Ergodic literature is a mode of textual organization in which nontrivial effort is required for the reader to traverse the text, beyond ordinary eye movement or turning pages. Espen J. Aarseth appropriated the term from physics, deriving it from the Greek ergon (“work”) and hodos (“path”). He introduced the term within his broader concept of cybertext, which he presents not as a literary genre but as a perspective on textual machines that compute or permute outputs. In Aarseth’s framing, the cybertextual process includes a semiotic sequence produced through the user’s material actions, which conventional notions of “reading” do not fully capture. Although frequently compared to “nonlinearity” in physics, Aarseth treats such nonlinearity in hypertext as a topological, graph-theoretic property of nodes and links rather than a concept imported from physics.
Concept Aarseth's book contains the most commonly cited definition of ergodic literature:
In ergodic literature, nontrivial effort is required to allow the reader to traverse the text. If ergodic literature is to make sense as a concept, there must also be nonergodic literature, where the effort to traverse the text is trivial, with no extranoematic responsibilities placed on the reader except (for example) eye movement and the periodic or arbitrary turning of pages.
His aim in providing this definition is to offer a radical break from earlier theories of Hypertext, which he says fall trap to 'an essentialist idea of "the computer medium" as a singular structure of well-defined properties,' and thus remain blind both to the plurality of types of media that a computer can support, and of the forms that text can take outside of a digital medium. Additionally, without free text search, many hypertext works end not less, but more linear than the conventional codex, in stark contrast with the technoutopist visions of hypertext and its capabilities. To this end, he aims to develop a model of literature organised not around the medium in which the work is contained, but instead around the manner in which the text itself functions and interacts with its reader.
Aarseth divides the text into three core constituent elements: the medium itself, the human operator, and the strings of signs which make up the text itself. The last of these he divides into two distinct categories: textons – strings of signs as they appear (or are encoded) in the text itself – and scriptons – strings of signs as they appear to the reader or user of the text. From there, Aarseth develops a typology of possible media positions for texts, organised around seven axes:
Dynamics: Whether scriptons are constant and unchanging (Static), or variable: in which case scriptons may vary in content or number, whilst textons remain fixed (Intratextonic Dynamics, or IDT), or both scriptons and textons may vary (Textonic Dynamics, or TDT). As an example, he gives Michael Joyce's afternoon, a story, which is a hypertext, but its scriptons and textons both remain fixed, and contrasts it with Multi-User Dungeons, where neither the number, nor amount of scriptons and textons can be known
Determinability: Whether a scripton's adjacent scriptons are always the same (Determinable), or whether they may change (Indeterminable). As an example, he gives adventure games, where sometimes the same action will always produce the same result, making it predictable, or it may involve some degree of randomness, such as a dice throw, making it unpredictable
Transiency: Whether the text requires direct input from the reader for the text to change (Intransient), or if it may change on its own (Transient). As an example, he gives William Gibson's Agrippa (A Book of the Dead), where the text scrolls past the screen at its own pace, regardless of user input.
Perspective: Whether the text requires its reader to play a strategic, intradiegetic role within the text (Personal), or not (Impersonal).These also later appear as Permanent and Impermanent As an example, he gives Italo Calvino's If on a winter's night a traveler, where he says there is little for the reader to actually, strategically, do, in contrast with Multi-User Dungeons, where the player is (at least partially) responsible for their character
Access: Whether scriptons are readily available to the user (Random) or not (Controlled). As an example, he gives the typical codex, which may be opened to any page, at any time, in contrast with hypertexts such as Stuart Moulthrop's Victory Garden, where accessing a specific passage may require from the reader to follow a specific path through other passages.
Linking: Whether the text is organised by links that are available to the reader at any point (Explicit), or only under certain circumstances (Conditional), or whether there are no links in the text whatsoever (None). As an example, he again gives Michael Joyce's afternoon, where certain parts of the text are accessible only once the reader has visited specific other passages.
User functions: Whether the reader only interprets the text (Interpretative) - a function present in all texts - or if they also decide what paths to take (Explorative), chooses or creates scriptons during a specific reading session (Configurative), or adds permanently to the body of the text (Textonic). As an example of a text that the reader explores he gives Julio Cortázar's Rayuela, for a text that the reader configures he gives Raymond Queneau's A Hundred Thousand Billion Poems, and for a text the reader permanently alters, he gives Allen S. Firstenberg's Unending Addventure.
Variable Possible value Dynamics Static, Intratextonic Dynamics, Textonic Dynamics Determinability Determinable, Indeterminable Transiency Transient, Intransient Perspective Personal, Impersonal Access Random, Controlled Linking Explicit, Conditional, None User function Interpretative, Explorative, Configurative, Textonic
These seven variables are capable of producing a total of 576 unique media positions (3 x 2 x 2 x 2 x 2 x 3 x 4), of which conventional hypertext and literature represent only a small part. Among these, Aarseth defines an 'ergodic text' as any one which requires at least one of the three other user functions beyond mere interpretation. Thus, both paper-based and electronic texts can be ergodic: "The ergodic work of art is one that in a material sense includes the rules for its own use, a work that has certain requirements built in that automatically distinguishes between successful and unsuccessful users."
Cybertext
Cybertext is Aarseth’s perspective on dynamic textual machines; he proposes reserving ‘cybertext’ for texts where ‘calculation’ produces scriptons. Thus, works of hypertext fiction of the simple node and link variety may be ergodic literature but not cybertext: it is the presence of conditional links and computation that shift them toward being cybertext. A chat bot such as ELIZA is considered a cybertext because when the reader types in a sentence, the text-machine actually calculates directly based on the user's input to generate a textual response. Likewise, the I Ching is cited as an example of cybertext: though it is the reader that carries out the calculation, the rules to generate responses are clearly embedded in the text itself.
It has been argued that these distinctions are not entirely clear and scholars still debate the fine points of the definitions. The concepts of cybertext and ergodic literature were of seminal importance to new media studies, in particular literary approaches to digital texts and to game studies.
Cybertext has also been suggested as a tool to enhance learner engagement and motivation in language education.
Examples
Aarseth gives two major lists of examples of ergodic literature throughout the work – first in the opening chapter, then in the third, where a possible typology is discussed. The major examples listed throughout the work include:
{| class="wikitable" |+ Examples given by Espen Aarseth |- ! Title !! Creator !! Format or Description |- | (No specific example or location identified) || Ancient Egyptians || Stone wall inscriptions of the temples in ancient Egypt that are connected two-dimensionally (on one wall) or three dimensionally (from wall to wall or room to room). |- | afternoon: a story || Michael Joyce || Published by Eastgate Systems, it is one of the earliest examples of hypertext literature, along with Victory Garden. |- | Agrippa|| William Gibson || A 300-line semi-autobiographical novel. It exists in two forms, either on a 3.5" floppy disk, where the text scrolls by automatically, and encrypts itself once it has passed, or as an artist's book, of which the pages have been treated with photosensitive chemicals, such that the first exposure to light would effect gradual fading. |- | Book Unbound || John Cayley || A "holographic sentence generator that merges and mutates other texts, inviting readers to feed their favorite results back into the system." |- | Calligrammes || Guillaume Apollinaire || A collection of poems, which "fork out on the page" – the "words and sentences on [the] page are spread out in many directions," and invite "linking/jumping.", with no clear right sequence. It can be considered alternatively an example of concrete or visual poetry. |- | Cent mille milliards de poèmes || Raymond Queneau || A set of ten poems, printed on card with each line on a separate strip. As all ten sonnets have not just the same rhyme scheme but the same rhyme sounds, any lines from a sonnet can be combined with any from the nine others, allowing for 1014 (= 100,000,000,000,000) different poems. |- | Colossal Cave Adventure || William Crowther & Don Woods || A text-based adventure game, where the player interacts with locations and objects via simple word prompts, interpreted by the game's natural language input system. |- | Composition No. 1, Roman || Marc Saporta || The historically first "book in a box," consisting of a bundle of pages that can be shuffled and read in any order. Conceptually similar to B. S. Johnson's (1969) The Unfortunates, which consists of 27 unbound "sections" of varying lengths, with only the first and last specified as such. |- | ELIZA || Joseph Weizenbaum || An early natural language processing computer program created from 1964 to 1966 at MIT. |- | Falcon 5: The Dying Sun || Mark Smith & Jamie Thomson || A game book, similar to The Money Spider. Differs from it in that it adds dice-rolling as a mechanic, and thus a degree of indeterminacy. |- | Hopscotch || Julio Cortázar || A stream-of-consciousness novel which can be read either linearly, or according to an alternative chapter order listed at the start. |- | I Am Awake at the Place Where Women Die || Jenny Holzer || A temporary LED installation, wherein a "linear electronic text [...] endlessly repeats its short, painful messages." |- | The I Ching || China, as old as 900 BCE || A divination text in which bundles of yarrow stalks are arranged to form numbers. |- | The Money Spider || Robin Waterfield & Wilfred Davies || A "typical gamebook in which the reader must solve a puzzle by choosing the right path through the many fragments of the text." |- | MUD1 || Richard Bartle & Roy Trubshaw || The first multi-user dungeon, a text-based multiplayer real-time virtual world. |- | Night of January 16th|| Ayn Rand || A play about a trial, where "members of the audience are picked to be the jury," with two possible endings depending on their verdict. |- | Norisbo || Randi Strand || An artist's book, which "folds from all four sides, so the reader reads a unique sequence folded by the last reader and then folds the pages to leave a unique combination for the next reader." |-