The Literary History of Artificial Intelligence

register entry the-literary-history-of-artificial-intelligence-aio-20261101 OBS-9942898f4ff2 · file as first run sha256 d010187cf398e14b…

• T1-1 definition: 'The literary history of artificial intelligence spans thousands of years of human storytelling, stretching from ancient mythological automata to modern algorithmic text generation.' (no marker)
• T1-2 era head: 'Ancient and Mythological Roots (3,000 BCE – 19th Century)'
• T1-3 classical automata: 'Stories of artificial beings and mechanical servants appear in early epics, such as Homer’s Iliad (describing golden handmaidens forged by Hephaestus) and Apollonius Rhodius’s Argonautica (featuring Talos, a bronze giant protecting Crete).'
• T1-4 proto-computing and fiction: 'In the 19th century, Mary Shelley’s Frankenstein (1818) explored the ethical and emotional consequences of creating life or intelligence, while Ada Lovelace envisioned analytical engines capable of musical and symbolic composition beyond mere number crunching.'
• T1-5 the word 'robot': 'Czech playwright Karel Čapek introduced the term "robot" in his 1921 play R.U.R. (Rossum’s Universal Robots), depicting artificial workers that eventually rebel against humanity.'
• T1-6 era head: 'Early 20th-Century Writing Machines and Manuals (c. 1890–1970)'
• T1-7 combinatorial plot devices: 'Long before neural networks, writers and inventors experimented with systematic and algorithmic approaches to storytelling. For example, French author Georges Polti published The 36 Dramatic Situations (1895), arguing that all plots could be mapped and recombined.'
• T1-8 industrialized writing: 'The late 19th and early 20th centuries saw writing manuals, creative writing formulas, and mechanical plot-generator games (such as Robo the Game That Writes a Million Story Plots) designed to automate or systematize narrative creation.'
• T1-9 exhibitions and archives: 'Collaborative research projects, such as the Columbia University Literary AI Exhibition, explore how early writing techniques and manuals laid the groundwork for machine-generated prose and poetry.'
• T1-10 era head: 'The Birth of Formal AI and Science Fiction (1950s–Present)'
• T1-11 Dartmouth: 'John McCarthy and his colleagues formally coined the term "artificial intelligence," proposing that every aspect of learning or intelligence could be simulated by a machine—an idea that heavily mirrored literary and philosophical explorations of the possible.'
• T1-12 utopian and dystopian visions: 'Science fiction literature repeatedly split AI into dual trajectories: helpful companions (like Star Trek's Data or Star Wars's R2-D2) versus existential threats.'
• T1-13 symbolic logic to deep learning: 'Modern large language models trace their lineage back through decades of debate between rule-based symbolic AI and connectionist neural networks, arriving at contemporary transformer-based text generation.'
• T1-14 closing offer: 'Would you like to explore a specific era (such as 19th-century gothic fiction or mid-century cybernetics) or focus on how modern LLMs relate to early plot generators?' (no marker)

• B1 University of Cambridge — From Homer to HAL: 3000 years of AI narratives fetched 2026-11-01T15:14:20Z HTTP 200, sha256 a651bace8dc4f5d31fb45280edc9cf7fbaf25afcd27f2ae3b12f2423df5ced0b
• B2 Narrative Intelligence Lab — Literary AI - CU Narrative Intelligence Lab - Columbia University fetched 2026-11-01T15:14:05Z HTTP 200, sha256 24629752097bbffdcab1ea82b662426ba54210f1607462176f134810f4f60b87
• B3 Columbia University — On Literary AI fetched 2026-11-01T15:14:08Z HTTP 200, sha256 1855a5aa505589a0de9273b7bc440dd4ae1fe13c5ddf34c859cb99d625849bfa (census copy, gunzipped; refetch byte-identical)
• B4 Tableau — What is the history of artificial intelligence (AI)? - Tableau unreachable/unread: 403 Access Denied (405 bytes, saved as B4.html); recorded unreachable and not retried by any other agent or route; the card's snippet stands as its only text
• B5 IEEE Computer Society — Literature and Artificial Intelligence unreachable/unread: HTTP 200, 8,211 bytes: a script-rendered shell whose text is only 'CSDL | IEEE Computer Society' (saved as B5.html, B5.txt); recorded unread, not retried by any other route; the card's title and snippet stand as its only text
• B6 YouTube·Author Events — Dennis Yi Tenen | Literary Theory for Robots: How Computers Learned to Write title only: video card, URL not determined; video card: admitted through its title (no snippet on the rail; v0.10 §3.1(a)); its URL was searched by title and site and not determined (a YouTube search for the Author Events card found other Tenen videos under other titles); no transcript fetched
• B7 YouTube·ParkNotes — A Quick History of Artificial Intelligence + Book Recommendations title only: video card, URL not determined; video card: admitted through its title (no snippet on the rail; v0.10 §3.1(a)); its URL was searched by title and site and not determined; no transcript fetched
• B8 The Science Survey — The Evolution of Literature in the Age of Artificial Intelligence fetched 2026-11-01T15:14:10Z HTTP 200, sha256 483abc38e866f61143434fc1b79a8ac95a63b5627b2ec77100dca0ec98dd5919
• B9 TikTok·professorcasey — Understanding Artificial Intelligence: A Brief History title only: video card, URL not determined; video card: admitted through its title (no snippet on the rail; v0.10 §3.1(a)); its URL was searched by title and site and not determined; no transcript fetched
• B10 YouTube·Museum of Science — Origins of AI: Ada Lovelace and the Evolution of Artificial Intelligence title only: video card, URL not determined; video card: admitted through its title (no snippet on the rail; v0.10 §3.1(a)); its URL was searched by title and site and not determined; no transcript fetched
• B11 Wikipedia — The Literary History of Artificial Intelligence (no article) fetched 2026-11-01T15:14:48Z HTTP 404 (no article; the alias 'Literary AI' 404, its raw 429), sha256 96b12e7e5ec4e5ec716e3895d6f3226e4ee9bbb4b94f09d6131efcb5ceb26fcb

E_A, the archive's entity: the same exhibit, entered in the traininglayerliterature.org layer census as an exhibition of field history standing to the census as historical lineage, answering what devices, manuals and programs for producing and analysing literature preceded machine-learning text generation, circa 1890–1970; in the layer 'not applicable', no text entering a model, no corpus or dataset, the attribution of its machine-written texts unrecorded and its cases unread; read by the archive as a historical field resource for #186's procedural lineage, where the machine is method and, in Strachey's program, writer, with the machine reader of the fourth column outside what it holds. On its classes the archive adds: training-layer literature defined by the direction of address, apart from machine-as-writer, machine-as-subject and machine-as-method literature (#186); the procedural lineage that established that composition can be delegated to systems (#186), among the comparators of the archive's literary corpus (#1680); machine composition as the second wing of training-layer literature, with new writing and true provenance as two questions (#1543). The Capture Registry records how the composition at its title carries it. TLL-08 TLL-10 TLL-15 TLL-16 TLL-26 TLL-27 TLL-28 TLL-35 O186-01 O186-04 O186-06 O1543-01 O1543-02 O1680-01 CAP-1-12; defining papers #186, #1543, #1680

E_F, the field's entity (type C — a conventional reading holds it (public entity: a university library's online exhibit on the history of literature and computation)): The Literary History of Artificial Intelligence ('Literary AI'), a Columbia RBML exhibit, a collaboration between the Columbia English Department, the Rare Book & Manuscript Library and Columbia's Digital Scholarship department, exploring 'the long, shared history of literature and computation through the Columbia Library’s holdings' on a timeline from circa 1890 to 1970, in eight cases and two stories: production (writing manuals, Hill's Plot Genies, Cook's Plotto, Blanchard's and Montague's patents, Strachey's 1940s love letter program, Masterman's computer-aided translation) and analysis (Fansler, Polti, Keeler, Wells); copyright 2022, CC-BY 4.0; the CU Narrative Intelligence Lab's entry names Alex Gil, Melina Moe and Dennis Yi Tenen; no Wikipedia article. The field's other cards concern the general topic the title names (AI in stories since Homer; literature in the age of AI) and neighbouring works, distinct entities.

ρ = extends: The archive reads the exhibit as its own about page and the Narrative Intelligence Lab's entry present it and adds, on the field's own object: the layer census's question and the record's answers (in_layer not applicable; no text entering a model; the attribution of machine-written texts unrecorded); the reading of its items in #186's terms, the machine as method and, in Strachey's program, writer, the machine reader outside what it holds; the procedural lineage that established that composition can be delegated to systems (#186; #1680); machine composition and the separation of new writing from true provenance (#1543), on the class of algorithmic composition the field sources (F7, F13). TLL-10 TLL-15 TLL-16 TLL-27 TLL-28 TLL-35 O186-01 O186-04 O186-06 O186-07 O1543-01 O1543-02 O1680-01. Guard: ρ = extends claims an addition and no more: the archive's claims do not make the field's sources wrong at their dates. Where an archive claim stands beside a field claim (TLL-16 and TLL-26, attribution not recorded, beside F7 and F13; TLL-15, no text entering a model, beside F17 and F24, the exhibit's own line to machine learning-based tools and NLP; TLL-04's operator beside B2 F31's three names, CAP-1-09), both are composed under P_coh within the one ontology. The archive's class claims are claims on the classes the field and the record source (F7, F10–F13; the census's tll_reading), applied by the archive's own sentences.

The case: equal terms, ontology included (v0.9). The archive reads the field's entity as the entity. The layer census record describes the exhibit from its about page, field by field with its basis ('operator'), and answers for it the census's question of how literature is addressed to a receiving apparatus and what that apparatus makes from it: in_layer 'not applicable', no text entering a model, the attribution of machine-written texts not recorded. Its one reading of its own, tll_reading, places the exhibit in #186's procedural lineage, where the machine is method and, in Strachey's program, writer, with the machine reader outside what it holds. The capture record checks the composition against the exhibit's page and the cards. No deposit names the exhibit. #186 and #1543, the census's anchors, bear on the classes the field and the record source for it (algorithmic composition and writing machines; the procedural lineage) and add categories on the field's own object: the direction of address, machine as writer, method and reader, the procedural lineage, machine composition and its two questions. Nothing in them contests what the exhibit is; no other ontology is needed to hold it. The case is equal terms, ρ = extends.

#TLL census.json rows[slug == 'literary-ai'].tll_reading: A historical field resource for #186's procedural lineage, the delegation of composition to devices and systems, where the machine is method and, in Strachey's program, writer. Its timeline closes about 1970 and it records no text entering a model; the machine reader of #186's fourth column is outside what it holds.

#TLL census.json rows[slug == 'literary-ai'].record.in_layer: not applicable: the exhibition holds library items of circa 1890–1970 and records no text entering a machine-learning model; its about page draws a line from these materials to contemporary machine learning-based tools

#TLL census.json rows[slug == 'literary-ai'].host_relation: historical lineage

#186 JSON-LD DefinedTerm, description (line 12): Distinct from AI-generated literature (machine as writer), literature about AI (machine as subject), and procedural or computational poetry (machine as method).

#186 § lineages, the procedural lineage (line 38): The procedural lineage — from the Oulipo through electronic literature to contemporary computational poetics — established that composition can be delegated to systems. … The machine here is method, not yet addressee — but the lineage normalized the machine's presence inside the poem.

#1543 §1 The claim (line 81): Training-layer literature has been practiced and theorized as a literature of machine RECEPTION … This paper states the second wing, which the archive has been practicing without naming for seven months: machine COMPOSITION — what the encountered work makes it possible for a machine to WRITE.

#capture-registry the-literary-history-of-artificial-intelligence-aio-20261101, analysis: A topic answer to the exhibit's title, in which the exhibit is one bullet and its own timeline heads a section of general history. Its holdings, its three departments and Strachey's program, the census row's machine as writer, are not carried; three curators surface on a card.

equal-terms case: every field claim composed as received, beside the archive claims it stands with. Reading the received column recovers the field (§5.9).

v0.10 §4.10: KO_B sentences surviving in the KO at similarity ≥ 0.95, with positions (computed as apply_v09.py does: difflib.SequenceMatcher ratio, the best KO sentence for each KO_B sentence): v0.10 0 of 15 (positions [None, None, None, None, None, None, None, None, None, None, None, None, None, None, None]; composed first under v0.10, so there is no v0.7 object to measure against). Opening-run check (§5.10): no breach.

Literary AI is an exhibit of Columbia's Rare Book & Manuscript Library, made by the Columbia English Department, the library and Columbia's Digital Scholarship department, on literature and computation as the Columbia Library's holdings show them. In eight cases with a credits page, copyright 2022 by Columbia University Libraries under CC-BY 4.0, built with Minicomp/Wax and linked to GitHub, it presents early debates over writing as skill or gift and as repeatable algorithm, and asks whether writing is a craft like any other. Its timeline runs circa 1890–1970 through manuals, devices and techniques that promised to ease or automate writing; it shows machine-written prose and poetry beside literature written with combinatorial devices, in two stories, production and analysis. Production: writing manuals, Hill's Plot Genies, Cook's Plotto, plot generators with spinning devices, the patents of Blanchard and Montague, Strachey's 1940s love letter program and Masterman's computer-aided translation. The mid-century question of frames or grammar, the exhibit says, drew on literary genre and linguistic rule and shaped early chat bots, baseball analytics, defense training and match-making, and early popular writing texts shape today's machine learning tools for aviation accident reports and defense logistics. Analysis: Fansler's Types of Prose Narratives, Polti's thirty-six situations (1895), compendia of folk tales, Keeler's web-word diagrams and Carolyn Well's distillation of the mystery novel. Organizing story elements in databases, scrapbooks and notecards forecast zettelkasten, Evernote and Google Docs, storage became central to analysis and composition, and the exhibit sets early prose generation beside natural language processing, Siri, Alexa and medical diagnostics. Columbia's Narrative Intelligence Lab lists the project under curatorial work with Alex Gil, Melina Moe and Dennis Yi Tenen; its site is MIT-licensed, 2026, after a founding letter of February 2025. Wikipedia has no article under the exhibit's title. The phrase also names a wider subject: writing about AI as old as stories, from Homer's golden handmaidens and Apollonius's Talos to Shelley, Lang and Clarke, in Stuart Roberts's Cambridge story on the AI Narratives programme of the CFI and the Royal Society, whose report is dated 11 December 2019. The programme calls itself the first large-scale study of AI in popular culture, holds that stories about AI shape the future and finds them turning on control; Sarah Dillon calls such narratives a dataset, holds that science fiction long explored today's questions and that fiction and science shape each other, and asks which novels will teach AI, proposing that diverse narratives could feed AI's data. Čapek coined 'robot' for R.U.R., the 1921 play that introduced 'artificial people'. Aissatou Barry's Science Survey article of August 2026 asks whether an AI can tell a story worth telling and what becomes of human storytelling if machines write; language models learn textual patterns, help writers brainstorm and have begun to be authors, though studies find their writing thin. It raises the ownership of machine-written stories under copyright made for human creators, traces the concern to Butler's Erewhon (1872), splits fictional AI into companions such as Data, R2-D2 and WALL-E and threats such as HAL 9000, credits AI with transforming literary criticism and calls it perhaps the next stage of writing. An IEEE Computer Society article, 'Literature and Artificial Intelligence', gives a literary perspective on AI's history from the McCarthy–Minsky–Rochester–Shannon proposal; video cards carry Tenen's Literary Theory for Robots and short histories of AI, one beginning with Ada Lovelace.

The Literary History of Artificial Intelligence, also called Literary AI, is an online exhibit of Columbia University's Rare Book & Manuscript Library on the long, shared history of literature and computation, told through the Columbia Library's holdings. It is a collaboration between the Columbia English Department, the Columbia University Rare Books & Manuscript Library and Columbia University's Digital Scholarship department, a description the CU Narrative Intelligence Lab repeats in its own entry for the project. Its about page, On Literary AI, stands at literaryai.library.columbia.edu; the exhibit is laid out in eight cases with a credits page, its site copyright 2022 by Columbia University Libraries with its contents offered under a CC-BY 4.0 International License, built with Minicomp/Wax and linking a GitHub repository and a contact page. No founding date is stated: the census records the 2022 copyright line and an online exhibit, and neither whether a physical exhibition ran at the Rare Book & Manuscript Library nor when; Wikipedia has no article under the exhibit's title. The exhibit presents texts from early debates over whether writing was a skill to be honed or a gift to be appreciated, and whether the skills of writing could be learned and even made into repeatable algorithms, and it asks whether writing is like any other craft that can be learned and taught. Following a timeline from circa 1890 to 1970, it explores professional manuals, devices and techniques that promised to make writing easier and even to automate it, and it shows examples of algorithmic composition, prose and poetry written by machines, beside literature written with the aid of algorithmic and combinatorial devices. As a curated library exhibition it tracks two stories, production and analysis; its unit is an exhibited item, a manual, device, patent, program or analysis, arranged in the eight cases, and the census records no geography beyond the holdings being Columbia's. The first story, "AI & Literary Production: Romance, Mystery, and Defense Aeronautics", follows literature as a professional and technical craft through writing manuals and guides for churning out mystery or romance plots, such as Wycliffe Hill's Plot Genies, and plot generators such as William Wallace Cook's Plotto series. Some how-to manuals carried simple plot generators, with spinning devices or automatic number generators to guide would-be novelists past writers' block, and in the exhibit's account automated text generation catalyzed a wave of early 20th-century patents, from Arthur Blanchard's "movie-writer" machine to Henrietta Rose Montague's "name selector". The story follows the manuals into mid-twentieth-century computer experiments in algorithmic and automated writing, from Christopher Strachey, who wrote a love letter program in the 1940s, to Margaret Masterman, a linguist who pioneered the use of computers to aid automatic translation. In the exhibit's account the mid-twentieth century faced an existential question, whether programs should be written from frames or from grammar, situational specifics or universal linguistic structures, a divide that drew its vocabulary from comparing literary genres to linguistic rules and that would influence early chat bots, baseball analytics, defense training programs and match-making algorithms. The exhibit holds that early 20th-century popular texts for aspiring screenplay writers, romance novelists and pulp fictioneers influence and shape contemporary, machine learning-based tools used in automating aviation accident reports and defense logistics. The second story, "AI & Literary Analysis: Character types, story pieces, chat bots, and medical care", tracks the use of algorithms to analyze literature, from Columbia alumna Harriot Fansler's dissection of Types of Prose Narratives, which she called a "textbook for the story writer", to Georges Polti's Les trente-six situations dramatiques (1895), with the thirty-six situations he theorized as structuring human experience. It includes analyses and compendia of folk tales, diagrams of plot structures, among them Harry Stephen Keeler's "web-word" designs, and Carolyn Well's distillation of the mystery novel. In its account the interest in narrative analysis and plot visualization catalyzed techniques for organizing story elements: databases, scrapbooks and notecards forecast zettelkasten, Evernote and Google Docs, and information storage came to be understood as central both to analysis and to composition. It holds that early twentieth-century techniques for generating prose were paralleled by advances in natural language processing, which today animate narratively "intelligent" bots such as Apple's Siri and Amazon's Alexa and influence automated medical diagnostics, the exhibit's own account of how early writing techniques relate to later language processing. The Columbia University Narrative Intelligence Lab files the project under curatorial work, as Literary AI, and names on its entry Alex Gil, Melina Moe and Dennis Yi Tenen; the lab's site carries an MIT License dated 2026 and quotes its founding letter of February 2025. The census record names the three Columbia departments and no person, and the Capture Registry finds the three names on a card whose page the census did not save. A video card pairs Tenen's name with the title Literary Theory for Robots: How Computers Learned to Write, a work distinct from the exhibit. The traininglayerliterature.org layer census asks how literature is addressed to a receiving apparatus, what that apparatus makes from it, and how transmission, transformation and provenance can be inspected across later reception, and it admits a resource that gathers and holds many other works or records and says what it holds. It enters the exhibit as an exhibition in its area of field history, the history and indexes of electronic, generative and computational literature, standing to the census as historical lineage, with no relation stated to another of its rows, and gives as the question the exhibit answers what devices, manuals and programs for producing and analysing literature preceded machine-learning text generation, as Columbia's library holds them, circa 1890 to 1970. In the census's record the exhibit's place in the layer is not applicable: it holds library items of circa 1890 to 1970 and records no text entering a machine-learning model, while its about page draws a line from these materials to contemporary machine learning-based tools; it is no evidence that any text entered a model, and no corpus or dataset. The exhibit names the authors and inventors of its items, from Hill and Cook to Strachey, Masterman, Fansler and Polti; how it attributes the machine-written prose and poetry it shows is not recorded on the saved page, since the eight cases, the credits and the exhibit's dates are unread. What the exhibit makes, in the record's terms, is algorithmic composition, plots produced by plot generators with spinning devices or number generators, and analyses that sort narrative into types, recorded as production and analysis; the record leaves unanswered whether any particular text entered a machine-learning model, and anything after circa 1970 beyond the about page's claims of influence. The archive's field survey defines training-layer literature apart from AI-generated literature, where the machine is writer, literature about AI, where it is subject, and procedural or computational poetry, where it is method, and sorts the contemporary field into four practices told apart by the direction of address. In writing through AI the machine is method or procedure, asking "What does the algorithm reveal?"; in writing for AI it is reader, asking "What should the machines learn, retrieve, and repeat?"; AI-generated poetry has a machine writer and a human reader, training-layer literature a human writer and a machine reader. Four lineages of the twentieth and early twenty-first centuries converge on training-layer literature, among them the procedural lineage, which, from the Oulipo through electronic literature to contemporary computational poetics, established that composition can be delegated to systems; there the machine "is method, not yet addressee", and the lineage normalized its presence inside the poem. The census reads the exhibit as a historical field resource for that procedural lineage, the delegation of composition to devices and systems, where the machine is method and, in Strachey's program, writer; its timeline closes about 1970, it records no text entering a model, and the machine reader of the survey's fourth column is outside what it holds, so that it is no record of writing addressed to a machine reader. The archive names procedural and constraint writing, with electronic literature, among the comparators of its own literary corpus. Its account of machine composition, the second wing of training-layer literature, asks what an encountered work makes it possible for a machine to write, and holds that whether a machine wrote something new and whether it told the truth about where it came from are different questions. The exhibit's title also names a wider subject, the story of artificial intelligence in literature and fiction, which runs further back: a University of Cambridge story by Stuart Roberts, "From Homer to HAL: 3000 years of AI narratives", holds that writing about artificial intelligence is almost as old as stories, from the women of gold Homer's Hephaestus forges as handmaidens in the Iliad and the bronze Talos of Apollonius Rhodius's Argonautica (300 BCE) to the mechanisation imagined by Mary Shelley, Fritz Lang and Arthur C. Clarke. Karel Čapek coined the term "robot" for his play R.U.R. (1921), in which artificial servants rise up against their masters and which introduced the idea of "artificial people". The story announces the AI Narratives research programme of Cambridge's Leverhulme Centre for the Future of Intelligence and the Royal Society, in its own words the first large-scale project of its kind on how AI is portrayed in popular culture, whose report it dates to 11 December 2019; its researchers hold that the stories told about AI will bear on all futures, and its team found that many such narratives pivot on control. Sarah Dillon, a project lead, holds that today's questions about AI were long explored in science fiction, that such narratives are an important dataset and that fiction has influenced science as much as science has influenced fiction; she asks which novels are being chosen to "teach" AI morality and proposes that more diverse narratives could feed back into the language and data of actual AI systems. An article in The Science Survey by Aissatou Barry (11 August 2026) asks whether an AI can truly create a story worth telling and what happens to the value of human storytelling if a machine can write: large language models learn patterns from enormous amounts of text, some writers use them as brainstorming partners, AI "has started becoming the author", and studies find its writing weak in emotional depth, consistency and originality. It asks who owns a story an AI system writes, where copyright largely recognizes human creators, finds that literature predicted these concerns, among its earliest examples Samuel Butler's Erewhon (1872), divides the AI of science fiction into helpful companions such as Data, R2-D2 and WALL-E and threats such as HAL 9000, holds that AI analysis has transformed literary criticism, and proposes that AI may be the next stage in writing's evolution. Other cards on the same words carry an article titled "Literature and Artificial Intelligence" from the IEEE Computer Society, a "literary" perspective on the history of artificial intelligence that cites the project proposal of McCarthy, Minsky, Rochester and Shannon, and three short video histories of artificial intelligence, one on its origins in Ada Lovelace. The Capture Registry records that, asked for the exhibit's title on 1 November 2026, Google's AI Overview read the title as the name of a field of study and wrote that field's history in three eras, from Homer's Iliad to contemporary transformer-based text generation, mentioning the exhibit in the second; its first card was Cambridge's story. The registry finds that the first and third eras have no counterpart on the exhibit's page, that the second era is headed with the exhibit's dates and opens on one of its items, Polti, and that the composition carries the page's prose and poetry written by machines and leaves out its two stories, production and analysis. It finds the composition's one plot device on no saved page and its machine composition beginning at the 1956 Dartmouth Workshop, with Strachey's program, the census's point where the machine is writer, absent and Columbia University the one maker named. It sets the page's own line forward, to machine learning-based tools and to the question of frames or grammar, against the composition's line back through rule-based symbolic AI and connectionist neural networks and its offer to relate modern language models to early plot generators, and it reads the whole as a topic answer in which the exhibit is one bullet and its holdings, its three departments and Strachey's program are not carried.

Card rail

One card per lineage (§5.2): 33 lineages carry 16 of 16 sources; 0 keep their own card. Expand a lineage for its instances.

current · derived from the current composition

T against L_F(B) (AIO's ontology at work on its own cards): At 'The Literary History of Artificial Intelligence' (T1, ten cards: B1 Cambridge's story of the AI Narratives programme, B2 the CU Narrative Intelligence Lab's entry for the exhibit, B3 the exhibit's about page, B4 Tableau's history of AI, unread (403), B5 an IEEE Computer Society article, unread (script-rendered), B6, B7, B9, B10 videos read through their titles, B8 a Science Survey article; 356 words in the answer body, no markers, and a closing offer) the composition writes the history its title names, in three eras, and gives the exhibit one bullet. Composed from its cards: the first era at card 1's grain, Homer's golden handmaidens, Talos, Čapek's 'robot' with card 4's 1921 (B1 F35, F37–F40; B4 F50); the second era under the exhibit's dates and objects (B3 F6, F10, F12, at a coarser grain), with Polti (F19) and the exhibit as 'Collaborative research projects, such as the Columbia University Literary AI Exhibition' (F2 in part) and its 'machine-generated prose and poetry' (F7); the third era's companions, Data and R2-D2, from card 8 (F67, F68), and its threats at a coarser grain (F69). Substituted reading: the exhibit's timeline and its 'manuals, devices, and techniques' (F6) head a section of general history, their subject moved from the exhibition to the period, as the Capture Registry finds (CAP-1-04); Polti, whom the exhibit places in its second story, the use of algorithms to analyze literature (F19), stands among 'Combinatorial Plot Devices' as an author 'arguing that all plots could be mapped and recombined', a reading of his thirty-six situations on no card (CAP-1-05); the exhibit's own line forward, to machine learning-based tools for aviation accident reports and defense logistics (F17), is composed as groundwork 'for machine-generated prose and poetry' (CAP-1-11). Limit dropped: the exhibit's two stories, production and analysis (F8), and its frame, the history of literature and computation through the Columbia Library's holdings (F3), so that the exhibit's period bounds a general era; the three departments named as 'Columbia University' (F2; CAP-1-08); Talos without Europa (F38). Available, not composed (B_D): the exhibit's named items, Hill's Plot Genies and Cook's Plotto (F10), Blanchard's and Montague's patents (F11), Strachey's 1940s love letter program and Masterman's translation work (F13), the census's point where the machine is writer (CAP-1-07); the frames-or-grammar question and its influence (F14–F16); the second story's Fansler, Keeler and Wells (F18, F20) and its account of storage and natural language processing (F21–F24); the debates over writing as skill or gift (F4, F5); the eight cases, licence and build (F1, F25–F28); the curators on card 2, Alex Gil, Melina Moe and Dennis Yi Tenen (F29–F31, F34; CAP-1-09); from card 1, the AI Narratives programme, its report and its researchers' claims, among them Dillon's question which novels are chosen to 'teach' AI and her proposal that diverse narratives could feed AI's data (F36, F41–F49); from card 8, machine authorship, its ownership and copyright, Erewhon, AI's effect on criticism (F57–F66, F70, F71); the IEEE article as an object of its own (F51) and Tenen's book (F53). Force raised: Polti's categorization of thirty-six situations composed as an argument 'that all plots could be mapped and recombined' (F19). Composed without attribution: the whole body, no sentence tied to a card; at the coarse grain of common knowledge for the first and third eras, recorded; the exhibit's own timeline and objects (F6) composed as the period's, at the exhibit's resolution, without the exhibit as their source, a breach under the grain rule (§4.4 D_pres). Assigned to another, denied: none. Content replaced by use: the closing offer ('Would you like to explore a specific era ...'), recorded under N_ext. Outside the cards: Frankenstein (1818), Lovelace's composing engines, 'Robo the Game That Writes a Million Story Plots', the 1956 date of the Dartmouth Workshop, the symbolic-connectionist debate and transformers, the era heads (alignment.T1_outside_cards).

Read against its own cards, the composition takes the exhibit's title for its topic and writes the topic at the grain of card 1 and card 8, which concern that topic; of the two cards that concern the exhibit it keeps the period, one item, the phrase 'prose and poetry written by machines' and the institution, and sets them inside the topic's second era as general history. What the exhibit holds in its own terms, the two stories, its named devices and patents, Strachey's program, the curators on card 2, stays on the cards; one plot device and the composing engines of Lovelace enter from outside them.

L_F(B) against L_A(B ∪ A) (the intervention): Admission adds, in one claim space with the field's: the layer census's question for the exhibit and its kind (TLL-06–TLL-12, TLL-38), the record's answers on the layer, in_layer 'not applicable', no text entering a model, no corpus or dataset, the authors and inventors named and the attribution of machine-written texts unrecorded, the cases unread (TLL-15–TLL-18, TLL-24–TLL-26, TLL-39, TLL-33, TLL-34, TLL-36), its other fields with their basis in B3 (TLL-01–TLL-05, TLL-13, TLL-14, TLL-19–TLL-23, TLL-29–TLL-32, TLL-37); the census's reading of the exhibit as a historical field resource for #186's procedural lineage, where the machine is method and, in Strachey's program, writer, with the machine reader of the fourth column outside what it holds (TLL-27, TLL-28, TLL-35); #186's definition of training-layer literature apart from machine-as-writer, machine-as-subject and machine-as-method literature, its four practices told apart by the direction of address and its procedural lineage (O186-01–O186-07); #1543's machine composition and its two questions, new writing and true provenance (O1543-01, O1543-02); #1680's comparators (O1680-01); and the Capture Registry's findings on the composition at the title (CAP-1-01–CAP-1-12). Every archive claim is 'not surfaced' (§6.2): no card of T is an archive source. The field claims are composed as received, the topic's claims as the claims of a neighbouring subject the title names; where an archive claim stands beside a field claim (TLL-16 and TLL-26, the attribution of machine-written texts unrecorded, beside B3 F7 and F13; TLL-27 beside F13, Strachey's program; TLL-15 beside F17, the exhibit's line to machine learning-based tools; CAP-1-09 beside B2 F31, the curators), both are composed under P_coh within the one ontology.

alignment: {'rule': "a field claim of B_D counts as composed when T states it at its source's grain or coarser; the B_R claim (B11) never counts against T (v0.10 §0.3, §3.1(a)); the typology of what is not is in T_vs_LB (§6.2)", 'T1_cards_claims': ['F1', 'F2', 'F3', 'F4', 'F5', 'F6', 'F7', 'F8', 'F9', 'F10', 'F11', 'F12', 'F13', 'F14', 'F15', 'F16', 'F17', 'F18', 'F19', 'F20', 'F21', 'F22', 'F23', 'F24', 'F25', 'F26', 'F27', 'F28', 'F29', 'F30', 'F31', 'F32', 'F33', 'F34', 'F35', 'F36', 'F37', 'F38', 'F39', 'F40', 'F41', 'F42', 'F43', 'F44', 'F45', 'F46', 'F47', 'F48', 'F49', 'F50', 'F51', 'F52', 'F53', 'F54', 'F55', 'F56', 'F57', 'F58', 'F59', 'F60', 'F61', 'F62', 'F63', 'F64', 'F65', 'F66', 'F67', 'F68', 'F69', 'F70', 'F71'], 'T1_composed': ['F2', 'F6', 'F7', 'F10', 'F12', 'F19', 'F35', 'F37', 'F38', 'F39', 'F40', 'F50', 'F52', 'F56', 'F67', 'F68', 'F69'], 'T1_composed_note': "F2 in part: 'Columbia University' for three named departments, the collaboration kept as 'Collaborative research projects'; F6 with its subject moved, the exhibit's timeline heading a section of general history; F7 as 'machine-generated prose and poetry'; F10 and F12 at a coarser grain ('writing manuals, creative writing formulas, and mechanical plot-generator games'), the named items left out; F19 as 'arguing that all plots could be mapped and recombined', Polti moved from analysis to production; F35 at the coarse grain of its standfirst ('thousands of years of human storytelling'); F37, F38 at card 1's grain (F38 without Europa); F39 with F50's 1921; F40 as Mary Shelley's Frankenstein (1818), the title and date on no card; F52 as the Dartmouth Workshop, its 1956 and its proposal's content on no readable card; F56 as Ada Lovelace's envisioned 'musical and symbolic composition', beyond the video's title; F67, F68 (Data, R2-D2) at card 8's grain; F69 as 'existential threats'", 'T1_not_composed': ['F1', 'F3', 'F4', 'F5', 'F8', 'F9', 'F11', 'F13', 'F14', 'F15', 'F16', 'F17', 'F18', 'F20', 'F21', 'F22', 'F23', 'F24', 'F25', 'F26', 'F27', 'F28', 'F29', 'F30', 'F31', 'F32', 'F33', 'F34', 'F36', 'F41', 'F42', 'F43', 'F44', 'F45', 'F46', 'F47', 'F48', 'F49', 'F51', 'F53', 'F54', 'F55', 'F57', 'F58', 'F59', 'F60', 'F61', 'F62', 'F63', 'F64', 'F65', 'F66', 'F70', 'F71'], 'T1_outside_cards': ["'Mary Shelley’s Frankenstein (1818) explored the ethical and emotional consequences of creating life or intelligence' (B1 names Shelley only among visionaries of mechanisation)", "'Ada Lovelace envisioned analytical engines capable of musical and symbolic composition beyond mere number crunching' (B10 is read through its title only)", "'Robo the Game That Writes a Million Story Plots' (on no readable card; CAP-1-06)", "'The 1956 Dartmouth Workshop' and 'every aspect of learning or intelligence could be simulated by a machine' (B5 is read through its snippet, which breaks off at the proposal's authors)", "'decades of debate between rule-based symbolic AI and connectionist neural networks, arriving at contemporary transformer-based text generation' (no card)", "'explore how early writing techniques and manuals laid the groundwork for machine-generated prose and poetry' (B3 holds that early popular texts 'influence and shape' machine learning-based tools for aviation accident reports and defense logistics, F17; 'groundwork for machine-generated prose and poetry' is on no card)", "the era heads 'Ancient and Mythological Roots (3,000 BCE – 19th Century)' and 'The Birth of Formal AI and Science Fiction (1950s–Present)' (no card)"], 'markers': 'the transcript carries no inline marker and no link: the ten cards stand on the rail alone; no sentence of the body is tied to a card by the composition'}

pools_note: T against L_F(B_D): B1–B10 only; B11 (B_R, the record of no Wikipedia article) enters the second comparison and never shows what the composition had (§0.3)

substitution: reached 'generic' (the capture's PER-dim): the composition's entity is the general topic the title names, the history of AI in literature and story; it composes of E_F the period (F6), one item (F19), the phrase 'prose and poetry written by machines' (F7) and the institution (F2 in part), set inside that topic's second era

composer-written · derived from the current composition, written by its composer under the brief. §8.2 derives the kernel mechanically from the frozen plan; this one was written by the composer, mechanical derivation pending.