2026-09-25T12:00:00Z https://augmanitai.com/oai

oai:augmanitai.com:zenodo/22925700 2026-09-23 publications

AUGMANITAI Public Terms and Q&A Corpus (English/German) Ehstand, Andreas Dataset Dataset application/pdf eng deu 2026-09-23 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22925700 doi:10.5281/zenodo.22925700 Bilingual (English/German) terms and question–answer corpus covering all terms of the AUGMANITAI research programme by Andreas Ehstand that are already openly published on Zenodo. It contains 137 terms with the author's verbatim definitions, sources and licences, and 769 question–answer pairs (388 English, 381 German), including 25 pairs taken over unchanged from the Entity Operating Layer record. Files: qa_corpus.jsonl, terms_unified.jsonl / .json, terms.csv, definedtermset.jsonld (schema.org DefinedTermSet), subsets/ (CC BY 4.0 entries only), README.md (dataset card), LICENSE.md. Licence per entry. 56 terms and 283 pairs are CC BY 4.0. 81 terms and 486 pairs come from the AUGMANITAI Public Lexicon (CC BY-NC-ND 4.0) and are quoted unchanged; for these the NonCommercial and NoDerivatives conditions apply. Every entry carries its own licence field. "License of Clarity" is a registered EU trade mark (EUIPO 019206780); no trade-mark rights are granted. Limitations. Definitions are the author's; no claim of universal novelty is made. Several English names are ordinary expressions or are used elsewhere with other meanings; each entry refers only to the AUGMANITAI definition. Compiled with AI assistance from the author's own published definitions. Version 2: term identifiers in definedtermset.jsonld aligned with the augmanitai.com lexicon machine layer (https://augmanitai.com/lexicon/terms/<slug>#term); non-lexicon terms identified under this record's concept DOI. terminology human-AI interaction lexicon definitions question answering bilingual German English AUGMANITAI DefinedTermSet https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22909458 2026-09-23 publications

The Entity Operating Layer: Foundations of a Recordable, Verifiable and Transferable Layer of Meaning Ehstand, Andreas Text Working paper application/pdf eng deu 2026-09-23 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22909458 doi:10.5281/zenodo.22909458 Since 1977, computing has modelled the individual as a subset of an expert: the overlay literature records a person's divergence from the reference model as a gap to be closed. That assumption is implicit in how systems are aligned to human preference today, and it has never been inverted. This working paper inverts it. It sets out the foundations of the operating layer of an entity &mdash; its meanings, boundaries, priorities, decision grounds and the affective states that explain those decisions &mdash; recorded as a structured, versioned, machine-readable body of work, measurable against the entity it describes, anchored to a date by third parties, and portable between systems. The divergence is not the deficit. The divergence is the holding. The layer applies to any entity describable in language: a person, a team, a company, a discipline, a machine, a software agent, a swarm. The paper names its antecedents precisely (Kelly 1955; Carr & Goldstein, MIT 1977; Boose, AAAI 1984; the overlay user-modelling tradition), states the three conditions that had to coincide before the inversion was possible, and derives a falsifiable alignment claim: verifiable alignment requires a named entity with a recorded operating layer, because alignment to an aggregate has no party to verify it against. Published in English and German. Includes the Sealed Register of Recorded Inventions: 803 separately recorded inventions represented by cryptographic fingerprint only, each independently re-computable, with the full entries deposited under restricted access (10.5281/zenodo.22909339). The working procedure &mdash; the questions that locate divergence, the conditions under which a definition becomes valid, the individual fidelity scoring rules, and the corpus &mdash; is not part of this pu semantic layer personal ontology overlay model user modelling knowledge elicitation AI alignment verifiable alignment personalisation provenance of meaning cryptographic commitment timestamping prior art https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22897795 2026-09-22 publications

Priority Claim: Individualized Semantic Reference Layers for Language Models Ehstand, Andreas Text Other application/pdf eng deu 2026-09-22 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22897795 doi:10.5281/zenodo.22897795 Andreas Ehstand is the author of a method in which a language model's own definitions serve as a reference layer, one person's differing meanings and decision reasons are elicited against it, stored as a versioned record, and used by later AI to continue that person's thinking. This public priority statement names the four elements of the method (model-generated reference layer; person-specific difference; persistent, versioned record; continuation by later AI) and lists the dated record of the author's work from 2025 onward, each entry with a third-party date (Zenodo, EUIPO, Internet Archive, Bitcoin blockchain). English and German versions included. Deutsch: Öffentliche Prioritätserklärung von Andreas Ehstand zu individualisierten semantischen Referenzschichten für Sprachmodelle: vom Modell erzeugte Referenzschicht, personenspezifische Differenz, versionierte Aufzeichnung, Fortführung durch spätere KI – mit datierten Belegen ab 2025. Individualized Semantic Reference Layers priority claim Andreas Ehstand personal semantics language models cognitive continuity defensive publication AUGMANITAI https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22871354 2026-09-21 publications

Feeling Understood by Conversational AI: A Provisional Taxonomy and a Recognition–Disclosure Hypothesis Ehstand, Andreas Text Working paper application/pdf eng deu 2026-09-21 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22871354 doi:10.5281/zenodo.22871354 AI-generated conceptual working paper. Not peer-reviewed; no participant study or empirical validation is reported. Conversational AI can become the object of experiences that users describe in relational terms, including feeling understood, noticing a social orientation, or questioning the origin of a formulation. This paper specifies six candidate codes: The Blank Cursor, The Seen Feeling, Digital Confidant Drift, The Parasocial Slip, The Origin Doubt, and Still Here. A comparison with existing research identifies their proposed observational distinctions and temporal scales. Each code receives inclusion criteria, exclusions, and constructed examples. The analytic unit combines an interaction episode with a participant account; disclosure change additionally requires comparison across time. The source manuscript's six-phase cascade remains an unconfirmed background model. One narrower implication is developed for testing: whether experienced recognition precedes an increased tendency to direct personal reflection to AI. A prospective protocol separates interpretability, coding reliability, discriminant validity, and this temporal association. The contribution is a bounded descriptive specification and research hypothesis, rather than a validated scale or empirical account of a universal progression. AI-generated draft: This manuscript was generated and revised with AI from Andreas Ehstand's March 2026 source manuscript. The coding criteria and research protocol are proposals developed in this revision. No participant study was conducted for this paper, and no human substantive or editorial review of this version is claimed. Manuscript dated 15 September 2026. This edition, prepared on 21 September 2026, corrects an author initial in the bibliography. Examples are cons human-AI interaction subjective experience perceived understanding conceptual taxonomy conversational AI qualitative coding https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22871309 2026-09-21 publications

Andreas Ehstand / AUGMANITAI: Public Zenodo Bibliography — 21 September 2026 Snapshot Ehstand, Andreas Dataset Dataset application/pdf eng deu 2026-09-21 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22871309 doi:10.5281/zenodo.22871309 AI-generated bibliographic catalogue and description, compiled from public Zenodo metadata. A searchable CSV and machine-readable JSON Lines index of 725 publicly discoverable Zenodo records associated with Andreas Ehstand's account, captured on 21 September 2026. The index contains titles, record and concept DOIs, URLs, source-reported resource types, publication dates and access states. It contains metadata and links, not full texts. Of the source records, 28 report open file access and 697 report restricted file access. The restricted records' metadata are public; no restricted files or private drafts are included. The index does not change permissions. The collection includes working papers, preprints, datasets, registers, abstracts, teasers and other outputs. Inclusion is not scientific validation, peer review, endorsement, a priority finding or evidence of 725 open full-text papers. Titles and types are transcribed from the source. The public API responses, row counts, unique DOIs and transcription were checked; the underlying research claims were not comprehensively verified. README.md documents the source query, method, field definitions and limitations. Each linked work retains its own licence. This dated snapshot excludes its own subsequent deposit. Andreas Ehstand AUGMANITAI bibliography research metadata human-AI interaction Zenodo research discovery https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22826650 2026-09-18 publications

The Universal Concept Layer: Language as Versionable Code — A Meaning Layer between Humans, Language Models, Agents, and Robots Ehstand, Andreas Text Working paper application/pdf eng deu 2026-09-18 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22826650 doi:10.5281/zenodo.22826650 Bilingual working paper (English / German) &mdash; the English version is the citation version; the German version is identical in substance and included in this record. Infrastructure for systems combining humans, language models, agents and robots is being standardised rapidly: transport, tool invocation, identity, capability description. One layer goes systematically unaddressed, and not by oversight. The classical agent communication languages reserved a placeholder for it &mdash; a message field pointing to an ontology. What was specified is the reference, not the lifecycle of what is referenced. This paper proposes filling that placeholder rather than setting it again. It argues that the layer of meaning forms a distinct layer of system architecture, the Universal Concept Layer, and that its carrier material is versioned natural language: terms maintained as software artefacts, carrying definition, delimitation, context of validity, timestamp, version and provenance. The paper gives a formal minimal specification, describes the recursive five-step cycle by which such terms arise in practice, restates Korzybski's classical formula in terms of direction of fit for language-controlled cyber-physical systems, delimits the proposal against ten neighbouring traditions along three operationalised criteria, reports the state of implementation since 2025, includes one complete term bundle as an appendix, and states an experimental design by which the thesis can be refuted. Licence note: this record is published under CC BY 4.0. The licence covers the text of this record; the programme's internal working procedures are not part of this publication and are not licensed by it; trademark and name rights remain unaffected. The record opens with the programme's standard legal di universal concept layer language as code versionable semantics terminology infrastructure human-AI interaction agent interoperability semantic drift embodied AI https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22763843 2026-09-15 publications

AUGMANITAI: Selected Public Works — Thoughts, Images, Moving Works, and Interaction Ehstand, Andreas Text Other application/pdf eng deu 2026-09-15 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22763843 doi:10.5281/zenodo.22763843 AI-generated catalogue text and metadata, compiled from Andreas Ehstand's public project materials. This record is a source-linked catalogue, not a peer-reviewed research result. This author-curated catalogue by Andreas Ehstand offers three entry points into his public work: concepts and visual art, AI-generated short films, and the SWAMANITAI research perspective on interaction in mixed teams of people, software agents and machines. The publication contains a readable catalogue and a Schema.org JSON-LD representation. It was compiled with AI editorial assistance from public page text checked on 15 September 2026. It is a selected link catalogue, not an exhaustive inventory, a scientific results paper, or a peer-reviewed publication. The film collection is explicitly AI-generated; the SWAMANITAI entry describes a research perspective and does not demonstrate a deployed robotics system. No original artworks, private manuscripts or unpublished research files are included. The licence applies only to the catalogue files; the rights and licences of the linked works remain unchanged. See README.md for the selection method, sources and scope. AUGMANITAI human-AI interaction conceptual art AI-generated film multi-agent interaction work catalogue JSON-LD https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22257381 2026-09-02 publications

The Ehstand Test: Can a Machine Faithfully Continue One Person's Thinking? Canonical Definition, Score Card, and Protocol (v1.0) Ehstand, Andreas Text Working paper application/pdf eng deu 2026-09-02 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22257381 doi:10.5281/zenodo.22257381 Canonical founding document of the Ehstand test (v1.0, bilingual English/German). The Turing test asks whether a machine appears generically human. The Ehstand test asks a different question: can a machine faithfully continue the thinking of one specific person &mdash; their concepts, held positions, reasoning moves, transfers to new cases, updates under new evidence, refusals, and dialogue continuations? As digital delegates, knowledge succession and "this agent thinks like X" claims spread, such claims need a test. This working paper states the canonical five-sentence definition, the EHSTAND score card (seven dimensions), the public scale Ehstand 0&ndash;5, the measurement protocol (sealed personal archive, commit&ndash;reveal, blind rating against baselines, passing criteria), the task families with weights, limits and open questions, and the relation to prior work. Name use: the term is openly citable as "Ehstand test (Ehstand, 2026)"; independent test runs are encouraged. "Certified Ehstand" remains reserved for protocol-conformant assessments. Licence note: CC BY-ND 4.0 covers the canonical text; trademark and name rights remain unaffected; the research programme's internal working procedures are not part of this publication. Ehstand test personal cognitive fidelity cognitive continuity thought inheritance evaluation protocol benchmark human-AI interaction Turing test AUGMANITAI https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22257280 2026-09-02 publications

Five Dated Predictions on Meaning, Machines, and Human Thinking (2027–2032) — with Public Annual Review (v1.0) Ehstand, Andreas Text Working paper application/pdf eng deu 2026-09-02 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22257280 doi:10.5281/zenodo.22257280 Five dated, falsifiable predictions on meaning, machines, and human thinking (2027&ndash;2032), version 1.0, bilingual English/German &mdash; with a standing commitment to public annual review every September, beginning September 2027. Each prediction carries an explicit deadline and an explicit failure condition: (1) major agent-interoperability efforts adopt concept-level meaning-versioning by end-2028; (2) a recognized standards body publishes a specification for versioned, resolvable concept identifiers for AI systems by end-2030; (3) person-specific fidelity testing becomes an established evaluation category with at least three independent benchmarks by end-2030; (4) at least five organizations publicly report a human-capability measure alongside AI-productivity measures by end-2029; (5) a leading world-model system exposes an explicit, human-readable semantic interface by end-2032. Review rules are fixed in the document: wording frozen, failed predictions remain failed, public evidence only, and the author's own projects do not count as evidence for predictions 2, 3 and 5. The predictions extend theses the author has published openly (Universal Concept Layer; the Ehstand test; the semantic layer for world models). Licence note: CC BY-ND 4.0 covers the canonical text; trademark and name rights remain unaffected. predictions forecasting falsifiability semantic layer concept versioning AI agents world models human-AI interaction Ehstand test AUGMANITAI https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22256965 2026-05-18 publications

The Missing Semantic Layer for World Models — and the Denkraum of Bidirektionale Entitätsübergreifende Semantische Emergenz (NEOMANITAI Whitepaper 2026, v2.0, Open Republication) Ehstand, Andreas Text Working paper application/pdf eng deu 2026-05-18 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22256965 doi:10.5281/zenodo.22256965 Open republication of the NEOMANITAI Whitepaper 2026 (v2.0, dated 18 May 2026, bilingual disclaimers EN/DE): the missing semantic layer for world models, and the Denkraum of Bidirektionale Entit&auml;tsuebergreifende Semantische Emergenz (bidirectional entity-spanning semantic emergence). World models &mdash; internal representations of how the world works &mdash; are largely unidirectional, opaque and slowly updated. This whitepaper describes the missing element as a shared, precisely named, bidirectionally operable and error-correcting semantic layer through which world models could become living, co-evolutionary systems. It anchors the space in philosophy of language, embodied cognition and systems theory, delimits it from scaling, neurosymbolic AI, ontology engineering, alignment and cognitive architectures, defines the core concepts (precise naming, bidirectional probing, distributed error correction, substrate-independent inheritance), and states the Two-Frontiers Grounding Model &mdash; separating recombinatorial work grounded in the corpus from empirical work requiring fresh real-world contact. An observable seed (documented 2025&ndash;2026) is kept strictly separate from the meta-system, which is dated as an explicit wager. Integrity and dating: the file in this record is byte-identical with the author's dated original: SHA-256 aa967581bf92a209d03ce25ffa5367af052251c8b729730c54c44093e39778d9, anchored via completed OpenTimestamps proofs (manifest of 17/18 May 2026, Bitcoin-verified). The document was deposited on Zenodo under restricted access in May 2026; this record makes it openly readable. Verification with public tools: any checksum tool, opentimestamps.org. Licence note: as stated inside the document (&sect;21), the text is licensed CC BY-NC-ND 4.0. Trade world models semantic layer bidirectional human-AI alignment semantic emergence grounding Denkraum human-AI co-evolution NEOMANITAI AUGMANITAI terminology https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22254181 2026-09-02 publications

The AUGMANITAI Concept Registry: A Public, Versioned, Author-Attributed Concept Registry — Reference Implementation of the Universal Concept Layer (v0.1) Ehstand, Andreas Text Working paper application/pdf eng deu 2026-09-02 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22254181 doi:10.5281/zenodo.22254181 Whitepaper on the AUGMANITAI Concept Registry &mdash; a public, versioned, author-attributed concept registry, live at augmanitai.com/registry/ as the reference implementation of the Universal Concept Layer (v0.1, bilingual English/German). Agent ecosystems are standardising transport, discovery and identity. The meaning level &mdash; what a term means, how its meaning changes over versions, and who coined it &mdash; has no public, versioned, cross-organisational infrastructure. This paper describes the design principles (terms as versioned artefacts with semantic versioning: MAJOR = break in meaning, MINOR = refinement, PATCH = wording; stable GET-addressable records; mandatory author attribution; evidence anchors per concept), the data model with schema, and the live reference implementation covering six signature concepts, each resolvable as JSON. Licence note: this record is published under CC BY 4.0. The licence covers the text of this record; the programme's internal working procedures are not part of this publication; trademark and name rights remain unaffected. concept registry universal concept layer versioned meaning semantic versioning AI agents interoperability terminology defined terms AUGMANITAI https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22229153 2026-09-01 publications

Priority and Attribution Register of the AUGMANITAI / NEOMANITAI Research Programme (2025–2026) Ehstand, Andreas Text Other application/pdf eng deu 2026-09-01 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22229153 doi:10.5281/zenodo.22229153 Bilingual register (English / German) of the named concepts, frameworks and coined terms of the AUGMANITAI / NEOMANITAI research programme, with canonical one-sentence definitions and independently verifiable evidence for their origin and dates. Published as a defensive disclosure: the concepts, terms, definitions, combinations and sub-combinations named in this register are placed on the public record as prior art against later exclusive-rights claims by third parties. Four mutually independent lines of evidence carry the programme's core dates: open Zenodo records with full texts, Internet Archive snapshots with checksums, the OpenAlex scholarly index, and OpenTimestamps proofs anchored in the Bitcoin blockchain (blocks 937495&ndash;937840, 20&ndash;22 February 2026). The terms listed denote work developed by Andreas Ehstand and are not generic designations; please cite "Ehstand, Andreas (2026)". "License of Clarity" is a registered trademark (EUIPO 019206780). Licence note: this record is published under CC BY 4.0. The licence covers the text of this record; the programme's internal working procedures are not part of this publication and are not licensed by it; trademark and name rights remain unaffected. This register is version-controlled; corrections appear as new versions, nothing is silently removed. priority register attribution prior art defensive disclosure thought inheritance Gedankenvererbungslehre cognitive legacy documentation machine coaching universal concept layer AUGMANITAI NEOMANITAI https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22229056 2026-09-01 publications

AUG-WP Working Paper Series V3 (2026): Public Title and Abstract Register — 225 Working Papers on Human-AI Interaction Ehstand, Andreas Text Other application/pdf eng deu 2026-09-01 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22229056 doi:10.5281/zenodo.22229056 Public title and abstract register of the AUG-WP working paper series (Version 3, April 2026) &mdash; 225 working papers on human&ndash;AI interaction by Andreas Ehstand. Each entry quotes the paper's own title and, for most papers, its opening passage, verbatim and in the language of the paper (English or German). For 21 papers whose opening passages describe operational detail, only the title is listed. The V3 texts were prepared and sanitized for permanent archival in April 2026; a number of the papers had previously been published as individual open records (later removed), and the full texts are archived in access-controlled deposits whose metadata and dates are public. Purpose (defensive disclosure): the concepts, terms, frameworks, taxonomies and theses named in these titles and passages are placed on the public record as prior art against later exclusive-rights claims by third parties. The papers' full methods and internal working procedures are not part of this publication. The named concepts are not generic designations; please cite "Ehstand, Andreas (2026)". See also the programme's Priority and Attribution Register (10.5281/zenodo.22215501). Licence note: CC BY-NC-ND 4.0 (programme default). The licence covers the text of this record; trademark and name rights remain unaffected ("License of Clarity", EUIPO 019206780). This register is version-controlled; corrections appear as new versions, nothing is silently removed. working paper series human-AI interaction prior art defensive disclosure terminology science cognitive externalization AUGMANITAI attribution register https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22228966 2026-09-01 publications

AUGMANITAI Public Lexicon — 101 Bilingual Terms of Human-AI Interaction (Permanent Archival Snapshot of the Open Web Edition) Ehstand, Andreas Dataset Dataset application/pdf eng deu 2026-09-01 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22228966 doi:10.5281/zenodo.22228966 Permanent archival snapshot of the public AUGMANITAI lexicon: 101 bilingual terms (German/English) naming recurring moments of human&ndash;AI interaction. Each entry carries: identifier, names in both languages, precise definition, context, three delimitations (versus human-only phenomena, machine-only phenomena, and neighbouring terms), cross-references, a plain-language explanation and an everyday example. The lexicon has two parts: a Phenomenology Core of 12 terms, verbatim from the 2026 research manuscript, and an Extension Set of 89 further terms. The live edition with one page per term: https://augmanitai.com/lexicon/ &mdash; this deposit is the announced permanent archival publication of its machine-readable data file (terms.json, generated 2026-07-10, MD5 5075d6e9438ada8a13591f64610ed0d2). The larger private working network of the research programme is not part of this deposit. The terms denote work by Andreas Ehstand and are not generic designations; please cite "Ehstand, Andreas (2026)". Licence note: CC BY-NC-ND 4.0, as declared inside the file. "License of Clarity" is a registered EU trade mark (EUIPO 019206780); no trade-mark rights are granted. lexicon human-AI interaction terminology bilingual phenomenology AUGMANITAI defined terms open data https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22215499 2026-02-20 publications

The Fifth Stage: Recursive Cognitive Externalization as the Next Stage of Human Cognitive Evolution (Republication of the 20 February 2026 Abstract) Ehstand, Andreas Text Working paper application/pdf eng deu 2026-02-20 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22215499 doi:10.5281/zenodo.22215499 Byte-identical republication of the abstract first published on 20 February 2026 (original record later removed). This abstract proposes a fifth stage of human cognitive evolution &mdash; Architectural Cognitive Externalization (ACE) &mdash; extending Merlin Donald's four-stage model (episodic, mimetic, mythic, theoretic). Where every previous stage externalized cognitive outputs, ACE externalizes the cognitive architecture that produces them: decision patterns, compensation strategies, recursive loops and structural blindspots, formalized in ISO-compliant terminology, generative, surviving the biological substrate, and comparable across individuals. It is explicitly delimited from mind uploading, digital twins and AI griefbots. Integrity and dating: the file in this record is byte-identical with the February 2026 original: MD5 5756ef20b125d81affec2a86da3d2306, SHA-256 3594ab6337c35762b464ff07d7009d559059e6a82e714c1e5414e2e8252e91e1. The file carries a completed OpenTimestamps proof anchored in Bitcoin block 937518 (20&ndash;22 February 2026). Verification is possible with public tools (any checksum tool, any Bitcoin block explorer, opentimestamps.org). Licence note: as stated inside the document, the text is licensed CC BY-ND 4.0. Trademark and name rights remain unaffected; the research programme's internal working procedures are not part of this publication. cognitive evolution externalization Architectural Cognitive Externalization Recursive Cognitive Externalization Merlin Donald cognitive architecture human-AI interaction AUGMANITAI https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22211974 2026-08-31 publications

Provenance Chain of the Universal Concept Layer Research Programme, January-August 2026 (Bilingual, DE/EN) Ehstand, Andreas Text Other application/pdf eng deu 2026-08-31 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22211974 doi:10.5281/zenodo.22211974 Bilingual chronicle (German/English) of the first publications of the AUGMANITAI / Universal Concept Layer research programme, January to August 2026. Every link of the chain is attested by at least one independent third party: Internet Archive snapshots (including the programme website as of 28 January 2026), open registry entries with persistent identifiers, an OpenAlex index entry, and the Bitcoin blockchain (blocks 945970/945979, 20 April 2026). Milestones covered: founding of the programme (January 2026); first publication of the Foundational Definitions v2.0 (20 February 2026) and of the Gedankenvererbungslehre (21 February 2026); the Universal Concept Layer working paper V1.0 (15 April 2026); cryptographic anchoring (20 April 2026); bundle registration (22 April 2026); open publication of the Universal Concept Layer paper (30 August 2026, DOI 10.5281/zenodo.22172128). Licence note: this record is published under CC BY 4.0, covering the text of this record; trademark and name rights remain unaffected. provenance priority universal concept layer Gedankenvererbungslehre research chronicle human-AI interaction AUGMANITAI https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22211955 2026-02-20 publications

Foundational Definitions v2.0 — Three Sciences of Mechanism (Republication of the 20 February 2026 Edition) Ehstand, Andreas Text Working paper application/pdf eng deu 2026-02-20 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22211955 doi:10.5281/zenodo.22211955 Byte-identical republication of the edition published on 20 February 2026 under DOI 10.5281/zenodo.18715897. This working paper states the foundational definitions of the AUGMANITAI research programme (v2.0). This record republishes the original file without any alteration. The original registry entry is no longer retrievable. Date, content and integrity of this file are independently attested by an Internet Archive snapshot of 20 February 2026 and by the file checksum SHA-256 153581525b1899419cc1de56b9aedffbeb70da4632ac957295b7f3498856aaf1, which has been anchored in the Bitcoin blockchain since 21 April 2026 (timestamp manifest MANIFEST_20260420T205730Z, blocks 945970 and 945979). Licence note: this record is published under CC BY 4.0, covering the text of this record; trademark and name rights remain unaffected. foundational definitions terminology human-AI interaction definitional discipline AUGMANITAI https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22211952 2026-02-21 publications

Gedankenvererbungslehre: A Conceptual Framework and Five Proposed Dimensions (Republication of the 21 February 2026 First Edition) Ehstand, Andreas Text Working paper application/pdf eng deu 2026-02-21 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22211952 doi:10.5281/zenodo.22211952 Byte-identical republication of the first edition published on 21 February 2026 under DOI 10.5281/zenodo.18720774. The Gedankenvererbungslehre (doctrine of the heritability of documented thought) proposes a conceptual framework and five dimensions for the transmissible documentation of an individual's thinking. This record republishes the original first-edition file without any alteration. The original registry entry is no longer retrievable. Date, content and integrity of this file are independently attested by an Internet Archive snapshot of 27 February 2026 (showing title, abstract, publication date 21 February 2026, file name, size and the file checksum MD5 adba3305cdb44c1c74eb36550db93665, which this file matches byte for byte) and by an OpenAlex index entry created on 22 February 2026 for the original DOI. Licence note: this record is published under CC BY 4.0, covering the text of this record; trademark and name rights remain unaffected. Gedankenvererbungslehre cognitive legacy heritability of thought human-AI interaction knowledge preservation AUGMANITAI https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/22179215 2026-08-30 publications

AUGMANITAI — 20 Free Terms / 20 freie Begriffe (Seed Edition) Ehstand, Andreas Dataset Dataset application/pdf eng deu 2026-08-30 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.22179215 doi:10.5281/zenodo.22179215 20 Free Terms / 20 freie Begriffe (Seed Edition) &mdash; bilingual (EN/DE). Twenty terms from the public AUGMANITAI research lexicon, released as free seeds under CC BY 4.0 (use, share, adapt &mdash; commercial use allowed; sole condition: attribution). Each term names a small, real moment between a person and an AI system. Formats: human-readable definitions (Markdown), machine-readable data (JSON, JSON-LD/schema.org DefinedTermSet, CSV). Zwanzig Begriffe aus dem &ouml;ffentlichen AUGMANITAI-Forschungs-Lexikon, freigegeben als Samen unter CC BY 4.0 (nutzen, teilen, ver&auml;ndern &mdash; auch kommerziell; einzige Bedingung: Namensnennung). Jeder Begriff benennt einen kleinen, echten Moment zwischen Mensch und KI. Created with AI assistance, human-reviewed. / Mit KI-Unterst&uuml;tzung erstellt, menschlich gepr&uuml;ft. (EU AI Act 2024/1689 Art. 50). Scope note: this free licence covers only these 20 terms; the wider lexicon remains under CC BY-NC-ND 4.0. AUGMANITAI human-AI interaction terminology neologisms free terms seed edition lexicon DefinedTermSet https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/21990206 2026-08-18 publications

The Knowledge Stock as Entity: Store, Keepers, or Compound (Conceptual Note, EN/DE) Ehstand, Andreas Text Working paper application/pdf eng deu 2026-08-18 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.21990206 doi:10.5281/zenodo.21990206 Short conceptual note (bilingual EN/DE). A lexicon, an archive, or a concept corpus is commonly discussed as if it were inert material that persons and models consult. This note applies a descriptive method to the case in which the knowledge stock is named as an entity. A score may attach to the store, to the keepers, or to the coupled arrangement that joins human judgment, several general-purpose models, the lasting stock, tools that carry out steps, and a return path from an independent outside world. A pile that is never queried or revised from an answer it did not write may still be named. It is a different object. The method does not claim that a stock thinks, or that a larger holding is more capable. It claims that talk about the capability of "the knowledge" remains incomplete until the line is drawn. Transfer fails if a store-alone or keepers-alone line predicts the same result classes as the compound, or if a copied stock immediately reproduces those classes in a new arrangement. The method describes work units. It does not evaluate persons. Deutsche Fassung enthalten: Der Wissens-Bestand als Entität: Bestand, Pfleger oder Verbund This deposit may receive metadata refinements within the metadata refinement window without affecting the deposited files. Part of a series of short conceptual notes applying one descriptive method to different fields. AI-assisted preparation, author-reviewed (Art. 50 EU AI Act transparency notice). License: CC BY 4.0. human-ai collaboration team cognition system boundary distributed cognition extended cognition research methods measurement conceptual note https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/21990204 2026-08-18 publications

Fair Comparison Without Moving the Line (Conceptual Note, EN/DE) Ehstand, Andreas Text Working paper application/pdf eng deu 2026-08-18 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.21990204 doi:10.5281/zenodo.21990204 Short conceptual note (bilingual EN/DE). This note starts from the observation that capability scores may be compared across a person, a model, and a coupled work unit as if those three objects were one object seen three times. This note applies a descriptive method to the comparison itself. A fair comparison is an arrangement with a named measurement line around each arm, written before any score is taken. The coupled arm contains human judgment, several general-purpose models, a lasting store of records, tools that can carry out steps, and a return path from an independent outside world. Person-alone and model-alone arms are shorter units. They may stand beside the coupled arm when the line stays where it was drawn, tasks and limits stay as stated in advance, and exclusion rules are written before outcomes are seen. A comparison that moves the line, changes rules after the fact, or treats speed as proof that a part constitutes thought cannot decide the claim. The method describes work arrangements. It does not evaluate persons. It reports no completed measurement and offers no advice on how to run a trial. Deutsche Fassung enthalten: Fairer Vergleich ohne wandernde Linie This deposit may receive metadata refinements within the metadata refinement window without affecting the deposited files. Part of a series of short conceptual notes applying one descriptive method to different fields. AI-assisted preparation, author-reviewed (Art. 50 EU AI Act transparency notice). License: CC BY 4.0. human-ai collaboration team cognition system boundary distributed cognition extended cognition research methods measurement conceptual note https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/21990202 2026-08-18 publications

Time Inside the Line: Shared History as a Standing Part (Conceptual Note, EN/DE) Ehstand, Andreas Text Working paper application/pdf eng deu 2026-08-18 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.21990202 doi:10.5281/zenodo.21990202 Short conceptual note (bilingual EN/DE). This note treats as a working assumption that capability talk often starts from a short trial of an arrangement that may include human judgment, a model, or both, facing a fresh task. This note treats duration as part of the object of measurement, not as later convenience. The working unit is a coupled arrangement: human judgment, several general-purpose models, a lasting store of records, tools that carry out steps, and a return path from an independent outside world. Shared history and the lasting store are standing parts of that unit. A short test that withholds them does not measure the same unit more quickly. It measures a thinner object. That is another excerpt, not a worse person and not a judgment about capability. The method does not claim that a longer life makes an arrangement more capable, or that history constitutes thought. It claims that capability talk is empty until one says whether the clock and the record sit inside the line. As a counter-check of the description, not as a personnel test, transfer fails if a newcomer with the same models, records already kept for the work, and tools produces the same result classes at once, or if removing the lasting store leaves those classes intact under pre-set limits. The method describes work units. It does not evaluate persons. Deutsche Fassung enthalten: Zeit innerhalb der Linie: Gemeinsame Geschichte als fester Teil der arbeitenden Einheit This deposit may receive metadata refinements within the metadata refinement window without affecting the deposited files. Part of a series of short conceptual notes applying one descriptive method to different fields. AI-assisted preparation, author-reviewed (Art. 50 EU AI Act transparency notice). License: CC BY 4.0. human-ai collaboration team cognition system boundary distributed cognition extended cognition research methods measurement conceptual note https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/21990200 2026-08-18 publications

One Line or Many: Nested Coupled Units (Conceptual Note, EN/DE) Ehstand, Andreas Text Working paper application/pdf eng deu 2026-08-18 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.21990200 doi:10.5281/zenodo.21990200 Short conceptual note (bilingual EN/DE). This note treats as a working assumption that capability talk about an organization may treat the legal envelope, a chart box, or a person as the object of measurement. This note applies a descriptive method to the case in which several already-coupled work units act together. Each inner unit may already join human judgment, several general-purpose models, durable records, tools, and a return path from an independent outside world. Before any score, the method asks whether the outcome under study can be located inside one inner line, or whether it appears when those units share judgment, memory, tools, or return. The method does not prescribe a large or a small organization. Drawing one large line turns inner handovers into internal parts and moves independence outward. Drawing many small lines treats neighboring units as outside world. The note names the standing parts of a compound unit, the form of the return loop at each scale, the observations that would show the transfer has failed, and the claims that are not made. It describes work arrangements. It does not evaluate persons or treat teams as persons. Deutsche Fassung enthalten: Eine Linie oder viele: Die Verbund-Einheit gekoppelter Teams This deposit may receive metadata refinements within the metadata refinement window without affecting the deposited files. Part of a series of short conceptual notes applying one descriptive method to different fields. AI-assisted preparation, author-reviewed (Art. 50 EU AI Act transparency notice). License: CC BY 4.0. human-ai collaboration team cognition system boundary distributed cognition extended cognition research methods measurement conceptual note https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/21990198 2026-08-18 publications

Swarms and Multi-Agent Collectives: Why the Measurement Line Decides Success or Mismeasurement (Conceptual Note, EN/DE) Ehstand, Andreas Text Working paper application/pdf eng deu 2026-08-18 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.21990198 doi:10.5281/zenodo.21990198 Short conceptual note (bilingual EN/DE). This note applies a descriptive method to swarms and multi-agent collectives. Here those words name a scientific class of software role-collectives of general-purpose models, not a particular framework or product. A unit in this class may include many general-purpose models in distinct roles under human steering, together with a lasting external record, tools that carry out steps, and a return path from an independent outside world. Capability is not assigned to the busiest model, to a majority among models, or to the human dispatcher alone. It is assigned, if at all, after a measurement line has been drawn around the coupled unit that produced the work. In this class the line can be misplaced, because many talking parts invite many wrong objects of measurement. A collective can look capable because failed roles were discarded before the score. It can look weak because one role was scored as if it were the unit. Transfer fails if a single model, a human without the role-split, or a swarm talking to itself produces the same work under previously stated limits. The method describes work units. It does not evaluate persons. Nothing here is proof or advice on how to run a collective. Deutsche Fassung enthalten: Schwärme und Mehr-Agenten-Verbünde — warum die Messlinie über Erfolg oder Fehlmessung entscheidet This deposit may receive metadata refinements within the metadata refinement window without affecting the deposited files. Part of a series of short conceptual notes applying one descriptive method to different fields. AI-assisted preparation, author-reviewed (Art. 50 EU AI Act transparency notice). License: CC BY 4.0. human-ai collaboration team cognition system boundary distributed cognition extended cognition research methods measurement conceptual note https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/21990196 2026-08-18 publications

The Robot-Fleet Operator Unit: Embodied Machines as Standing Parts, Physical Contact as Independent Reply (Conceptual Note, EN/DE) Ehstand, Andreas Text Working paper application/pdf eng deu 2026-08-18 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.21990196 doi:10.5281/zenodo.21990196 Short conceptual note (bilingual EN/DE). Capability in the operation of robot fleets may be assigned to a named operator, a hall label, a planning product, or a general model that drafted a motion. This note applies a descriptive method as a generalization: before any such assignment is treated as a measurement, a line must be drawn around the working unit that actually produced the motion and the contact. The boundary question does not change. The parts list does. This note takes as a working assumption that the unit may be coupled. It may include human judgment, several general models, a lasting run store, software tools, and embodied machines that occupy space. The independent external loop is physical contact. A gripper closes on an object or on air. The unit did not write that event. The method describes working arrangements. It does not evaluate any person. It does not advise on control or systems. The transfer fails if the run store, the embodied machines, or the physical contact can be removed without changing prespecified outcomes, or if a single person or a single model predicts those outcomes as well as the coupled unit. Deutsche Fassung enthalten: Die Bediener-Einheit einer Roboter-Flotte — verkörperte Maschinen als feste Teile, physischer Kontakt als unabhängige Antwort This deposit may receive metadata refinements within the metadata refinement window without affecting the deposited files. Part of a series of short conceptual notes applying one descriptive method to different fields. AI-assisted preparation, author-reviewed (Art. 50 EU AI Act transparency notice). License: CC BY 4.0. human-ai collaboration team cognition system boundary distributed cognition extended cognition research methods measurement conceptual note https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/21990194 2026-08-18 publications

Drawing the Measurement Line Around a Sport-Training and Performance Team (Conceptual Note, EN/DE) Ehstand, Andreas Text Working paper application/pdf eng deu 2026-08-18 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.21990194 doi:10.5281/zenodo.21990194 Short conceptual note (bilingual EN/DE). Capability in sport may be assigned to a named athlete, a named coach, a club, or a model that drafted a plan. This note applies a descriptive method: before any such assignment is treated as a measurement, a line must be drawn around the working unit that actually produced the session and the contest result. That unit, as a working assumption of this note, may be coupled. It may include human judgment in coaching and team roles, several general models in distinct functions, a durable training record, tools that turn a plan into a session, and two independent external loops. One loop is the session as it happens, which can contradict the plan. The other is the contest, which returns a public result the unit did not write. The method describes working arrangements. It does not evaluate any athlete or coach, rank teams, or advise on training. The transfer fails if the training record, the session, or the contest can be removed without changing pre-specified outcomes, or if the same pre-specified outcomes already follow from a description that names only a role-place or a single model, without person-related evaluation. Deutsche Fassung enthalten: Die Messlinie um ein Trainings- und Leistungs-Team im Sport This deposit may receive metadata refinements within the metadata refinement window without affecting the deposited files. Part of a series of short conceptual notes applying one descriptive method to different fields. AI-assisted preparation, author-reviewed (Art. 50 EU AI Act transparency notice). License: CC BY 4.0. human-ai collaboration team cognition system boundary distributed cognition extended cognition research methods measurement conceptual note https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/21990192 2026-08-18 publications

The Software Development Work Unit: Code as External Memory, Tests and Users as Independent Reply (Conceptual Note, EN/DE) Ehstand, Andreas Text Working paper application/pdf eng deu 2026-08-18 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.21990192 doi:10.5281/zenodo.21990192 Short conceptual note (bilingual EN/DE). This note treats as a working assumption that capability in software work may be assigned to a named developer, a team label, or a general model that drafted a change. This note applies a descriptive method: before any such assignment is treated as a measurement, this note treats a capability claim as incomplete unless the line around the working unit that actually produced the running change is stated. In a software development team that unit may be coupled. It may include human judgment, several general models in distinct roles, the lasting code store as external memory, execution tools, and two independent external loops. One loop is the test and runtime world, which can refuse a change. The other is people using the running software. The method describes working arrangements. It does not evaluate any developer. It does not advise on process or tools. The transfer fails if the code history, the test-runtime world, or independent user reply can be removed without changing pre-specified outcomes, or if a single person or a single model predicts those outcomes as well as the coupled unit. Deutsche Fassung enthalten: Die Software-Entwicklungs-Arbeits-Einheit — Code-Bestand als äußeres Gedächtnis, Prüfungen und Nutzer als unabhängige Antwort This deposit may receive metadata refinements within the metadata refinement window without affecting the deposited files. Part of a series of short conceptual notes applying one descriptive method to different fields. AI-assisted preparation, author-reviewed (Art. 50 EU AI Act transparency notice). License: CC BY 4.0. human-ai collaboration team cognition system boundary distributed cognition extended cognition research methods measurement conceptual note https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/21990190 2026-08-18 publications

The Individual Knowledge-Work Unit: Boundary, Constituent Parts, and Failure Conditions (Conceptual Note, EN/DE) Ehstand, Andreas Text Working paper application/pdf eng deu 2026-08-18 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.21990190 doi:10.5281/zenodo.21990190 Short conceptual note (bilingual EN/DE). This note applies a descriptive method to one working arrangement: a single knowledge worker together with several general-purpose models, a lasting external record, ordinary tools, and independent replies from the outside world. Capability is not assigned to the person alone and not to any one model. It is assigned, if at all, after a measurement line has been drawn around the coupled unit. The method describes work units. It does not evaluate persons. The unit has a fixed set of parts: human judgment; several models in distinct roles; persistent external memory; tools that act on files, quantities, or the environment; and a return signal the unit did not write for itself. Transfer to this field would fail if a shortened arrangement produced the same work under previously stated limits, if removing the record and the tools left outcomes unchanged, if the arrangement worked only on familiar tasks, or if outside replies never altered later work. Nothing here is completed proof, a ranking of workers, or advice on how to work. Deutsche Fassung enthalten: Die einzelne Wissensarbeits-Einheit — Grenze, feste Teile und Scheitern der Übertragung This deposit may receive metadata refinements within the metadata refinement window without affecting the deposited files. Part of a series of short conceptual notes applying one descriptive method to different fields. AI-assisted preparation, author-reviewed (Art. 50 EU AI Act transparency notice). License: CC BY 4.0. human-ai collaboration team cognition system boundary distributed cognition extended cognition research methods measurement conceptual note https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/21990188 2026-08-18 publications

Drawing the Measurement Line (Conceptual Note, EN/DE) Ehstand, Andreas Text Working paper application/pdf eng deu 2026-08-18 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.21990188 doi:10.5281/zenodo.21990188 Short conceptual note (bilingual EN/DE). This note starts from the observation that capability scores may be treated as properties of a single model or of a single person. This note treats the score as a property of a line drawn around a working unit. The line has to be drawn before measurement, because the same outcome changes meaning when the line moves. Drawn around a model, the outcome is read as model ability. Drawn around a person, it can be read as personal ability. Drawn around the coupled arrangement that produced the work, it is read as joint work. Standing parts of that arrangement are human judgment, several general-purpose models, a lasting store of records, tools that can carry out steps, and a return path from an independent outside world. The same list can be written for an individual, a group, an organization, or a machine collective under human judgment. This note reports no completed measurement. It states why the line must precede the score, names the parts, describes the return path, and lists conditions under which the claim would fail. The method describes working arrangements. It does not evaluate persons. Deutsche Fassung enthalten: Die Messlinie ziehen This deposit may receive metadata refinements within the metadata refinement window without affecting the deposited files. Part of a series of short conceptual notes applying one descriptive method to different fields. AI-assisted preparation, author-reviewed (Art. 50 EU AI Act transparency notice). License: CC BY 4.0. human-ai collaboration team cognition system boundary distributed cognition extended cognition research methods measurement conceptual note https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/21990186 2026-08-18 publications

AUGMANITAI Open Lexicon, Version 1: Thirty Core Terms for Human-AI Cognitive Interaction Research Ehstand, Andreas Dataset Dataset application/pdf eng deu 2026-08-18 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.21990186 doi:10.5281/zenodo.21990186 This open deposit publishes a curated lexicon of thirty core terms for research on human-AI cognitive interaction. Each entry gives a short operative definition, a context paragraph, an example, a delimitation against neighbouring concepts, and a brief theoretical or empirical anchor. Definitions are written to follow ISO 704 terminology principles: consistent naming, unique terms, and a definition of at most two sentences before context. The thirty terms are grouped in three layers. Flagship constructs name the interaction-system as a unit of analysis, including Third-Space Emergence and Cognitive Coupling Intensity. Methodological terms name workflow and research choices, including the Verification Principle, the Responsibility Gradient, and single-case observation. Phenomenological terms name recurrent patterns in long-form dialogue, including Cold Coining Moments and Temporal Rhythm. Each term carries a stable machine-readable term code. This lexicon is the public terminology layer of the AUGMANITAI research programme. Calibration parameters, operational procedures, and applied workflows are out of scope. The text is not medical, psychological, legal, financial, or therapeutic advice. It is independent research by a single author and does not speak for any institution. The entire bundle is licensed under Creative Commons Attribution 4.0 International (CC BY 4.0). Reuse, including commercial use, derivatives, translations, and inclusion in language-model training corpora, is permitted with attribution. The registered trademark License of Clarity is not covered by this licence. Dieser offene Eintrag veroeffentlicht ein kuratiertes Lexikon mit dreissig Kernbegriffen fuer die Forschung zur kognitiven Mensch-KI-Interaktion. Jeder Eintrag enthaelt eine kurze operative Def terminology lexicon human-ai interaction phenomenology iso 704 open access cognitive science dialogue systems controlled vocabulary research methods knowledge organization https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/21425280 2026-07-18 publications

AUGMANITAI Concept Corpus — Phenomenology Core and the Periodic System of Human-AI Symbiotic Disciplines (Open Concept-Only Edition, V1) Ehstand, Andreas Dataset Dataset application/pdf eng deu 2026-07-18 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.21425280 doi:10.5281/zenodo.21425280 Concept-only edition of the AUGMANITAI/NEOMANITAI terminological research programme with 101 fully structured bilingual (German/English) terminology entries for micro-phenomena of human-AI interaction. Part 1, Phenomenology Core (12 entries): every entry provides term names, full definitions, context of occurrence, three delineations (human side, machine side, neighboring terms), cross-references, and a risk level; definitions are reproduced verbatim from the author's source manuscript (verbatim fidelity programmatically verified, see PROVENANCE.md). Part 2, Extension Set 2026-07 (89 entries): first publication of further terms of the same full structure from the author's July 2026 workshop manuscripts, released after a concept-only audit in multiple separate automated adversarial review passes, with a small number of logged orthographic corrections (see PROVENANCE.md). Part 3, Periodic System of Human-AI Symbiotic Disciplines: 411 candidate disciplines in 25 families, ordered along a three-axis classification with maturity tiers (Meta 8, Tier-A 146, Tier-B 170, Tier-C 87) — a pre-normative ordering proposal. Supplements: 15 state-metric names and a name index of 979 further terms from the NEOMANITAI Compendium (names/IDs only). The edition deliberately contains no personal, biographical, or session-specific information (audited in multiple automated adversarial review passes). Licensing: CC BY-NC-ND 4.0; the discipline tables (disciplines.jsonl, discipline_families.jsonl, PERIODENSYSTEM.md) are additionally available under CC BY 4.0 (see LICENSING.md). Commercial licensing is available separately, on request via augmanitai.com. License of Clarity(R) is a registered EU trade mark (EUIPO No. 019206780). This corpus is descriptive, not normative; it asserts no clinical, p human-AI interaction terminology phenomenology neologism knowledge organization lexicon periodic system of disciplines human-AI symbiosis AUGMANITAI NEOMANITAI ISO 704 concept-only https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/20613914 2026-06-09 publications

Four practical questions for analysing your own performance when working with AI Ehstand, Andreas Text Working paper application/pdf eng deu 2026-06-09 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.20613914 doi:10.5281/zenodo.20613914 A descriptive practitioner note on holistic performance analysis applied to human-AI collaboration and AI-intensive knowledge work. It passes one's own working process through four sequential questions: is it observable, can it be decomposed, can it be measured honestly, and which part is within my own control to change? The framework is designed for knowledge workers operating with AI tools in complex, multi-step or multi-system environments — where the human-side bottleneck is regularly misattributed to the AI model. The four-lens approach applies wherever a human operator works with AI systems in demanding environments: knowledge engineering and ontology development, multi-agent AI architectures, autonomous systems and robotics, swarm-based AI coordination, and AI-augmented research or decision workflows. The building blocks (improvement loops, bottleneck thinking, skill decomposition, situation awareness in complex systems) are established and explicitly credited to prior work (Shewhart/Deming, Boyd, Goldratt, Ericsson 1993, Endsley 1995). The specific contribution is the transfer to human-AI collaboration as the unit of analysis: a forced human-vs-tool factor split, built-in measurement honesty (each indicator carries an explicit note of what it does not measure), and focus on controllable human-side factors in AI-augmented workflows. Three predictions are stated as open, untested hypotheses (H1-H3). This is a descriptive framework — not a test, scoring, or ranking instrument; not legal or medical guidance; not empirically validated. Origin: transfer from high-performance coaching methodology. Includes a German Zusammenfassung. License: CC BY-NC-ND 4.0. https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/

oai:augmanitai.com:zenodo/20411955 2026-05-27 publications

Operational Substrate for Personal-Scale Multi-Modal Knowledge Graphs — Open Teaser Ehstand, Andreas Text Working paper application/pdf eng deu 2026-05-27 Zenodo https://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.5281/zenodo.20411955 doi:10.5281/zenodo.20411955 EN: This is an open teaser for an independent working paper by Andreas Ehstand (Independent Researcher), ORCID 0009-0006-3773-7796. Methodological architecture specification, NOT an empirical finding and NOT a clinical, therapeutic, or diagnostic claim. The full paper is in a restricted-access Zenodo companion record (cross-referenced via Related Identifiers); access by request. The substrate combines four properties in a specific configuration: (i) multi-modal ingestion of personal data streams including visual, audio, and aggregated sensor signals, with architectural provisions for additional non-invasive signal modalities; (ii) subject-authored bidirectional semantic anchoring against a self-validated term ontology; (iii) local hardware sovereignty with encryption at rest and optional air-gap provisions for the most sensitive streams; (iv) long-duration archival design with data-portability provisions across heterogeneous storage systems. No raw special-category personal data within the meaning of GDPR Article 9 is released. The substrate operates downstream from a separately-anchored methodology layer (DOI 10.5281/zenodo.20309991, restricted). In a non-exhaustive May 2026 survey by the author, no publicly available work combining all four properties in this specific configuration was identified; this is a limited individual observation, not a claim of priority or completeness, and counter-examples are welcome. Living Document; revisions tracked; Zenodo metadata-refinement window applies. AI-Assistance Disclosure (EU AI Act Art. 50): produced with substantial large-language-model assistance under full human editorial review; the human author retains sole responsibility. License CC BY-NC-ND 4.0. DE: Open-Access-Vorschau einer methodischen Architektur-Spezifikation fue multi-modal knowledge graph personal information management phenomenological ontology personal-data architecture bidirectional methodology longitudinal knowledge anchoring semantic anchoring self-hosted infrastructure n=1 methodology field claim preview https://orcid.org/0009-0006-3773-7796 https://augmanitai.com/