N. Katherine Hayles English Department Los Angeles CA ...

N. Katherine HaylesEnglish DepartmentUniversity of California, Los AngelesLos Angeles CA 90095-1530

Narrating Bits: Encounters between Humans and Intelligent Machines

Thanks to Lev Manovich’s seminal discussion in The Language of New Media,

database and narrative are now often considered to be in a competitive relation with one

another. Important as Manovich’s analysis has been in launching a productive line of

inquiry, his construction of the narrative/database dyad is nevertheless plagued by certain

intractable problems. The centrality of the narrative/database dynamic to new media, and

especially to electronic literature, makes rethinking it an urgent concern. Addressing the

problems in Manovich’s analysis opens the way to a different view of the relation and a

crucial re-positioning of it. In this essay I propose an alternative interpretation of the

narrative/database configuration as a dynamic between narrative and a new term that I

call possibility space. I develop the epistemological and ontological implications of this

configuration through close attention to narrative theory and expand the theoretical scope

of interactions in which the dynamic can engage. The result, I argue, is a more nuanced

and powerful set of terms with the potential significantly to revise contemporary

understandings of narrative. In addition, the re-defined and re-positioned dynamic proves

a flexible, wide-ranging framework through which to understand experimental

contemporary practices in both print and electronic literature.

One set of problems arises from slippages in the way Manovich deploys the term

“database.” His initial usage references database in its technical sense: “database is

defined as a structured collection of data. The data stored in a database is organized for

fast search and retrieval by a computer and therefore, it is anything but a simple

2

collection of items” (218). As his argument progresses, however, he moves from this

technical definition to a much broader sense of the term, as when he acknowledges that

“not all new media objects are explicitly databases” (221), but he nevertheless treats

them as such by conflating databases with “data structures” (223). The ambiguities

deepen when he announces that “as a cultural form, the database represents the world as a

list of items, and it refuses to order this list” (225), a characterization at odds with his

earlier recognition that a database must order its terms according to specified parameters

to be searchable. In his argument’s later stages, databases become indistinguishable from

data, as when he claims that narratives within new media environments “on the level of

material organization . . . are all databases” (228). As a result of these slippages, he is

able to claim that databases can be found in any computer environment whatever, making

them virtually ubiquitous.1

Narrative, by contrast, is by and large limited in Manovich’s description to its

appearance in formal art objects such as novels, films, and to a lesser extent computer

games. Only so is he able to conclude that the surprise is not the pervasiveness of

databases but “why the other end of the spectrum—narratives—still exist in new media”

(228). This construction ignores the pervasiveness of narrative in everyday life. Jerome

Bruner, for example, highlights research indicating that a mother typically tells her young

child some kind of narrative about twenty times an hour, or roughly once every three

minutes. Narrative figures prominently virtually everywhere humans communicate and

congregate—in TV news, Sunday sermons, family reunions, coffee breaks, city buses,

gossip, and much more, so it is no surprise to find it everywhere on the Web as well,

3

from chat rooms and blogs to news sites and email. Moreover, Manovich claims that

new media objects “on the level of material organization” “are all databases” (11).

In human history, narratives are much older than electronic databases and

considerably older than databases in any form. Many anthropologists argue that

narratives may be nearly as old as humanity itself, as suggested by the narratives

structures emerging from pre-history through myth, epic poetry, and creation stories.2

One of the purposes narrative may have served in human evolution is its implicit ability

to create models of other minds and, more generally, models of the world. In its

emphasis on causality, agency and temporal progression, narrative provides not just

specific explanations but frameworks for explanations that allow people to understand

and predict how other people and the world around them will act and react. Seen in this

light, narrative in its historical and evolutionary role was a powerful tool for shaping as

well as expressing human subjectivity and sociality—and it remains so today.

By contrast, databases are much newer and depend for their widespread

importance on the ability to collect and organize data as well as transmit, search and

retrieve it. Although one could point to encyclopedias dating back to the classical period

as early forms of databases, as Manovich does, it was not until the eighteenth century and

the birth of statistics that databases in their modern sense emerged in census records,

medical data, and the like.3 Moreover, it was not until the twentieth century, and

especially the advent of the digital computer, that databases became central cultural

forms.4 Although databases are indeed pervasive in the contemporary period, as

Manovich says, they should not be seen as identical with data. On the contrary, the very

qualities that make them indispensable for some problems also virtually insure they will

4

be more or less unsatisfactory for others. No sooner is a database established than users

discover its limitations, for it allows searches only according to the specified parameters,

and users invariably discover other ways in which they want to interrogate the data in

ways impossible with the given database form.5 Connolly and Begg, for example, in

their influential textbook Database Systems give “some cold facts about database

software development projects by corporations,” including that 80-90% do not meet their

performance goals, 80% are delivered late and over budget, 40% fail or are abandoned,

and less than 40% fully address training and skills requirements (270).

Although humans design and build databases, databases are brought into the

world as artifacts through technologies of data compilation, storage, transmission and

retrieval, which have their own constraints and possibilities deeply affecting how

databases are built. In this sense the database has two parents, as it were, and both affect

the offspring’s nature—the humans whose needs the database serves, and the machines

that create and implement it.6 One way to look at the contemporary database, then, is as a

site where the encounter between humans and intelligent machines is performed and

enacted in precisely defined ways that are, moreover, designed for mass consumption

rather than creative uses. The database is a cognitive framework for dialogue between its

two parents, a staging ground where two very different modes of thought interact,

sometimes productively, sometimes with frustrating miscommunication and inefficiency.

But databases are only part of the story. Paraphrasing Donald Rumsfeld’s

notorious remark, we can characterize information using three categories: data that are

known (in his redundant formulations, known knowns), data that are known to exist but

about which no information is currently available (known unknowns), and data that are

5

neither foreseen in general outline nor known in particular (unknown unknowns).7 A

database contains only data that is known and instantiated in a particular form. It does

not contain other information expressed in modes of organization foreign to its specific

structure, nor does it contain anything currently thought but unknown in particular, or

even more revealingly, anything that is both unthought and unknown.8 The database, in

short, represents only a small slice of actual and potential data.

A more capacious term, possibility space, allows room for all three of these

epistemological categories. In a narrative context, knowns correspond to the content of

the narration. Known unknowns refer to outcomes possible within the world the

narrative delineates but not articulated within a given narrative. For example, someone

chooses the door concealing the lady, which leads to a certain outcome; although he does

not choose the door concealing the tiger and so that outcome is not narrated, it is a known

possibility and hence a known unknown. Unknown unknowns refer to possibilities that

could appear within a narrative system but are neither explicitly articulated nor

anticipated as possible outcomes. Any narrative system capable of generating emergent

phenomena can create unknown unknowns, for the very idea of emergence implies that

the interaction of multiple agents and components can result in unanticipated behaviors at

the global level that can not be predicted from the parts alone. Computer simulations

generating narrative content are examples of narrative systems that can produce emergent

phenomena if appropriately structured. These systems are data generators, not

databases.

Shifting the emphasis from database to possibility space enables both a more

precise and deeper comparison with narrative. Because the term is capacious, it allows us

6

to understand in broader ways its epistemological categories and their ontological

implications. In addition, it encourages us to think of narrative not as a natural enemy but

a co-evolution partner in creating complex conceptual ecologies in which many different

kinds of relationships are possible. These include competition, of course, but also

cooperation, symbiosis, parasitism, hyperparasitism, and so on. Having sketched out the

functions that the term “possibility space” can serve, I turn now to compare its

implications with the assumptions embedded in traditional narrative theory. As we shall

see, juxtaposing possibility space with narrative suggests crucial modifications when it is

conceived as the “other” of narrative.

Assumptions Embedded in Narrative Theory

The binary established by the Russian formalists of fabula and sjuzhet followed

the distinction, dating back to what Gerard Genette calls the “pre-history” of narratology,

of story and plot.9 Mieke Bal defines fabula as the “material or content that is worked

into a story,” while the story itself is “defined as a series of events” (7). This definition is

more or less echoed by Genette, Seymour Chatman, Shlomith Rimmon-Kenan, and

others. The sjuzhet, on the other hand, is the order of appearance of the events in the

work itself, or at Chatman, quoting Boris Tomaskevsky, puts it, “’how the reader

becomes aware of what happened’” (20). Different theorists transpose these older terms

into binaries with slightly different inflections, including story and discourse (Chatman),

fabula and story (Bal), and story and narrative (Genette), who sees both these terms

deriving from a third term, narrating. As these examples show, there is no consistent

terminology; Chatman’s story corresponds to Bal’s fabula, while her story corresponds to

7

his discourse. In the interest of clarity, therefore, I will use the older formalist terms

fabula and sjuzhet, which at least have the virtue of avoiding confusion.

Virtually every major theorist of narratology has emphasized that only the sjuzhet

is literally present in the text; the fabula is a contingent construction created by the

competent reader, aided by her tacit knowledge of narrative and cultural codes.

Complementing the claim by structuralists that fabulas show consistent deep structures

across genres and time periods (folk tales, myths, dramas, etc. in different eras),

narratologists tend to focus on sjuzhet, either implying or explicitly claiming that the

form of the narrative comprises a system capable of signifying meaning, over and above

what the fabula or content of the narrative conveys. Genette is cautiously emphatic on

this point: while many critics concern themselves with the analysis of narrative content,

narratology as such, considered as a theoretical discourse, “remains (provisionally?) the

property solely of the analysts of narrative mode” (1998, 16). So much is this the case

that Rimmon-Kenan, writing a chapter on the fabula (in her terms, story), feels compelled

to explain her decision to treat the fabula in isolation. “Users cannot produce or decipher

stories without some (implicit) competence in respect of narrative structure, i.e. in

something which survives paraphrase or ‘translation,’” which fact she uses to justify as a

“working hypothesis” the “preliminary assumption that story-structure or narrativity is

isolatable.” She is quick to add, however, that this “does not amount to granting any

undisputed priority, whether logical or ontological, to story over text (if forced to decide,

I would rather opt for the latter)” (8).

Nevertheless, as Rimmon-Kenan’s “working hypothesis” suggests, the readerly

activity of constructing a fabula from the sjuzhet is ubiquitous in literary criticism and in

8

reading generally. Once constructed in the reader’s imagination, the fabula is then, in a

backward feedback loop, used to interpret and understand the sjuzhet. Consider Benjy’s

narrative in Faulkner’s The Sound and the Fury; has anyone made sense of this sjuzhet

without constructing a fabula that allows a translation of sorts to take place? Many other

narrative fictions, particularly postmodern and experimental novels, illustrate the point,

including such major works as Thomas Pynchon’s Gravity’s Rainbow and Joseph

McElroy’s Plus. Even narratologists rely on this dynamic; the very fact that they

continually remind their readers that the fabula does not exist in the text testifies to the

strength of the illusion. Less disciplined by narratological theory than these critics, many

readers derive from their work of fabula-construction the inference that the sjuzhet is

somehow the “distortion” of the fabula that must be straightened out fully to comprehend

the text. In this sense, the fabula is seen as more robust, more ontologically grounded,

than the sjuzhet, which is interpreted as a view of the fabula seen through a lens that

never renders it exactly as it would be “in reality.”

The scare quotes are meant to gesture toward the obvious, that the “reality” of a

narrative fiction is an invention. Yet in the emphasis on causality, agency, and

consistency of the relation between narrating and narrator, narratives mimic those

qualities of reality that make it susceptible to systemic analysis, which is to say, those

qualities that make reality “real.” Chatman is eloquent on this characteristic of

narratives. Asserting that narratives have qualities of wholeness, transformation, and

self-regulation, he argues that every transformation must yield another possibility within

the narrative system and not something outside that system. “Only certain possibilities

can occur. Further, the narrative will not admit events or other kinds of phenomena that

9

do not ‘belong to it and preserve its laws.’ Of course certain events or existents that are

not immediately relevant may be brought in. But at some point their relevance must

emerge, otherwise we object that the narrative is ‘ill-formed’” (21-22). Contrasting

narratives with random bits of cocktail chatter, he says that if we assembled such

excerpts, we “would clearly not have a narrative (unless we insisted upon inferring

one . . . ). The events in a true narrative, on the other hand, ‘come on the scene as

already ordered,’ in Piaget’s phrase. Unlike a random agglomerate of events, they

manifest a discernible organization” (21). In this view, orderliness is not merely inherent

in the form of narratives but actually constitutes a narrative as such. Whatever a

narrative’s content, then, its form implies a stable ontology for the arena in which it

operates and in this sense generates a more or less stable ontology expressed through the

reader’s construction of its fabula.

Moreover, the inextricable entwining of fabula and sjuzhet guarantees that this

will always be an ontology constituted through, and indissociable from, the subjectivity

(or subjectivities) “seeing” and “saying” the world. In the very act of interpreting the

“distortions” that the sjuzhet implies, readers interpolate another assumption, namely that

those “distortions” shed light on the perceptions and mind of the narrator and/or

focalizer, as appropriate to the narrative situation. Indeed, Mieke Bal puts this activity at

the center of her theory of narrative. “To talk about narrators . . . is to impute agency to a

subject of narration, even if this subject is not to be identified with the narrator. Actors,

in the fabula, are the subjects of action. This attention paid to subjectivity is, indeed, the

basic tenet of the theory presented in this book. It is meant to insist on the complex

manner in which narrative communicates” (11). Moreover—and this is so obvious that

10

it falls below the radar of most narratologists—the subjectivity constituted through the act

of narrating is presumed to be human (leaving aside those rare narratives that purport to

be written with a non-human narrator and/or focalizer, in which cases it could be argued

that they inevitably import anthropomorphic assumptions and indeed must to so to be

legible to human readers). The human sensibility informing narratives comes close to the

surface in Genette when he puts himself on record as denying the possibility that there

could be a narrative without a narrator.

Narrative without a narrator, the utterance without an uttering,

seems to me pure illusion. . . . I can only therefore set against its

devotees this regretful confession: ‘Your narrative without a narrator

may perhaps exist, but for the forty-seven years during which I have been reading

narratives, I have never met one.’ Regretful is, moreover, a term of pure

politeness, for if I were to meet such a narrative, I would flee as quickly

as my legs could carry me: when I open a book, whether it is a narrative or

not, I do so to have the author speak to me (1998, 101).

To recapitulate where this (de)tour through narratology has taken us, we can say

that narrative through its conjunction of fabula and sjuzhet implies the existence of a

more or less stable ontology, whatever its content; that ontology is inextricably tied up

with the subjectivities of the minds seeing and saying the world; and those subjectivities

are human (or human-inflected). No doubt these are some of the qualities Manovich had

in mind when he argued that database and narrative are natural enemies. What happens

to this dynamic, and the assumptions embedded in narrative theory and narrative itself, if

we juxtapose narrative not to database but to the possibility space? As I shall argue, the

11

juxtaposition puts pressure on all three of the major assumptions noted above. At the

same time, it creates new kinds of transformations that break open the “closed” system of

narrative as it has been conceived in narrative theory. The result? Either a revision of

narrative theory, or the admission of hybrid forms of which narrative theory has not taken

sufficient account. Genette gestures toward this possibility when he reminds us, “Every

day Nature and Culture breed thousands of ‘monsters’ that are as fit as fiddles” (1998,

129).

Versions of Possibility Spaces

As indicated earlier, one version of a possibility space contains “known knowns.”

Perhaps the most pervasive instance of this version is the database. When database is

combined with narrative, it distorts the narrative form and pushes it toward the threshold

of being “ill-formed,” at which point theorists such as Seymour Chatman question

whether it can remain a narrative at all. Recall Chatman’s suggestion that the far side of

that threshold can be represented by the randomness of cocktail chatter bits that have no

relevance to each other. Suppose that we form a database of such conversational

fragments and put them into dynamic interplay with narratives. The result would be

something like Stuart Moulthrop’s Reagan Library, a hybrid form between narrative,

game, and database. In Reagan Library, a user’s first visit to a screen results in a “word

salad,” a mélange of fragments. Most of the fragments have been randomly assigned to

the screen by an algorithm drawing on a database of such fragments. Some of the

fragments, however, are anchored to that screen and become part of a coherent narrative

that gradually emerges as the user visits the screen a second, third, and fourth time (small

12

squares at the screen’s bottom indicate how many times that screen has been visited).

Upon the fourth visit, the screen consists of a coherent narrative that will not change

further during that play of the work. This technique creates a kind of half-full/ half-

empty ambiguity, for one can either view the “word salad” as richly suggestive noise

stimulating creative possibilities, or an annoyance to be zipped through as soon as

possible to as to get to comprehensible narrative. The characters narrate their stories in

split screens, the bottom half of which are devoted to text, while the top halves are

navigable images displaying clickable icons amidst enigmatic landscapes. There is no

indication which icon leads to which of the four world levels (designated by the colors

red, black, white, and blue), so navigation has a random element recapitulated verbally by

the “word salad” of the database fragments. In performing a boundary between database

and narrative that the user crosses and re-crosses many times in the course of viewing the

work, Reagan Library foregrounds the respective pleasures and functionalities that

database and narrative can offer. The database fragments gesture toward a large body of

cultural material and invite the reader to speculate on its significance to the work; the

narratives give the pleasure of participating in the virtual worlds of the characters.

Since the database fragments operate according to an algorithm, the work also

enacts an encounter between the human desire for stories and machine intelligence. The

dynamic is suggestive of a zero-sum game, for as the narrative coherence increases, the

database fragments decrease. This dynamic is complicated by the “Notes” screens, which

give crucial information about the cultural, informatic, and linguistic contexts of the

narratives. The more the user visits the “Notes” screen, the more the information on the

screen disappears, thereby setting up an economic exchange in which fullness of

13

information must be paid for by replacing replicable machine memory with fallible

human recall (that is, as the information disappears users are forced to fall back on their

recollections of previous screens rather than have the computer “remember” the

information for them). It is not a coincidence that individual and cultural memories are

crucial issues in this work. Named for a victim of Alzheimer’s, a national institution of

remembrance, and a storehouse of cultural documents, the work has many narrator-

characters whose memories are impaired or altogether gone, as if miming for us the

erasures of cultural and collective memory that the Reagan era represented for many

progressive thinkers.

Other versions of possibility spaces are hypermedia/hypertext fictions using links

and layers to create multiple reading pathways. A fairly conventional link strategy is

employed in Caitlin Fisher’s These Waves of Girls, winner of the 2001 Fiction Award

from the Electronic Literature Organization, which narrates coming-of-age stories for

lesbian girls. Linking goes between text passages used in combination with frame

graphics, sound files, and images; navigation is controlled by sidebar menus and linked

words and images within the text. Strictly speaking, the work does not use a database as

such to provide text, images and sounds, but rather Flash files embedded within html

encoding. The effect is primarily narrative; machine intelligence enters only to

complicate and extend the narrative possibilities by allowing multiple reading pathways

and different narrative orderings. Significantly, the narrative content is also

conventional, in the sense that the narratives are focalized on characters whose

humanness is never in doubt.

14

Fisher’s text in this sense contrasts with Moulthrop’s, for in Reagan Library the

characters, with their memory losses and gradual emergence amidst “noise” that seems to

come from outer space, already gesture toward the posthuman. Texts like Moulthrop’s

are cousins, if not siblings, to Oulipo works that use algorithms and textual constraints to

generate narratives. Whether implemented in the computer or not, algorithms embody

mathematical procedures and thus are foreign to the kinds of human perceptions

represented in traditional narratives. It is no accident that writers interested in breaking

out of the circumference of human perception generate texts that are often hybrids

between database and narrative. Georges Perec’s Life: A User’s Manual (La vie mode

d’emploi), with its exhaustive database-like catalogues of the contents of apartments,

bookshelves, coffee tables, wall decorations and basement storehouses, employs a

narrator who is made to disappear as much as possible into the anonymous, unlocatable

voice that Genette identifies with zero focalization (73). Another example is John Cage’s

use of “chance operations” in his linguistic compositions. Murearu, for example, was

composed by subjecting passages from Henry Thoreau’s journals (particularly those

having to do with music, silence, and sound) to I Ching operations, which were also used

to determine the pronouns of the focalizers and the varying size and nature of the

typefaces. The point of such “chance operations,” from Cage’s perspective, was to do an

end-run around human intentions and thereby open the composition, and the composer, to

non-human forces greater than human desire and agency could encompass. Even so

human-oriented a critic as Genette feels the appeal of such procedures. Quoting from

narrator’s comment in Borges’ Library of Babel that “’It is enough that a book be

possible for it to exist’” (121), Genette contemplates working out a table of combinations

15

in which all possibilities for narrative would be exhausted. After commenting that “This

proposition will not, I hope, be taken too literally,” he continues: “For me, what is

important about it is not this or that actual combination but the combinatorial principle

itself . . . [creating] constellations in which every parameter can a priori come into play

with every other . . . without being in too much of a hurry to proclaim definitive

incompatibilities” (1998, 129).

The implication that narrative is human-bound while database gestures toward the

non-human can serve as a transition from works using “known knowns” to those focusing

on “known unknowns” and, at still further remove, “unknown unknowns.” For example,

Talan Memmott’s Lexia to Perplexia crosses and re-crosses the boundary between

“known knowns” and “known unknowns.”10 Unlike the frame and link structures of

These Waves of Girls, where controlling the text is easy and links are clearly identified by

different colored text, Lexia to Perplexia uses multiple layers and animations to overlay

text and graphics onto the initial textual surface, often obscuring text underneath. The

cursor is very “nervous,” responding to minute motions that may not be intentional on the

user’s part, evoking the feeling that this text is not entirely under human control.

Moreover, the layered and “hidden” nature of many passages makes it likely that any

given user will not be able to discover all of the text, introducing the possibility of

“known unknowns”—that is, passages that are knowable in principle but not in practice.

As the text moves beyond narrative into a hybrid form combining narrative fragments

with graphics, cryptic announcements, and elaborate screen designs (all signifying textual

components but not narrative in the usual sense), the textual content plays with voices,

still recognizably human, anticipating a transformation into hybrid subjectivities where

16

“local” bodies connect to “remote” bodies by mingling with the computer apparatus in a

process appropriately called “remotional” (Lexia to Perplexia/Double Funnels).

Other examples of texts generating “known unknowns” are simulations that use

interactive functionalities to create new (and therefore unknown) combinations of plot

developments and outcomes. The simulations differ from the multiple reading paths

created by links because they are not prescripted in advance but rather emerge

interactively through encounters between virtual agents. Jonathan Gratch has developed

software to create such emergent outcomes, one version of which generates on-the-fly

narrative dialogues between two virtual roommates using algorithms to satisfy their own

goals while also responding to each other’s desires. Such variables as

competitiveness/cooperativeness and aggression/passivity can be adjusted to achieve

different outcomes via different routes. Other versions of the software are used in

military training simulations created by Gratch and his co-authors under the auspices of

the Institute for Creative Technologies, a joint venture of the University of Southern

California and the U. S. Army, in which live participants interact with virtual agents to

arrive at different outcomes for various military scenarios.

Michael Mateas’s interactive drama Façade, a collaboration with Andrew Stern,

creates similar “known unknowns” by using knowledge-based conversational agents

leading to multiple emergent possibilities. Scripted in ABLE, a programming language

Mateas developed, Façade’s narrative progresses through a series of dramatic “beats,”

with each “beat” comprising a narrative unit. Façade is an especially interesting example

because it embodies the design constraint of following a traditional narrative arc with a

clear beginning, middle, and end. Contrasting his model with work like Gratch’s that

17

allows strong autonomy for the virtual agents, Mateas notes that with strong autonomy,

“agents chose their next actions based on local perception of the environment plus

internal state corresponding to the goals and possibly the emotional state of the agent”

(2003, 2; see also Mateas, “Expressive AI; A Hybrid Art and Science Practice,” 2001).

The stronger the agent’s autonomy, the less likely it is that the story as a whole will be

well-formed as a narrative. Mateas made the decision to create instead “weakly

autonomous characters which are tightly coupled to a more deliberative story-guidance

framework while still maintaining the reactivity necessary for real-time interaction”

(2001, 2). The agents, in other words, can act only within an overall AI architecture that

ensures the narrative will have a recognizable and indeed rather traditional shape. The

decision is part of his overall project of “building an image of the human in the machine”

(2001, 1). “Expressive AI” is in his view “a way of exploring what it means to be

human by building systems,” a perspective that assumes from the outset narrative

coherence and form (2001, 1). Rather than push the boundaries of what it means to be

human by creating hybrid forms that enact the encounter of humans with intelligent

machines, his work constrains the boundaries within which the virtual agents can act so

as to guarantee that recognizably human narratives will emerge.

Further along the spectrum are works using self-evolving programs to create

“unknown unknowns,” where even the AI architecture can evolve in ways not foreseen or

predicted by its creators. With genetic programs able to change their own algorithms,

there is no guarantee that anything like traditional narrative forms will emerge. Indeed,

part of the point of creating such simulations is to open the possibility that something

entirely unexpected will be created. At present there are, to my knowledge, no verbal

18

simulations that meet this criterion. There are, however, visual narratives that have

elicited verbalizations of the narrative progression constituted by animated images (and

underneath the images, the evolution of the code on which the images are based).

Perhaps the most successful is Karl Sim’s video animation Evolving Virtual Creatures

showing virtual creatures engaging in game-like activities to control an object that

resembles a hockey puck.11 Both the morphology and neurology of the creatures have

self-evolved, with neither being explicitly determined by the creator. The program uses

artificial selection to test each generation of creatures as they strive to achieve some goal.

Specified fitness criteria determine which of the creatures is most successful. Variations

of this creature’s “genetic code” are then calculated to create the next generation of

creatures, which undergoes similar testing and selection. When the different species of

creatures are then put in competition with one another, the result is a contest that follows

a recognizable narrative arc of having a beginning, middle, and end. Unlike Mateas’

Façade, however, the narrative emerges spontaneously, its recognizable trajectory being

a result of the game-like context. Moreover, in Sims’ video compilation of this work as a

whole, a larger narrative arc focusing on the creatures’ development emerges as they self-

evolve to become increasingly efficient and sophisticated in their designs.

Part of the difficulty of creating a verbal parallel to Sims’ Evolving Virtual

Creatures comes from linguistic constraints on syntax and grammar.12 Whereas any

motion that works in the virtual environment is permitted in Sims’ simulation, for a

language-based simulation, variants resulting in verbal nonsense would far, far outweigh

linguistically competent utterances, making undirected evolution a much trickier

proposition than it is for Sims’ virtual creatures. Nevertheless, the idea of an intelligent

19

machine that could self-evolve along lines unforeseen by its human creators, while still

retaining the capacity to communicate to humans, has been imagined by Stanislaw Lem

in his novella “Golem XIV.” Indeed, the whole point of that narrative is to create the

simulacrum of an intelligence whose nature cannot be grasped by its human interlocutors,

although the text’s language must necessarily assert rather than perform this fact.

As the technology moves closer to being able to make the transition from known

unknowns to unknown unknowns, the possibility space grows larger and, as such,

potentially puts more pressure on traditional narrative form. Even in its more

conservative guise as a space of known knowns and/or known unknowns, however, it still

has the capacity to interact with narrative in ways that significantly shift the grounding

assumptions constituted by the traditional critical dyad of fabula and sjuzhet. Having

fleshed out the concept of possibility space by the preceding examples, I turn now to

consider its epistemological and ontological implications.

Contrasting Possibility Space with Narrative

In contrast to narrative, possibility space is likely to be governed by

combinatorics rather than linear causalities. In hybrid or complex forms both linear

causality and combinatorics are usually in play, so the distinction is one of emphasis and

priority rather than absolute separation. In an evolving adaptive simulation, for example,

there is generally some kind of causal relation between one generation of virtual creatures

and another, but combinatorics also enter strongly into the algorithms. In Sims’ Evolving

Virtual Creatures, for example, the genotype of one generation is used as the starting

point for variants—determined by combinatorics—of the second generation. Narrative,

20

on the other hand, is fundamentally and centrally causal. Brian Richardson in Unlikely

Stories: Causality and the Nature of Narrative voices a consensus view when he writes,

“Cause is one of the most significant and fundamental aspects of narrative” (14), even if,

as he goes on to say, its subtleties and complexities are under-theorized in narrative

criticism. When narratives “plot against probability,” as Richardson notes of postmodern

fictions such as Angela Carter’s Wise Children and Bharati Mukherjee’s Jasmine, they

approach the threshold of chance events from the presupposition most readers will have

that somehow these events will turn out to be causally related after all. In general, we

can say that the more works move along the spectrum from known knowns to unknown

unknowns, the more important combinatorics are likely to be in generating the possibility

spaces in which a large number of unforeseen variants, events and outcomes can arise.

Related to the central role of causality in narrative are its evolutionary and

historical functions of creating frameworks that allow humans to interact robustly with

their environments, including crafting models of how other minds and the world in

general work. Whereas narrative capitalizes on and reinforces human presuppositions

that make the world make sense, possibility space carries the scent of the non-human, the

algorithmic, the procedural, the machinic. The contrast can be seen by recalling yet again

the narrator’s comment in Borges “Library of Babel,” “It is enough that a book be

possible for it to exist.” Unlike narratives in the real world, a book in the Library of

Babel need not make sense (as indeed the overwhelming majority do not), and it need not

convey to readers anything about how the world works. Considered as a representation

of a possibility space in which all the possible combinations of alphabetic symbols are

21

instantiated in books, the Library of Babel shows with cruel clarity how inhuman such a

combinatorics would be for human purposes and desires.

As long as narratology is dominated by the theoretical dyad fabula-sjuzhet

(including the variants discussed earlier of story, discourse, narrative, narrating, etc.), it

retains an almost unavoidable presupposition of realism embedded in narrative form

(whatever the narrative content), for the fabula is related to the sjuzhet through the

assumption that the sjuzhet takes place within the storyworld expressed through the

fabula. Of course, although the fabula logically precedes the sjuzhet, the sjuzhet factually

and chronologically must precede the reader’s construction of the fabula. Nevertheless,

according to the testimony of the innumerable readers and most writers, the construction

of the fabula remains an indispensable aid in creating, representing, and interpreting the

sjuzhet.

What happens, then, if the sjuzhet is understood to be generated from a possibility

space rather than a fabula, as in the (admittedly extreme) case of the hypothesized books

in the Library of Babel? Readers may and probably will continue to construct a fabula

(as the narrator in Borges’ fiction does to explain the existence of the Library), but the

power of this construction to convey an ontology is weakened. Underwriting the

existence of the sjuzhet is not the assumption of a prior storyworld, with its more or less

convincing ontology, but merely the operations of a possibility space running through all

possible permutations, some of which are understood as being realized in the sjuzhet.

Consequently, as the possibility space cooperates, competes, and otherwise engages with

narrative to create fictions, readers move from the relative ontological security of the

fabula◊sjuzhet inference to the more ontologically unstable progression: possibility

22

space◊sjuzhet(◊fabula). Unlike trying to imagine an infinite storyworld—a project

before which the imagination trembles-- it is quite easy, as Borges shows, to imagine an

infinite possibility space. Given the information the narrator of that fiction provides, the

number of books in the Library of Babel can be calculated to be of the order of

magnitude of 1075, a quantity greater than the number of atoms in the universe. Although

one may easily write this number and calculate it almost as easily, it is quite impossible

for the imagination to encompass what such a magnitude means in human terms, because

it so far exceeds the compass of normal life and experience. No wonder, then, that many

inhabitants of the Library choose to throw themselves down the infinite stairwells in

despair, for they are trapped in a truly inhuman space.

Although the possibility space can have such ominous overtones, it can also be

understood as opening the human to the unthought and unrecognized otherness of a

universe much bigger than human conception can hold. This is the promise of emergent

adaptive simulations—that we can program our machines to create what we ourselves did

not conceive as such. Even databases embody a weak form of this promise, for they

order and make searchable and retrievable amounts of information that would overwhelm

unaided human perception. An optimistic interpretation of the infiltration of narrative by

possibility space, then, is that humans are now able to go beyond their evolutionary

inheritance into realms that cannot easily be imagined or represented by the human mind

alone. When the ancient technology of natural language is combined with the capabilities

of intelligent machines, ontology may become unstable but the scope of epistemology is

increased, potentially immeasurably so.

23

Complex Ecologies

As mentioned earlier, the range of interactions between narrative and possibility

space extend far beyond the antagonistic and competitive relation that Manovich suggests

for narrative and database. Jason Nelson’s Dreamphage illustrates the complexity of

these interactions and the subtleties of their dynamics. The work is introduced by an

initial screen, narrated by Dr. Bomar Felt, setting up the work’s premise: an unknown

virus is infecting the human population, its telltale symptom being monologic dreams the

victims experience. Night after night, an infected subject dreams the same dream, which

appears with increasing frequency; within a few weeks, the disease progresses to coma

and the death. The clue to the virus’s nature, Dr. Felt suggests, may lie within the

dreams.

Proceeding from this screen, the user encounters a series of framed rectangles

within rectangles within rectangles, within which swirl images of pages sometimes

partially obscured by irregular polygons, which vaguely suggest viral fragments floating

in the fluid interior of a cell. Since the images that appear in this dynamic display are

interfaces to other parts of the work, this interactive animation functions as a graphic

realization of a possibility space. The user can bring the swirling pages into visibility

either by clicking on small icons at the corners of the rectangles or holding down the

cursor and moving it up or down the screen. The pages present brief clinical notes

written by doctors who have examined infected patients.

Deeper layers can be accessed by clicking on the pages, which takes the user to

the “books” in which the patients’ dreams are recorded. The “books” are images

representing print documents, with turnable pages that the user manipulates by

24

laboriously catching a corner with a cursor and “flipping” the page over. This

navigational strategy performs what Jay Bolter and Richard Grusin call remediation, the

re-presentation of one medium in another (here a book in a digital medium). The

character-focalizers in the books describe surrealistic scenes focusing on coupons for one,

office furniture for another, love and cattle for a third. Since the dreams are expressions

of the virus, it is not clear whether the narrator is the infected human or the virus, or

perhaps a hybrid of the two, in which case the voice can be understood as a remediation

of viral thoughts articulated through a human subject. Moreover, the two remediations

here performed, the remediation of the book by the computer and the virus by the human,

are teasingly connected by some of the links contained within the books. One link leads

to a passage suggesting that “single cells or even biological tissue becomes like a ‘brain-

stuff’ made out of Cray supercomputers,” thereby establishing a connection between the

biological activities of neural cells and the data processing of intelligent machines. The

series of associations then goes something like this: brain cells are like computers; brain

cells direct the behaviors that create the books; computers remediate the books and so

indirectly remediate the articulations associated with the brain cells; the virus infects the

brain cells and causes the host to write the books; in indirect fashion, the virus thus

inhabits the computer as well as the brain; the host expresses herself/himself through the

computer, which in this sense becomes the medium in which the infected patient, now

part human and part virus, becomes a legible subject of the narrating, which itself

becomes an irresolvable compound of human, virus, and machine. On one of the

swirling pages there is even the suggestion that this hybrid condition is the virus.

25

As the machinic becomes entwined with the human, narrative moves from the

causal sequences typical of traditional storytelling toward the combinatoric algorithms

typical of possibility space. The results are quasi-narratives that strive to capture all

possible combinations within causal sequences, as if unwilling to abandon either

causality or combinatorics. Everything is connected to everything else through the

combinatorics of the dreams, which form for any given patient a sequence that repeats

without being a repetition, for as one of the doctors notes, patients do not experience the

dreams as copies, even though in some sense they are all the same dream. Another

doctor notes that “all of the dreams I have recorded refer to ‘Connections’ between ropes

of the lives. A cure might be found by untying these knots, burning their frayed ends.”

As the comment suggests, in addition to the connections between a single person’s

dreams, there are also connections between one person’s dreams and those of all the other

infected people. If interpreted as a causal chain (a probable inference, given the

centrality of causality to narrative), this structure leads to propositions like the following

from one of the dream books:

And by sunlight I mean those sparkling particles the super-intelligent viruses

manipulating the fiery burst we call the sun use to control our, deceivingly

harmless, aquarium fish. But then that’s another story now isn’t it. Moving on,

this substance holds our world and all the other worlds together. It makes us sad

and happy and hungry for humping. Sometimes this goo [next page] but love has

nothing to do with goo. Instead love is governed by a complex system of ropes

and wires haphazardly connected to cattle in the Texas panhandle. Lucky for us it

26

seems the cattle2 haven’t yet discovered their power over love (Dreamphage,

Book 4, “Angrybo/vine/dis/ease).

The footnote2 leads to a window with the cryptic comment, “makes cows sleep,” which

opens to another window telling us “Henry lost his herd to paranoid software.”

The ontological instability of the fabula that a reader might construct from this

bizarre series of connections is palpable, from viruses in the sun to fish in the aquarium to

human sexual desire to Texas panhandle cattle (and this is only a partial list, as the

narrative interjection “but that’s another story now isn’t it” indicates). Through its

(sometimes very funny) absurdities, the fabula is revealed as subordinate to some other

kind of ordering principle, which the work visually and verbally associates with the

possibility space. Moreover, the trope of a mysterious virus spreading by unknown

transmission mechanisms and expressing itself through dream-narratives suggests that the

possibility space is infecting the fabula. In this case, the transmission mechanism is

nothing other than the hybrid human-computer subjectivity that is both the subject and

object of the narrating.

How might we describe the relation of narrative and possibility space in this

work? Partly antagonistic, insofar as the infection of the fabula by the possibility space

distorts the presumed shape of narrative and causes it to be, by conventional standards,

“ill-formed”; partly symbiotic, in the sense that this infection lets the work do more than

it otherwise could; partly parasitic, in that the virus both expresses and transmits itself

through the narratives, which may serve the virus’s own purposes rather than those of the

host; partly mimicry, for as the dreams progressively invade waking life, they mime what

they displace, so that (as one of the doctors notes) life itself becomes a kind of dream;

27

and partly hyperparasitic, in the sense that the work as a whole can be seen to parasitize

the parasitism of the virus to re-define what a narrative can be.

To flesh out the argument, I have constructed a list of other works in which the

interactions between narrative and possibility space exemplify the dynamics of

competition, cooperation, symbiosis, mimicry, parasitism and hyperparasitism. Whether

or not the reader agrees with these classifications is somewhat beside the point, for most

of the works partake of several dynamics, which in any event can be interpreted in

various ways. Rather, I want to illustrate the broader claim that the complex ecology

generated by the interaction of narrative and possibility space in Dreamphage is

characteristic of many contemporary print and electronic works.

I mentioned earlier that reliance on the fabula/sjuzhet dyad, along with the

presuppositions it entails, are ubiquitous in literary criticism, as in reading more

generally. The juxtaposition of narrative and possibility space not only creates new

opportunities for narratology but also for critical inquiry. Normally literary criticism is

published in journal or book form; other critics (sometimes) read these publications and

reply, usually in other venues and diverse time frames varying from several months to

several years after the initial publication. As a result, there is no possibility for

interactions that would change the original piece in view of readers’ reactions, objections,

extrapolations, queries, and modifications. In this sense each work exists in isolation

from the others, crystallized in the form in which it reached print. With electronic media,

however, new versions can easily be created and put to alongside or in place of the

original piece. Moreover, the time lag between electronic publication and a reader’s

reply is cut from months and years to hours or days. Already online publications such as

28

the Electronic Book Review are taking advantage of these digital capabilities to re-

envision how scholarly publication can take place. As this first issue of Vectors

demonstrates, the journal is trying to push the envelope further through its commitment to

publish multimedia criticism, that is, criticism in digital media that engages in practices

that would not be possible in print, and this article partakes of that experiment.

Criticism sometimes takes narrative form, and when it does, it can rightly be said

to imply a fabula, a universe of understanding and discourse. The fabulas implied by

different pieces of criticism can have substantial overlap, but they can also have

important differences—differences set into stone (i.e., wood pulp) by the publishing

conditions described above. What if we imagine criticism proceeding not through a

sjuzhet that implies a fabula, but as a possibility space? Suppose a the possibility space

capacious enough to allow a large number of instantiations, as well as modifications that

change its nature, making it an emergent adaptive system. Further suppose that an

expression of the possibility space is subject to instant re-arrangement, re-interpretation,

and re-engineering by other participants. It becomes, in other words, a kind of playing

field on which many different games may be played and (more significantly) different

kinds of games can evolve.

This is the idea that inspired “Narrating Bits” as an electronic work, done in

collaboration with Erik Loyer, a Los-Angeles-based Web designer and creative writer

who has authored a number of electronic literary works, among them Marrow Monkey

and Chroma, and the Stamen design studio, particularly Eric Rodenbeck who worked in

close collaboration with Loyer to develop the concept, create a visual design, and

implement it in conjunction with the Stamen team. In the design, the possibility space for

29

“Narrating Bits” is defined by keywords listed on the left side of the screen. Also shown

is an abstracted representation of the article as a segmented straight line stretching across

the screen. Selecting a keyword triggers the display of a new, curved line tracing a

trajectory that represents where the keyword appears in the article. The display allows

the user intuitively to grasp the keyword’s relation to the essay’s conceptual space. Each

keyword also defines a space for users to comment on the article and read comments by

other users. Thus the keyword list works as a trailhead into the article, with each

keyword tracing a distinctive path that winds through the article’s possibility space,

which in turn is made mutable, extendable, and adaptable by the addition and collection

of users’ comments.

The point of the experiment is not so much to use criticism in an entirely new

way, for critics have always built upon, modified, and transformed each other’s work.

Rather, it is to create a flexible interface that would allow critical exchanges to be

visualized and experienced as dynamic interplays that resemble the give-and-take of a

good conversation rather than staking out “positions” and waiting months or years for

responses. The larger goal is to open new possibilities for understanding the changing

roles of narrative in a digital age, when the age-old ability of narrative to shape and

express human subjectivity is coming into intimate contact with the capacity of intelligent

machines to store, process, and generate massive amounts of data.

I cannot imagine a human world without narrative, but I can imagine narratives

transformed and enriched by their interactions with possibility space in the complex

ecologies of contemporary media and culture.

30

Complex Ecologies Generated by the Interactions of Narrative and Possibility Space

Cooperation: Georges Perec, Life: A User’s Manual

Possibility space defined by the apartment building, its inhabitants

and contents; narrative emerges amidst listing of data from

possibility space. They cooperate in creating a meta-narrative.

Milorad Pavi_, Dictionary of the Khazars

Alphabetic entries in three books of the Dictionary provide

the data for the possibility space; narrative emerges from

correlations between characters and actions in the different books

and different entries. They cooperate in creating a meta-narrative.

Caitlin Fisher, These Waves of Girls

Typical of narrative patterns that come from following links;

possibility space defined by manifold of all possible narrative

trajectories. Narrative remains fragmented, connections are

thematic, to which the architecture of the possibility space

contributes.

M. D. Coverley, The Book of Going Forth By Day

Possibility space defined by ancient and modern Egypt, as well

as tripartite division of the soul in ancient Egyptian religion.

Coherent narratives united by concepts instantiated in graphics,

navigation and architecture of possibility space.

31

Competition: Stuart Moulthrop, Reagan Library

Possibility space defined as the four world levels plus the Notes;

narrative emerges through focalizations on the different characters;

suggestion of meta-narrative in symbolic connections between

themes of different levels.

Jonathan Lethem, Motherless Brooklyn

Narrative coherent but interrupted by Tourette’s syndrome

outbursts from protagonist Lionel Essrog. These fragmentary and

often obscene involuntary tics define a linguistic and psychological

possibility space that competes with controlled articulations.

Parasitism: Jason Nelson, Dreamphage

The dream-virus infects humans and starts hijacking their dreams

for its own (unknown and therefore ambiguous) purposes.

MEZ, ][(ad)] [Dressed in a Skin Code

Parasitizes email messages, chat rooms, etc. for narrative

fragments that are then “mezanglled” by interjecting coding

expressions and punctuations to create a hybrid discourse.

Hyperparasitism: Talan Memmott, Lexia to Perplexia

The narrative fragments focus on hybrid machinic-human

subjectivities in which the human has been parasitized by the

32

machine, but it then weaves these fragments into a meta-narrative

accessible to persistent and astute users.

Mark Hansen, Listening Post

An art installation that exhibits a random series of sentences from

internet chat rooms across a grid of 252 LCD screens as they are

simultaneously read by voice synthesizer. The installation thus

first parasitizes human narratives and converts them into database

form, then hyperparasitizes the database to re-deploy them as

narrative collages.

Mimicry

Jason Nelson, Dreamphage

Computer screen images that mimic the appearance and

functionalities of a book.

Chomskybot

An applet, created by John Lawler, Anthony Aristar, and John

Sowa, that generates sentences solely on their syntactical

properties by randomly combining phrases into a sentence. The

passages are taken from Noam Chomsky’s

Syntactic Structures (1957). The applet thus mimics

Chomsky’s prose to generate passages that sometimes appear to

make darkly enigmatic sense.

33

Works Cited

Argyros, Alexander J. A Blessed Rage for Order: Deconstruction, Evolution,and Chaos. Ann Arbor: University of Michigan Press, 1991.

Bal, Mieke. Narratology: Introduction to the Theory of Narrative, 2nd ed.Toronto: University of Toronto Press, 1997.

Borges, Jorge Luis. “The Library of Babel,” Ficciones. Translated by AnthonyKerrigan. New York: Grove Press, 1969. Pp. 79-88.

Bruner, Jerome. Acts of Meaning. Cambridge: Harvard University Press, 1990.Cage, John. “Mureau,” M: Writings, ’67-72. Middleton: Wesleyan University Press,

1974. Pp. 35-56.Chatman, Seymour. Story and Discourse: Narrative Structure in Fiction and

Film. Ithaca: Cornell University Press, 1978.Connolly, Thomas and Carolyn Begg. Database Systems. New York: Harlow,

Essex: Pearson Education Limited, 2002 (1995).Coverley, M. D. The Book of Going Forth By Day. <http://califia.hispeed.com/Egypt>.Fisher, Caitlin. These Waves of Girls. <http://www.yorku.ca/caitlin/waves>.Genette, Gérard. Narrative Discourse: An Essay in Method. Translated by Jane E.

Lewin. Ithaca: Cornell University Press, 1980.----------. Narrative Discourse Revisited. Translated by Jane E. Lewin. Ithaca:

Cornell University Press, 1988.Gottschall, Jonathan and David Sloan Wilson, eds. The Literary Animal:

Evolution and the Nature of Narrative. Evanston: NorthwesternUniversity Press, 2005.

Grusin, Richard and Jay David Bolter. Remediation: Understanding NewMedia. Cambridge: MIT Press, 2000.

Hayles, N. Katherine. "Simulated Narratives: What Virtual Creatures Can Teach Us." Critical Inquiry 26, no. 1 (Autumn 1999): 1-26.

----------. Writing Machines. Cambridge: MIT Press, 2001.Headrick, Daniel R. When Information Came of Age: Technologies of Knowledge in the Age of Reason and

Revolution, 1700-1850. New York: Oxford University Press, 2002.Landau, Misia. “Human Evolution as Narrative.” American Scientist 72 (1984): 262-268.----------. Narratives of Human Evolution. New Haven: Yale University Press, 1993.Lawler, Jon, Anthony Aristar, and John Sowa. Chomskybot.

<http://rubberducky.org/cgi-bin/chomsky.pl>.Lethem, Jonathan. Motherless Brooklyn. New York: Vintage, 2000.Loyer, Erik. Chroma (2001). <http://www.marrowmonkey.com/chroma>.----------. The Lair of the Marrow Monkey (1998).

<http://www.marrowmonkey.com/lair>.Manovich, Lev. The Language of New Media. Cambridge: MIT Press, 2001.Martin, James. Computer Data-Base Organization. Second Edition. Englewood

Cliffs, NJ: Prentice-Hall, 1977 (1975).Mateas, Michael. “Expressive AI: A Hybrid Art and Science Practice.” Leonardo:

Journal of the International Society for Arts, Sciences, and Technology 34(2)(2001): 147-153.

----------. “Research Statement.” <http://www-2.cs.cmu.edu/~michaelm/research.pdf>.

34

Memmott, Talan. Lexia to Perplexia.<http://www.uiowa.edu/~iareview/tirweb/hypermedia/talan_memmott/>.

MEZ (Mary-Anne Breeze). ][(ad)] [Dressed in a Skin Code.<http://www.cddc.vt.edu/host/netwurker>.

Moulthrop, Stuart. Reagan Library. <http://iat.ubalt.edu/moulthrop/hypertexts/rl>.

Nelson, Jason. Dreamphage.<http://www.heliozoa.com/dreamaphage/diseaseinterface.html>.

Perec, Georges. Life: A User’s Manual. Translated by David Bellos. Boston: David R. Godine, 1988.Porter, Theodore M. The Rise of Statistical Thinking, 1820-1900. Princeton:

Princeton University Press, 1988.Richardson, Brian. Unlikely Stories: Causality and the Nature of Modern

Narrative. Newark: University of Delaware Press, 1997.Rimmon-Kenan, Shlomith. Narrative Fiction: Contemporary Poetics, 2nd ed. New

York: Routledge, 2002.Sims, Karl. Evolving Virtual Creatures. [paper in pdf]

http://portal.acm.org/citation.cfm?id=192167.----------. [video of animation] <http://www.biota.org/ksims/blockies/>.Turner, Mark. The Literary Mind: The Origins of Thought and Language. New York:

Oxford University Press, 1998.Werner, G. M. and Michael G. Dyer. “Evolution of Communication in Artificial

Organisms,” C. G. Langton, C. Taylor, J. D. Farmer, and S. Rasmussen, eds.Artificial Life II. Redwood City CA: Addison Wesley Publications, 1991. Pp.659-687.

35

Endnotes

1In her study of database textbooks over the last quarter century, Sarah

Richardson makes this observation: “The 1977 text [James Martin, Computer Data-Base

Organization] teaches theory that has not yet been realized beyond specialized

applications and opines about the larger importance of the database in the future tense.

The 2002 text [Connolly and Begg, Database Systems], in contrast, addresses an

audience whose world is so imbued with the everyday uses of the database, including

everyday purchases on credit cards, store inventory maintenance, travel reservation tools,

internet searching and shopping, university student information systems, and library

indexing, that the importance of the database to society is seen more in its ordinariness,

its ubiquity in our day-to-day lives, than in its extraordinary potential implications,”

unpublished manuscript. I am indebted to Sarah Richardson for her help in researching

the evolution of textbooks on databases from the 1970’s to the present.

2 Among the theorists arguing for a central role of narrative in human evolution

are Alexander J. Argyros, A Blessed Rage for Order; Mark Turner, The Literary Mind,

focusing on the narrative genre he calls the parable; and Misia Landau, “Human

Evolution as Narrative,” which both reviews some of the claims for the importance of

narrative in human evolution and also treats the technical literature of evolution as itself a

field rich with narratives, and her book-length study on the same topic, Narratives of

Human Evolution. Also relevant is the forthcoming essay collection, The Literary

Animal: Evolution and the Nature of Narrative, edited by Jonathan Gottschall and David

Sloan Wilson.

36

3 See Daniel R. Headrick, When Information Came of Age: Technologies of

Knowledge in the Age of Reason and Resolution, 1700-1850, especially on the

development of demographic data and statistics; a useful synthetic study that carries the

story into the nineteenth century is Theodore M. Porter, The Rise of Statistical Thinking,

1820-1900.

4 In 1977, James Martin already could see that databases would become

ubiquitous, with significant consequences for social and cultural practices: “In all walks

of life and in all areas of industry, data banks will change the realms of what it is possible

for man to do. In centuries hence, historians will look back to the coming of computer

data banks and their associated facilities as a step which changed the nature of the

evolution of society, perhaps eventually having a greater effect on the human condition

than even the invention of the printing press” (2).

5 Hence the growing emphasis on data-mining, for it is designed to interact with

data not formatted so as to allow searches of the kind it wants to carry out.

6 Another way to make this distinction is between those humans who understand

and have access to machine architectures and those who do not. In her comparison of

database textbooks, Sarah Richardson notes the strong distinction made between

“sophisticated” and “naïve” users. She writes, “the relational model found its success in

part by offering, in addition to the general practicalities of data independence for

application development, a conceptual approach to databasing that restricts the general

user’s view and potential to meddle with the data and its structure and systems. In

database rhetoric, the user is regarded warily, and a great deal of work is directed towards

ensuring that the user’s interaction with the data is constrained to a range of anticipated

37

operations and views, and further, that the user is unaware that this is the case. The

language of ‘internal’ and ‘external’ views, definitions, and realities is one way that both

textbook authors conceptualize this dual-reality approach to databasing” (28).

7 Donald Rumsfeld, “As we know, there are known knowns. There are things we

know we know. We also know there are known unknowns. That is to say, we know

there are some things we do not know. But there are also unknown unknowns, the ones

we don’t know we don’t know” (Department of Defense news briefing, February 12,

2002), cited in “The Poetry of D. H. Rumsfeld” by Hart Seely,

<http://slate.msn.com/id/2081042/>.

8 I am indebted to Nicholas Gessler for this observation.

9Gérard Genette, Narrative Discourse Revisited, p 14. In this book Genette

replies to criticisms of his earlier text Narrative Discourse: An Essay in Method.

Although his theory is therefore contained almost wholly in the first volume, the second

book is useful in that he more fully explains his rationale and reasoning for the theory. I

will therefore quote more frequently from the second book, although my comments

presuppose an understanding of his first book as well.

10 For a fuller discussion of Lexia to Perplexia, see N. Katherine Hayles, Writing

Machines.

11 For a more detailed analysis of Sims’ work and its solicitation of

anthropomorphic responses by viewers, see N. Katherine Hayles, "Simulated Narratives:

What Virtual Creatures Can Teach Us.”

38

12 For an indication of the difficulties, see G. M. Werner and Michael Dyer,

“Evolution of Communication in Artificial Organisms.”