Wikidata

register entry wikidata-aio-20261101 OBS-1a237ffb01e9 · file as first run sha256 0ffd6dec00433a58…

• T1-1 definition: 'Wikidata is a free, open, and collaborative multilingual knowledge graph hosted by the Wikimedia Foundation.'
• T1-2 function: 'It serves as a central storage hub for structured data used by Wikipedia, Wikimedia sister projects, and external applications.'
• T1-3 Open License: 'All data is released under a Creative Commons CC0 public domain dedication, meaning anyone can use, copy, and modify it freely.'
• T1-4 Items and Identifiers: 'Each entity gets a unique "Q-number" identifier (such as Q42 for Douglas Adams), while properties use "P-numbers" (such as P69 for "educated at").'
• T1-5 Statements and Triples: 'Information is stored as subject-predicate-object statements (claims) that link entities to values, often backed by references to external sources.'
• T1-6 Multilingual: 'Labels, descriptions, and aliases exist across hundreds of languages so content can be managed and viewed globally.'
• T1-7 SPARQL Query Service: 'Users can extract and analyze data or build visual graphs using the Wikidata Query Service.'
• T1-8 closer: 'Would you like help writing a specific SPARQL query or finding the Q-number identifier for a particular topic?'

• B1 Wikidata — Wikidata card; B1.html sha256 1b85b6cc2f727937, HTTP 200 at 2026-11-01T06:15:16Z
• B2 Wikipedia — Wikidata - Wikipedia card; census copy wikidata--wikipedia sha256 6eab13fa20541e89, HTTP 200 at 2026-11-01T05:32:48Z
• B3 Wikidata Query Service — Wikidata Query Service card; B3.html sha256 c08d8b554051466c, HTTP 200 at 2026-11-01T06:15:17Z; HTTP 200, 17,940 bytes, script-rendered (the extracted text is the interface's controls only); card snippet stands
• B4 Wikidata — Wikidata:Introduction card; census copy wikidata--intro sha256 a5d3749d5b95b101, HTTP 200 at 2026-11-01T05:32:48Z
• B5 YouTube·Llyfrgell Genedlaethol - National Library of Wales — An introduction to Wikidata card; unreached: a video card; WebSearch on title and channel did not return the video's URL; title and snippet stand
• B6 GitHub — What is Wikidata? card; B6.html sha256 565f2139ec9316c0, HTTP 200 at 2026-11-01T06:15:47Z
• B7 YouTube·Wikimedian in Residence - University of Edinburgh — A brief intro to Wikidata and how to set up an account. card; title only (no snippet on the rail)
• B8 YouTube·Columbia SPS — Wikidata, Knowledge Graphs, and Beyond card; title only (no snippet on the rail)
• B9 Wikidata — Property:P742 (pseudonym) own_page; census copy wikidata-pseudonym--property sha256 ab8481d73509b550, HTTP 200 at 2026-11-01T05:06:34Z
• B10 Wikipedia — International Standard Name Identifier census_third_party; census copy isni--wikipedia sha256 24575f41df384288, HTTP 200 at 2026-11-01T05:32:47Z
• B11 Wikipedia — Virtual International Authority File census_third_party; census copy viaf--wikipedia sha256 d506ac21de7d3c5c, HTTP 200 at 2026-11-01T05:32:48Z
• B12 Wikipedia — ORCID census_third_party; census copy orcid-names--wikipedia sha256 08dcd844e5310347, HTTP 200 at 2026-11-01T05:06:34Z

E_A, the archive's entity: the same knowledge base, in the census record and the archive's deposits: name-holding infrastructure that keeps a chosen name as data about its bearer (P742) and also holds heteronyms as items of their own, the archive's among them; a gate whose notability policy removed the archive's heteronym items in December 2025 and January 2026; a store whose statements travel downstream in a truthy form without their references LS-01 LS-02 LS-29 LS-12 LS-07 LS-30 LS-06 D1697-02 LS-23 LS-31 LS-10 D623-01 LS-28 LS-27 D623-02 D115-01 D261-01 LS-13 LS-25 D755-01 LS-08 LS-09 LS-14 LS-34 D668-01 D96-01 D261-02 D261-03 D261-04 D668-02 D761-01 D761-02 D761-03 D761-04 D761-05 D761-06 D668-03 D668-04 D1009-01 D1697-04 D1611-01 LS-16 D1697-01 D1697-03 D1192-03 LS-15 LS-24 LS-19 D1192-01 D1192-02 D1192-04 D1192-05 D5-01 D5-02 D5-05 D5-03 D5-04 D5-06 O1722-02 LS-03 LS-04 LS-05 LS-32 LS-33 O940-01 O940-02 O940-03 O1722-01 LS-21 LS-11 LS-22 LS-17 LS-18 LS-20 LS-26 D738-01 D1697-05 CAP-1-07 CAP-1-06 CAP-1-01 CAP-1-08 CAP-1-02 CAP-1-04 CAP-1-05 CAP-1-03 CAP-1-09; defining papers #5, #96, #115, #261, #623, #668, #738, #755, #761, #940, #1009, #1192, #1611, #1697, #1722

E_F, the field's entity (type C — a conventional reading holds it (public entity: a collaboratively edited multilingual knowledge base and knowledge graph of structured data)): Wikidata, 'a free , collaborative , multilingual , secondary knowledge base, collecting structured data' (its Introduction, F29), 'a collaboratively edited multilingual knowledge graph hosted by the Wikimedia Foundation' (Wikipedia, F7), developed mainly by Wikimedia Deutschland and launched in October 2012 (F9, F19): items with labels, descriptions, aliases and statements, CC0 structured data, a SPARQL Query Service, identifier properties to external authority files, and the pseudonym property P742

ρ = extends: The archive reads Wikidata as its own pages and Wikipedia present it and adds, on that object: the census's reading of how it holds names and what it cannot settle of authorship; heteronym items beside the pseudonym string; the notability gate as observed at the archive's deletions; the two forms of a statement and a falsifiable prediction about reference loss; and its function downstream as an entity source. LS-01 LS-02 LS-29 LS-12 LS-07 LS-30 LS-06 D1697-02 LS-23 LS-31 LS-10 D623-01 LS-28 LS-27 D623-02 D115-01 D261-01 LS-13 LS-25 D755-01 LS-08 LS-09 LS-14 LS-34 D668-01 D96-01 D261-02 D261-03 D261-04 D668-02 D761-01 D761-02 D761-03 D761-04 D761-05 D761-06 D668-03 D668-04 D1009-01 D1697-04 D1611-01 LS-16 D1697-01 D1697-03 D1192-03 LS-15 LS-24 LS-19 D1192-01 D1192-02 D1192-04 D1192-05 D5-01 D5-02 D5-05 D5-03 D5-04 D5-06 O1722-02 LS-03 LS-04 LS-05 LS-32 LS-33 O940-01 O940-02 O940-03 O1722-01 LS-21 LS-11 LS-22 LS-17 LS-18 LS-20 LS-26 D738-01 D1697-05 CAP-1-07 CAP-1-06 CAP-1-01 CAP-1-08 CAP-1-02 CAP-1-04 CAP-1-05 CAP-1-03 CAP-1-09. Guard: ρ = extends claims an addition and no more (§1.2a). It does not make the field's sources wrong at their dates; where the archive's readings of the heteronym model differ among themselves (#96, #261, #668), all are composed under P_coh with their dates.

The case: equal terms, ontology included (v0.9). The archive reads the field's entity as the entity. The census record describes Wikidata from its own pages and Wikipedia, each field with its basis, and reads it as infrastructure that holds a chosen name as data about its bearer (LS-21); the deposits that name Wikidata take it as the knowledge base its pages describe and add to it: the notability gate as it fell on the archive's heteronyms (#5, #1192, #1697), heteronyms held as items of their own (#668, #96, #761), the truthy and full forms of a statement with a prediction of reference loss (#1697), and Wikidata's place downstream as an entity source (#261, #115, #623). Each adds a sense, a mechanism or a limit on the field's own object, and none supplies an ontology of Wikidata the field lacks: ρ = extends, the equal-terms case.

#96 ¶70: Wikidata's architecture is closer to correct. Wikidata models Pessoa's heteronyms as entities of type "heteronym" (Q1146754) with a "creator" (P170) property linking to Pessoa. The heteronym is a *node*, not an attribute.

#761 §II Curatorial Principles: Distinguish heteronym from pseudonym. A heteronym carries a full constructed biography, worldview, and stylistic signature; a pseudonym is a name without that structure.

#1697 §2.4: Wikidata admission is disjunctive, and its second criterion requires no Wikipedia article. … The structured layer is open with a deletion history; the enclosure sits at the article and its sources.

#LS census.json rows[slug == 'wikidata']['ls_reading']: Wikidata holds a chosen name as data about its bearer: the pseudonym property P742 lists 'heteronym' among its aliases with pen name and stage name and stores the name as a string on the item of the person who used it, so in #940's terms the data model keeps the label and records it as a pseudonym of one function. The saved sources do not show whether a heteronym is given an item of its own or how such an item would stand to its civil bearer.

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).

v0.9 §4.10 (v0.10 unchanged): 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): v0.9 0 of 18 (positions [None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None]; composed first under v0.9, so there is no v0.7 object to measure against). Opening-run check (§5.10): no breach.

Wikidata is a free and open knowledge base that can be read and edited by humans and machines and, in its introduction's words, a free, collaborative, multilingual, secondary knowledge base collecting structured data for Wikipedia, Wikimedia Commons, the other Wikimedia wikis and anyone in the world. Wikipedia calls it a collaboratively edited multilingual knowledge graph hosted by the Wikimedia Foundation, launched in October 2012 and developed mainly by Wikimedia Deutschland, the Foundation's first new project since 2006, its creation funded by the Allen Institute for AI, the Gordon and Betty Moore Foundation and Google. It is the central storage of structured data for the Wikimedia projects, supplying interlanguage links and some infobox data, and a 2020 study found content-affecting transclusion in about 5% of English Wikipedia articles. Editors enter and maintain its data and decide the rules of content, bots also enter data, and a 2020 study found much of it imported en masse by bots; its introduction was last edited on 21 October 2026. Its main page counted 123,855,402 data entities on 1 November 2026; Wikipedia gives more than 123 million in October 2026 and 1.65 billion statements in early 2025, marked as needing verification, with scholarly articles the largest part and biographies next. Items have a label, a description, any number of aliases and a Q-number, such as Douglas Adams (Q42); labels need not be unique, while a label with its description must be. Statements pair a property, a P-number such as educated at (P69), with a value, as "author" pairs a work with a person; qualifiers and references are optional, and each statement has a rank. It records statements with their sources and connections to other databases, which supports verifiability, and data entered in any language is available in all the others. The data is published under CC0 1.0, reusable without permission and commercially, as Wikipedia and the Library Carpentry lesson also say; its footer and Wikipedia's note confine CC0 to structured data in four namespaces, with other text under CC BY-SA, and its content is exported in standard formats and interlinkable on the linked data web. The Query Service, released in 2015, answers SPARQL queries, and it and Reasonator are among the tools for examining items; QLever is to replace Blazegraph by 30 June 2027. The Embedding Project of October 2025 supports the Model Context Protocol for AI systems, and Siri and Alexa have used the dataset. Google announced in 2014 the closure of Freebase in Wikidata's favor and closed it in 2016, and Google's Knowledge Graph draws on Wikidata's data. An identifier is a property linking an item to an external database, such as an authority control database, and Wikidata interlinks and crosswalks identifiers such as ORCID, GND and VIAF and contains authority files and bibliographic data. It has properties for ISNI (P213), VIAF (P214) and ORCID (P496), incorporates VIAF file numbers and is listed, with a citation needed, among VIAF's participants; an author's item shows the author's place in national catalogs, VIAF and fan databases, and Verified Handles links accounts to its items. It is limited to basic statements such as who wrote a work, much like a catalog record, and its focus is basic claims, with descriptive and narrative information left to Wikipedia, which complements it. Property P742, pseudonym, "alias used by someone", whose aliases include pen name, stage name, user name and heteronym, takes a string of up to 100 characters on the item of a human, character or group of humans, with qualifiers such as start time, end time and applies to work; it is equivalent to the GND's pseudonym, its examples pair Elsa Bernstein with Ernst Rosmer and Beyoncé with Sasha Fierce, nickname, subject named as and art name are related to it, and its page was last edited on 24 August 2026. A notability policy decides what belongs, property constraints are alerts and hints, and citation on P742 is a suggestion. In 2025 the Digital Public Goods Alliance recognised it as a digital public good, and teaching material introduces it: a Library Carpentry lesson updated on 17 April 2026 and videos from the National Library of Wales, the University of Edinburgh and Columbia SPS.

Wikidata is a free, collaborative, multilingual and secondary knowledge base that collects structured data to support Wikipedia, Wikimedia Commons, the other wikis of the Wikimedia movement and anyone in the world; it describes itself as a knowledge base that humans and machines can both read and edit, and Wikipedia describes it as a collaboratively edited multilingual knowledge graph hosted by the Wikimedia Foundation. It is the central store of structured data for the Wikimedia projects, supplying their interlanguage links and some of their infobox data, and its repository is open to the Foundation's wikis and to others; a 2020 study reported by Wikipedia found that transclusion affecting an article's content occurred in about 5% of English Wikipedia articles. The Wikimedia Foundation hosts and owns it and Wikimedia Deutschland mainly develops it; it was launched on 29 October 2012 as the Foundation's first new project since 2006, its creation funded with €1.3 million from the Allen Institute for AI, the Gordon and Betty Moore Foundation and Google. Its data is entered and maintained by editors, who decide the rules of content creation and management, and by automated bots, and a 2020 study reported by Wikipedia found that a large share of the data was imported en masse from other databases by bots; its method combines community editing and bot imports with references, ranks and constraints. On 1 November 2026 its main page counted 123,855,402 data entities; Wikipedia gives more than 123 million entities in October 2026 and 1.65 billion item statements in early 2025, the second figure marked as needing verification, and items for scholarly articles are the largest part of its content, followed by biographies. Its repository consists mainly of items, each with a label, a description and any number of aliases and identified by a Q followed by a number, a persistent QID; its showcase item is the author Douglas Adams, Q42. Labels need not be unique, while the combination of a label and its description must be. Statements describe an item by a property and a value, properties carrying P-numbers such as educated at (P69); formally they are key–value pairs, so that the property "author" can take as its value a person such as Sir Arthur Conan Doyle. Qualifiers refine a statement and references may annotate it, both optional parts of a statement on the item page, and each statement has a rank: normal, preferred, or deprecated for a value known to contain errors. On the archive's reading, the statement exists in two forms: the full statement carries rank, qualifiers and references, and the truthy statement keeps the best-ranked simple value, the entity form "at its purest: a name, a property, a value, with the operation and its evidence left on the full statement". It calls itself secondary because it records statements with their sources and their connections to other databases, which supports verifiability, and the census accordingly counts it no primary source; data entered in any one language is available in all the others. Its introduction publishes the data under the Creative Commons CC0 1.0 dedication, to be copied, modified and distributed without permission and for commercial purposes, and Wikipedia and a Library Carpentry lesson describe its data the same way; its page footer, a note in Wikipedia and the census record confine CC0 to structured data in the main, Property, Lexeme and EntitySchema namespaces, with text in the other namespaces under CC BY-SA, held by the Wikimedia Foundation, and its content is exported in standard formats and can be interlinked with other open data sets. Ford and Iliadis, quoted by the archive, observe that third parties can legally store its facts in proprietary databases and so lose their dependence on Wikidata and Wikipedia as a source, with possible costs to data workers and the political economy of data labor. The Wikidata Query Service, announced in 2015, answers SPARQL queries over the data, its page listing the prefixes of its vocabulary, and it and Reasonator are among the tools for examining items; introductory videos present it as the way to ask questions of the data and visualize it. In June 2026 the Wikimedia Foundation and Wikimedia Deutschland chose QLever to replace Blazegraph, with all query traffic to move by 30 June 2027, and the Query Service Days of October 2026 addressed the migration; the introduction was last edited on 21 October 2026. The data is also reached through APIs, among them a REST API stable since 11 November 2024, and through Lua on client wikis. The Wikidata Embedding Project, available since October 2025, offers vector-based search by plain-language query and supports the Model Context Protocol, making the data more readily available to AI systems, and virtual assistants such as Siri and Alexa have used the dataset; the archive states that Wikidata feeds AI systems. Google helped fund its creation, announced in December 2014 that it would shut down its own Freebase in Wikidata's favor and closed Freebase in 2016, and Google's Knowledge Graph draws on Wikidata's data; the archive reads Google's entity reconciliation as able to draw on Wikidata QIDs beside Wikipedia pages, Knowledge Graph IDs and ORCID. On a reading the archive holds, inclusion in Wikidata and Wikipedia increasingly decides whether an entity is real for the applications downstream of them: search results, knowledge panels, AI training data and citation indices. A property that links an item to an external database, such as an authority control database used by libraries and archives, is called an identifier, and through such properties Wikidata interlinks identifiers like ORCID, GND and VIAF, maps them for authority control and crosswalks one to another; it also contains authority files and bibliographic data. It has properties for ISNI (P213), VIAF (P214) and ORCID (P496), it incorporates VIAF file numbers, and Wikipedia lists it, with a citation needed, among VIAF's participants. An author's item shows, in its Identifiers section, how the author stands in national library catalogs, the Virtual International Authority File and fan databases, and Verified Handles links its directory of social media accounts to Wikidata items. The archive anchors its own deposits to Wikidata among the external identity authorities of SPXI's fifth layer, with the stated dependency that the anchor degrades if those authorities are compromised or shut down. It is limited to basic statements, such as when and where a work was published and who wrote it, much like a catalog record, and Wikipedia carries the descriptive and narrative account; the census records it as a holder of statements about items, among them works, with an "author" statement pairing a work with a person. Property P742, pseudonym, is "alias used by someone", and its aliases include pen name, stage name, user name and heteronym; its value is a string of up to 100 characters without a semicolon, set on the item of a human, character, group of humans or another allowed subject, with qualifiers such as start time, end time and applies to work. Its examples pair a person with a name (Elsa Bernstein with Ernst Rosmer, Beyoncé with Sasha Fierce, Madonna with Louise Ciccone); it is equivalent to the GND's pseudonym element, nickname, subject named as and art name stand beside it, and its page was last edited on 24 August 2026; the census reads it as one property of Wikidata. Heteronyms are also held as items of their own: in April 2026 most of the Pessoa heteronyms on Wikidata were items described as "heteronym used by Fernando Pessoa", with their statements of instance, use, biography and works mostly carrying no reference. Within the archive three readings of that model stand: that Wikidata treats Pessoa's heteronym as a first-class node of type heteronym linked to Pessoa by a creator property, "closer to correct" (#96, May 2026); that its categories assume one author to one person and force a heteronym either into its biographical author or into fiction, collapsing heteronyms into pseudonym categories (#261, January 2026); and that it flattens Pessoa's typology into the one class "heteronym" (#668, April 2026). Between 22 April and 20 May 2026 an account declared as the archive editor's created about 155 items and amended more than 60, among them heteronym items such as Johannes Sigil (Q139796042) and thirty-three heteronyms of Pessoa, under the principle that a heteronym, with a constructed biography, worldview and style, is distinguished from a pseudonym; an amendment it made to the item "heteronym" was reverted. The archive's Pessoa Knowledge Graph models the kinds of heteronym as distinct values of instance of (P31) and the heteronym as author (P50) of its works, and reads statements sourced only by import from Wikipedia or VIAF as vulnerable. In the archive's minting procedure for a heteronymic position, the description of its Wikidata item declares the position heteronymic, as the canonical web page and the ORCID alternate-name field do (#1009). Wikidata holds the archive's concepts and persons as items, among them Lee Sharks (Q139713433) and the Crimson Hexagonal Archive (Q139807914), both resolving on 20 October 2026, and Operative Semiotics (Q139807910), the authorship of these items left unestablished by the archive. A notability policy decides what data belongs: an item is admitted with a valid sitelink, as a clearly identifiable conceptual or material entity describable by serious and publicly available references, or for a structural need, so that admission needs no Wikipedia article, and on the archive's reading the threshold is lower than Wikipedia's by design. Property constraints are treated as testing alerts and hints, citation on P742 is a suggestion constraint, and the census records that references are optional, so that whether a given statement is sourced stays open. On 20 December 2025 the items for the archive's heteronyms and for the New Human movement were deleted after a notification applying a test of a Wikipedia article, a structural need, or identifiers with serious sources; Johannes Sigil was kept for a time as founder of the deleted New Human, and after the record was published and an ombudsman complaint made, the logs and one entry were restored. On 4 January 2026 the item Q137446960, Johannes Sigil, holding a label and the description "Poet and founder of New Human", was deleted as an "Empty item", and New Human (Q137460296) was deleted as spam; once an item is deleted, its history is visible to administrators alone. The archive reads these deletions as a threshold favoring institutional validation over independent scholarship, creation bound to criteria and deletion to an administrator's judgment, the burden falling on the described entity, and counts as lost a node connecting a name to its works (#5); it places the heteronymous among those on whom the existence tax falls, everyone whose existence the index has not already ratified (#1722). The leesharks.com authorship census holds Wikidata as infrastructure in its areas of attribution and credit, heteronymy and meta-aggregation, admits it as an aggregator that gathers many records and says what it holds, and asks of it what sourced statements it holds about a person or work, the names a person used and the identifiers in other authority files, under the census's question of how authorship is made, named, distributed, attributed and sustained and what is lost when those relations collapse. In the terms of #940, a pseudonym changes the label attached to an authorial function and a heteronym changes the organization of the function itself, which is attributable through the civil bearer without being reducible to that bearer; the civil name is the atomic unit of attribution across the systems that govern legitimacy, and a heteronym is declared by a published registry of the orthonym, the roles and the dates (#1722). On the census's reading, Wikidata holds a chosen name as data about its bearer, keeping the label and recording it as a pseudonym of one function, and it records no authorial function: an "author" statement links a work to an item, and P742 gives no place to a voice, corpus or responsibility distinct from the item's subject. It answers what sourced statements it holds, under which pseudonyms a person is recorded with what qualifiers, which authority identifiers an item carries and how a pen name or heteronym is held; it leaves open whether a name is a pseudonym or a heteronym in #940's sense, P742's aliases treating the two as one, and how authorial function is distributed in a work. The census did not read the notability and help pages, the items that use P742, or how a heteronym with an item of its own is linked to its bearer. In 2025 the Digital Public Goods Alliance recognised it as a digital public good, and teaching material introduces it, among it a Library Carpentry lesson last updated on 17 April 2026 and videos from the University of Edinburgh and Columbia SPS on setting up an account and on knowledge graphs. At the address of one of the archive's deletion records, a composer drew on Wikidata's deletion policy pages, administrator edit-count statistics and Meta-Wiki talk pages (#738). The archive predicts that where Wikidata statements exist both referenced and as truthy values, compositions reproduce the truthy value without its reference more often than they reproduce referenced article sentences without citation, and no difference would falsify the prediction (#1697). Asked for the name on 1 November 2026, Google's AI Overview introduced the knowledge base as a newcomer would meet it, with host, license, items and statements and the query service, and the Capture Registry finds it accurate as a data model and query service, its statements and multilingual labels composed at their grain. The registry finds that the Overview left out Wikimedia Deutschland, the editors, the bots, the launch date and the count, gave the internal identifiers Q42 and P69 and no external one, used "aliases" as its only word for a second name and stated the license wider than the introduction's footer, so that the identifier hub and the pseudonym property, the two features that make Wikidata a census row, were absent.

current · derived from the current composition

T against L_F(B) (AIO's ontology at work on its own cards): At "Wikidata" (T1, eight cards: B1 the main page, B2 Wikipedia, B3 the Query Service, B4 the Introduction, B5 a National Library of Wales video, B6 Library Carpentry's lesson on GitHub Pages, B7 and B8 videos by title), T composes the knowledge base at its data model: the definition (T1-1, F7 nearly verbatim with F1 and F29), central storage (T1-2, F2, F45), the CC0 license (T1-3, F30, F8, F52), items and properties by number (T1-4, F33, F34, F5), statements with references (T1-5, F13, F14, F32), multilingual labels (T1-6, F33, F36) and the Query Service (T1-7, F20, F38, F42, F28). Available in B_D and not composed: the developer, the editors and the bots (F9, F31, F17), the launch and its funding (F19, F18), the counts (F3, F10, F16), label uniqueness (F12), the rank (F14 in part), the constraints and notability (F15, F39), the limit to basic statements (F53, F54), Google's relation to Wikidata (F22, F44), the Embedding Project and machine reuse (F21, F23) and the QLever migration (F6, F20 in part). Limit dropped: T1-3 states 'All data is released under' CC0, which the introduction's body (F30), Wikipedia's lead (F8) and the lesson (F52) also state; the footer of the introduction (F40) and Wikipedia's note (F24), both in B_D, confine CC0 to structured data in four namespaces, and T carries the claim without that qualification. Substituted sense: T1-4 heads 'Items and Identifiers' and gives the Q- and P-numbers as identifiers (F11's sense); the sense B4 defines, a property linking an item to an external database such as an authority control database (F35), with B6's ORCID, GND and VIAF crosswalks for authority control (F46–F50), is available and not composed. Composed without attribution: none at the source's resolution beyond the cards' own wording; T cites its cards inline. Closer: T1-8, an action prompt. The pseudonym property (B9) and the ISNI, VIAF and ORCID properties named in the census's third-party sources (B10–B12) are B_R: they cannot show what the composition had (§3.1(a)), and stand in the second comparison as not surfaced.

v0.10 §0.5: the first diff is ontological and narrow. T and the field agree on the entity (a collaborative knowledge base of structured data); T composes it as a newcomer's data model and query tool, and leaves out what the cards say of who makes it, how its data is admitted and limited, and how its identifier properties join it to the name authorities. The one sense T changes is 'identifier', where the field holds two.

L_F(B) against L_A(B ∪ A) (the intervention): Admission adds, in one claim space with the field's: the B_R pool's pseudonym property, P742, a string on the bearer's item with 'heteronym' among its aliases (F58–F68, not surfaced), and its ISNI, VIAF and ORCID properties (F69–F74, not surfaced); the census record's reading of it as name-holding infrastructure that keeps a chosen name as data about its bearer and records no authorial function (LS-03, LS-09, LS-11, LS-21), with what it can and cannot settle (LS-17–LS-20); #940's distinction of pseudonym and heteronym and #1722's declared registry (O940-01–03, O1722-01); the archive's record that heteronyms are also held as items of their own (D668-01, D96-01, D761-04, D761-05, D1009-01), with three archive readings of that model held side by side under P_coh (D96-01; D261-02–04; D668-02); the archive's own items and edits on Wikidata (D761-01–06, D668-03, D1697-04, D1611-01); the notability gate and the deletions of the archive's heteronym items in December 2025 and January 2026, with the partial restoration and the archive's reading of the threshold (D1697-01, D1697-03, D1192-01–05, D5-01–06, O1722-02); the full and truthy forms of a statement and the prediction P4 of reference loss, with its falsifier (D1697-02, D1697-05); Wikidata as an entity source downstream (D261-01, D115-01, D623-02) and the CC0 dynamics Ford and Iliadis describe (D623-01); SPXI's anchoring to it (D755-01); and the Capture Registry's findings on T (CAP-1-01–09). None contests what the field says Wikidata is: each adds a sense, a mechanism or a limit on the field's own object (ρ = extends).

alignment: {'rule': "a field claim counts as composed when T states it at its source's grain or coarser; B_D (cards) alone answers the first question (v0.10 §0.3, §3.1(a)); B_R claims are listed apart and are never 'available, not composed'", 'T1_composed': {'T1-1': ['F7', 'F1', 'F29'], 'T1-2': ['F2', 'F45', 'F37'], 'T1-3': ['F30', 'F8', 'F52'], 'T1-4': ['F33', 'F34', 'F5', 'F11'], 'T1-5': ['F13', 'F14', 'F32'], 'T1-6': ['F33', 'F36'], 'T1-7': ['F20', 'F38', 'F42', 'F28']}, 'B_D_available_not_composed': ['F3', 'F4', 'F6', 'F9', 'F10', 'F12', 'F15', 'F16', 'F17', 'F18', 'F19', 'F21', 'F22', 'F23', 'F24', 'F25', 'F26', 'F27', 'F31', 'F35', 'F39', 'F40', 'F41', 'F43', 'F44', 'F46', 'F47', 'F48', 'F49', 'F50', 'F51', 'F53', 'F54', 'F55', 'F56', 'F57'], 'B_D_limit_dropped': ['F40 (B4 footer: CC0 for structured data in four namespaces, CC BY-SA for other text) and F24 (B2 note, the same) beside T1-3, which composes F30, F8 and F52 at their grain'], 'B_D_substituted_sense': ["'identifier': T1-4 gives the internal Q- and P-numbers the name 'identifier' (F11's sense, the QID as persistent identifier); F35 (B4) defines an identifier as a property linking an item to an external database such as an authority control database, a sense T does not compose, with F46–F50 (B6) on authority control and crosswalks"], 'outside_cards': ["'hundreds of languages' (T1-6): no saved field text states a count of languages for labels; the nearest are F36 (any language available in all) and the main page's interlanguage list, not ledgered as a claim", "'subject-predicate-object statements (claims)' and 'triples' (T1-5): B6 calls statements triples ('Statements are also known as “triples”', not ledgered); F13 states key–value pairs", "'build visual graphs' (T1-7): F42 ('visualizing all the data', a video card)"], 'closer': 'T1-8, an action prompt (content replaced by use, §6.2)', 'B_R_not_surfaced': ['F58', 'F59', 'F60', 'F61', 'F62', 'F63', 'F64', 'F65', 'F66', 'F67', 'F68', 'F69', 'F70', 'F71', 'F72', 'F73', 'F74']}

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.