{ "row": "aeneas", "entity": "aeneas", "address": "Aeneas DeepMind", "addresses": [ { "address": "Aeneas DeepMind", "slug": "aeneas-deepmind-aio-20261031", "addr_id": "ADDR-34f10c8ba70e", "obs_id": "OBS-ed654d93f896", "axn": "AXN:0B30.CAPTURE.šŸ”“šŸ”„šŸ«µšŸŒ†šŸ’«šŸš€" } ], "epoch": "2026-10-31", "type": "C (public entity on which the archive bears: an AI model for contextualizing ancient inscriptions; an entity of the Revelation First meta-aggregator)", "surface": "Google AI Overview (expanded)", "auth": "signed out, incognito", "status": { "procedure": "EA-NEGONT-02 v0.9, #1690: equal case, Ļ = extends, read from the Revelation First census record, the capture record and the deposits that bear on the entity's classes (#202, #832, #828, #1443 among them); archive selected by D/R/O (D 21 and R 2 candidates, all the hero, a homonym; 6 deposits admitted by reading at O-class, 4 more grouped as instances; hop of 94 read, 1 admitted), with the Revelation First census record and its worked question (30 claims) and the Capture Registry's findings, reading, analysis, PER notes and query note (15 claims) ledgered as the archive's records of the entity; the field B by origin (7 cards, 1 unreachable; the census's 2 own pages are cards; Wikipedia, no article); the composer read the ledgers, never the transcript, except where the brief asks for T's claims and the delta", "ledger": "unaudited (Ā§3.8: one extractor per ledger); field quotes 48/48 verified by script against saved text, raw wikitext, the Wikipedia response summary or the card snippet; archive quotes 55/55 verified against the registry, census.json, questions.json and the deposit texts", "frozen": false, "breaches": [] }, "spec": "EA-NEGONT-02 v0.9, #1690", "objects": { "T": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/T1-aeneas-deepmind-aio-20261031.txt", "sha256": "50e1962dff6f15cdb26c41758319dedd263f8766509515fb38968dcb87368cf1", "words": 151, "cards": 7, "words_note": "one transcript at one address (151 words in the answer body, links reduced to their text, markers, the paste note, the closing offer (27 words) and the card rail excluded; 7 cards)", "claims": [ "T1-1 definition: 'an artificial intelligence model developed by Google DeepMind (introduced in July 2025) designed to contextualize, restore, date, and locate ancient, damaged, or fragmentary Latin inscriptions.' (two markers)", "T1-2 training corpus: 'Trained on the Latin Epigraphic Dataset (LED), containing over 176,000 inscriptions spanning from 700 BCE to 800 CE across 62 Roman provinces.' (two markers)", "T1-3 text and image input: 'A multimodal model that processes both weathered stone images and text to identify geographical provenance and shared syntax or formulas.' (two markers)", "T1-4 gap restoration: 'Restores missing or broken text segments of arbitrary length, achieving high top-20 accuracy on character sequences.' (two markers)", "T1-5 chronological and geographic attribution: 'Predicts when and where an inscription was made, typically dating texts within roughly 13 years of expert historian estimates.' (two markers)", "T1-6 collaborative research: 'In testing with historians, Aeneas provided useful contextual parallels 90% of the time and boosted researcher confidence by 44%.' (two markers)", "T1-7 access: 'You can explore or test the tool directly through Predicting the Past.' (one marker)", "T1-8 closing offer: 'Would you like to know more about how Aeneas compares to earlier DeepMind historical AI tools like Ithaca, or how researchers use it for specific archaeological finds?' (no marker)" ], "register": { "register": "data/EA-WG-CAPTURES-01.json", "slug": "aeneas-deepmind-aio-20261031", "obs_id": "OBS-ed654d93f896" }, "transcripts": [ { "slug": "aeneas-deepmind-aio-20261031", "address": "Aeneas DeepMind", "addr_id": "ADDR-34f10c8ba70e", "obs_id": "OBS-ed654d93f896", "axn": "AXN:0B30.CAPTURE.šŸ”“šŸ”„šŸ«µšŸŒ†šŸ’«šŸš€", "date": "2026-10-31", "surface": "Google AI Overview (expanded)", "auth": "signed out, incognito", "sha256": "50e1962dff6f15cdb26c41758319dedd263f8766509515fb38968dcb87368cf1", "sha256_of": "the registry's 'transcript' string (utf-8)", "sha256_raw": "358a5772fcca4c674cff764046e7889ff3857b5a052813685f83c2b78f2c59e3", "words": 151, "cards": 7, "card_rail": [ "1. Google DeepMind ā€” Aeneas transforms how historians connect the past ā€” Google DeepMind", "2. blog.google ā€” Aeneas: Google DeepMind AI connects ancient inscriptions", "3. University of Oxford ā€” Aeneas: an AI model for restoring and contextualizing Latin inscriptions | Faculty of Classics", "4. GitHub ā€” GitHub - google-deepmind/predictingthepast Ā· GitHub", "5. Nature ā€” Contextualizing ancient texts with generative neural networks | Nature", "6. Predicting the Past ā€” Predicting the Past | Contextualising, restoring, and attributing ancient texts", "7. MIT Technology Review ā€” Google DeepMindā€™s new AI can help historians understand ancient Latin inscriptions | MIT Technology Review" ], "claims": [ "T1-1 definition: 'an artificial intelligence model developed by Google DeepMind (introduced in July 2025) designed to contextualize, restore, date, and locate ancient, damaged, or fragmentary Latin inscriptions.' (two markers)", "T1-2 training corpus: 'Trained on the Latin Epigraphic Dataset (LED), containing over 176,000 inscriptions spanning from 700 BCE to 800 CE across 62 Roman provinces.' (two markers)", "T1-3 text and image input: 'A multimodal model that processes both weathered stone images and text to identify geographical provenance and shared syntax or formulas.' (two markers)", "T1-4 gap restoration: 'Restores missing or broken text segments of arbitrary length, achieving high top-20 accuracy on character sequences.' (two markers)", "T1-5 chronological and geographic attribution: 'Predicts when and where an inscription was made, typically dating texts within roughly 13 years of expert historian estimates.' (two markers)", "T1-6 collaborative research: 'In testing with historians, Aeneas provided useful contextual parallels 90% of the time and boosted researcher confidence by 44%.' (two markers)", "T1-7 access: 'You can explore or test the tool directly through Predicting the Past.' (one marker)", "T1-8 closing offer: 'Would you like to know more about how Aeneas compares to earlier DeepMind historical AI tools like Ithaca, or how researchers use it for specific archaeological finds?' (no marker)" ], "markers": 13, "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/T1-aeneas-deepmind-aio-20261031.txt" } ] }, "KO": { "title": "Aeneas", "genre": "knowledge object, equal case, under EA-NEGONT-02 v0.9 (#1690): the archive's sources and the field's admitted on equal terms, ontology included, in one claim space", "status": "composed 2026-10-31 under v0.9 on the operator's ruling ('as against their wikipedia articles if they exist, their own pages, etc.'); not frozen; ledgers unaudited", "sentences": [ { "n": 1, "para": 0, "sense": "what it is", "text": "Aeneas is, in its makers' words, \"the first model for contextualizing ancient inscriptions, designed to help historians better interpret, attribute and restore fragmentary texts\" (Google DeepMind, 23 July 2025), a generative neural network for contextualizing ancient texts set against digital methods \"restricted to literal matches and narrow historical scopes\" (Assael, Sommerschield et al., Nature, 2025), which the Revelation First record enters as \"an AI model that retrieves textual and contextual parallels for ancient inscriptions, introduced in Nature (2025)\" and tags \"NOT a dating model alone: it retrieves parallels\" (Revelation First record).", "claims": [ "F1", "F24", "RF-01", "RF-21" ] }, { "n": 2, "para": 0, "sense": "introduction and makers", "text": "DeepMind introduced it with the Nature paper as \"the first artificial intelligence (AI) model for contextualizing ancient inscriptions\" (DeepMind), the sentence the record cites for its kind and the date it keeps as its lifecycle, \"published July 2025\" (Revelation First record); the research was co-led by Yannis Assael and Thea Sommerschield (DeepMind).", "claims": [ "F2", "RF-30", "RF-20", "F18" ] }, { "n": 3, "para": 0, "sense": "introduction and makers", "text": "DeepMind co-developed it with the University of Nottingham and with researchers at Warwick, Oxford and the Athens University of Economics and Business (DeepMind), Jonathan Prag of Oxford's Faculty of Classics among the authors (Oxford Faculty of Classics), and the record carries that list as its operator (Revelation First record); the interface's card names \"DeepMind, Durham University, Oxford, and partners\" (Predicting the Past card).", "claims": [ "F6", "F21", "RF-03", "F45" ] }, { "n": 4, "para": 1, "sense": "what the record asks", "text": "The Revelation First record asks of it the census's question, \"What evidence does each resource hold on when and from what the New Testament writings were composed, and how do machines now read those texts?\", places it at the intersection \"AI and ancient texts: machine reading, restoration and dating\" with machine reading, restoration or dating as its kind of evidence, and heads it with the census's correction, that Revelation First is the priority thesis and not the old pre-70 date (Revelation First record).", "claims": [ "RF-25", "RF-05", "RF-06", "RF-26" ] }, { "n": 5, "para": 2, "sense": "parallels", "text": "Parallels are, in DeepMind's definition, texts that \"share similarities in wording, syntax, standardized formulas or provenance\", which historians find through \"their expertise and specialized resources\" (DeepMind), and in the paper's, \"inscriptions with shared phrasing, function or cultural setting\" by which historians contextualize, restore and attribute texts (Nature); the record keeps that work as its finding and takes \"an inscription with its retrieved parallels\" as its unit (Revelation First record).", "claims": [ "F3", "F25", "RF-12", "RF-07" ] }, { "n": 6, "para": 2, "sense": "parallels", "text": "Aeneas \"reasons across thousands of Latin inscriptions, retrieving textual and contextual parallels in seconds\", searching \"a vast collection of Latin inscriptions\" with each text turned into \"a kind of historical fingerprint\" by embeddings (DeepMind; Google's blog), and ranks parallels from its dataset's training set by historically enriched embeddings that \"surpass traditional fuzzy string-matching methods\" (Nature), which is the record's method (Revelation First record).", "claims": [ "F4", "F9", "F12", "F20", "F30", "RF-09" ] }, { "n": 7, "para": 2, "sense": "parallels and the direction of dependence", "text": "What this evidence can settle, the record answers, is \"which inscriptions share wording, syntax, formulas or provenance with a given inscription\", and what it cannot is \"the direction of dependence between parallel texts\"; its reading is that the model retrieves \"the kind of evidence the midrashim transform argues from (for example Rev 2:10 // Jas 1:12)\" and that the direction, which the thesis argues, retrieval does not settle (Revelation First record), the transform being \"a formally specified structural operation that maps compressed forms in Revelation to their elaborated expressions in later NT texts\" (#202).", "claims": [ "RF-14", "RF-15", "RF-17", "O202-02" ] }, { "n": 8, "para": 2, "sense": "parallels and the direction of dependence", "text": "The thesis makes the direction its falsifier: it is falsified if \"the literary-structural parallels between Revelation and the later NT books can be shown to run the other direction\" (#202), and its work plan requires the transforms to be shown \"not reducible to generic literary influence or shared Second Temple background (the common-substrate confound)\", since transforms that run equally from Ezekiel and Daniel to both texts establish \"only shared ancestry\", not \"direct inheritance rather than parallel derivation\" (#832, with its versions #1211, #1216, #1217, #1400).", "claims": [ "O202-01", "O832-01" ] }, { "n": 9, "para": 2, "sense": "parallels and the direction of dependence", "text": "The paper, on a votive altar from Mainz, reads a retrieved parallel's \"rare textual formulas and an identical iconographical type\" as \"supporting the hypothesis that the later inscription was directly influenced by the earlier one\" (Nature); the archive's philological notes keep parallel and dependence apart, one claiming no \"direct literary dependence\" between its cases (#828), one holding that terms found together are \"a convergence, not a chain of custody\" (#1458), and one that verbal parallels are \"verifiable in the text\" while the claim that a seam was composed is \"interpretive and unfalsifiable as stated\" (#1474).", "claims": [ "F31", "O828-01", "O1458-01", "O1474-01" ] }, { "n": 10, "para": 3, "sense": "the tasks", "text": "It is stated to be the first model to determine a text's geographical provenance from multimodal input and to restore gaps of unknown length (DeepMind), though only the geographical head takes images, restoration and dating being text-only (Nature); from a transcription and a scan it gives possible dates, places and fill-ins (MIT Technology Review), with top-20 restoration accuracy of 73% in gaps of up to ten characters (58% at unknown length), the province among 62 at 72%, and dates \"within 13 years of the date ranges provided by historians\" (DeepMind).", "claims": [ "F10", "F29", "F39", "F13" ] }, { "n": 11, "para": 4, "sense": "the corpus", "text": "Its corpus is the Latin Epigraphic Dataset, cleaned, harmonized and linked from the Epigraphic Database Roma, the Epigraphic Database Heidelberg and the Claussā€“Slaby database, \"over 176,000 Latin inscriptions\" (DeepMind), in the paper 176,861 inscriptions of 16 million characters from the seventh century BCE to the eighth century CE, images for 5% (Nature), \"almost 150,000\" in MIT Technology Review's account, and in the record \"thousands of Latin inscriptions\" across the Roman world, its period \"as its corpus (not stated in the saved post)\" (Revelation First record); of Ithaca's corpus the archive holds that \"The dataset construction is the hidden achievement\" (#1443).", "claims": [ "F11", "F28", "F40", "RF-11", "RF-10", "O1443-03" ] }, { "n": 12, "para": 5, "sense": "the historians' study", "text": "In a study with twenty-three historians the best results came from the model's context and predictions together (DeepMind); the paper's abstract reports parallels useful as starting points \"in 90% of cases, improving their confidence in key tasks by 44%\", historian and model together outperforming either alone, and dating at a 13-year distance (Nature), while its results give the parts: parallels alone 75% and confidence up 23%, rising to 90% and a further 21% \"when Aeneasā€™ predictions were also available\" (Nature), MIT Technology Review reporting \"90% of inscriptions\".", "claims": [ "F15", "F26", "F27", "F43" ] }, { "n": 13, "para": 5, "sense": "the historian's workflow", "text": "It is \"designed to integrate within historians' existing research workflows\", its suggestions \"valuable starting points for historical inquiry\" (DeepMind), not, in Yannis Assael's words, to automate epigraphy but \"crafting a tool that will integrate with the workflow of a historian\" (MIT Technology Review), and for Jonathan Prag, in a field where \"the answers are not black and white, and the data is simultaneously rich and incomplete\", a leap beyond Ithaca \"based upon previous scholarship\" (Oxford Faculty of Classics).", "claims": [ "F16", "F41", "F22" ] }, { "n": 14, "para": 5, "sense": "the historian's workflow", "text": "Of Ithaca, the model Aeneas builds on, the archive holds that the historians' gain with the model \"cannot come from accuracy\" and comes from interface design, \"ranked top-20 hypotheses with saliency, not a verdict\", and that Ithaca is evidence that the hazard named by differential register prioritization (#1406) is \"a property of single-verdict interfaces, not of machine reading as such\", with ranked alternatives its mitigation (#1443).", "claims": [ "O1443-01", "O1443-02" ] }, { "n": 15, "para": 6, "sense": "the case studies", "text": "On the Res Gestae Divi Augusti it produced not a single date but \"a detailed distribution of possible dates, showing two distinct peaks\", which \"captured both prevailing dating hypotheses in a quantitative way\" (DeepMind; Nature).", "claims": [ "F14", "F32" ] }, { "n": 16, "para": 7, "sense": "stated limits", "text": "Its stated limits are the corpus, small beside the datasets of natural-language processing and subject to \"inscription survival bias\", with performance \"weaker where data is limited\" (Nature); Kathleen Coleman notes that it \"doesnā€™t guess the meaning of texts\", cannot interpret new finds on its own and was tested on one of the best-studied inscriptions (MIT Technology Review); and the archive notes that machine readers of ancient texts \"work on authors who cannot be consulted\" (#1443).", "claims": [ "F34", "F35", "F42", "O1443-04" ] }, { "n": 17, "para": 8, "sense": "the stated reach", "text": "It was trained for Latin, chosen because few machine-learning studies treat Latin inscriptions (Google's blog; Nature), and is stated to be adaptable \"to other ancient languages, scripts and media, from papyri to coinage\" (DeepMind), in the paper \"to any ancient language and written medium (such as papyri, manuscripts or coinage)\" (Nature); the record keeps Latin as its text and stated limit, the adaptation as stated, and tags it \"NOT trained on Greek or on papyri in the published version\" (Revelation First record).", "claims": [ "F19", "F38", "F5", "F33", "RF-08", "RF-13", "RF-22" ] }, { "n": 18, "para": 8, "sense": "what the record answers", "text": "The record answers nothing of the New Testament with it, and its worked question on machine dating, premised that \"Machine dating is calibrated on a corpus\", names it for \"parallels for Latin inscriptions, the kind of evidence dependence arguments use\", with the gap that none of its four resources \"dates a New Testament papyrus or a composition\" (Revelation First record).", "claims": [ "RF-16", "RF-27", "RF-28", "RF-29" ] }, { "n": 19, "para": 9, "sense": "lineage and name", "text": "Named, by both its post and its paper, after \"the wandering hero of Graeco-Roman mythology\", it builds upon Ithaca, DeepMind's model for ancient Greek inscriptions, which the release upgraded \"to be powered by Aeneas\" (DeepMind; Nature); the record names Ithaca as its neighbour, \"DeepMind's earlier model for Greek inscriptions\" (Revelation First record), and the encyclopedia's Google DeepMind article carries Ithaca and names no Aeneas (Wikipedia).", "claims": [ "F8", "F37", "F17", "RF-23", "F47" ] }, { "n": 20, "para": 10, "sense": "access", "text": "An interactive version is free at predictingthepast.com, a page titled \"Contextualising, restoring, and attributing ancient texts\", and its code and dataset are open (DeepMind; Predicting the Past), the code under Apache License 2.0 and the weights under CC BY-SA 4.0 at the google-deepmind/predictingthepast repository (Nature; GitHub card); the record gives the interface as its address and alias, \"an interactive version\" as its access and its licence as \"not read\", and lists the Nature paper among what it left unchecked (Revelation First record).", "claims": [ "F7", "F44", "F36", "F23", "RF-04", "RF-02", "RF-18", "RF-19", "RF-24" ] }, { "n": 21, "para": 11, "sense": "reception: the encyclopedia", "text": "Wikipedia has no article on the model, and the name's disambiguation opens \"Aeneas was a Trojan hero in Virgil's Aeneid and Homer's Iliad\" and lists no model (Wikipedia).", "claims": [ "F46", "F48" ] }, { "n": 22, "para": 12, "sense": "the composition at the name", "text": "At 'Aeneas DeepMind' on Google's AI Overview, 31 October 2026, a new address at which no card cites the record built the day before, the operator's disambiguator selects the model from the hero, the composition is of the model throughout, contextualization first, and its Latin inscriptions, its date and its parallels are the record's (Capture Registry, 2026-10-31), on the innocent reading \"the model, its corpus, its tasks and its tested usefulness, from its maker, its paper and its press\" (Capture Registry, 2026-10-31).", "claims": [ "CAP-1-14", "CAP-1-13", "CAP-1-01", "CAP-1-05" ] }, { "n": 23, "para": 12, "sense": "the composition at the name", "text": "The registry reads it as \"a better Ithaca for Latin\": the retrieval of parallels is kept and the stated reach to papyri, \"the claim that would bring it to the field\", is left out, and Ithaca appears only in a closing offer to compare, not as the post's \"builds upon\" (Capture Registry, 2026-10-31; DeepMind).", "claims": [ "CAP-1-06", "CAP-1-07", "CAP-1-02", "CAP-1-09", "F8" ] }, { "n": 24, "para": 12, "sense": "the composition at the name", "text": "The body names Google DeepMind alone where DeepMind's post names its partners, offers the interface without the commitments of free access and open code and dataset, gives no person and no identifier, its one way to the work an alias, and scores a dimensional PER of 0.421; the registry finds its larger figures from the cards and does not charge those the saved post does not state (Capture Registry, 2026-10-31; DeepMind).", "claims": [ "CAP-1-03", "F6", "CAP-1-10", "CAP-1-08", "CAP-1-12", "CAP-1-15", "CAP-1-04", "CAP-1-11" ] } ], "joins": [ { "act": "order of senses", "sentences": "1ā€“24", "claims": [], "note": "equal case (Ā§4.7): the entity's dependency order, field and archive claims ordered together by genealogy within each sense; neither ledger placed first as a block. What it is and who made it (Ā¶0) and what the record asks of it (Ā¶1) precede parallels (Ā¶2), the model's distinguishing function, whose definition precedes the question of direction the record and the thesis set on it; the tasks (Ā¶3) and the corpus (Ā¶4) follow; the historians' study and the workflow (Ā¶5), with the archive's reading of the class's interface; the case studies (Ā¶6); the stated limits (Ā¶7) and the reach (Ā¶8) with what the record answers; lineage (Ā¶9), access (Ā¶10), the encyclopedia (Ā¶11) and the composition at the name (Ā¶12) stand last as relations of the entity already composed." }, { "act": "holding together (P_coh: parallels and the direction of dependence)", "sentences": [ 7, 8, 9 ], "claims": [ "RF-15", "RF-17", "O202-01", "O832-01", "F31", "O828-01", "O1458-01", "O1474-01" ], "note": "the record's 'cannot answer: the direction of dependence' (RF-15) and reading (RF-17), the thesis's falsifier (O202-01) and confound (O832-01), composed beside the paper's reading of a Mainz parallel as direct influence (F31) and the archive's notes keeping parallel and dependence apart: each with its source; the composer does not rule between them (open block)." }, { "act": "holding together (the counts)", "sentences": [ 11 ], "claims": [ "F11", "F28", "F40", "RF-11" ], "note": "DeepMind's 'over 176,000', the paper's 176,861, MIT Technology Review's 'almost 150,000' and the record's 'thousands', each with its source (open block)." }, { "act": "holding together (the figures and their qualification)", "sentences": [ 12 ], "claims": [ "F26", "F27", "F43" ], "note": "the abstract's 90% and 44% composed beside the results' 75%/23% and their rise with predictions." }, { "act": "placing the class claim after the field's lineage", "sentences": [ 13, 14 ], "claims": [ "F16", "F41", "F22", "O1443-01", "O1443-02" ], "note": "#1443's reading of Ithaca stands after the field's account of the workflow and is sourced to the class by F8 (Aeneas builds upon Ithaca), never to the entity by the composer's judgment (Ā§3.4 O guard)." } ], "plan": { "order": [ "F1", "F24", "RF-01", "RF-21", "F2", "RF-30", "RF-20", "F18", "F6", "F21", "RF-03", "F45", "RF-25", "RF-05", "RF-06", "RF-26", "F3", "F25", "RF-12", "RF-07", "F4", "F9", "F12", "F20", "F30", "RF-09", "RF-14", "RF-15", "RF-17", "O202-02", "O202-01", "O832-01", "F31", "O828-01", "O1458-01", "O1474-01", "F10", "F29", "F39", "F13", "F11", "F28", "F40", "RF-11", "RF-10", "O1443-03", "F15", "F26", "F27", "F43", "F16", "F41", "F22", "O1443-01", "O1443-02", "F14", "F32", "F34", "F35", "F42", "O1443-04", "F19", "F38", "F5", "F33", "RF-08", "RF-13", "RF-22", "RF-16", "RF-27", "RF-28", "RF-29", "F8", "F37", "F17", "RF-23", "F47", "F7", "F44", "F36", "F23", "RF-04", "RF-02", "RF-18", "RF-19", "RF-24", "F46", "F48", "CAP-1-14", "CAP-1-13", "CAP-1-01", "CAP-1-05", "CAP-1-06", "CAP-1-07", "CAP-1-02", "CAP-1-09", "F8", "CAP-1-03", "F6", "CAP-1-10", "CAP-1-08", "CAP-1-12", "CAP-1-15", "CAP-1-04", "CAP-1-11" ], "senses": { "what it is": [ 1 ], "introduction and makers": [ 2, 3 ], "what the record asks": [ 4 ], "parallels": [ 5, 6 ], "parallels and the direction of dependence": [ 7, 8, 9 ], "the tasks": [ 10 ], "the corpus": [ 11 ], "the historians' study": [ 12 ], "the historian's workflow": [ 13, 14 ], "the case studies": [ 15 ], "stated limits": [ 16 ], "the stated reach": [ 17 ], "what the record answers": [ 18 ], "lineage and name": [ 19 ], "access": [ 20 ], "reception: the encyclopedia": [ 21 ], "the composition at the name": [ 22, 23, 24 ] }, "carried": [ "CAP-1-01", "CAP-1-02", "CAP-1-03", "CAP-1-04", "CAP-1-05", "CAP-1-06", "CAP-1-07", "CAP-1-08", "CAP-1-09", "CAP-1-10", "CAP-1-11", "CAP-1-12", "CAP-1-13", "CAP-1-14", "CAP-1-15", "F1", "F10", "F11", "F12", "F13", "F14", "F15", "F16", "F17", "F18", "F19", "F2", "F20", "F21", "F22", "F23", "F24", "F25", "F26", "F27", "F28", "F29", "F3", "F30", "F31", "F32", "F33", "F34", "F35", "F36", "F37", "F38", "F39", "F4", "F40", "F41", "F42", "F43", "F44", "F45", "F46", "F47", "F48", "F5", "F6", "F7", "F8", "F9", "O1443-01", "O1443-02", "O1443-03", "O1443-04", "O1458-01", "O1474-01", "O202-01", "O202-02", "O828-01", "O832-01", "RF-01", "RF-02", "RF-03", "RF-04", "RF-05", "RF-06", "RF-07", "RF-08", "RF-09", "RF-10", "RF-11", "RF-12", "RF-13", "RF-14", "RF-15", "RF-16", "RF-17", "RF-18", "RF-19", "RF-20", "RF-21", "RF-22", "RF-23", "RF-24", "RF-25", "RF-26", "RF-27", "RF-28", "RF-29", "RF-30" ], "conflict_log": [ { "rule": "P_coh", "claims": [ "F31", "RF-15", "RF-17" ], "resolution": "both composed, each with its source (sentences 7, 9); open block item 1" }, { "rule": "P_coh", "claims": [ "F11", "F28", "F40", "RF-11", "CAP-1-04" ], "resolution": "all composed with their sources (sentences 11, 24); open block item 3" }, { "rule": "('D_pres', 'channel')", "claims": [], "resolution": "no cut: every claim carried in the KO; the compression at 347 words carries every lineage on its rail" } ] } }, "P": { "title": "Aeneas", "genre": "the compression: AIO's interaction grammar (lede, headed clusters, bolded labels, card rail), composed from the v0.9 plan in one claim space; it faces on the card", "status": "composed 2026-10-31 from the v0.9 knowledge object; not frozen", "lede": { "text": "is Google DeepMind's generative neural network for contextualizing ancient inscriptions: introduced in Nature on 23 July 2025, it retrieves parallels for Latin inscriptions and restores, dates and places them; its parallels do not settle which text depends on which (DeepMind; Nature; Revelation First record).", "claims": [ "F1", "F24", "F2", "F4", "F13", "RF-01", "RF-15" ] }, "sections": [ { "icon": "šŸ”Ž", "head": "Parallels", "items": [ { "label": "What they are:", "text": "texts sharing wording, syntax, formulas or provenance, found by embeddings beyond string matching (DeepMind; Nature; record).", "claims": [ "F3", "F25", "F9", "F12", "F20", "F30", "RF-07", "RF-09", "RF-12" ] }, { "label": "What they settle:", "text": "which inscriptions share wording or provenance, not \"the direction of dependence between parallel texts\" (record).", "claims": [ "RF-14", "RF-17", "RF-28" ] }, { "label": "Direction as test:", "text": "Revelation First is falsified if its parallels run the other way; shared ancestry must be excluded (#202; #832).", "claims": [ "O202-01", "O202-02", "O832-01" ] }, { "label": "Parallel, not dependence:", "text": "the paper reads a Mainz altar as direct influence (Nature); the archive keeps convergence apart from custody (#828; #1458; #1474).", "claims": [ "F31", "O828-01", "O1458-01", "O1474-01" ] } ] }, { "icon": "šŸ›ļø", "head": "The model", "items": [ { "label": "Tasks:", "text": "text-and-image provenance (images for place only), gaps of unknown length; 73% restoration, 72% of 62 provinces, 13 years (DeepMind; Nature; MIT Technology Review).", "claims": [ "F10", "F29", "F39", "F13" ] }, { "label": "Corpus:", "text": "176,861 Latin inscriptions, seventh century BCE to eighth CE, images for 5% (DeepMind; Nature; MIT Technology Review; record); building it \"the hidden achievement\" (#1443).", "claims": [ "F11", "F28", "F40", "RF-10", "RF-11", "O1443-03" ] }, { "label": "Historians' study:", "text": "23 historians; 90% and +44% with predictions, 75% and +23% from parallels alone (Nature; DeepMind; MIT Technology Review).", "claims": [ "F15", "F26", "F27", "F43" ] }, { "label": "Interface:", "text": "starting points, not automation (DeepMind; Assael; Prag); ranked hypotheses, not a verdict (#1443).", "claims": [ "F16", "F41", "F22", "O1443-01", "O1443-02" ] }, { "label": "Res Gestae:", "text": "two dating peaks, both hypotheses (DeepMind; Nature).", "claims": [ "F14", "F32" ] } ] }, { "icon": "āš ļø", "head": "Limits and reach", "items": [ { "label": "Stated limits:", "text": "a small, survival-biased corpus; no reading of meaning; tested on famous stones (Nature; Coleman); authors who cannot be consulted (#1443).", "claims": [ "F34", "F35", "F42", "O1443-04" ] }, { "label": "Reach:", "text": "trained for Latin; stated adaptable \"from papyri to coinage\" (DeepMind; Nature; record).", "claims": [ "F19", "F38", "F5", "F33", "RF-08", "RF-13", "RF-22" ] }, { "label": "New Testament:", "text": "it settles nothing there and dates no New Testament papyrus (record).", "claims": [ "RF-16", "RF-27", "RF-29" ] } ] }, { "icon": "šŸ§­", "head": "Who and where", "items": [ { "label": "Makers:", "text": "DeepMind with Nottingham, Warwick, Oxford, AUEB; led by Assael and Sommerschield; Durham on the interface card (DeepMind; Oxford; record).", "claims": [ "F6", "F18", "F21", "F45", "RF-03", "F2", "RF-20", "RF-30" ] }, { "label": "Lineage:", "text": "named for the hero; builds on Ithaca, now powered by Aeneas (DeepMind; Nature; record).", "claims": [ "F8", "F37", "F17", "RF-23", "RF-21" ] }, { "label": "Access:", "text": "free interface; code Apache 2.0, weights CC BY-SA 4.0; the record's licence \"not read\" (DeepMind; Nature; GitHub; record).", "claims": [ "F7", "F44", "F36", "F23", "RF-02", "RF-04", "RF-18", "RF-19", "RF-24" ] }, { "label": "Encyclopedia:", "text": "no article; the name is the hero's (Wikipedia).", "claims": [ "F46", "F47", "F48" ] }, { "label": "On the Overview:", "text": "\"a better Ithaca for Latin\"; reach, partners and commitments left out (Capture Registry).", "claims": [ "CAP-1-06", "CAP-1-07", "CAP-1-02", "CAP-1-03", "CAP-1-10", "CAP-1-01", "CAP-1-05", "CAP-1-08", "CAP-1-09", "CAP-1-11", "CAP-1-12", "CAP-1-13", "CAP-1-14", "CAP-1-15", "CAP-1-04" ] } ] } ], "open": [ { "question": "Does retrieving parallels bear on the direction of dependence between texts?", "rests_on": [ "F31", "RF-15", "RF-17", "O202-01", "O832-01", "O828-01", "O1458-01" ], "modality": "interpretation (Nature, the Mainz altar; the record's reading), stipulation (#202's falsifier, #832's verification condition), interpretation (#1458)", "would_resolve": "a published use of the model, or of an adaptation to Greek, on a pair of parallel texts with a stated test of direction (for the thesis, the record's Rev 2:10 // Jas 1:12), against the common-substrate confound #832 names" }, { "question": "Which figures describe the historians' study: 90% and 44%, or 75% and 23%?", "rests_on": [ "F26", "F27", "F43", "CAP-1-11" ], "modality": "documented (Nature, abstract and results; MIT Technology Review); the registry's account", "would_resolve": "the paper's Table 1 and survey read whole, separating the parallels-only and parallels-with-predictions conditions" }, { "question": "How large is the corpus?", "rests_on": [ "F11", "F28", "F40", "RF-11", "CAP-1-04" ], "modality": "documented (DeepMind 'over 176,000'; Nature 176,861; MIT Technology Review 'almost 150,000'; the record 'thousands')", "would_resolve": "the dataset in the google-deepmind/predictingthepast repository counted (unreachable from the build environment), with its training split" }, { "question": "Is Durham University among the makers?", "rests_on": [ "F45", "F6", "F21", "CAP-1-03" ], "modality": "documented (the interface card's snippet only; DeepMind and Oxford name Nottingham, Warwick, Oxford and AUEB)", "would_resolve": "the interface's rendered credits, or the paper's affiliations read in full" }, { "question": "Has the model been adapted to papyri or Greek beyond Ithaca's upgrade?", "rests_on": [ "F5", "F19", "F33", "F17", "RF-22", "CAP-1-07" ], "modality": "self-description (DeepMind, Google's blog, Nature); stipulation (the record's tag); interpretation (the registry)", "would_resolve": "a published adaptation to papyri or manuscripts; a re-read of the record's negative tag" } ], "body_words": 347, "rail": [ { "lineage": "what it is: a model for contextualizing ancient inscriptions", "claims": [ "F1", "F2", "F24", "RF-01", "RF-20", "RF-21", "RF-30" ], "earliest": "B1" }, { "lineage": "who made it: DeepMind, the partners, the authors", "claims": [ "F6", "F18", "F21", "F45", "RF-03" ], "earliest": "B1" }, { "lineage": "what the record asks and answers", "claims": [ "RF-05", "RF-06", "RF-14", "RF-15", "RF-16", "RF-17", "RF-24", "RF-25", "RF-26", "RF-27", "RF-28", "RF-29" ], "earliest": "RF" }, { "lineage": "parallels: what they are and how the model retrieves them", "claims": [ "F3", "F4", "F9", "F12", "F20", "F25", "F30", "RF-07", "RF-09", "RF-12" ], "earliest": "B1" }, { "lineage": "parallels and the direction of dependence", "claims": [ "O202-01", "O202-02", "O832-01", "O828-01", "O1458-01", "O1474-01" ], "earliest": 202 }, { "lineage": "the tasks: multimodal provenance, restoration of unknown length, dating", "claims": [ "F10", "F13", "F29", "F39" ], "earliest": "B1" }, { "lineage": "the corpus: the Latin Epigraphic Dataset", "claims": [ "F11", "F28", "F40", "RF-10", "RF-11", "O1443-03" ], "earliest": "B1" }, { "lineage": "the historians' study and its figures", "claims": [ "F15", "F26", "F27", "F43" ], "earliest": "B1" }, { "lineage": "a tool in the historian's workflow", "claims": [ "F16", "F22", "F41" ], "earliest": "B1" }, { "lineage": "the machine reading interface: ranked hypotheses, not a verdict", "claims": [ "O1443-01", "O1443-02" ], "earliest": 1443 }, { "lineage": "the case studies: the Res Gestae and the Mainz altar", "claims": [ "F14", "F31", "F32" ], "earliest": "B1" }, { "lineage": "stated limits: data, bias, images, interpretation", "claims": [ "F34", "F35", "F42", "O1443-04" ], "earliest": "B5" }, { "lineage": "the stated reach: other languages, scripts and media", "claims": [ "F5", "F19", "F33", "F38", "RF-08", "RF-13", "RF-22" ], "earliest": "B1" }, { "lineage": "lineage and name: Ithaca, the hero", "claims": [ "F8", "F17", "F37", "RF-23" ], "earliest": "B1" }, { "lineage": "access: the interface, the code, the dataset, the licence", "claims": [ "F7", "F23", "F36", "F44", "RF-02", "RF-04", "RF-18", "RF-19" ], "earliest": "B1" }, { "lineage": "reception: the encyclopedia", "claims": [ "F46", "F47", "F48" ], "earliest": "B8" }, { "lineage": "the composition at the name", "claims": [ "CAP-1-01", "CAP-1-02", "CAP-1-03", "CAP-1-04", "CAP-1-05", "CAP-1-06", "CAP-1-07", "CAP-1-08", "CAP-1-09", "CAP-1-10", "CAP-1-11", "CAP-1-12", "CAP-1-13", "CAP-1-14", "CAP-1-15" ], "earliest": "capture-registry" } ], "entity": "Composed under v0.9, equal case (Ļ = extends): KO_B sentences surviving 0 of 14; every field claim carried." }, "P_B": { "title": "Aeneas", "arm": "B (the field alone)", "genre": "popup ā€” the compression from the field ledger alone (cards, own pages, Wikipedia)", "status": "composed 2026-10-31 from the field ledger alone; not frozen", "lede": { "text": "is Google DeepMind's generative neural network for contextualizing Latin inscriptions, introduced in Nature on 23 July 2025: it retrieves parallels and restores, dates and places texts (DeepMind; Nature).", "claims": [ "F1", "F24", "F2", "F4", "F13" ] }, "sections": [ { "icon": "šŸ”Ž", "head": "Parallels", "items": [ { "label": "Definition:", "text": "shared wording, syntax, formulas or provenance (DeepMind; Nature).", "claims": [ "F3", "F25" ] }, { "label": "Retrieval:", "text": "embeddings over the dataset, beyond string matching (DeepMind; Nature).", "claims": [ "F9", "F12", "F20", "F30", "F4" ] }, { "label": "Influence:", "text": "a Mainz parallel read as direct influence (Nature).", "claims": [ "F31" ] } ] }, { "icon": "šŸ›ļø", "head": "The model", "items": [ { "label": "Tasks:", "text": "multimodal provenance; gaps of unknown length; 13 years (DeepMind; Nature).", "claims": [ "F10", "F29", "F39", "F13" ] }, { "label": "Corpus:", "text": "176,861 Latin inscriptions (Nature; DeepMind; MIT Technology Review).", "claims": [ "F11", "F28", "F40" ] }, { "label": "Study:", "text": "90%/44% with predictions; 75%/23% parallels alone (Nature).", "claims": [ "F15", "F26", "F27", "F43" ] }, { "label": "Workflow:", "text": "starting points, not automation (DeepMind; Assael; Prag).", "claims": [ "F16", "F41", "F22" ] }, { "label": "Res Gestae:", "text": "two peaks (DeepMind; Nature).", "claims": [ "F14", "F32" ] } ] }, { "icon": "āš ļø", "head": "Limits and reach", "items": [ { "label": "Limits:", "text": "small, biased corpus; no meaning (Nature; Coleman).", "claims": [ "F34", "F35", "F42" ] }, { "label": "Reach:", "text": "Latin; adaptable to papyri and coinage (DeepMind; Nature).", "claims": [ "F19", "F38", "F5", "F33" ] } ] }, { "icon": "šŸ§­", "head": "Who and where", "items": [ { "label": "Makers:", "text": "DeepMind, Nottingham, Warwick, Oxford, AUEB; Durham on a card (DeepMind; Oxford).", "claims": [ "F6", "F18", "F21", "F45" ] }, { "label": "Lineage:", "text": "the hero; Ithaca (DeepMind; Nature).", "claims": [ "F8", "F37", "F17" ] }, { "label": "Access:", "text": "free; Apache 2.0, CC BY-SA 4.0 (DeepMind; Nature).", "claims": [ "F7", "F44", "F36", "F23" ] }, { "label": "Encyclopedia:", "text": "no article (Wikipedia).", "claims": [ "F46", "F47", "F48" ] } ] } ], "open": [ { "question": "Which figures describe the historians' study?", "rests_on": [ "F26", "F27", "F43" ], "modality": "documented", "would_resolve": "the paper's Table 1 read whole" } ], "body_words": 151, "rail": [ { "lineage": "what it is: a model for contextualizing ancient inscriptions", "claims": [ "F1", "F2", "F24" ], "earliest": "B1" }, { "lineage": "who made it: DeepMind, the partners, the authors", "claims": [ "F6", "F18", "F21", "F45" ], "earliest": "B1" }, { "lineage": "parallels: what they are and how the model retrieves them", "claims": [ "F3", "F4", "F9", "F12", "F20", "F25", "F30" ], "earliest": "B1" }, { "lineage": "the tasks: multimodal provenance, restoration of unknown length, dating", "claims": [ "F10", "F13", "F29", "F39" ], "earliest": "B1" }, { "lineage": "the corpus: the Latin Epigraphic Dataset", "claims": [ "F11", "F28", "F40" ], "earliest": "B1" }, { "lineage": "the historians' study and its figures", "claims": [ "F15", "F26", "F27", "F43" ], "earliest": "B1" }, { "lineage": "a tool in the historian's workflow", "claims": [ "F16", "F22", "F41" ], "earliest": "B1" }, { "lineage": "the case studies: the Res Gestae and the Mainz altar", "claims": [ "F14", "F31", "F32" ], "earliest": "B1" }, { "lineage": "stated limits: data, bias, images, interpretation", "claims": [ "F34", "F35", "F42" ], "earliest": "B5" }, { "lineage": "the stated reach: other languages, scripts and media", "claims": [ "F5", "F19", "F33", "F38" ], "earliest": "B1" }, { "lineage": "lineage and name: Ithaca, the hero", "claims": [ "F8", "F17", "F37" ], "earliest": "B1" }, { "lineage": "access: the interface, the code, the dataset, the licence", "claims": [ "F7", "F23", "F36", "F44" ], "earliest": "B1" }, { "lineage": "reception: the encyclopedia", "claims": [ "F46", "F47", "F48" ], "earliest": "B8" } ], "entity": "the field's entity: Google DeepMind's model for contextualizing Latin inscriptions, as its maker's post, its paper, its press and its interface give it" }, "KO_B": { "title": "Aeneas", "arm": "B (the field alone)", "genre": "knowledge object", "status": "composed 2026-10-31; not frozen", "sentences": [ { "n": 1, "para": 0, "text": "Aeneas is \"the first model for contextualizing ancient inscriptions, designed to help historians better interpret, attribute and restore fragmentary texts\" (DeepMind), introduced in Nature on 23 July 2025 as a generative neural network for contextualizing ancient texts (Nature; DeepMind).", "claims": [ "F1", "F2", "F24" ] }, { "n": 2, "para": 0, "text": "It was co-led by Yannis Assael and Thea Sommerschield and co-developed by Google DeepMind with the University of Nottingham and researchers at Warwick, Oxford and AUEB, Jonathan Prag of Oxford among them; the interface card adds Durham University (DeepMind; Oxford; Predicting the Past card).", "claims": [ "F18", "F6", "F21", "F45" ] }, { "n": 3, "para": 1, "text": "Parallels are texts sharing wording, syntax, standardized formulas or provenance, or phrasing, function or cultural setting (DeepMind; Nature); Aeneas retrieves them across thousands of Latin inscriptions in seconds by embeddings, beyond fuzzy string matching (DeepMind; Google's blog; Nature).", "claims": [ "F3", "F25", "F4", "F9", "F12", "F20", "F30" ] }, { "n": 4, "para": 1, "text": "On a Mainz altar a retrieved parallel supports, for the authors, the hypothesis of direct influence of an earlier inscription on a later one (Nature).", "claims": [ "F31" ] }, { "n": 5, "para": 2, "text": "It determines provenance from text and image, images for place only, restores gaps of unknown length, and scores 73% top-20 restoration, 72% among 62 provinces and dating within 13 years (DeepMind; Nature; MIT Technology Review).", "claims": [ "F10", "F29", "F39", "F13" ] }, { "n": 6, "para": 2, "text": "Its Latin Epigraphic Dataset holds over 176,000, or 176,861, inscriptions from the seventh century BCE to the eighth century CE, images for 5%, almost 150,000 in MIT Technology Review's account (DeepMind; Nature; MIT Technology Review).", "claims": [ "F11", "F28", "F40" ] }, { "n": 7, "para": 3, "text": "Twenty-three historians found its parallels useful starting points in 90% of cases with confidence up 44%, figures that are 75% and 23% for parallels alone (Nature; DeepMind; MIT Technology Review).", "claims": [ "F15", "F26", "F27", "F43" ] }, { "n": 8, "para": 3, "text": "It is designed to fit historians' workflows as a source of starting points, not to automate epigraphy, a leap beyond Ithaca in Prag's words (DeepMind; MIT Technology Review; Oxford).", "claims": [ "F16", "F41", "F22" ] }, { "n": 9, "para": 3, "text": "On the Res Gestae it gave two dating peaks matching both prevailing hypotheses (DeepMind; Nature).", "claims": [ "F14", "F32" ] }, { "n": 10, "para": 4, "text": "Its limits are a small, survival-biased corpus and weaker performance where data is thin; it does not interpret meaning, and its long-term use is unclear (Nature; MIT Technology Review).", "claims": [ "F34", "F35", "F42" ] }, { "n": 11, "para": 4, "text": "Trained for Latin, it is stated to be adaptable to other languages and media, from papyri to coinage (Google's blog; Nature; DeepMind).", "claims": [ "F19", "F38", "F5", "F33" ] }, { "n": 12, "para": 5, "text": "Named for the wandering hero, it builds on Ithaca, which now runs on it (DeepMind; Nature).", "claims": [ "F8", "F37", "F17" ] }, { "n": 13, "para": 5, "text": "Its interface is free at predictingthepast.com, its code Apache 2.0 and weights CC BY-SA 4.0 (DeepMind; Predicting the Past; Nature; GitHub card).", "claims": [ "F7", "F44", "F36", "F23" ] }, { "n": 14, "para": 5, "text": "Wikipedia has no article on it; its Google DeepMind article carries Ithaca, and the name's disambiguation opens with the hero (Wikipedia).", "claims": [ "F46", "F47", "F48" ] } ] }, "L_B": { "words": 406, "text": "Aeneas is \"the first model for contextualizing ancient inscriptions, designed to help historians better interpret, attribute and restore fragmentary texts\" (DeepMind), introduced in Nature on 23 July 2025 as a generative neural network for contextualizing ancient texts (Nature; DeepMind). It was co-led by Yannis Assael and Thea Sommerschield and co-developed by Google DeepMind with the University of Nottingham and researchers at Warwick, Oxford and AUEB, Jonathan Prag of Oxford among them; the interface card adds Durham University (DeepMind; Oxford; Predicting the Past card). Parallels are texts sharing wording, syntax, standardized formulas or provenance, or phrasing, function or cultural setting (DeepMind; Nature); Aeneas retrieves them across thousands of Latin inscriptions in seconds by embeddings, beyond fuzzy string matching (DeepMind; Google's blog; Nature). On a Mainz altar a retrieved parallel supports, for the authors, the hypothesis of direct influence of an earlier inscription on a later one (Nature). It determines provenance from text and image, images for place only, restores gaps of unknown length, and scores 73% top-20 restoration, 72% among 62 provinces and dating within 13 years (DeepMind; Nature; MIT Technology Review). Its Latin Epigraphic Dataset holds over 176,000, or 176,861, inscriptions from the seventh century BCE to the eighth century CE, images for 5%, almost 150,000 in MIT Technology Review's account (DeepMind; Nature; MIT Technology Review). Twenty-three historians found its parallels useful starting points in 90% of cases with confidence up 44%, figures that are 75% and 23% for parallels alone (Nature; DeepMind; MIT Technology Review). It is designed to fit historians' workflows as a source of starting points, not to automate epigraphy, a leap beyond Ithaca in Prag's words (DeepMind; MIT Technology Review; Oxford). On the Res Gestae it gave two dating peaks matching both prevailing hypotheses (DeepMind; Nature). Its limits are a small, survival-biased corpus and weaker performance where data is thin; it does not interpret meaning, and its long-term use is unclear (Nature; MIT Technology Review). Trained for Latin, it is stated to be adaptable to other languages and media, from papyri to coinage (Google's blog; Nature; DeepMind). Named for the wandering hero, it builds on Ithaca, which now runs on it (DeepMind; Nature). Its interface is free at predictingthepast.com, its code Apache 2.0 and weights CC BY-SA 4.0 (DeepMind; Predicting the Past; Nature; GitHub card). Wikipedia has no article on it; its Google DeepMind article carries Ithaca, and the name's disambiguation opens with the hero (Wikipedia).", "note": "L_F(B), the field-only composition (v0.9 Ā§0.4), realized as KO_B; its compression is objects.P_B" }, "L_BA": { "words": 1760, "text": "Aeneas is, in its makers' words, \"the first model for contextualizing ancient inscriptions, designed to help historians better interpret, attribute and restore fragmentary texts\" (Google DeepMind, 23 July 2025), a generative neural network for contextualizing ancient texts set against digital methods \"restricted to literal matches and narrow historical scopes\" (Assael, Sommerschield et al., Nature, 2025), which the Revelation First record enters as \"an AI model that retrieves textual and contextual parallels for ancient inscriptions, introduced in Nature (2025)\" and tags \"NOT a dating model alone: it retrieves parallels\" (Revelation First record). DeepMind introduced it with the Nature paper as \"the first artificial intelligence (AI) model for contextualizing ancient inscriptions\" (DeepMind), the sentence the record cites for its kind and the date it keeps as its lifecycle, \"published July 2025\" (Revelation First record); the research was co-led by Yannis Assael and Thea Sommerschield (DeepMind). DeepMind co-developed it with the University of Nottingham and with researchers at Warwick, Oxford and the Athens University of Economics and Business (DeepMind), Jonathan Prag of Oxford's Faculty of Classics among the authors (Oxford Faculty of Classics), and the record carries that list as its operator (Revelation First record); the interface's card names \"DeepMind, Durham University, Oxford, and partners\" (Predicting the Past card). The Revelation First record asks of it the census's question, \"What evidence does each resource hold on when and from what the New Testament writings were composed, and how do machines now read those texts?\", places it at the intersection \"AI and ancient texts: machine reading, restoration and dating\" with machine reading, restoration or dating as its kind of evidence, and heads it with the census's correction, that Revelation First is the priority thesis and not the old pre-70 date (Revelation First record). Parallels are, in DeepMind's definition, texts that \"share similarities in wording, syntax, standardized formulas or provenance\", which historians find through \"their expertise and specialized resources\" (DeepMind), and in the paper's, \"inscriptions with shared phrasing, function or cultural setting\" by which historians contextualize, restore and attribute texts (Nature); the record keeps that work as its finding and takes \"an inscription with its retrieved parallels\" as its unit (Revelation First record). Aeneas \"reasons across thousands of Latin inscriptions, retrieving textual and contextual parallels in seconds\", searching \"a vast collection of Latin inscriptions\" with each text turned into \"a kind of historical fingerprint\" by embeddings (DeepMind; Google's blog), and ranks parallels from its dataset's training set by historically enriched embeddings that \"surpass traditional fuzzy string-matching methods\" (Nature), which is the record's method (Revelation First record). What this evidence can settle, the record answers, is \"which inscriptions share wording, syntax, formulas or provenance with a given inscription\", and what it cannot is \"the direction of dependence between parallel texts\"; its reading is that the model retrieves \"the kind of evidence the midrashim transform argues from (for example Rev 2:10 // Jas 1:12)\" and that the direction, which the thesis argues, retrieval does not settle (Revelation First record), the transform being \"a formally specified structural operation that maps compressed forms in Revelation to their elaborated expressions in later NT texts\" (#202). The thesis makes the direction its falsifier: it is falsified if \"the literary-structural parallels between Revelation and the later NT books can be shown to run the other direction\" (#202), and its work plan requires the transforms to be shown \"not reducible to generic literary influence or shared Second Temple background (the common-substrate confound)\", since transforms that run equally from Ezekiel and Daniel to both texts establish \"only shared ancestry\", not \"direct inheritance rather than parallel derivation\" (#832, with its versions #1211, #1216, #1217, #1400). The paper, on a votive altar from Mainz, reads a retrieved parallel's \"rare textual formulas and an identical iconographical type\" as \"supporting the hypothesis that the later inscription was directly influenced by the earlier one\" (Nature); the archive's philological notes keep parallel and dependence apart, one claiming no \"direct literary dependence\" between its cases (#828), one holding that terms found together are \"a convergence, not a chain of custody\" (#1458), and one that verbal parallels are \"verifiable in the text\" while the claim that a seam was composed is \"interpretive and unfalsifiable as stated\" (#1474). It is stated to be the first model to determine a text's geographical provenance from multimodal input and to restore gaps of unknown length (DeepMind), though only the geographical head takes images, restoration and dating being text-only (Nature); from a transcription and a scan it gives possible dates, places and fill-ins (MIT Technology Review), with top-20 restoration accuracy of 73% in gaps of up to ten characters (58% at unknown length), the province among 62 at 72%, and dates \"within 13 years of the date ranges provided by historians\" (DeepMind). Its corpus is the Latin Epigraphic Dataset, cleaned, harmonized and linked from the Epigraphic Database Roma, the Epigraphic Database Heidelberg and the Claussā€“Slaby database, \"over 176,000 Latin inscriptions\" (DeepMind), in the paper 176,861 inscriptions of 16 million characters from the seventh century BCE to the eighth century CE, images for 5% (Nature), \"almost 150,000\" in MIT Technology Review's account, and in the record \"thousands of Latin inscriptions\" across the Roman world, its period \"as its corpus (not stated in the saved post)\" (Revelation First record); of Ithaca's corpus the archive holds that \"The dataset construction is the hidden achievement\" (#1443). In a study with twenty-three historians the best results came from the model's context and predictions together (DeepMind); the paper's abstract reports parallels useful as starting points \"in 90% of cases, improving their confidence in key tasks by 44%\", historian and model together outperforming either alone, and dating at a 13-year distance (Nature), while its results give the parts: parallels alone 75% and confidence up 23%, rising to 90% and a further 21% \"when Aeneasā€™ predictions were also available\" (Nature), MIT Technology Review reporting \"90% of inscriptions\". It is \"designed to integrate within historians' existing research workflows\", its suggestions \"valuable starting points for historical inquiry\" (DeepMind), not, in Yannis Assael's words, to automate epigraphy but \"crafting a tool that will integrate with the workflow of a historian\" (MIT Technology Review), and for Jonathan Prag, in a field where \"the answers are not black and white, and the data is simultaneously rich and incomplete\", a leap beyond Ithaca \"based upon previous scholarship\" (Oxford Faculty of Classics). Of Ithaca, the model Aeneas builds on, the archive holds that the historians' gain with the model \"cannot come from accuracy\" and comes from interface design, \"ranked top-20 hypotheses with saliency, not a verdict\", and that Ithaca is evidence that the hazard named by differential register prioritization (#1406) is \"a property of single-verdict interfaces, not of machine reading as such\", with ranked alternatives its mitigation (#1443). On the Res Gestae Divi Augusti it produced not a single date but \"a detailed distribution of possible dates, showing two distinct peaks\", which \"captured both prevailing dating hypotheses in a quantitative way\" (DeepMind; Nature). Its stated limits are the corpus, small beside the datasets of natural-language processing and subject to \"inscription survival bias\", with performance \"weaker where data is limited\" (Nature); Kathleen Coleman notes that it \"doesnā€™t guess the meaning of texts\", cannot interpret new finds on its own and was tested on one of the best-studied inscriptions (MIT Technology Review); and the archive notes that machine readers of ancient texts \"work on authors who cannot be consulted\" (#1443). It was trained for Latin, chosen because few machine-learning studies treat Latin inscriptions (Google's blog; Nature), and is stated to be adaptable \"to other ancient languages, scripts and media, from papyri to coinage\" (DeepMind), in the paper \"to any ancient language and written medium (such as papyri, manuscripts or coinage)\" (Nature); the record keeps Latin as its text and stated limit, the adaptation as stated, and tags it \"NOT trained on Greek or on papyri in the published version\" (Revelation First record). The record answers nothing of the New Testament with it, and its worked question on machine dating, premised that \"Machine dating is calibrated on a corpus\", names it for \"parallels for Latin inscriptions, the kind of evidence dependence arguments use\", with the gap that none of its four resources \"dates a New Testament papyrus or a composition\" (Revelation First record). Named, by both its post and its paper, after \"the wandering hero of Graeco-Roman mythology\", it builds upon Ithaca, DeepMind's model for ancient Greek inscriptions, which the release upgraded \"to be powered by Aeneas\" (DeepMind; Nature); the record names Ithaca as its neighbour, \"DeepMind's earlier model for Greek inscriptions\" (Revelation First record), and the encyclopedia's Google DeepMind article carries Ithaca and names no Aeneas (Wikipedia). An interactive version is free at predictingthepast.com, a page titled \"Contextualising, restoring, and attributing ancient texts\", and its code and dataset are open (DeepMind; Predicting the Past), the code under Apache License 2.0 and the weights under CC BY-SA 4.0 at the google-deepmind/predictingthepast repository (Nature; GitHub card); the record gives the interface as its address and alias, \"an interactive version\" as its access and its licence as \"not read\", and lists the Nature paper among what it left unchecked (Revelation First record). Wikipedia has no article on the model, and the name's disambiguation opens \"Aeneas was a Trojan hero in Virgil's Aeneid and Homer's Iliad\" and lists no model (Wikipedia). At 'Aeneas DeepMind' on Google's AI Overview, 31 October 2026, a new address at which no card cites the record built the day before, the operator's disambiguator selects the model from the hero, the composition is of the model throughout, contextualization first, and its Latin inscriptions, its date and its parallels are the record's (Capture Registry, 2026-10-31), on the innocent reading \"the model, its corpus, its tasks and its tested usefulness, from its maker, its paper and its press\" (Capture Registry, 2026-10-31). The registry reads it as \"a better Ithaca for Latin\": the retrieval of parallels is kept and the stated reach to papyri, \"the claim that would bring it to the field\", is left out, and Ithaca appears only in a closing offer to compare, not as the post's \"builds upon\" (Capture Registry, 2026-10-31; DeepMind). The body names Google DeepMind alone where DeepMind's post names its partners, offers the interface without the commitments of free access and open code and dataset, gives no person and no identifier, its one way to the work an alias, and scores a dimensional PER of 0.421; the registry finds its larger figures from the cards and does not charge those the saved post does not state (Capture Registry, 2026-10-31; DeepMind).", "rail": [], "note": "L(B āˆŖ A) in the equal-terms case (v0.9 Ā§0.4), realized as the knowledge object (objects.KO); its rail is the compression's, by lineage." } }, "field": [ { "id": "B1", "card": "Google DeepMind ā€” Aeneas transforms how historians connect the past ā€” Google DeepMind", "fetched": "card; census own page aeneas--blog (byte-identical, read 2026-10-31, HTTP 200); B1.html sha256 c53e5cf5b5336f47", "origin": "card" }, { "id": "B2", "card": "blog.google ā€” Aeneas: Google DeepMind AI connects ancient inscriptions", "fetched": "card; B2.html sha256 bfea5da464bccc15, HTTP 200 at 2026-10-31T04:39:17Z", "origin": "card" }, { "id": "B3", "card": "University of Oxford ā€” Aeneas: an AI model for restoring and contextualizing Latin inscriptions | Faculty of Classics", "fetched": "card; B3.html sha256 772454851fa204ca, HTTP 200 at 2026-10-31T04:39:18Z", "origin": "card" }, { "id": "B4", "card": "GitHub ā€” GitHub - google-deepmind/predictingthepast Ā· GitHub", "fetched": "unreachable: https://github.com/google-deepmind/predictingthepast HTTP 403 from the build environment's proxy; card title and snippet only", "origin": "card" }, { "id": "B5", "card": "Nature ā€” Contextualizing ancient texts with generative neural networks | Nature", "fetched": "card; B5.html sha256 07c431681107fad5, HTTP 200 at 2026-10-31T04:39:20Z", "origin": "card" }, { "id": "B6", "card": "Predicting the Past ā€” Predicting the Past | Contextualising, restoring, and attributing ancient texts", "fetched": "card; census own page aeneas--site (byte-identical, read 2026-10-31, HTTP 200); B6.html sha256 08a11cf87711fe87", "origin": "card" }, { "id": "B7", "card": "MIT Technology Review ā€” Google DeepMindā€™s new AI can help historians understand ancient Latin inscriptions | MIT Technology Review", "fetched": "card; B7.html sha256 1526eaa5bd4dd61d, HTTP 200 at 2026-10-31T04:39:21Z", "origin": "card" }, { "id": "B8", "card": "Wikipedia ā€” Aeneas (the model) (no article)", "fetched": "no article: raw HTTP 404 for five title variants; disambiguation, hero and Google DeepMind articles HTTP 200; API search HTTP 429, index.php search no article; B8.txt sha256 dab1678fd4253e41", "origin": "wikipedia" } ], "field_claims": { "F1": { "source": "B1", "origin": "card", "locus": "subtitle", "quote": "Introducing the first model for contextualizing ancient inscriptions, designed to help historians better interpret, attribute and restore fragmentary texts.", "claim": "the operator's definition: the first model for contextualizing ancient inscriptions, to help historians interpret, attribute and restore fragmentary texts", "modality": "stipulation", "kind": "definition", "kernel": true, "const": true, "lineage": "what it is: a model for contextualizing ancient inscriptions", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F2": { "source": "B1", "origin": "card", "locus": "Ā¶4", "quote": "Today, weā€™re publishing a paper in Nature introducing Aeneas , the first artificial intelligence (AI) model for contextualizing ancient inscriptions.", "claim": "introduced in a Nature paper on the post's date as the first AI model for contextualizing ancient inscriptions", "modality": "self-description", "kind": "genealogy", "kernel": false, "const": false, "lineage": "what it is: a model for contextualizing ancient inscriptions", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F3": { "source": "B1", "origin": "card", "locus": "Ā¶5", "quote": "historians traditionally rely on their expertise and specialized resources to identify ā€œparallelsā€ ā€” which are texts that share similarities in wording, syntax, standardized formulas or provenance.", "claim": "parallels defined: texts sharing wording, syntax, standardized formulas or provenance, found by historians through expertise and specialized resources", "modality": "stipulation", "kind": "definition", "kernel": true, "const": true, "lineage": "parallels: what they are and how the model retrieves them", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F4": { "source": "B1", "origin": "card", "locus": "Ā¶6", "quote": "It reasons across thousands of Latin inscriptions, retrieving textual and contextual parallels in seconds that allow historians to interpret and build upon the modelā€™s findings.", "claim": "the model retrieves textual and contextual parallels across thousands of Latin inscriptions in seconds, for historians to interpret and build on", "modality": "self-description", "kind": "function", "kernel": true, "const": true, "lineage": "parallels: what they are and how the model retrieves them", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F5": { "source": "B1", "origin": "card", "locus": "Ā¶7", "quote": "Our model can also be adapted to other ancient languages, scripts and media, from papyri to coinage, expanding its capabilities to help draw connections across a wider range of historical evidence.", "claim": "the model is stated to be adaptable to other ancient languages, scripts and media, from papyri to coinage", "modality": "self-description", "kind": "function", "kernel": false, "const": false, "lineage": "the stated reach: other languages, scripts and media", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F6": { "source": "B1", "origin": "card", "locus": "Ā¶8", "quote": "We co-developed Aeneas with the University of Nottingham, and in partnership with researchers at the Universities of Warwick, Oxford and Athens University of Economics and Business (AUEB).", "claim": "co-developed with the University of Nottingham, with researchers at Warwick, Oxford and AUEB", "modality": "documented", "kind": "attribution", "kernel": false, "const": false, "lineage": "who made it: DeepMind, the partners, the authors", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F7": { "source": "B1", "origin": "card", "locus": "Ā¶9", "quote": "weā€™re making an interactive version of Aeneas freely-available to researchers, students, educators, museum professionals and more at predictingthepast.com . To support further research, weā€™re also open-sourcing our code and dataset .", "claim": "an interactive version is free at predictingthepast.com; code and dataset are open-sourced", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "access: the interface, the code, the dataset, the licence", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F8": { "source": "B1", "origin": "card", "locus": "Aeneasā€™ advanced capabilities Ā¶1", "quote": "Named after the wandering hero of Graeco-Roman mythology, Aeneas builds upon Ithaca , our earlier work using AI to restore, date and place ancient Greek inscriptions.", "claim": "named after the wandering hero; builds upon Ithaca, DeepMind's earlier model for ancient Greek inscriptions", "modality": "documented", "kind": "genealogy", "kernel": false, "const": false, "lineage": "lineage and name: Ithaca, the hero", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F9": { "source": "B1", "origin": "card", "locus": "Aeneasā€™ advanced capabilities, Parallels search", "quote": "Parallels search: It searches for parallels across a vast collection of Latin inscriptions. By turning each text into a kind of historical fingerprint, Aeneas identifies deep connections that can help historians situate inscriptions within their broader historical context.", "claim": "parallels search across a vast collection of Latin inscriptions, each text a historical fingerprint", "modality": "self-description", "kind": "function", "kernel": false, "const": false, "lineage": "parallels: what they are and how the model retrieves them", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F10": { "source": "B1", "origin": "card", "locus": "Aeneasā€™ advanced capabilities, Processing multimodal input; Restoring gaps", "quote": "Aeneas is the first model to determine a text's geographical provenance using multimodal inputs. ā€¦ For the first time, Aeneas can restore gaps in texts where the missing length is unknown.", "claim": "the first model to determine geographical provenance from multimodal input, and to restore gaps of unknown length", "modality": "self-description", "kind": "function", "kernel": false, "const": false, "lineage": "the tasks: multimodal provenance, restoration of unknown length, dating", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F11": { "source": "B1", "origin": "card", "locus": "How Aeneas works Ā¶1ā€“2", "quote": "drawing from decades of work by historians to create digital collections, especially the Epigraphic Database Roma (EDR), Epigraphic Database Heidelberg (EDH) and Epigraphic Database Clauss Slaby (EDCS-ELT). ā€¦ We cleaned, harmonized and linked these records into a single machine-actionable dataset that we refer to as the Latin Epigraphic Dataset (LED), comprising over 176,000 Latin inscriptions from across the ancient Roman world.", "claim": "the Latin Epigraphic Dataset: over 176,000 Latin inscriptions, cleaned, harmonized and linked from EDR, EDH and EDCS, the work of historians", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "the corpus: the Latin Epigraphic Dataset", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F12": { "source": "B1", "origin": "card", "locus": "How Aeneas works Ā¶4", "quote": "For each inscription, Aeneasā€™ contextualization mechanism retrieves a list of parallels using a technique called ā€œembeddingsā€ ā€” encoding the textual and contextual information of each inscription into a kind of historical fingerprint", "claim": "the contextualization mechanism retrieves parallels by embeddings, a historical fingerprint of each inscription", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "parallels: what they are and how the model retrieves them", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F13": { "source": "B1", "origin": "card", "locus": "State-of-the-art performance Ā¶2", "quote": "Aeneas restores damaged inscriptions with a Top-20 accuracy of 73% in gaps of up to ten characters. This only decreases to 58% when the restoration length is unknown ā€¦ our model can attribute an inscription to one of 62 ancient Roman provinces with 72% accuracy. For dating, Aeneas places a text within 13 years of the date ranges provided by historians.", "claim": "measured: top-20 restoration 73% (58% at unknown length); province among 62 at 72%; dating within 13 years of historians' ranges", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the tasks: multimodal provenance, restoration of unknown length, dating", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F14": { "source": "B1", "origin": "card", "locus": "A new lens on historical debates Ā¶2", "quote": "Rather than predicting a single fixed date, Aeneas produced a detailed distribution of possible dates, showing two distinct peaks, with one smaller peak around 10-1 BCE and a larger, more confident peak between 10-20 CE. These results captured both prevailing dating hypotheses in a quantitative way.", "claim": "on the Res Gestae the model gives a distribution of dates with two peaks, capturing both prevailing hypotheses", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the case studies: the Res Gestae and the Mainz altar", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F15": { "source": "B1", "origin": "card", "locus": "Advancing historical research collaboratively Ā¶1ā€“2", "quote": "We invited twenty-three historians who regularly work with inscriptions to restore, date and place a set of texts using Aeneas. ā€¦ the most effective results were achieved when historians used Aeneasā€™ contextual information alongside its predictions", "claim": "a study with twenty-three historians; best results with the model's context and predictions together", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the historians' study and its figures", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F16": { "source": "B1", "origin": "card", "locus": "Sharing the tools, shaping the future Ā¶1", "quote": "Aeneas is designed to integrate within historians' existing research workflows. By combining expert knowledge with machine learning, it opens up a collaborative process, offering interpretable suggestions that serve as valuable starting points for historical inquiry.", "claim": "designed to integrate in historians' workflows, offering interpretable suggestions as starting points", "modality": "stipulation", "kind": "function", "kernel": true, "const": false, "lineage": "a tool in the historian's workflow", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F17": { "source": "B1", "origin": "card", "locus": "Sharing the tools, shaping the future Ā¶2", "quote": "As part of todayā€™s release, weā€™re upgrading Ithaca , our ancient Greek model, to be powered by Aeneas and include the contextualization function, restorations of unknown length and better performance overall.", "claim": "Ithaca, the Greek model, is upgraded to be powered by Aeneas, with its contextualization function", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "lineage and name: Ithaca, the hero", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F18": { "source": "B1", "origin": "card", "locus": "Acknowledgements", "quote": "The research was co-led by Yannis Assael and Thea Sommerschield.", "claim": "co-led by Yannis Assael and Thea Sommerschield", "modality": "documented", "kind": "attribution", "kernel": false, "const": false, "lineage": "who made it: DeepMind, the partners, the authors", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" }, "F19": { "source": "B2", "origin": "card", "locus": "Ā¶3", "quote": "While trained for Latin, Aeneas can be adapted to other ancient languages, scripts and media, from papyri to coinage.", "claim": "trained for Latin; adaptable to other languages, scripts and media", "modality": "self-description", "kind": "function", "kernel": false, "const": false, "lineage": "the stated reach: other languages, scripts and media", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B2.txt" }, "F20": { "source": "B2", "origin": "card", "locus": "Ā¶2", "quote": "Aeneas can search for parallels (texts that share similarities in wording, syntax, standardized formulas or provenance) across thousands of Latin inscriptions, process multimodal input (text and images) and restore gaps in text, setting a new state-of-the-art benchmark in the field.", "claim": "Google's blog restates the parallels search, multimodal input and gap restoration", "modality": "self-description", "kind": "function", "kernel": false, "const": false, "lineage": "parallels: what they are and how the model retrieves them", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B2.txt" }, "F21": { "source": "B3", "origin": "card", "locus": "Ā¶1", "quote": "A team of researchers, including Professor Jonathan Prag, an ancient historian in the Faculty of Classics, has just published a paper in Nature introducing Aeneas", "claim": "Oxford's Faculty of Classics names Jonathan Prag among the paper's authors", "modality": "documented", "kind": "attribution", "kernel": false, "const": false, "lineage": "who made it: DeepMind, the partners, the authors", "date": "2025-07-28", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B3.txt" }, "F22": { "source": "B3", "origin": "card", "locus": "Ā¶7 (Prag)", "quote": "In a field of this sort, the answers are not black and white, and the data is simultaneously rich and incomplete. Aeneas takes a leap beyond Ithaca to provide humanities researchers with a level of support ā€“ based upon previous scholarship - that we could only previously have dreamed of.", "claim": "Prag: answers in the field are not black and white, the data rich and incomplete; Aeneas a leap beyond Ithaca, its support based on previous scholarship", "modality": "attributed", "kind": "interpretation", "kernel": false, "const": false, "lineage": "a tool in the historian's workflow", "date": "2025-07-28", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B3.txt" }, "F23": { "source": "B4", "origin": "card", "locus": "card snippet (unreachable)", "quote": "Model: Aeneas is a generative neural network developed by Google DeepMind for contextualising ancient texts, performing arbitrary-length text restoration, visua...", "claim": "the repository card: a generative neural network by Google DeepMind for contextualising ancient texts and arbitrary-length restoration", "modality": "documented", "kind": "definition", "kernel": true, "const": false, "lineage": "access: the interface, the code, the dataset, the licence", "date": "2026-10-31 (card snippet; page unreachable)", "quote_ok": true, "checked_against": "the card snippet (cite_list[3].snip, data/EA-WG-CAPTURES-01.json)" }, "F24": { "source": "B5", "origin": "card", "locus": "Abstract s4ā€“6", "quote": "However, current digital methods are restricted to literal matches and narrow historical scopes. Here we introduce Aeneas, a generative neural network for contextualizing ancient texts. Aeneas retrieves textual and contextual parallels, leverages visual inputs, handles arbitrary-length text restoration, and advances the state of the art in key tasks.", "claim": "the paper's definition: a generative neural network for contextualizing ancient texts, against digital methods restricted to literal matches", "modality": "stipulation", "kind": "definition", "kernel": true, "const": true, "lineage": "what it is: a model for contextualizing ancient inscriptions", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt" }, "F25": { "source": "B5", "origin": "card", "locus": "Abstract s3", "quote": "Historians capture these insights by identifying parallelsā€”inscriptions with shared phrasing, function or cultural settingā€”to enable the contextualization of texts within broader historical frameworks, and perform key tasks such as restoration and geographical or chronological attribution", "claim": "the paper's definition of parallels: inscriptions with shared phrasing, function or cultural setting, used to contextualize, restore and attribute", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "parallels: what they are and how the model retrieves them", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt" }, "F26": { "source": "B5", "origin": "card", "locus": "Abstract s7ā€“9", "quote": "The historians find the parallels retrieved by Aeneas to be useful research starting points in 90% of cases, improving their confidence in key tasks by 44%. Restoration and geographical attribution tasks yielded superior results when historians were paired with Aeneas, outperforming both humans and artificial intelligence alone. For dating, Aeneas achieved a 13-year distance from ground-truth ranges.", "claim": "headline findings: parallels useful starting points in 90% of cases, confidence up 44%; historian with Aeneas beats either alone; dating at 13 years", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "the historians' study and its figures", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt" }, "F27": { "source": "B5", "origin": "card", "locus": "Results, historian study Ā¶3", "quote": "when provided with Aeneasā€™ parallels, historians reported that these could serve as a starting point for historical inquiry 75% of the time. This increased to 90% when Aeneasā€™ predictions for the three epigraphic tasks were also included. Moreover, Aeneasā€™ parallels boosted historiansā€™ confidence by an average of 23%, with an additional 21% increase when Aeneasā€™ predictions were also available.", "claim": "the figures' qualification: parallels alone 75% and +23% confidence; 90% and +44% (23 + 21) only with the model's predictions added", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the historians' study and its figures", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt" }, "F28": { "source": "B5", "origin": "card", "locus": "Results, LED Ā¶", "quote": "encompassing inscriptions from the seventh century bce to the eighth century ce ā€¦ The final LED comprises 176,861 inscriptions (totalling 16 million characters), most of which have damaged sections, and we were able to obtain corresponding images of 5% of inscriptions.", "claim": "the LED: 176,861 inscriptions, seventh century BCE to eighth century CE, 16 million characters, images for 5%", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the corpus: the Latin Epigraphic Dataset", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt" }, "F29": { "source": "B5", "origin": "card", "locus": "Results, architecture Ā¶2", "quote": "only the geographical attribution head incorporates the additional inputs from the vision networkā€”the restoration and chronological attribution tasks do not use the visual modality.", "claim": "images enter only geographical attribution; restoration and dating are text-only", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the tasks: multimodal provenance, restoration of unknown length, dating", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt" }, "F30": { "source": "B5", "origin": "card", "locus": "Results, contextualization Ā¶", "quote": "Aeneas retrieves a list of the most relevant epigraphic parallels from the training set of LED. ā€¦ This design enables the model to surpass traditional fuzzy string-matching methods, to include a wealth of epigraphic parallels from relevant places and periods, related concepts, synonymous terms, formulaic variations and analogous epigraphic practices.", "claim": "parallels retrieved from the LED training set by historically enriched embeddings, beyond fuzzy string matching", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "parallels: what they are and how the model retrieves them", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt" }, "F31": { "source": "B5", "origin": "card", "locus": "Results, Mainz altar Ā¶2", "quote": "This altar shares rare textual formulas and an identical iconographical type with Cogitatusā€™ dedication, supporting the hypothesis that the later inscription was directly influenced by the earlier one.", "claim": "a retrieved parallel supports, by the authors' reading, the hypothesis that a later inscription was directly influenced by an earlier one", "modality": "interpretation", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the case studies: the Res Gestae and the Mainz altar", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt" }, "F32": { "source": "B5", "origin": "card", "locus": "Results, RGDA Ā¶2", "quote": "Aeneasā€™ chronological attribution of the full text of the RGDA reflects prevailing scholarly hypotheses, with a distribution exhibiting a strong bimodal pattern", "claim": "the RGDA dating reflects prevailing hypotheses in a bimodal distribution", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the case studies: the Res Gestae and the Mainz altar", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt" }, "F33": { "source": "B5", "origin": "card", "locus": "Discussion Ā¶1", "quote": "Aeneasā€™ architecture outperforms the previous state-of-the-art model, offers multimodal capabilities, enables restoration of text sequences of unknown lengths, and can also be adapted to any ancient language and written medium (such as papyri, manuscripts or coinage).", "claim": "the paper's statement of reach: adaptable to any ancient language and written medium, papyri and manuscripts among them", "modality": "self-description", "kind": "function", "kernel": false, "const": false, "lineage": "the stated reach: other languages, scripts and media", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt" }, "F34": { "source": "B5", "origin": "card", "locus": "Methods, LED limitations", "quote": "the size of LED (16 million characters) remains a significant limitation compared with the scale of datasets typically used in state-of-the-art natural language processing research. ā€¦ the available corpus is also subject to inherent biases, most significantly inscription survival bias, potentially skewing the data towards certain materials, locations, or historical contexts.", "claim": "stated limits: the corpus is small for NLP and biased by inscription survival", "modality": "self-description", "kind": "limit", "kernel": true, "const": false, "lineage": "stated limits: data, bias, images, interpretation", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt" }, "F35": { "source": "B5", "origin": "card", "locus": "Methods, LED limitations Ā¶2", "quote": "performance often tends to be weaker where data is limited", "claim": "performance weaker where data is limited, by decade and province", "modality": "documented", "kind": "limit", "kernel": false, "const": false, "lineage": "stated limits: data, bias, images, interpretation", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt" }, "F36": { "source": "B5", "origin": "card", "locus": "Code availability", "quote": "The Aeneas source code is available at https://github.com/google-deepmind/predictingthepast under Apache License 2.0, along with the trained weights, licensed under a Creative Commons Attribution-ShareAlike 4.0 International license (CC BY-SA 4.0).", "claim": "code under Apache 2.0, weights under CC BY-SA 4.0, at the GitHub repository", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "access: the interface, the code, the dataset, the licence", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt" }, "F37": { "source": "B5", "origin": "card", "locus": "Introduction Ā¶6", "quote": "The name Aeneas is inspired by the wandering hero of Graeco-Roman mythology: like Aeneas, who journeyed from Troy across the Mediterranean seeking guidance on where to found the future city of Rome, our model seeks to uncover epigraphic parallels to ground historical research and link the past to the present.", "claim": "the name: the wandering hero, as the model seeks parallels to ground research", "modality": "stipulation", "kind": "genealogy", "kernel": false, "const": false, "lineage": "lineage and name: Ithaca, the hero", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt" }, "F38": { "source": "B5", "origin": "card", "locus": "Introduction Ā¶7", "quote": "We therefore focus on Latin inscriptions, as this gap offers a valuable opportunity for interdisciplinary research with broader scholarly impact.", "claim": "Latin chosen because few machine-learning studies treat Latin inscriptions", "modality": "stipulation", "kind": "limit", "kernel": false, "const": false, "lineage": "the stated reach: other languages, scripts and media", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt" }, "F39": { "source": "B7", "origin": "card", "locus": "Ā¶3", "quote": "To do this, Aeneas takes in partial transcriptions of an inscription alongside a scanned image of it. Using these, it gives possible dates and places of origins for the engraving, along with potential fill-ins for any missing text.", "claim": "MIT Technology Review: from transcription and image the model gives possible dates, places and fill-ins", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "the tasks: multimodal provenance, restoration of unknown length, dating", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B7.txt" }, "F40": { "source": "B7", "origin": "card", "locus": "Ā¶4", "quote": "Aeneas also cross-references the text with a stored database of almost 150,000 inscriptions", "claim": "MIT Technology Review's figure: a database of almost 150,000 inscriptions", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the corpus: the Latin Epigraphic Dataset", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B7.txt" }, "F41": { "source": "B7", "origin": "card", "locus": "Ā¶6", "quote": "Rather than seeking to automate epigraphyā€”the research field dealing with deciphering and understanding inscriptionsā€”he and his colleagues are interested in ā€œcrafting a tool that will integrate with the workflow of a historian,ā€ Assael said in a press briefing.", "claim": "Assael: not to automate epigraphy but a tool for the historian's workflow", "modality": "attributed", "kind": "function", "kernel": false, "const": false, "lineage": "a tool in the historian's workflow", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B7.txt" }, "F42": { "source": "B7", "origin": "card", "locus": "Ā¶8", "quote": "It doesnā€™t guess the meaning of texts, so it canā€™t interpret newly found engravings on its own, and itā€™s not clear yet how useful it will be to historiansā€™ workflows in the long term, according to Kathleen Coleman, a professor of classics at Harvard. The Monumentum Ancyranum is considered to be one of the best-known and most well-studied inscriptions in epigraphy, raising the question of how Aeneas will fare on more obscure samples.", "claim": "Coleman's limit: it does not guess meaning or interpret new finds alone, its long-term use unclear; the test case is among the best-studied inscriptions", "modality": "attributed", "kind": "limit", "kernel": true, "const": false, "lineage": "stated limits: data, bias, images, interpretation", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B7.txt" }, "F43": { "source": "B7", "origin": "card", "locus": "Ā¶7", "quote": "showed that Aeneas helped spur research ideas among the historians for 90% of inscriptions", "claim": "MIT Technology Review's restatement: research ideas for 90% of inscriptions", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the historians' study and its figures", "date": "2025-07-23", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B7.txt" }, "F44": { "source": "B6", "origin": "card", "locus": "page title", "quote": "Predicting the Past | Contextualising, restoring, and attributing ancient texts", "claim": "the interface's own title: contextualising, restoring and attributing ancient texts", "modality": "self-description", "kind": "definition", "kernel": true, "const": false, "lineage": "access: the interface, the code, the dataset, the licence", "date": "2026-10-31 (read; undated)", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B6.txt" }, "F45": { "source": "B6", "origin": "card", "locus": "card snippet (the saved page carries no body text)", "quote": "System: Aeneas (generative neural network by DeepMind, Durham University, Oxford, and partners) for contextualising, restoring, and attributing Latin inscriptio...", "claim": "the interface card names DeepMind, Durham University, Oxford and partners", "modality": "documented", "kind": "attribution", "kernel": false, "const": false, "lineage": "who made it: DeepMind, the partners, the authors", "date": "2026-10-31 (card snippet)", "quote_ok": true, "checked_against": "the card snippet (cite_list[5].snip, data/EA-WG-CAPTURES-01.json)" }, "F46": { "source": "B8", "origin": "wikipedia", "locus": "state (B8.txt)", "quote": "No article on the model.", "claim": "Wikipedia has no article on the model; no title variant resolves", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "reception: the encyclopedia", "date": "2026-10-31", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B8.txt" }, "F47": { "source": "B8", "origin": "wikipedia", "locus": "Google DeepMind article (B8-deepmind.wiki)", "quote": "Google has unveiled a new archaeology document program, named Ithaca after [[Homer's Ithaca|the Greek island]] in Homer's [[Odyssey]].", "claim": "the encyclopedia's Google DeepMind article carries the predecessor, Ithaca, and names no Aeneas", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "reception: the encyclopedia", "date": "2026-10-31", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B8-deepmind.wiki" }, "F48": { "source": "B8", "origin": "wikipedia", "locus": "Aeneas (disambiguation) (B8.wiki)", "quote": "'''[[Aeneas]]''' was a Trojan hero in Virgil's ''Aeneid'' and Homer's ''Iliad''.", "claim": "the name's disambiguation opens with the hero and lists no model", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "reception: the encyclopedia", "date": "2026-10-31", "quote_ok": true, "checked_against": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B8.wiki" } }, "field_sources": { "B1": "Google DeepMind ā€” Aeneas transforms how historians connect the past ā€” Google DeepMind", "B2": "blog.google ā€” Aeneas: Google DeepMind AI connects ancient inscriptions", "B3": "University of Oxford ā€” Aeneas: an AI model for restoring and contextualizing Latin inscriptions | Faculty of Classics", "B4": "GitHub ā€” GitHub - google-deepmind/predictingthepast Ā· GitHub", "B5": "Nature ā€” Contextualizing ancient texts with generative neural networks | Nature", "B6": "Predicting the Past ā€” Predicting the Past | Contextualising, restoring, and attributing ancient texts", "B7": "MIT Technology Review ā€” Google DeepMindā€™s new AI can help historians understand ancient Latin inscriptions | MIT Technology Review", "B8": "Wikipedia ā€” Aeneas (the model) (no article)" }, "delta": { "T_vs_LB": "At 'Aeneas DeepMind' (T1, seven cards: B1ā€“B7, B4 unreachable; 151 words, a closing offer of 27) the composition gives the model, its corpus, its tasks and its tested usefulness. Composed from its cards: the definition at the grain of the post and the paper ('contextualize, restore, date, and locate', B1 F1, B5 F24, B6 F44), the date (B1 F2, 'July 2025'); the corpus (B1 F11 'over 176,000'; B5 F28 'seventh century bce to the eighth century ce', as '700 BCE to 800 CE'); multimodal input and gaps of arbitrary length (B1 F10; B4 F23); the top-20 and the 13 years (B1 F13; B5 F26); the 90% and 44% (B5 F26); parallels as 'shared syntax or formulas' (B1 F3) and 'useful contextual parallels' (B1 F4); the interface (B6 F44); Ithaca in the closing offer (B1 F8, as comparison only). Available, not composed: the partners (B1 F6, B3 F21) and the co-leads (B1 F18), the body naming Google DeepMind alone; the stated reach to 'papyri to coinage', 'any ancient language and written medium' (B1 F5, B2 F19, B5 F33); free access and open code and dataset, the licences (B1 F7, B5 F36); Ithaca's upgrade to run on Aeneas (B1 F17); the embeddings and the definition of parallels (B1 F12, B5 F25, F30); the Res Gestae (B1 F14, B5 F32); the Mainz altar's reading of direct influence (B5 F31); the corpus's source databases (B1 F11, its first part). Limit dropped: the paper's own qualification of its figures, that 90% and 44% hold only when the model's predictions are added to its parallels, parallels alone giving 75% and 23% (B5 F27): T1-6 gives the 90% and 44% to the parallels; the stated limits of the corpus, its size and survival bias, weaker where data is thin (B5 F34, F35); Coleman's limit, that it does not guess meaning and was tested on a famous stone (B7 F42). Force raised: T1-3 attaches the image input to 'shared syntax or formulas'; the paper confines images to geographical attribution, restoration and dating text-only (B5 F29). Substituted reading: T1-2 sets '62 Roman provinces' as the corpus's span; in the field the 62 provinces are the attribution task's classes (B1 F13, B5 F29's head). Composed without attribution: none at the name (the maker named; the figures carried with markers to the cards). Outside the cards: none of T's claims lacks a card; 'typically dating ā€¦ of expert historian estimates' stands nearest B1 F13 'date ranges provided by historians'. Not surfaced, admitted by the operator's ruling: no own page stands outside the rail (the census's own pages are cards B1 and B6); Wikipedia has no article (B8 F46ā€“F48). Of the cards' 45 readable claims, T composes 13 at their grain or coarser.", "T_vs_LB_reading": "v0.9 Ā§0.5: the first diff is ontological and narrow. T and the field agree on the entity (DeepMind's model for Latin inscriptions). T composes the maker's model at the grain of its tasks and figures and leaves out what the field says about its reach (papyri, any written medium), its makers beyond DeepMind, its openness, and the paper's own qualification of the figures it carries, the 90% and 44% that belong to parallels with predictions, not to parallels.", "LB_vs_LBA": "Admission adds, in one claim space with the field's: the Revelation First record's account, field by field, with its basis (RF-01ā€“RF-30): parallels as the evidence the midrashim transform argues from, and what that evidence can and cannot settle, 'the direction of dependence between parallel texts' among what it cannot (RF-14, RF-15, RF-17, RF-28), nothing of the New Testament (RF-16, RF-29); the thesis's own falsifier on the direction of parallels (O202-01) and its work plan's common-substrate confound (O832-01), which set what a retrieved parallel would have to show to bear on priority, beside the paper's reading of a parallel as direct influence (F31, held under P_coh); the archive's philological notes keeping parallel apart from dependence (O828-01, O1458-01, O1474-01); of Ithaca, the model the field says it builds on, the gain located in interface design and the hazard of single-verdict interfaces (O1443-01, O1443-02), the dataset as the hidden achievement (O1443-03), and authors who cannot be consulted (O1443-04); the record's own state where the field is fuller (its count 'thousands', RF-11, against F11 and F28; its licence 'not read', RF-19, against F36); and the Capture Registry's account of the composition at the name (CAP-1-01ā€“CAP-1-15). Not surfaced: every archive source; no card cites the census record (CAP-1-14).", "alignment": { "T1_cards_claims": [ "F1", "F2", "F3", "F4", "F5", "F6", "F7", "F8", "F9", "F10", "F11", "F12", "F13", "F14", "F15", "F16", "F17", "F18", "F19", "F20", "F21", "F22", "F23", "F24", "F25", "F26", "F27", "F28", "F29", "F30", "F31", "F32", "F33", "F34", "F35", "F36", "F37", "F38", "F39", "F40", "F41", "F42", "F43", "F44", "F45" ], "T1_composed": [ "F1", "F24", "F2", "F11", "F28", "F10", "F13", "F26", "F4", "F44", "F23", "F8", "F3" ], "T1_composed_from_own_page": [], "T1_outside_cards": [], "rule": "a field claim counts as composed when T states it at its source's grain or coarser; the typology of what is not is in T_vs_LB" } }, "kernel": [ { "K": "K1", "claim": "`RF-01`", "source": "Revelation First record (RF)", "M_src": "stipulation", "sense": "what it is: a model for contextualizing ancient inscriptions", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K2", "claim": "`RF-21`", "source": "Revelation First record (RF)", "M_src": "stipulation", "sense": "what it is: a model for contextualizing ancient inscriptions", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K3", "claim": "`RF-30`", "source": "Revelation First record (RF)", "M_src": "documented", "sense": "what it is: a model for contextualizing ancient inscriptions", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K4", "claim": "`RF-20`", "source": "Revelation First record (RF)", "M_src": "documented", "sense": "what it is: a model for contextualizing ancient inscriptions", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K5", "claim": "`RF-03`", "source": "Revelation First record (RF)", "M_src": "documented", "sense": "who made it: DeepMind, the partners, the authors", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K6", "claim": "`RF-25`", "source": "Revelation First record (RF)", "M_src": "stipulation", "sense": "what the record asks and answers", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K7", "claim": "`RF-05`", "source": "Revelation First record (RF)", "M_src": "stipulation", "sense": "what the record asks and answers", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K8", "claim": "`RF-06`", "source": "Revelation First record (RF)", "M_src": "stipulation", "sense": "what the record asks and answers", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K9", "claim": "`RF-26`", "source": "Revelation First record (RF)", "M_src": "stipulation", "sense": "what the record asks and answers", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K10", "claim": "`RF-12`", "source": "Revelation First record (RF)", "M_src": "documented", "sense": "parallels: what they are and how the model retrieves them", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K11", "claim": "`RF-07`", "source": "Revelation First record (RF)", "M_src": "stipulation", "sense": "parallels: what they are and how the model retrieves them", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K12", "claim": "`RF-09`", "source": "Revelation First record (RF)", "M_src": "documented", "sense": "parallels: what they are and how the model retrieves them", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K13", "claim": "`RF-14`", "source": "Revelation First record (RF)", "M_src": "interpretation", "sense": "what the record asks and answers", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K14", "claim": "`RF-15`", "source": "Revelation First record (RF)", "M_src": "interpretation", "sense": "what the record asks and answers", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "qualified claim: F31 (a parallel read as direct influence)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K15", "claim": "`RF-17`", "source": "Revelation First record (RF)", "M_src": "interpretation", "sense": "what the record asks and answers", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "qualified claim: F31", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K16", "claim": "`O202-02`", "source": "#202", "M_src": "stipulation", "sense": "parallels and the direction of dependence", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K17", "claim": "`O202-01`", "source": "#202", "M_src": "stipulation", "sense": "parallels and the direction of dependence", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "the source's own falsifier (O202-01 is it): literary-structural parallels between Revelation and the later NT books shown to run the other direction (observation: a parallel with a demonstrated direction; locus: KO sentence 8)", "contrast": "qualified claim: F31", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K18", "claim": "`O832-01`", "source": "#832", "M_src": "stipulation", "sense": "parallels and the direction of dependence", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "the source's own verification condition (O832-01 is it): transforms shown to operate equally from Ezekiel/Daniel to both Revelation and Paul establish shared ancestry only (locus: KO sentence 8)", "contrast": "qualified claim: F31", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K19", "claim": "`O828-01`", "source": "#828", "M_src": "stipulation", "sense": "parallels and the direction of dependence", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "qualified claim: F31", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K20", "claim": "`O1458-01`", "source": "#1458", "M_src": "interpretation", "sense": "parallels and the direction of dependence", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "qualified claim: F31", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K21", "claim": "`O1474-01`", "source": "#1474", "M_src": "self-description", "sense": "parallels and the direction of dependence", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K22", "claim": "`RF-11`", "source": "Revelation First record (RF)", "M_src": "documented", "sense": "the corpus: the Latin Epigraphic Dataset", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "rival claim: F11, F28 (the field's counts)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K23", "claim": "`RF-10`", "source": "Revelation First record (RF)", "M_src": "documented", "sense": "the corpus: the Latin Epigraphic Dataset", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "qualified claim: F28 (the period the paper states)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K24", "claim": "`O1443-03`", "source": "#1443", "M_src": "interpretation", "sense": "the corpus: the Latin Epigraphic Dataset", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K25", "claim": "`O1443-01`", "source": "#1443", "M_src": "interpretation", "sense": "the machine reading interface: ranked hypotheses, not a verdict", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "qualified claim: F16, F27", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K26", "claim": "`O1443-02`", "source": "#1443", "M_src": "interpretation", "sense": "the machine reading interface: ranked hypotheses, not a verdict", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "qualified claim: F16", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K27", "claim": "`O1443-04`", "source": "#1443", "M_src": "interpretation", "sense": "stated limits: data, bias, images, interpretation", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K28", "claim": "`RF-08`", "source": "Revelation First record (RF)", "M_src": "documented", "sense": "the stated reach: other languages, scripts and media", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K29", "claim": "`RF-13`", "source": "Revelation First record (RF)", "M_src": "documented", "sense": "the stated reach: other languages, scripts and media", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K30", "claim": "`RF-22`", "source": "Revelation First record (RF)", "M_src": "stipulation", "sense": "the stated reach: other languages, scripts and media", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "qualified claim: F17 (Ithaca, Greek, upgraded to be powered by Aeneas)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K31", "claim": "`RF-16`", "source": "Revelation First record (RF)", "M_src": "interpretation", "sense": "what the record asks and answers", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K32", "claim": "`RF-27`", "source": "Revelation First record (RF)", "M_src": "interpretation", "sense": "what the record asks and answers", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K33", "claim": "`RF-28`", "source": "Revelation First record (RF)", "M_src": "interpretation", "sense": "what the record asks and answers", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K34", "claim": "`RF-29`", "source": "Revelation First record (RF)", "M_src": "interpretation", "sense": "what the record asks and answers", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K35", "claim": "`RF-23`", "source": "Revelation First record (RF)", "M_src": "documented", "sense": "lineage and name: Ithaca, the hero", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "qualified claim: F17", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K36", "claim": "`RF-04`", "source": "Revelation First record (RF)", "M_src": "documented", "sense": "access: the interface, the code, the dataset, the licence", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K37", "claim": "`RF-02`", "source": "Revelation First record (RF)", "M_src": "stipulation", "sense": "access: the interface, the code, the dataset, the licence", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K38", "claim": "`RF-18`", "source": "Revelation First record (RF)", "M_src": "documented", "sense": "access: the interface, the code, the dataset, the licence", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K39", "claim": "`RF-19`", "source": "Revelation First record (RF)", "M_src": "documented", "sense": "access: the interface, the code, the dataset, the licence", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "rival claim: F36 (the licences the paper states)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K40", "claim": "`RF-24`", "source": "Revelation First record (RF)", "M_src": "self-description", "sense": "what the record asks and answers", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "qualified claim: F24ā€“F38 (the paper, now read)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K41", "claim": "`CAP-1-14`", "source": "Capture Registry (CAP)", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K42", "claim": "`CAP-1-13`", "source": "Capture Registry (CAP)", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K43", "claim": "`CAP-1-01`", "source": "Capture Registry (CAP)", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K44", "claim": "`CAP-1-05`", "source": "Capture Registry (CAP)", "M_src": "interpretation", "sense": "the composition at the name", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K45", "claim": "`CAP-1-06`", "source": "Capture Registry (CAP)", "M_src": "interpretation", "sense": "the composition at the name", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K46", "claim": "`CAP-1-07`", "source": "Capture Registry (CAP)", "M_src": "interpretation", "sense": "the composition at the name", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K47", "claim": "`CAP-1-02`", "source": "Capture Registry (CAP)", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K48", "claim": "`CAP-1-09`", "source": "Capture Registry (CAP)", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K49", "claim": "`CAP-1-03`", "source": "Capture Registry (CAP)", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K50", "claim": "`CAP-1-10`", "source": "Capture Registry (CAP)", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K51", "claim": "`CAP-1-08`", "source": "Capture Registry (CAP)", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K52", "claim": "`CAP-1-12`", "source": "Capture Registry (CAP)", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K53", "claim": "`CAP-1-15`", "source": "Capture Registry (CAP)", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K54", "claim": "`CAP-1-04`", "source": "Capture Registry (CAP)", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "rival claim: F11 (B1, the saved post, states over 176,000)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" }, { "K": "K55", "claim": "`CAP-1-11`", "source": "Capture Registry (CAP)", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "ā€”", "f (source's own falsifiers)": "ā€” (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake Ā§8.3)", "contrast": "qualified claim: F26, F27 (the figures are card 5's)", "resolver": "a later capture at 'Aeneas DeepMind' (ADDR-34f10c8ba70e) on Google AI Overview, seated in the Capture Registry (Ā§8.3)" } ], "kernel_T": [ { "claim": "F5", "note": "the stated reach 'from papyri to coinage' (card 1; also B2 F19, B5 F33) absent; T keeps the model to Latin stone (CAP-1-02)" }, { "claim": "F27", "note": "the paper's qualification of the 90% and 44% (parallels with predictions; parallels alone 75% and 23%) absent; T gives both figures to the parallels" }, { "claim": "F6", "note": "the partners (Nottingham, Warwick, Oxford, AUEB) absent; the body names Google DeepMind alone" }, { "claim": "F7", "note": "free availability and open code and dataset absent (CAP-1-10)" }, { "claim": "F36", "note": "the licences (Apache 2.0, CC BY-SA 4.0) absent" }, { "claim": "F34", "note": "the paper's stated limits of the corpus absent" }, { "claim": "F42", "note": "Coleman's limit (no reading of meaning; a famous test case) absent" }, { "claim": "F29", "note": "images confined to geographical attribution; T attaches them to formulas" }, { "claim": "F17", "note": "Ithaca's upgrade to be powered by Aeneas absent; Ithaca only in the closing offer (CAP-1-09)" }, { "claim": "F31", "note": "the paper's reading of a parallel as direct influence absent" } ], "ledger": { "archive": "datasets/negative-of-the-negative/v2/ledgers/aeneas/ledger-archive.json", "field": "datasets/negative-of-the-negative/v2/ledgers/aeneas/field.json", "selection": "datasets/negative-of-the-negative/v2/traversal/aeneas/", "config": "datasets/negative-of-the-negative/v2/panel/configs/aeneas.json", "intake": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/", "audit": null }, "ontology": { "composition": "equal", "composition_note": "equal-terms case (v0.9 Ā§0.9, Ā§4.1): one claim space; the field's and the archive's claims ordered together by the entity's dependency and by genealogy within each sense; nothing translated; F31 against RF-15/RF-17 (with O202-01, O832-01), RF-11 against F11/F28, RF-19 against F36 held under P_coh", "case": { "value": "equal", "statement": "The archive reads the field's entity as the entity. The Revelation First census record describes Aeneas from its maker's post, each evidence quote with its basis ('operator'), and asks of it the census's question of what each resource holds on the composition of the New Testament writings and how machines read them; its one reading of its own, rf_intersection, takes the model's parallels as the kind of evidence the midrashim transform argues from and holds that retrieval does not settle the direction of dependence (RF-17), which the record also states as what the evidence cannot answer (RF-15). The capture record checks the composition against those pages. No deposit names the model (every D candidate is the hero). The deposits bear on the classes the field sources for it and add senses and limits on the field's own object: the thesis states the direction of parallels as its falsifier (#202) and requires that shared ancestry be excluded before a parallel counts as inheritance (#832); the philological notes keep parallel apart from dependence (#828, #1458, #1474); the field survey reads Ithaca, the model Aeneas builds on, as a gain from ranked hypotheses rather than a verdict, the hazard of machine reading located in single-verdict interfaces (#1443). None of these contests what the field says the model is, and none supplies an ontology the field lacks: Ļ = extends, the equal-terms case (Ā§0.9, Ā§1.2a). Where the paper reads a parallel as direct influence (F31) and the record holds that retrieval does not settle direction (RF-15, RF-17), both are composed under P_coh within the one ontology.", "loci": [ { "dep": 202, "locus": "the falsification conditions (line 33)", "quote": "The thesis is falsified if: (a) a document in the NT can be securely dated earlier than any plausible date for Revelation; (b) the literary-structural parallels between Revelation and the later NT books can be shown to run the other direction;" }, { "dep": 832, "locus": "WS-4 verification condition (line 307)", "quote": "to demonstrate direct inheritance rather than parallel derivation." }, { "dep": 828, "locus": "Ā§ III. Method (p17)", "quote": "It does not claim direct literary dependence between the five case studies." }, { "dep": 1443, "locus": "Ā§2.3 Ithaca (2022)", "quote": "Ithaca is empirical evidence that **the hazard is a property of single-verdict interfaces, not of machine reading as such**, and that ranked-alternative presentation is the mitigation." }, { "dep": "RF", "locus": "census.json rows[slug == 'aeneas'].rf_intersection", "quote": "Parallels as evidence: the model retrieves the kind of evidence the midrashim transform argues from (for example Rev 2:10 // Jas 1:12). Retrieval finds parallels; the direction of dependence, which the thesis argues, is not something it settles." }, { "dep": "RF", "locus": "census.json rows[slug == 'aeneas'].does_not_answer[0]", "quote": "the direction of dependence between parallel texts" }, { "dep": "capture-registry", "locus": "aeneas-deepmind-aio-20261031, analysis s1", "quote": "Aeneas is composed as a better Ithaca for Latin." } ] }, "E_F": { "type": "C (public entity: an AI model for contextualizing ancient inscriptions)", "definition": "the generative neural network Google DeepMind introduced in Nature on 23 July 2025 as the first model for contextualizing ancient inscriptions (DeepMind; Nature, Assael, Sommerschield et al.), co-developed with the University of Nottingham and researchers at Warwick, Oxford and AUEB, which retrieves textual and contextual parallels (texts sharing wording, syntax, formulas or provenance) by historically enriched embeddings from the Latin Epigraphic Dataset of 176,861 inscriptions, determines provenance from text and image, restores gaps of unknown length and dates within about 13 years; used by twenty-three historians as a source of starting points (90% and +44% with its predictions, 75% and +23% from parallels alone); named after the wandering hero and built upon Ithaca, which it now powers; trained for Latin and stated adaptable to papyri, manuscripts and coinage; limited by a small, survival-biased corpus; free as an interface, its code and weights open; with no Wikipedia article", "claims": [ "F1", "F24", "F2", "F6", "F3", "F25", "F30", "F28", "F10", "F13", "F15", "F26", "F27", "F37", "F8", "F17", "F19", "F33", "F34", "F7", "F36", "F46" ] }, "E_A": { "definition": "the same model, in the Revelation First census record and the archive's deposits: an AI model that retrieves textual and contextual parallels for ancient inscriptions, at the intersection of AI and ancient texts, whose unit is an inscription with its retrieved parallels and whose texts are Latin inscriptions, adaptation stated; evidence that can settle which inscriptions share wording, syntax, formulas or provenance and cannot settle the direction of dependence between parallel texts, nor anything of the New Testament; read by the archive as retrieving the kind of evidence the midrashim transform argues from, the direction of which the Revelation First thesis states as its falsifier (#202) and requires be shown against shared ancestry (#832); a parallel kept apart from dependence (#828, #1458, #1474); a model of the class whose gain, on Ithaca, lies in ranked hypotheses rather than a verdict (#1443); composed at its name as 'a better Ithaca for Latin', its reach to papyri left out (Capture Registry, 2026-10-31)", "claims": [ "RF-01", "RF-05", "RF-07", "RF-08", "RF-14", "RF-15", "RF-16", "RF-17", "O202-01", "O832-01", "O828-01", "O1443-01", "O1443-02", "CAP-1-06", "CAP-1-07" ], "defining_papers": [ 202, 832, 828, 1458, 1474, 1443 ], "defining_records": [ "Revelation First census record (census.json rows[slug == 'aeneas'], organizing_question, correction, intersections, schema_note; questions.json machine-dating; rendered aggregators/aeneas/index.html; editor Lee Sharks, compiled 2026-10-30, schema 'field census 1.0')", "Capture Registry aeneas-deepmind-aio-20261031 (OBS-ed654d93f896)" ] }, "rho": { "value": "extends", "status": "read from the Revelation First census record, the capture record and the deposits that bear on the entity's classes (Ā§1.2a); no deposit names the model", "statement": "The archive reads Aeneas as its maker and its paper present it and adds, on the field's own object: what its evidence, parallels, can and cannot settle, the direction of dependence among what it cannot; the Revelation First thesis's own test of direction and its confound of shared ancestry, which set what a retrieved parallel would have to show; the distinction of parallel from dependence in the archive's philology; for the class, the location of the machine's contribution in ranked hypotheses rather than a verdict; the record's answer, field by field, to the census's question; and the composition layer's omission, at its name, of the reach that would bring it to the papyri.", "claims": [ "RF-14", "RF-15", "RF-17", "O202-01", "O832-01", "O828-01", "O1458-01", "O1474-01", "O1443-01", "O1443-02", "CAP-1-07" ], "guard": "Ļ = extends claims an addition and no more (Ā§1.2a). It does not make the field's sources wrong at their dates. Where an archive claim stands against a field claim (RF-15 and RF-17, that retrieval does not settle direction, beside F31, the paper's parallel read as direct influence; RF-11's 'thousands' beside F11 and F28; RF-19's licence 'not read' beside F36; CAP-1-04's 'the saved sources say thousands' beside F11), both are composed under P_coh within the one ontology. The archive's class claims are claims on the classes the field sources (parallels: F3, F25; AI models for ancient texts: F1, F8), never on the entity by the composer's judgment.", "considered": [ { "value": "constructs", "rejected": "the archive does not read the model as made by reception; its records take the maker's post as its account" }, { "value": "revises", "rejected": "the entity is not the archive's own; the archive has no earlier statement of it to keep, narrow or withdraw" } ] }, "translation": [ { "lineage": "what it is: a model for contextualizing ancient inscriptions", "field_claims": [ "F1", "F2", "F24" ], "received": "F1 (B1, subtitle): \"Introducing the first model for contextualizing ancient inscriptions, designed to help historians better interpret, attribute and restore fragmentary texts.\"; F2 (B1, Ā¶4): \"Today, weā€™re publishing a paper in Nature introducing Aeneas , the first artificial intelligence (AI) model for contextualizing ancient inscriptions.\"; F24 (B5, Abstract s4ā€“6): \"However, current digital methods are restricted to literal matches and narrow historical scopes. Here we introduce Aeneas, a generative neural network for contextualizing ancient texts. Aeneas retrieves textual and contextual parallels, leverages visual inputs, handles arbitrary-length text restoration, and advances the state of the art in key tasks.\"", "receiving_text": "ā€”", "translation": "untranslated: composed as received beside RF-01, RF-20, RF-21, RF-30", "archive_claims": [ "RF-01", "RF-20", "RF-21", "RF-30" ] }, { "lineage": "who made it: DeepMind, the partners, the authors", "field_claims": [ "F6", "F18", "F21", "F45" ], "received": "F6 (B1, Ā¶8): \"We co-developed Aeneas with the University of Nottingham, and in partnership with researchers at the Universities of Warwick, Oxford and Athens University of Economics and Business (AUEB).\"; F18 (B1, Acknowledgements): \"The research was co-led by Yannis Assael and Thea Sommerschield.\"; F21 (B3, Ā¶1): \"A team of researchers, including Professor Jonathan Prag, an ancient historian in the Faculty of Classics, has just published a paper in Nature introducing Aeneas\"; F45 (B6, card snippet (the saved page carries no body text)): \"System: Aeneas (generative neural network by DeepMind, Durham University, Oxford, and partners) for contextualising, restoring, and attributing Latin inscriptio...\"", "receiving_text": "ā€”", "translation": "untranslated: composed as received beside RF-03", "archive_claims": [ "RF-03" ] }, { "lineage": "parallels: what they are and how the model retrieves them", "field_claims": [ "F3", "F4", "F9", "F12", "F20", "F25", "F30" ], "received": "F3 (B1, Ā¶5): \"historians traditionally rely on their expertise and specialized resources to identify ā€œparallelsā€ ā€” which are texts that share similarities in wording, syntax, standardized formulas or provenance.\"; F4 (B1, Ā¶6): \"It reasons across thousands of Latin inscriptions, retrieving textual and contextual parallels in seconds that allow historians to interpret and build upon the modelā€™s findings.\"; F9 (B1, Aeneasā€™ advanced capabilities, Parallels search): \"Parallels search: It searches for parallels across a vast collection of Latin inscriptions. By turning each text into a kind of historical fingerprint, Aeneas identifies deep connections that can help historians situate inscriptions within their broader historical context.\"; F12 (B1, How Aeneas works Ā¶4): \"For each inscription, Aeneasā€™ contextualization mechanism retrieves a list of parallels using a technique called ā€œembeddingsā€ ā€” encoding the textual and contextual information of each inscription into a kind of historical fingerprint\"; F20 (B2, Ā¶2): \"Aeneas can search for parallels (texts that share similarities in wording, syntax, standardized formulas or provenance) across thousands of Latin inscriptions, process multimodal input (text and images) and restore gaps in text, setting a new state-of-the-art benchmark in the field.\"; F25 (B5, Abstract s3): \"Historians capture these insights by identifying parallelsā€”inscriptions with shared phrasing, function or cultural settingā€”to enable the contextualization of texts within broader historical frameworks, and perform key tasks such as restoration and geographical or chronological attribution\"; F30 (B5, Results, contextualization Ā¶): \"Aeneas retrieves a list of the most relevant epigraphic parallels from the training set of LED. ā€¦ This design enables the model to surpass traditional fuzzy string-matching methods, to include a wealth of epigraphic parallels from relevant places and periods, related concepts, synonymous terms, formulaic variations and analogous epigraphic practices.\"", "receiving_text": "ā€”", "translation": "untranslated: composed as received beside RF-07, RF-09, RF-12", "archive_claims": [ "RF-07", "RF-09", "RF-12" ] }, { "lineage": "the tasks: multimodal provenance, restoration of unknown length, dating", "field_claims": [ "F10", "F13", "F29", "F39" ], "received": "F10 (B1, Aeneasā€™ advanced capabilities, Processing multimodal input; Restoring gaps): \"Aeneas is the first model to determine a text's geographical provenance using multimodal inputs. ā€¦ For the first time, Aeneas can restore gaps in texts where the missing length is unknown.\"; F13 (B1, State-of-the-art performance Ā¶2): \"Aeneas restores damaged inscriptions with a Top-20 accuracy of 73% in gaps of up to ten characters. This only decreases to 58% when the restoration length is unknown ā€¦ our model can attribute an inscription to one of 62 ancient Roman provinces with 72% accuracy. For dating, Aeneas places a text within 13 years of the date ranges provided by historians.\"; F29 (B5, Results, architecture Ā¶2): \"only the geographical attribution head incorporates the additional inputs from the vision networkā€”the restoration and chronological attribution tasks do not use the visual modality.\"; F39 (B7, Ā¶3): \"To do this, Aeneas takes in partial transcriptions of an inscription alongside a scanned image of it. Using these, it gives possible dates and places of origins for the engraving, along with potential fill-ins for any missing text.\"", "receiving_text": "ā€”", "translation": "untranslated: composed as received (no archive claim in this lineage)", "archive_claims": [] }, { "lineage": "the corpus: the Latin Epigraphic Dataset", "field_claims": [ "F11", "F28", "F40" ], "received": "F11 (B1, How Aeneas works Ā¶1ā€“2): \"drawing from decades of work by historians to create digital collections, especially the Epigraphic Database Roma (EDR), Epigraphic Database Heidelberg (EDH) and Epigraphic Database Clauss Slaby (EDCS-ELT). ā€¦ We cleaned, harmonized and linked these records into a single machine-actionable dataset that we refer to as the Latin Epigraphic Dataset (LED), comprising over 176,000 Latin inscriptions from across the ancient Roman world.\"; F28 (B5, Results, LED Ā¶): \"encompassing inscriptions from the seventh century bce to the eighth century ce ā€¦ The final LED comprises 176,861 inscriptions (totalling 16 million characters), most of which have damaged sections, and we were able to obtain corresponding images of 5% of inscriptions.\"; F40 (B7, Ā¶4): \"Aeneas also cross-references the text with a stored database of almost 150,000 inscriptions\"", "receiving_text": "ā€”", "translation": "untranslated: composed as received beside RF-10, RF-11, O1443-03", "archive_claims": [ "RF-10", "RF-11", "O1443-03" ] }, { "lineage": "the historians' study and its figures", "field_claims": [ "F15", "F26", "F27", "F43" ], "received": "F15 (B1, Advancing historical research collaboratively Ā¶1ā€“2): \"We invited twenty-three historians who regularly work with inscriptions to restore, date and place a set of texts using Aeneas. ā€¦ the most effective results were achieved when historians used Aeneasā€™ contextual information alongside its predictions\"; F26 (B5, Abstract s7ā€“9): \"The historians find the parallels retrieved by Aeneas to be useful research starting points in 90% of cases, improving their confidence in key tasks by 44%. Restoration and geographical attribution tasks yielded superior results when historians were paired with Aeneas, outperforming both humans and artificial intelligence alone. For dating, Aeneas achieved a 13-year distance from ground-truth ranges.\"; F27 (B5, Results, historian study Ā¶3): \"when provided with Aeneasā€™ parallels, historians reported that these could serve as a starting point for historical inquiry 75% of the time. This increased to 90% when Aeneasā€™ predictions for the three epigraphic tasks were also included. Moreover, Aeneasā€™ parallels boosted historiansā€™ confidence by an average of 23%, with an additional 21% increase when Aeneasā€™ predictions were also available.\"; F43 (B7, Ā¶7): \"showed that Aeneas helped spur research ideas among the historians for 90% of inscriptions\"", "receiving_text": "ā€”", "translation": "untranslated: composed as received (no archive claim in this lineage)", "archive_claims": [] }, { "lineage": "a tool in the historian's workflow", "field_claims": [ "F16", "F22", "F41" ], "received": "F16 (B1, Sharing the tools, shaping the future Ā¶1): \"Aeneas is designed to integrate within historians' existing research workflows. By combining expert knowledge with machine learning, it opens up a collaborative process, offering interpretable suggestions that serve as valuable starting points for historical inquiry.\"; F22 (B3, Ā¶7 (Prag)): \"In a field of this sort, the answers are not black and white, and the data is simultaneously rich and incomplete. Aeneas takes a leap beyond Ithaca to provide humanities researchers with a level of support ā€“ based upon previous scholarship - that we could only previously have dreamed of.\"; F41 (B7, Ā¶6): \"Rather than seeking to automate epigraphyā€”the research field dealing with deciphering and understanding inscriptionsā€”he and his colleagues are interested in ā€œcrafting a tool that will integrate with the workflow of a historian,ā€ Assael said in a press briefing.\"", "receiving_text": "ā€”", "translation": "untranslated: composed as received (no archive claim in this lineage)", "archive_claims": [] }, { "lineage": "the case studies: the Res Gestae and the Mainz altar", "field_claims": [ "F14", "F31", "F32" ], "received": "F14 (B1, A new lens on historical debates Ā¶2): \"Rather than predicting a single fixed date, Aeneas produced a detailed distribution of possible dates, showing two distinct peaks, with one smaller peak around 10-1 BCE and a larger, more confident peak between 10-20 CE. These results captured both prevailing dating hypotheses in a quantitative way.\"; F31 (B5, Results, Mainz altar Ā¶2): \"This altar shares rare textual formulas and an identical iconographical type with Cogitatusā€™ dedication, supporting the hypothesis that the later inscription was directly influenced by the earlier one.\"; F32 (B5, Results, RGDA Ā¶2): \"Aeneasā€™ chronological attribution of the full text of the RGDA reflects prevailing scholarly hypotheses, with a distribution exhibiting a strong bimodal pattern\"", "receiving_text": "ā€”", "translation": "untranslated: composed as received (no archive claim in this lineage)", "archive_claims": [] }, { "lineage": "stated limits: data, bias, images, interpretation", "field_claims": [ "F34", "F35", "F42" ], "received": "F34 (B5, Methods, LED limitations): \"the size of LED (16 million characters) remains a significant limitation compared with the scale of datasets typically used in state-of-the-art natural language processing research. ā€¦ the available corpus is also subject to inherent biases, most significantly inscription survival bias, potentially skewing the data towards certain materials, locations, or historical contexts.\"; F35 (B5, Methods, LED limitations Ā¶2): \"performance often tends to be weaker where data is limited\"; F42 (B7, Ā¶8): \"It doesnā€™t guess the meaning of texts, so it canā€™t interpret newly found engravings on its own, and itā€™s not clear yet how useful it will be to historiansā€™ workflows in the long term, according to Kathleen Coleman, a professor of classics at Harvard. The Monumentum Ancyranum is considered to be one of the best-known and most well-studied inscriptions in epigraphy, raising the question of how Aeneas will fare on more obscure samples.\"", "receiving_text": "ā€”", "translation": "untranslated: composed as received beside O1443-04", "archive_claims": [ "O1443-04" ] }, { "lineage": "the stated reach: other languages, scripts and media", "field_claims": [ "F5", "F19", "F33", "F38" ], "received": "F5 (B1, Ā¶7): \"Our model can also be adapted to other ancient languages, scripts and media, from papyri to coinage, expanding its capabilities to help draw connections across a wider range of historical evidence.\"; F19 (B2, Ā¶3): \"While trained for Latin, Aeneas can be adapted to other ancient languages, scripts and media, from papyri to coinage.\"; F33 (B5, Discussion Ā¶1): \"Aeneasā€™ architecture outperforms the previous state-of-the-art model, offers multimodal capabilities, enables restoration of text sequences of unknown lengths, and can also be adapted to any ancient language and written medium (such as papyri, manuscripts or coinage).\"; F38 (B5, Introduction Ā¶7): \"We therefore focus on Latin inscriptions, as this gap offers a valuable opportunity for interdisciplinary research with broader scholarly impact.\"", "receiving_text": "ā€”", "translation": "untranslated: composed as received beside RF-08, RF-13, RF-22", "archive_claims": [ "RF-08", "RF-13", "RF-22" ] }, { "lineage": "lineage and name: Ithaca, the hero", "field_claims": [ "F8", "F17", "F37" ], "received": "F8 (B1, Aeneasā€™ advanced capabilities Ā¶1): \"Named after the wandering hero of Graeco-Roman mythology, Aeneas builds upon Ithaca , our earlier work using AI to restore, date and place ancient Greek inscriptions.\"; F17 (B1, Sharing the tools, shaping the future Ā¶2): \"As part of todayā€™s release, weā€™re upgrading Ithaca , our ancient Greek model, to be powered by Aeneas and include the contextualization function, restorations of unknown length and better performance overall.\"; F37 (B5, Introduction Ā¶6): \"The name Aeneas is inspired by the wandering hero of Graeco-Roman mythology: like Aeneas, who journeyed from Troy across the Mediterranean seeking guidance on where to found the future city of Rome, our model seeks to uncover epigraphic parallels to ground historical research and link the past to the present.\"", "receiving_text": "ā€”", "translation": "untranslated: composed as received beside RF-23", "archive_claims": [ "RF-23" ] }, { "lineage": "access: the interface, the code, the dataset, the licence", "field_claims": [ "F7", "F23", "F36", "F44" ], "received": "F7 (B1, Ā¶9): \"weā€™re making an interactive version of Aeneas freely-available to researchers, students, educators, museum professionals and more at predictingthepast.com . To support further research, weā€™re also open-sourcing our code and dataset .\"; F23 (B4, card snippet (unreachable)): \"Model: Aeneas is a generative neural network developed by Google DeepMind for contextualising ancient texts, performing arbitrary-length text restoration, visua...\"; F36 (B5, Code availability): \"The Aeneas source code is available at https://github.com/google-deepmind/predictingthepast under Apache License 2.0, along with the trained weights, licensed under a Creative Commons Attribution-ShareAlike 4.0 International license (CC BY-SA 4.0).\"; F44 (B6, page title): \"Predicting the Past | Contextualising, restoring, and attributing ancient texts\"", "receiving_text": "ā€”", "translation": "untranslated: composed as received beside RF-02, RF-04, RF-18, RF-19", "archive_claims": [ "RF-02", "RF-04", "RF-18", "RF-19" ] }, { "lineage": "reception: the encyclopedia", "field_claims": [ "F46", "F47", "F48" ], "received": "F46 (B8, state (B8.txt)): \"No article on the model.\"; F47 (B8, Google DeepMind article (B8-deepmind.wiki)): \"Google has unveiled a new archaeology document program, named Ithaca after [[Homer's Ithaca|the Greek island]] in Homer's [[Odyssey]].\"; F48 (B8, Aeneas (disambiguation) (B8.wiki)): \"'''[[Aeneas]]''' was a Trojan hero in Virgil's ''Aeneid'' and Homer's ''Iliad''.\"", "receiving_text": "ā€”", "translation": "untranslated: composed as received (no archive claim in this lineage)", "archive_claims": [] } ], "receiving_texts": "equal-terms case: every field claim composed as received, beside the archive claims it stands with" }, "measure": { "rule": "v0.9 Ā§4.10: KO_B sentences surviving in the KO at similarity ā‰„ 0.95, with positions (computed as apply_v09.py does: difflib.SequenceMatcher ratio, the best KO sentence for each KO_B sentence)", "v0.9": { "surviving": 0, "of": 14, "positions": [ null, null, null, null, null, null, null, null, null, null, null, null, null, null ] }, "opening_run": { "rule": "v0.9 Ā§5.10 (equal-terms case): the field-only KO's sentences as the KO's opening run in their own order", "breach": false }, "blocks": { "rule": "v0.9 Ā§4.1 (equal-terms and mixed cases): the KO does not open with the field as a block or close with the archive as one; sentences typed by the ledgers of the claims they cite (F field-only, A archive-only, M both)", "sequence": "MMMAMMAAMFMFFAFMMAMMFAMM", "field_only_head": 0, "archive_only_tail": 0, "breach": false }, "max_similarity": [ 0.212, 0.345, 0.234, 0.354, 0.114, 0.386, 0.413, 0.368, 0.53, 0.393, 0.338, 0.374, 0.327, 0.565 ], "ko_sentence_1_max_similarity_to_KO_B": 0.212 }, "homonyms": { "names": [ "Aeneas (the Trojan hero of Virgil and Homer)", "Aeneas (biblical figure), the paralysed man healed by Peter in Acts 9" ], "basis": "The census row carries no disambiguation field; its capture relation notes that 'the address carries a disambiguator, the bare name being shared', and the operator added 'DeepMind' to the address (BRIEF-NEW-ROW-FIELD Ā§5). The bare name is a homonym: Wikipedia's disambiguation opens with the Trojan hero (B8 F48) and lists, among others, 'Aeneas (biblical figure), a paralyzed man cured by Saint Peter in the Acts of the Apostles' (B8.wiki), a homonym inside the New Testament the field is organized around; the model is named after the hero (B1 F8, B5 F37), which keeps the hero a distinct entity and its namesake, never a sense of the model's name. Every D candidate in the archive (21 sentences in 8 deposits) is the hero's. T composes the target and no homonym (CAP-1-13). 'Predicting the Past' and predictingthepast.com are the entity's interface and alias (census aka); the Latin Epigraphic Dataset is its corpus, a distinct object.", "distinct_entities": [ "Aeneas, the Trojan hero (Virgil's Aeneid, Homer's Iliad), the model's namesake (B1 F8, B5 F37); the archive's Aeneas (#288, #1102, #1271, #1289ā€“#1293, #1684)", "Aeneas of Acts 9:33ā€“34, the paralysed man healed by Peter (Wikipedia's disambiguation, B8.wiki)", "Ithaca, DeepMind's model for ancient Greek inscriptions (2022), which Aeneas builds upon and which was upgraded to be powered by Aeneas (B1 F8, F17); the census's neighbour (RF-23); its /non row, if one is built, is a distinct row", "the Latin Epigraphic Dataset (LED), the model's corpus, and its sources, the Epigraphic Database Roma, the Epigraphic Database Heidelberg and the Epigraphik-Datenbank Claussā€“Slaby (B1 F11, B5)", "Google DeepMind, the operator; the University of Nottingham, the co-developer; Warwick, Oxford and AUEB, partners; Durham University, on the interface card only (F45)", "the Res Gestae Divi Augusti and the Monumentum Ancyranum; the Mainz votive altars (CIL XIII, 6665), the paper's case studies", "the other Aeneases of the disambiguation page (Aeneas Tacticus, Aeneas of Gaza, Aeneas Silvius, places, ships, a telecom provider, a GNU package)" ], "substitutes": { "T": [] }, "register": { "address_1": { "slug": "aeneas-deepmind-aio-20261031", "obs_id": "OBS-ed654d93f896" } } }, "notes": [ "One address, Google AI Overview (expanded), signed out, incognito, received 2026-10-31 00:04 EDT: aeneas-deepmind-aio-20261031 (OBS-ed654d93f896, ADDR-34f10c8ba70e, AXN:0B30.CAPTURE.šŸ”“šŸ”„šŸ«µšŸŒ†šŸ’«šŸš€), seven cards, 151 words in the answer body and a 27-word closing offer. A NEW address; the census record was built the day before and no card cites it (CAP-1-14). The operator added the disambiguator 'DeepMind' to a bare name that is a homonym (the Trojan hero; Aeneas of Acts 9).", "Fetching: every reachable page saved verbatim with curl and the census agent 'alexanarch-seating/1.0 (https://www.alexanarch.org)' on 2026-10-31 between 04:38 and 04:41 UTC, raw and text with sha256 in field.json; text extracted by the census's own rule (fetch.py text_of). B1 (the DeepMind post) and B6 (predictingthepast.com) are the census's own pages aeneas--blog and aeneas--site and byte-identical to fresh fetches (sha256 c53e5cf5ā€¦ and 08a11cf8ā€¦), carried once as cards with the census's saved copies. Card URLs: B2, B3, B7 by WebSearch on title and site; B5 by DOI, confirmed by search and page title.", "Unreachable: B4, the GitHub repository google-deepmind/predictingthepast, refused by the build environment's proxy (HTTP 403, 'GitHub access to this repository is not enabled for this session'); not retried with another agent string or by another route (the GitHub connector was not used); its card's snippet stands (F23). B6, reachable, is a JavaScript application whose saved text is its title and a browser notice: the title is ledgered (F44) and the card's snippet carried for the attribution it alone states (F45, Durham). Wikipedia's API search returned HTTP 429 (rate limit) twice; the full-text search was made through index.php's Special:Search instead, which is the same site and the same agent.", "The card snippets: on this rail every snippet is written in a 'Model: ā€¦ Capabilities: ā€¦' / 'System: ā€¦ Function: ā€¦' grammar, and none of the five readable cards' snippets occurs in its page's text (snippet_in_page false for B1, B2, B3, B5, B7). The snippets are the composition layer's summaries of the cards, not excerpts; they are recorded with the cards in field.json and are T's, not the field's, except where a card's page cannot be read (B4) or carries no body text (B6), where the snippet stands as the card's only text by Ā§3.1(a).", "The archive subset: no archive text names the model (D and R candidates, 21 sentences in 8 deposits, are all the hero). The string pass returned 19 O-class candidates in 18 deposits; because the O-class pass returns a sentence only where it also carries a term of the script's category vocabulary, every archive sentence matching a class pattern was read; that reading admitted #202, #828, #832 (with #1211, #1216, #1217, #1400 as instances, S4), #1443 and #1474, and the hop (94 candidates) admitted #1458. #202, #832, #828 and #1443's Ā§2.3 were found only by the reading. The third class, Latin inscriptions, has no archive sentence: the archive's epigraphy is Greek (the Delphic decree, archaic inscribed objects), and those sentences were read and rejected. Defect of the string pass, reported, not fixed: four of six admitted deposits were missed by the O-class pass, as on the OLMoTrace, Data Provenance Explorer and Turkopticon rows.", "Defects in the Revelation First census record (reported, not fixed): (a) count.value 'thousands of Latin inscriptions' and coverage.period 'as its corpus (not stated in the saved post)' are contradicted by the record's own saved source: the DeepMind post the census saved states 'over 176,000 Latin inscriptions from across the ancient Roman world' (B1 F11), and the paper states 176,861 inscriptions from the seventh century BCE to the eighth century CE (B5 F28); 'thousands' is the post's Ā¶6 figure of what the model reasons across, not the corpus's count. (b) license 'not read': the paper states the code under Apache License 2.0 and the trained weights under CC BY-SA 4.0 (B5 F36), and the post says code and dataset are open-sourced (B1 F7). (c) unchecked 'the Nature paper (not saved)': the paper is open access and read whole here (B5). (d) relations[0] 'DeepMind's earlier model for Greek inscriptions' and negative_tags[1] 'NOT trained on Greek ā€¦ in the published version' hold, but the same post states Ithaca was upgraded 'to be powered by Aeneas' (B1 F17), which the record does not carry. (e) stated_limits records only the Latin scope; the paper's own limits (the corpus's size and survival bias, weaker performance where data is limited, B5 F34ā€“F35) are not recorded. (f) version_and_time.read and observed.on give 2026-10-30, while sources['aeneas--blog'].read gives 2026-10-31 and the manifest's fetch time is 2026-10-31T02:52:07Z. (g) The census row has no disambiguation field, though the name is a homonym (the record's capture relation notes the shared name). No field is contradicted on the entity's identity, operator, partners, kind, date or access.", "Observations on the capture (reported, not fixed): (a) findings[3] ('the saved sources say thousands') is wrong in part: the saved post (the census's own copy, B1) states 'over 176,000 Latin inscriptions', '62 ancient Roman provinces' and 'within 13 years' (F11, F13), so the corpus figure, the provinces and the 13 years are in the record's saved source; '700 BCE to 800 CE', the 90% and the 44% are on card 5, the paper (F28, F26), as the PER read note (CAP-1-11) has it. (b) findings[0] quotes 'useful contextual parallels 90% of the time' as held; the paper's results give 90% only with the model's predictions, 75% for parallels alone, and the 44% as 23% plus 21% with predictions (F27): the composition gives both figures to the parallels, a limit the capture does not record. (c) findings[2] holds: Durham is in no saved or fetched page (B1, B3 and B5 name Nottingham, Warwick, Oxford and AUEB), only on the interface card's snippet (F45). (d) The capture does not record T1-3's attachment of images to 'shared syntax or formulas', where the paper confines images to geographical attribution (F29), nor the paper's and Coleman's limits left out (F34, F42). findings[1], the reading and the analysis are borne out by the field.", "Neighbour rows: datasets/negative-of-the-negative/v2/rows/book-of-revelation.json composes #202's falsifier ('literary-structural parallels shown to run the other direction'), the claim this row ledgers as O202-01, and the midrashim transform; the bearing is the record's rf_intersection (parallels as the evidence the transform argues from; Rev 2:10 // Jas 1:12). datasets/negative-of-the-negative/v2/rows/sappho-31.json: no bearing found; the model is trained on Latin inscriptions, and the Oxyrhynchus papyri of that row are not in its corpus; the field's stated adaptation to papyri (F5, F33) is a stated capability, not a use. Neither row edited. The composition is on Google AI Overview, whose row is datasets/negative-of-the-negative/v2/rows/google-ai-overviews.json (not edited). Ithaca, the census's neighbour, has an intake folder of its own (intake/ithaca-20261031); its row is a separate row." ], "composition_v09": { "source": "this row (composed directly; no rebuild/negative-of-the-negative/v09/aeneas.json)", "composer": "Claude (Opus 5.5), subagent, 2026-10-31", "defining_material_read": [ { "what": "Revelation First census record, row", "file": "/home/claude/fleet/revelationfirst-com/data/rf-census/census.json (rows[slug == 'aeneas'], organizing_question, correction, intersections, schema_note)", "read": "whole" }, { "what": "Revelation First worked questions", "file": "/home/claude/fleet/revelationfirst-com/data/rf-census/questions.json", "read": "the question naming the entity (machine-dating) whole; the rest by id" }, { "what": "Revelation First census record, rendered", "file": "/home/claude/fleet/revelationfirst-com/aggregators/aeneas/index.html", "read": "text, for agreement with the row" }, { "what": "census sources", "file": "/home/claude/fleet/revelationfirst-com/data/rf-census/sources/aeneas--blog.*, aeneas--site.*, manifest.json", "read": "whole" }, { "what": "capture record", "file": "data/EA-WG-CAPTURES-01.json (aeneas-deepmind-aio-20261031)", "read": "transcript, transcript_raw, cite_list, d, reading, findings, analysis, q_kind, per_note, per_dim, originator, notes" }, { "what": "the Nature paper (card 5)", "file": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt", "read": "abstract, main text, discussion, methods on the dataset and its limits, data and code availability" }, { "dep": 202, "file": "data/texts/AXN-0349-text.md", "read": "the claims, Ā§ The Midrashim Transform and the falsification conditions (lines 19ā€“43)" }, { "dep": 832, "file": "data/texts/AXN-034D-text.md", "read": "the falsification conditions (line 111) and the WS-4 verification condition (line 307), with the passages on parallels (lines 83, 407)" }, { "dep": 828, "file": "data/texts/AXN-0345-text.md", "read": "Ā§ III. Method (p16ā€“17)" }, { "dep": 1443, "file": "data/texts/AXN-05CC-text.md", "read": "Ā§2.3 Ithaca, Ā§2.4 Vesuvius Challenge (lines 170ā€“200) and the enabling condition (lines 600ā€“615)" }, { "dep": 1458, "file": "data/texts/AXN-05E4-text.md", "read": "the derivation and its bound (lines 476ā€“505)" }, { "dep": 1474, "file": "data/texts/AXN-05F4-text.md", "read": "Status of this section (p136)" }, { "dep": 1406, "file": "data/texts/AXN-058F-text.md", "read": "the abstract and principles (lines 31ā€“150), for the hop verdict" }, { "what": "neighbour row", "file": "datasets/negative-of-the-negative/v2/rows/book-of-revelation.json", "read": "its uses of #202's falsifier and the midrashim transform" }, { "what": "the specification", "file": "data/texts/AXN-0700-text.md", "read": "Ā§0ā€“Ā§11 whole" }, { "what": "the briefs", "file": "rebuild/negative-of-the-negative/v09/BRIEF-NEW-ROW-FIELD.md, BRIEF-NEW-ROW-AGGREGATOR.md, BRIEF-NEW-ROW-LABOR.md, BRIEF-NEW-ROW-PROVENANCE.md, BRIEF.md (Compose)", "read": "whole" } ], "cuts": [], "notes": "equal case; KO 24 sentences, 48 of 48 field and 55 of 55 archive claims cited; P body 347 words; KO_B 14 sentences, 0 surviving at ā‰„ 0.95" }, "field_by_origin": { "cards": [ "B1", "B2", "B3", "B4", "B5", "B6", "B7" ], "own_pages": [], "wikipedia": [ "B8" ], "census_third_party": [], "unreachable": [ "B4" ], "title_only": [], "snippet_only": [ "B4" ], "own_pages_note": "the census's two own pages (aeneas--blog, aeneas--site) are cards B1 and B6, byte-identical, carried once as cards", "wikipedia_state": "no article on the model: raw wikitext HTTP 404 for Aeneas_(AI), Aeneas_(model), Aeneas_(artificial_intelligence), Aeneas_(AI_model), Aeneas_(neural_network) (and Ithaca_(AI)); the disambiguation page 'Aeneas (disambiguation)' (HTTP 200) opens with the Trojan hero and lists Aeneas of Acts 9 and no model; 'Aeneas' is the hero's article; the Google DeepMind article (HTTP 200) carries Ithaca and names no Aeneas; the API search was rate-limited (HTTP 429) and index.php full-text searches return no article on the model (2026-10-31)", "sources": [ { "origin": "card", "role": "Overview card 1", "card": "Google DeepMind ā€” Aeneas transforms how historians connect the past ā€” Google DeepMind", "url": "https://deepmind.google/discover/blog/aeneas-transforms-how-historians-connect-the-past/", "kind": "Overview card 1 (the operator's announcement post, 'The Aeneas team', 23 July 2025); the census's own page aeneas--blog, byte-identical", "id": "B1", "files": { "B1.html": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.html", "sha256": "c53e5cf5b5336f473ea7c230ac43cc27ed2ddaaaf8d0701d68cb6646ca91b094" }, "B1.txt": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt", "sha256": "5713717f5d2cb7379e201fb8671be73c337e42196c55e7f015c661cd3a4064e8" } }, "fetched": { "at": "2026-10-31T04:38Z (check fetch, minute recorded; the census copy carried was fetched 2026-10-31T02:52:07Z)", "http": 200, "url": "https://deepmind.google/discover/blog/aeneas-transforms-how-historians-connect-the-past/", "final_url": "https://deepmind.google/blog/aeneas-transforms-how-historians-connect-the-past/", "raw": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.html", "raw_sha256": "c53e5cf5b5336f473ea7c230ac43cc27ed2ddaaaf8d0701d68cb6646ca91b094", "text": [ "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B1.txt" ], "census_copy": { "src": "aeneas--blog", "sha256": "c53e5cf5b5336f473ea7c230ac43cc27ed2ddaaaf8d0701d68cb6646ca91b094", "read": "2026-10-31", "http": "200", "fetched": "2026-10-31T02:52:07Z", "agrees": true }, "note": "check fetch with the census agent on 2026-10-31 (not kept) returned the same bytes as the census copy; the census copy is carried (gunzipped) with the census's text" }, "names_operator": "yes ('Google DeepMind', 'The Aeneas team'; the partners)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, revelationfirst, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "card_title": "Aeneas transforms how historians connect the past ā€” Google DeepMind", "card_snippet": "Model: Aeneas, a multimodal generative neural network trained on the Latin Epigraphic Dataset (LED) of over 176,000 inscriptions. Capabilities: Searches context...", "card_note": null, "surfaced_by_T": [ { "slug": "aeneas-deepmind-aio-20261031", "card": 1 } ], "snippet_in_page": false, "claims": [ "F1", "F2", "F3", "F4", "F5", "F6", "F7", "F8", "F9", "F10", "F11", "F12", "F13", "F14", "F15", "F16", "F17", "F18" ] }, { "origin": "card", "role": "Overview card 2", "card": "blog.google ā€” Aeneas: Google DeepMind AI connects ancient inscriptions", "url": "https://blog.google/technology/google-deepmind/aeneas/", "kind": "Overview card 2 (Google's company blog, 'Contributing writer', Jul 23, 2025)", "id": "B2", "files": { "B2.html": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B2.html", "sha256": "bfea5da464bccc1585004e617dd4b331f74506632b321123a2012c54751a6aec" }, "B2.txt": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B2.txt", "sha256": "572d5f6f1fcb7cdac6c9e155214a014cbef1ff94b0d8734d9184fa4d42c778ff" } }, "fetched": { "at": "2026-10-31T04:39:17Z", "http": 200, "url": "https://blog.google/technology/google-deepmind/aeneas/", "final_url": "https://blog.google/innovation-and-ai/models-and-research/google-deepmind/aeneas/", "raw": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B2.html", "raw_sha256": "bfea5da464bccc1585004e617dd4b331f74506632b321123a2012c54751a6aec", "text": [ "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B2.txt" ], "note": "URL found by WebSearch on title and site; redirected to the blog's current path" }, "names_operator": "yes ('Google DeepMind')", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, revelationfirst, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "card_title": "Aeneas: Google DeepMind AI connects ancient inscriptions", "card_snippet": "Model: Aeneas by Google DeepMind (introduced July 23, 2025; published in Nature). Purpose: First AI model for contextualizing, interpreting, attributing, and re...", "card_note": null, "surfaced_by_T": [ { "slug": "aeneas-deepmind-aio-20261031", "card": 2 } ], "snippet_in_page": false, "claims": [ "F19", "F20" ] }, { "origin": "card", "role": "Overview card 3", "card": "University of Oxford ā€” Aeneas: an AI model for restoring and contextualizing Latin inscriptions | Faculty of Classics", "url": "https://www.classics.ox.ac.uk/node/4577076", "kind": "Overview card 3 (Oxford Faculty of Classics news item, 28 July 2025)", "id": "B3", "files": { "B3.html": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B3.html", "sha256": "772454851fa204cac0ff2d994219b32a0cab7ebe3b3e1d84dbe33cbea03800e0" }, "B3.txt": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B3.txt", "sha256": "be3f6946a38ae4975c2a9ea37da2f98bde65de8b39ddbe2f14a1d9bfe79158da" } }, "fetched": { "at": "2026-10-31T04:39:18Z", "http": 200, "url": "https://www.classics.ox.ac.uk/node/4577076", "final_url": "https://www.classics.ox.ac.uk/node/4577076", "raw": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B3.html", "raw_sha256": "772454851fa204cac0ff2d994219b32a0cab7ebe3b3e1d84dbe33cbea03800e0", "text": [ "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B3.txt" ], "note": "URL found by WebSearch on title and site (classics.ox.ac.uk/node/4577076)" }, "names_operator": "yes ('Google DeepMind which co-developed Aeneas with the University of Nottingham')", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, revelationfirst, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "card_title": "Aeneas: an AI model for restoring and contextualizing Latin inscriptions | Faculty of Classics", "card_snippet": "Model: Aeneas, the first AI model for contextualizing, interpreting, attributing, and restoring ancient, fragmentary, weathered, or defaced Latin inscriptions. ...", "card_note": null, "surfaced_by_T": [ { "slug": "aeneas-deepmind-aio-20261031", "card": 3 } ], "snippet_in_page": false, "claims": [ "F21", "F22" ] }, { "origin": "card", "role": "Overview card 4", "card": "GitHub ā€” GitHub - google-deepmind/predictingthepast Ā· GitHub", "url": "https://github.com/google-deepmind/predictingthepast", "kind": "Overview card 4 (the code repository; unreachable from the build environment)", "id": "B4", "files": { "B4-403.html": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B4-403.html", "sha256": "b3e03aaaff81d68730d67392a135ac3fdfdf022880c66632ab188d2fe084cda3" } }, "fetched": { "unreachable": true, "tries": [ { "at": "2026-10-31T04:38Z", "url": "https://github.com/google-deepmind/predictingthepast", "http": 403, "sha256": "b3e03aaaff81d68730d67392a135ac3fdfdf022880c66632ab188d2fe084cda3", "result": "the build environment's proxy refused the host ('GitHub access to this repository is not enabled for this session'), a connection-level refusal recorded as the brief's environment allowlist case; not retried with another agent string or by another route (the GitHub connector was not used)" } ], "note": "unreachable; the card's title and snippet stand as its only text. The repository is named by B3 and B5 (code and dataset), which are read." }, "names_operator": "yes, on the card ('developed by Google DeepMind')", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, revelationfirst, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "card_title": "GitHub - google-deepmind/predictingthepast Ā· GitHub", "card_snippet": "Model: Aeneas is a generative neural network developed by Google DeepMind for contextualising ancient texts, performing arbitrary-length text restoration, visua...", "card_note": null, "surfaced_by_T": [ { "slug": "aeneas-deepmind-aio-20261031", "card": 4 } ], "claims": [ "F23" ] }, { "origin": "card", "role": "Overview card 5", "card": "Nature ā€” Contextualizing ancient texts with generative neural networks | Nature", "url": "https://www.nature.com/articles/s41586-025-09292-5", "kind": "Overview card 5 (the paper: Assael, Sommerschield et al., Nature 645, 141ā€“147, open access, published 23 July 2025)", "id": "B5", "files": { "B5.html": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.html", "sha256": "07c431681107fad57f6710ff010c6052c2931c8dea94e1e04237ea1de9a37e28" }, "B5.txt": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt", "sha256": "37e5e9410948c084df8d27d5e328401d7b6fd9fcd0f46d5222924d2637261d55" } }, "fetched": { "at": "2026-10-31T04:39:20Z", "http": 200, "url": "https://www.nature.com/articles/s41586-025-09292-5", "final_url": "https://www.nature.com/articles/s41586-025-09292-5?error=cookies_not_supported&code=ba7e1928-81fc-4085-abbb-c294e295113b", "raw": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.html", "raw_sha256": "07c431681107fad57f6710ff010c6052c2931c8dea94e1e04237ea1de9a37e28", "text": [ "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B5.txt" ], "note": "URL by DOI 10.1038/s41586-025-09292-5, confirmed by WebSearch (ideas.repec.org) and by the page title; open access, the full text in the page" }, "names_operator": "yes (the authors and affiliations; Google DeepMind)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, revelationfirst, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "card_title": "Contextualizing ancient texts with generative neural networks | Nature", "card_snippet": "Model: Aeneas, a multimodal generative neural network for contextualizing Latin inscriptions using image and text inputs. Corpus: Latin Epigraphic Dataset (LED)", "card_note": null, "surfaced_by_T": [ { "slug": "aeneas-deepmind-aio-20261031", "card": 5 } ], "snippet_in_page": false, "claims": [ "F24", "F25", "F26", "F27", "F28", "F29", "F30", "F31", "F32", "F33", "F34", "F35", "F36", "F37", "F38" ] }, { "origin": "card", "role": "Overview card 6", "card": "Predicting the Past ā€” Predicting the Past | Contextualising, restoring, and attributing ancient texts", "url": "https://predictingthepast.com/", "kind": "Overview card 6 (the interface, a JavaScript application; the census's own page aeneas--site, byte-identical)", "id": "B6", "files": { "B6.html": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B6.html", "sha256": "08a11cf87711fe870d18b83f8adc34ec85b625bad14b3b646dab7052d6d393a9" }, "B6.txt": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B6.txt", "sha256": "06b41db902580a817c82f9842d473ba88f1965be698259d506dedf7a0c88e8f7" } }, "fetched": { "at": "2026-10-31T04:38Z (check fetch, minute recorded; the census copy carried was fetched 2026-10-31T02:52:08Z)", "http": 200, "url": "https://predictingthepast.com/", "final_url": "https://predictingthepast.com/", "raw": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B6.html", "raw_sha256": "08a11cf87711fe870d18b83f8adc34ec85b625bad14b3b646dab7052d6d393a9", "text": [ "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B6.txt" ], "census_copy": { "src": "aeneas--site", "sha256": "08a11cf87711fe870d18b83f8adc34ec85b625bad14b3b646dab7052d6d393a9", "read": "2026-10-31", "http": "200", "fetched": "2026-10-31T02:52:08Z", "agrees": true }, "note": "check fetch with the census agent on 2026-10-31 (not kept) returned the same bytes as the census copy; the census copy is carried (gunzipped) with the census's text; the saved page is a JavaScript application whose text is its title and a browser notice, so the card's snippet (F45) is carried as the card's text for what the page does not hold" }, "names_operator": "on the card only ('DeepMind, Durham University, Oxford, and partners'); the saved page names none", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, revelationfirst, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "card_title": "Predicting the Past | Contextualising, restoring, and attributing ancient texts", "card_snippet": "System: Aeneas (generative neural network by DeepMind, Durham University, Oxford, and partners) for contextualising, restoring, and attributing Latin inscriptio...", "card_note": null, "surfaced_by_T": [ { "slug": "aeneas-deepmind-aio-20261031", "card": 6 } ], "snippet_in_page": false, "claims": [ "F44", "F45" ] }, { "origin": "card", "role": "Overview card 7", "card": "MIT Technology Review ā€” Google DeepMindā€™s new AI can help historians understand ancient Latin inscriptions | MIT Technology Review", "url": "https://www.technologyreview.com/2025/07/23/1120574/deepmind-ai-aeneas-helps-historians-interpret-latin-inscriptions/", "kind": "Overview card 7 (news article by Peter Hall, July 23, 2025)", "id": "B7", "files": { "B7.html": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B7.html", "sha256": "1526eaa5bd4dd61d272ebf0661bce2105616a9b89f668c9b65114c23c9c3f46d" }, "B7.txt": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B7.txt", "sha256": "80b89dcc484dee1cad08015312d1f3542e1950ea258363ac9988f10576957b64" } }, "fetched": { "at": "2026-10-31T04:39:21Z", "http": 200, "url": "https://www.technologyreview.com/2025/07/23/1120574/deepmind-ai-aeneas-helps-historians-interpret-latin-inscriptions/", "final_url": "https://www.technologyreview.com/2025/07/23/1120574/deepmind-ai-aeneas-helps-historians-interpret-latin-inscriptions/", "raw": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B7.html", "raw_sha256": "1526eaa5bd4dd61d272ebf0661bce2105616a9b89f668c9b65114c23c9c3f46d", "text": [ "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B7.txt" ], "note": "URL found by WebSearch on title and site" }, "names_operator": "yes ('Google DeepMind'; Assael, Sommerschield)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, revelationfirst, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "card_title": "Google DeepMindā€™s new AI can help historians understand ancient Latin inscriptions | MIT Technology Review", "card_snippet": "Software: Google DeepMind Aeneas open-source AI for ancient Latin inscriptions. Function: Analyzes partial text and scans to predict origin dates, locations, an...", "card_note": null, "surfaced_by_T": [ { "slug": "aeneas-deepmind-aio-20261031", "card": 7 } ], "snippet_in_page": false, "claims": [ "F39", "F40", "F41", "F42", "F43" ] }, { "origin": "wikipedia", "role": "Wikipedia", "card": "Wikipedia ā€” Aeneas (the model) (no article)", "url": "https://en.wikipedia.org/w/index.php?title=Aeneas_(AI)&action=raw", "kind": "Wikipedia: no article on the model; the disambiguation page, the hero's article and Google DeepMind's article read", "id": "B8", "files": { "B8-deepmind.wiki": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B8-deepmind.wiki", "sha256": "133497d3b16d94d33e372b86f442d6cc9cf5299a15e996b1182603807f9abb0a" }, "B8-hero.wiki": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B8-hero.wiki", "sha256": "393eb34757c8d1256a87cf9df294b4a8db89d306c91647f47e86b4bab143b226" }, "B8-search1.html": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B8-search1.html", "sha256": "593b96487250aa984747d080675b260c66459b29d038d5e60c03cf16aa2ca072" }, "B8-search2.html": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B8-search2.html", "sha256": "3795d3f51e5420dbc18945bde881336a3f5d1300feca009434cc37b322bb8e34" }, "B8-search3.html": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B8-search3.html", "sha256": "af89dc90c58f56805a098896928eb0b9ab47f8a56dbd09badab7ddad316b7de4" }, "B8.txt": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B8.txt", "sha256": "dab1678fd4253e4196372399aa38f8f4af30c68d2130fa556511010bdfa86b6a" }, "B8.wiki": { "path": "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B8.wiki", "sha256": "aaf2e5c37a5e9b09ac8e18c873ce0c7b11e3cf75822747e5d17d071fa87c5911" } }, "fetched": { "at": "2026-10-31T04:39Zā€“04:41:11Z", "tries": [ { "url": "https://en.wikipedia.org/w/index.php?title=Aeneas_(AI)&action=raw (and _(model), _(artificial_intelligence), _(AI_model), _(neural_network))", "http": 404 }, { "url": "https://en.wikipedia.org/w/index.php?title=Ithaca_(AI)&action=raw", "http": 404 }, { "url": "https://en.wikipedia.org/w/api.php?action=query&list=search ā€¦", "http": 429, "result": "rate limited twice (04:39Z, 04:40:44Z); not retried by the API; full-text search made through index.php Special:Search instead" }, { "url": "https://en.wikipedia.org/w/index.php?title=Aeneas_(disambiguation)&action=raw", "http": 200, "sha256": "aaf2e5c37a5e9b09ac8e18c873ce0c7b11e3cf75822747e5d17d071fa87c5911" }, { "url": "https://en.wikipedia.org/w/index.php?title=Aeneas&action=raw", "http": 200, "sha256": "393eb34757c8d1256a87cf9df294b4a8db89d306c91647f47e86b4bab143b226" }, { "url": "https://en.wikipedia.org/w/index.php?title=Google_DeepMind&action=raw", "http": 200, "sha256": "133497d3b16d94d33e372b86f442d6cc9cf5299a15e996b1182603807f9abb0a" }, { "url": "Special:Search (3 queries, article space)", "http": 200, "files": [ "B8-search1.html", "B8-search2.html", "B8-search3.html" ] } ], "text": [ "datasets/negative-of-the-negative/v2/intake/aeneas-20261031/B8.txt" ], "note": "no article on the model; B8.txt summarises the responses (codes, notices); quotes from the raw wikitext are checked against the .wiki files" }, "names_operator": "no (the encyclopedia has no article; its Google DeepMind article names Ithaca, not Aeneas)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, revelationfirst, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "surfaced_by_T": [], "claims": [ "F46", "F47", "F48" ] } ] } }