of N D s , for instanw, should not see as its primary object the rituals and interactions between the characters inside; but rather the relation between the outside participants (the users) and their inside symbolic actions. Literary theory, on the other hand should not fonts on the sacial behavior made possibIe by textual symbIs, but on how the sign system is used to c o n s m and explore the possibility of a text-based representation of identity. If a cooperation between anthropology and textual criticism is to be achieved, the two &pbnes slloufd not try to do each other's work or mistake the other's ontology for its own.

After these speculations the question remains: What will the study of nonlinearity and q h t u a l i t y do to literary theory? At this point there can be no clear answer. Between the bluny promises of technology and the sharp edges of political reality t h m is, in the words of Jacques Den&, "as

yet no ex-we." This essay has attempted to create a usable terminology for the study of a wider range of textualitits than has hitherto been acknowledged by the field of Ii terary study and to point to some current problems and challenges in the study of computer-mediated textudities. As we have

seen, fundamental shctural terms We story, plot, fiction, and numtiw, are not always suitable to describe die nonhear kxhtalities. To use them without quahfication is clearly imsponsible. The figures of nonlinearity suggest that one must revise literary terminology and pwtics in order to avoid further confusion and unnecessary ambiguity. Some of my reconfigurations of these literivy and theoretical concepts might turn out to be unnecessary, and others are pobably not radical enough. As I have shown, in addition to

hypertext there is a wealth of nonlinear text types, from ancient inscriptions to sophisticated computer programs based on the latest semantic research. I have not tried to present an exhaustive ernpirid survey of such types or to give a detailed historical exposition of the development and spread of textual nonlinearity. Others are very welcome to either of these tasks, I have no intention of taking them on. Nor do I believe that there is any need to construct a historid tradition of nonlinear literature, as the specimens I have seen so far seern to be different from and isolated from each other rather than belong to anything that a n reasonably be characterized as a common genre. There are undoubtedly local wadi tions, But nodnear strategies appear to rise out of a prevalent and tram-historic need to compose a practical effect, perpendidar to linear t e x t d r y , but usually with a specific and constructive or subversive rather than sensationalistic or frivolous objective.

When confronted with new data that is recognized as relevant but unusual an academic hcipline such as literary studies can employ at least two different tactics to harmonize the situation. The existing theories may be used to grasp and focus the new material (the intruder is tam&, or the new material can be used to reevaluate and mo* the old perspectives (the field i s changed). Hcre I have focused not on

the effects and insights by the various branches of

literary thmry when applied to nonlinear texts but on the ptential for new perspectives an literature in general h a t the study of nonlinear d i t y might bring us. Nonlinear texts and lterary theories may haw a lot to say to each other,

but we should not let only one side do al l the taking

Notes 1. Terry Eagleton, Lilemry Theory: An lntrodu~'~~ (Minneapolis:

University of Minnesota Press. 1983). 2. Norbert Wiener, Cybemetr'a; Or; Control and Communication in

the Animal and the Machine (New Kork: Technology, 1948).

3. N. Katherine Hayles, Choos Bound: Orderly Disorder in Contemporury L~temture and Science (Ithaca: Cornell University Press, 1990).

4. Roland Barthes, S/Z (Paris: SeuiI, 1970). 20. 5. George P. Landow, Hypertext: The Convergence of Contempomy

Critical Theoy and TechnoIogy (Bakimore: Johns Hopkim University Press, 1992).

5. Espen barseth. Terb: of Change: Towards a Poeh-a oj Nonlinearity [c.phil. diss., unpublished, University of Bergen, Department of Comparative Literature, 1991). 7. See Michae! J. Greenacre, Theory and dpplicutions of

Conespondence Analysis (London: Academic Press, 19%).

8, Wifliam Gibson, Agrippa: A Rook 01 the Deod (New York: Kevin Begos, 1992). 9. Kathy Acker, Blood and Gutr in High Schod (New York: Grove

Press, 1978), 46-51. 10. Rotf Gundlach, 'Tempefwlief," in l&mn der ~gyptologie (Wiesbaden: Otto Flarrassowitz, 1985). 6:407-11, 11. James legge. I Ching: Bmk of Change5 (1888; Secaucus, W.J.: Citadel Press, 1964), 7.

12. C. G. Jung, foreword to I Ching or Bwk of Changes, trans. Cay F. Baynes from a German translation by Richard Wilhelrn (1950; London: Arkanapenguin, 1989). Ivii-twi.

13. Raymond Queneau, Cent MilIe Milliards de Pohes (Paris: Gallimard, 1961).

14 .William Paulson, "Computers, Minds, and Texts: Pretiminaly Reflections," New literory History 20 (1989): 297. 15. For critical views of political claims about etectronic media i n general, see James W. Carey, Communication As Culture: Essays on Media and Society (Boston: Unwin Hyman, 1988). especially "The Mythos of the Electronic Revolution," and about hypertext, Stuart Moulthrop, "You Say You Want a Revolution?: Hypertext and the Laws of Media," Postmodern Culture 1 (May 1991). -

16. Theodor Holm Nelson, Litera@ Machines, ed. 87,1 (Swarthmore, Pa.: Theodor H. Nelson, 1987). and Vannevar Bush, "As We May Think," Atlantic Monthly 176 (July 1945): 101-8. 17. Michael Joyce, Afternoon, a stay (Cambridge, Mass.: Eastgate Systems, 1990). For discussions of Aflernoon, see Stuart Moulthrop, "Hypertext and 'the Hyperrear Hypertext '89 (New York: Association of Computing Machinety, 1989), 259-67; 3. David Bolter, Wtiting Space: The Computer, Hypertext, and the History of Writing (Hitlsdale, N.J.: Lawrence Erlbaum, 1991); and Landow,

18. Moulthmp, "Hypertext and "the HyperreaL* 239.

29. Inger Christensen, The Meaning of Metofirn'on: A Cn'tical study of SeIected Novels by S t e m Nabokov. B ~ r t h , and Becket (Berg, Universitetsforlaqet, 1981). 1-41-43.

20. Bolter, Wrih'ng S p a 7.

21. Jon Lanestedt, Episk Pmgmmvare--En Pittemr 50p91are-R Literary Text Type?] (c.phiL diss.. University of oil*, -

198s). William Crowther and Don W o d s , Adventurn, this version implemented by Gordon Letwin (I'BM/Microsoft, 1981).

en:

h i r

22. Anthony J. Niesz and Norman N- Holland. 'Intemckive h '~on," Critic01 Inquiry 11 (1984): 113. 23. Robin WaterField and Wilfred Da*. The M o w Spider (London: Penguin 1988). 24. See James Richard Meehan, The M e t o m k Wrih'ng Storjes ~ry. Computer (Ph.D. diss., Yale University, University Micdtms International 1976); Roger C. Schank and Peter Childers, The Cognitiw Computer: On Lorrguup, Learning, and Am'fitiol h:eNigence (Reading, Mass.: Addison-Wesley, 1984); and Michael hbowi'tz, "Creating Characters in a Story-telling Universe," pwficr 13 (1984): 171-94.

25. Brenda LaureL Computers as T k t r e (Reading, Mass.: Add&"- Wesley, 1991). 26. Cf. Carey, Communication AF Cullure, ch. 8. 27. Marshall M d u han. Undmtanding Meditx The Extensions of Man (New York: Penguin Books/Merrtor, 1964), 225.

28. Elizabeth M. Reid, Eledmpolis: Commwnr'wtion and Communrty on Internet Rely Chut (honors thesis. University of Metbourne, 1991). 29. Richard Bartle, Interactive Muh-User Computer Gumes (parts from a research report cornmissiorred by Bn'hh Telecom, disseminated by the author, 1990).

30. Thomas Pynchon, The Cy'ng oflot 49 (London: Picador, 1967).

31.Umberto ECO, A Theow of Semiotia (Blmmington: Indiana University Press, 1979), 9.

3 2. Louis Hjelrnslw, Pdeqom~omeoo lo a Thwy of language (1943). trans. Francis J. Whiffield (Madison: University of Wisconsin Press, 1961). 30. 33. Pierre Fontanier, Les Figures du Discours (1821-30: Paris: FLarnmarion, 1968), 271.

34. Howard Rheingotd, Virtual ReatiQ (New York. Summit Book. 1991)- 35. See also Richard Ziegfeld, "Interactive Fich'on: A New Literary Genre?" New Literary History 20 (1989): 341-72. 36. George E, Marcus and Michael M. J. fischer, Aothropolog as Cultural Critique: An Experimental Moment in the Human Science* (Chicago: University of Chicago Press, 1986).

37. Clifford Geertz, Works and Lives: The ~ n t h r o p o l o g i ~ t 0s Author (Stanford: Stanford University Press, 1988). 38. David Tomas, "Old Rituak for New Space," in Cyberspace: First Steps, ed. MFchaeZ Benedik (Cambridge: M I 1 Press, 1991)~ 31-47n

Nomadic Rower and Cultural Resistance Thc C r i t i d Art Ensemble turned the rhizome on its root. ThC nomadology, fashioned by Gilles Deleuze and Fklix Guattari (C inspiration for d t u r e workers-from the netart mailing list

critique. Rhizome language has been adopted as 3 alternative to the Mew Left that in retrospect seems to have d essay the CAE posits the rhimme and nomad-as used by ma network andh~pertcxtud technologies-as figures of torpor,

That the applications of rhizome language to new media ta

misreading of Deleuze and Guastari is o f little importance in t meaning through use. By co-opting the Bangrage of the rhizon technohhtionists. CAE cut to the heart of the argument th; inherently incompatible with the power relations of t h e indus the idea that it is impossible for power to co-opt network and technologies have a manifest destiny of freedom The CAE awed: that the power elite are now the primary I

that these technologies allow them. as well as the sites of indu to make resistance in physical space inefectud: that the cornr the elite are now Mly cyberspad. so cyberspace becomes the that the only potentially effective cyberspace resistance is "dis information technolcgies and the potenkid panic created by t resulting destalsiilization of the safe harbors cxeatec

employ). Since the mid-1990s, significant flaws have been found in

these is that corporate power continues to feel a need to press World Trade Organization). and continues to have its meetin! Meanwide, protest movements haw found effective means t( their own communication, and have developed the nomadic a the street wherever it chooses to gather. The power elite have trapped in meeting-mom bunkers, rather than free to move a

Yet the question of how, or whether, to use new media tech rather than organizing and reporting (physical) action, remail work that positiom the network as a space for action has bee and "cyberhippies." But is unclear whether this work i s actuau gathering signatures on a petition. It does not close down spa public presence the way physical sit-ins have. It does ~rovidc ; people to express personal dissatisfaction with power, as pe6' of personal (physical) risk. Yet this work goes beyt~nd the peti which exist both for the participants and for their audience-

CAE, for their part, have weighed in to the debate by declar even the physical sifiin has almost no tactical value, and that main tactically negative (even if they may possess small &a: