cuneiform
Enable an AI agent to recognise cuneiform writing, assess the evidence and uncertainty behind its identification and interpretation, and choose justified next steps for documenting or reading it.
Research draft, second pass
A second pass drafted this model: the structure a model of this thing needs, and what is known about it in the world. The line under this one says how the second half was obtained - researched against sources, or recalled without web access, in which case nothing here was read anywhere and every claim is a lead to verify. Unreviewed either way.
Researched by: Codex + Grok
Purpose and description
Enable an AI agent to recognise cuneiform writing, assess the evidence and uncertainty behind its identification and interpretation, and choose justified next steps for documenting or reading it.
Cuneiform is a family of writing systems whose signs are impressed as wedges in clay (or carved in stone and other media) and that originated in southern Mesopotamia in the late fourth millennium BCE for Sumerian, later adapted to Akkadian and other languages of the ancient Near East.
It can be Identify candidate cuneiform passages and request targeted examination of diagnostically useful wedges.; Compare sign forms against explicitly identified sign lists and securely attributed parallels.; Produce evidence-linked sign identifications and transliterations while preserving alternative readings.; Assess whether language identification or translation is sufficiently supported to proceed.; Align corresponding passages across witnesses without erasing graphic or textual differences.; Prepare a digital representation that distinguishes surviving marks, uncertain readings and editorial restorations..
Distinguishing features
Test whether apparent wedges form recurrent, organised sign configurations rather than isolated tool impressions or decorative patterns.
Test whether sign shapes and their arrangement fit an identifiable cuneiform tradition; wedge-like appearance alone is insufficient.
Distinguish the writing system from the language: recognising cuneiform must not automatically assign a Sumerian, Akkadian or other linguistic reading.
Distinguish logosyllabic, alphabetic and other proposed sign-system assignments through inventory and use, rather than treating all cuneiform as one encoding system.
Distinguish an inscription's visible marks from a modern transliteration, normalised font rendering or reconstructed reading, each of which represents different evidence.
Scope
+ Recognition of cuneiform writing and discrimination from incidental marks, decoration and imitations
+ Sign forms, variants and inventories within an identified cuneiform tradition
+ Writing direction, sign order, line organisation and reading sequence
+ Relationships between signs, linguistic values, transliterations and interpretations
+ Legibility, competing readings and evidential limits of a particular inscription
+ Compatibility and uncertainty when comparing cuneiform witnesses or digital representations
- The carrier object's excavation, ownership, conservation and physical treatment
- Complete grammars and lexicons of languages written in cuneiform
- Historical reconstruction of the people, institutions and events mentioned in texts
- Collection management, acquisition and legal title to artefacts
- General photography, scanning and digital file preservation
- Independent models of other writing systems or modern typefaces
Characteristics
- Cuneiform identification status
- unassessed | candidate | supported | disputed | rejected, with evidence Prevents wedge-like marks from being accepted as writing without checking their configuration and context.
- Cuneiform tradition
- Named tradition or competing assignments, with period and regional qualifications where supported Constrains which sign inventories, variants and reading conventions are relevant.
- Sign-system organisation
- logosyllabic | alphabetic | other specified | unresolved Determines what kinds of sign-to-language mapping an agent may attempt.
- Language association
- Text or passage linked to one or more candidate languages, with supporting evidence Separates graphic recognition from language identification and permits multilingual or unresolved passages.
- Graphic execution
- impressed | incised | painted or drawn | digitally rendered | other | unresolved Helps distinguish meaningful sign-form differences from differences caused by production or representation.
- Reading order
- Supported or alternative sequences linking signs, lines, columns and inscribed surfaces A valid sign identification can still produce a wrong reading if textual order is mistaken.
- Sign legibility
- clear | partial | ambiguous | lost | inaccessible in available evidence, assessed by location Separates damage to the text from uncertainty caused by inadequate images or examination.
- Readable sign proportion
- Percentage of assessable sign positions under a stated counting method; unknown when positions cannot be established Supports a qualified assessment of how much of a passage can be read without treating supplied restorations as surviving signs.
- Interpretive support
- observed | proposed | independently corroborated | contested, separately for sign identity, value and passage interpretation Prevents confidence in a visible sign from being transferred automatically to its pronunciation or meaning.
Also called
Where this came from
wikidata · CC0 1.0
Drafted structure
Bundle to layer to finding to question, as the second pass will find it: 6 bundles · 11 layers · 18 findings · 28 questions.
Cuneiform recognition Establish whether the marks constitute cuneiform and which sign tradition could account for them.
An agent must identify the writing before applying sign lists or linguistic interpretations.
Wedge and sign evidence
Examine the configurations that make marks candidates for cuneiform signs.
Organised sign configurations
Record the visible basis for grouping marks into signs, including alternatives involving damage, decoration or imitation.
- Which combinations of wedge heads, strokes, orientations and spacing support segmentation into cuneiform signs? definition
- What evidence distinguishes these configurations from accidental impressions, decoration or a modern imitation? boundary
Tradition assignment
Constrain identification through compatible sign inventories and writing conventions.
Candidate sign traditions
Record candidate traditions and the diagnostic evidence for each without deriving language solely from appearance.
- Which sign forms and conventions support each proposed cuneiform tradition, and which contradict it? definition
- Which examined sign lists or attributed inscriptions supply the comparisons, and how closely do they match the proposed period and region? provenance
Sign forms and identities Connect observed wedge configurations to sign identities while retaining variation and ambiguity.
Cuneiform reading depends on distinguishing actual graphic evidence from normalised sign identification.
Graphic segmentation
Establish where signs begin and end and which marks belong to them.
Sign boundaries
Record alternative segmentations where crowding, overlapping impressions or damage obscure sign boundaries.
- Which visible marks belong to each proposed sign, and where could the same marks support a different segmentation? boundary
- Which wedge orientations, relative positions or spacing measurements would discriminate between those segmentations? measurement
Variants and normalisation
Relate an observed form to candidate sign identities and tradition-specific variants.
Form-to-sign mapping
Preserve the observed form alongside sign-list identifiers and any uncertainty introduced by normalisation.
- Which sign identities remain compatible with the surviving wedges, using which named sign list and edition? provenance
- Would normalising this form erase a distinction relevant to this tradition, scribal practice or reading? boundary
Textual order and functions Establish reading sequence and distinguish linguistic signs from numerical, metrological or organising functions.
Recognised signs cannot be interpreted reliably without their order and local textual roles.
Inscription layout
Represent how the inscription's spatial arrangement supports a reading sequence.
Supported reading sequence
Record sign, line, column and surface order, retaining unresolved transitions.
- What supports the proposed orientation and sequence of signs, lines, columns and inscribed surfaces? definition
- Where do rulings, changes of orientation or discontinuities require separate reading sequences or alternative continuations? boundary
Sign function in context
Determine what role a sign plays in its passage before assigning a linguistic value.
Functional classification
Record whether a sign is being treated as phonetic, logographic, determinative, numerical, metrological or otherwise functional, where those distinctions apply.
- Which contextual evidence supports the proposed function of this sign in the identified tradition? definition
- For a proposed numerical or metrological expression, which notation and unit conventions must be established before calculating a quantity? action
Readings and interpretation Connect sign identities to language-dependent readings and passage interpretations through explicit evidence.
A sign's identification, transliteration and interpretation are distinct judgments that can have different levels of support.
Language and sign values
Assess candidate languages and the values signs may have within them.
Contextual reading candidates
Retain competing values where graphic, grammatical or lexical evidence does not select a single reading.
- What linguistic evidence supports the language assignment independently of the script identification? definition
- Which candidate sign values fit the local spelling and grammatical context, and which alternatives remain unresolved? boundary
Transliteration and restoration
Keep visible text, editorial representation and inferred missing text distinguishable.
Evidence-linked transliteration
Connect each transliterated or restored element to its location, editorial convention and justification.
- Which transliteration elements represent visible signs, uncertain identifications, supplied signs or unresolved gaps, under which conventions? provenance
- What examined parallel or linguistic constraint supports each restoration, and is that support sufficient to use it in a translation? action
Witnesses and reading readiness Assess whether available representations support a reading and how that reading relates to other witnesses.
Images, hand copies, editions and digital encodings can omit or alter distinctions needed to evaluate cuneiform.
Observation quality
Distinguish missing inscriptional evidence from limitations of the available representation.
Legibility and next observation
Locate ambiguous wedges and identify the observation most likely to resolve them.
- For each uncertain sign, does the limitation arise from surface loss, an obscured impression, image lighting, viewing angle or conflicting copies? measurement
- Which targeted image, three-dimensional view or specialist collation could resolve the specific competing readings? action
Witness and encoding correspondence
Control comparisons between inscriptions, editions and digital representations.
Preserved and lost distinctions
Record whether aligned passages and encoded signs preserve the distinctions required by the intended use.
- Are compared records separate ancient witnesses, reproductions of one inscription or derivative editions, and what establishes that relationship? provenance
- Which graphic variants, uncertain signs or editorial distinctions would be lost by the proposed character encoding or normalised transcription? boundary
Evidence and external alignment What the world already says about this thing, gathered so the model can be checked against it.
A model that cannot be lined up against existing standards, identifiers and practice cannot be adopted by anyone who already uses them.
Reported evidence
Findings from the breadth pass, kept separate from the structural claims.
Kinds and varieties
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- pictographic-proto-cuneiform (Uruk IV-III / Jemdet Nasr administrative tablets)
- Sumero-Akkadian logo-syllabic cuneiform (the main Mesopotamian tradition)
- Hittite (and other Anatolian) cuneiform
- Elamite cuneiform
- Hurrian and Urartian cuneiform
- Ugaritic alphabetic cuneiform
- Old Persian semi-alphabetic cuneiform
- cuneiform numerals, metrological signs, and paleographic period-styles (ED, OA/OB, NA/NB, etc.)
- Which of these kinds and varieties hold for the sense of cuneiform this model covers, and on what evidence? provenance
Identifiers and schemes
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- Wikidata - Q401 (cuneiform) - Item for the writing system; related items include Q269159 (cuneiform script / Sumero-Akkadian), Q724230 (Old Persian cuneiform), Q822656 (Ugaritic alphabet).
- ISO 15924 - Xsux (440) Sumero-Akkadian Cuneiform; Xpeo (030) Old Persian; Ugar (900) Ugaritic - Three separate script codes; do not collapse Old Persian or Ugaritic into Xsux.
- ISO 15924 / Unicode alias - Xsux → Unicode script property Cuneiform; blocks U+12000-U+123FF, U+12400-U+1247F, U+12480-U+1254F - Ugaritic is U+10380-U+1039F; Old Persian is U+103A0-U+103DF.
- CDLI P-number - P plus 6 digits (e.g. P000001) - Artifact-level identifier in the Cuneiform Digital Library Initiative; the usual public key for a tablet or inscribed object.
- Museum / collection accession - collection prefix + number (e.g. BM 92687, VAT 12553, AO 19938, K. 4375) - British Museum (BM, K.), Vorderasiatisches Museum (VAT), Louvre (AO), etc.; still the legal and catalogue identity of physical objects.
- Which of these identifiers and schemes hold for the sense of cuneiform this model covers, and on what evidence? provenance
Standards and regulation
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- ISO 15924 (ISO / Unicode Consortium as registration authority): script codes Xsux, Xpeo, Ugar.
- The Unicode Standard (Unicode Consortium): encoding of Sumero-Akkadian cuneiform, Ugaritic, and Old Persian, including names, code charts, and rendering notes.
- UNESCO 1970 Convention on the Means of Prohibiting and Preventing the Illicit Import, Export and Transfer of Ownership of Cultural Property: export/import of cuneiform tablets as cultural property.
- National antiquities laws of Iraq, Syria, Turkey, Iran and neighbouring states (e.g. Iraq Law No. 55 of 2002 for the Antiquities and Heritage of Iraq): excavation, ownership, and export of inscribed objects.
- Museum documentation standards as applied to tablets (CIDOC CRM / Spectrum practice in holding institutions), plus community identifiers from CDLI and ORACC rather than a single legal 'cuneiform standard'.
- Which of these standards and regulation hold for the sense of cuneiform this model covers, and on what evidence? provenance
Real-world use
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- Primary historical use: administrative accounts, legal contracts, royal inscriptions, letters, lexical lists, omens, literature, and scholarly texts on clay tablets, envelopes, bricks, cylinders, stelae, and seals in Mesopotamia, Elam, Anatolia, the Levant, and (for Old Persian) Achaemenid Iran, c. 3400 BCE-1st century CE.
- Modern scholarly use: reading, edition, and lemmatization in Assyriology, Hittitology, and related fields; teaching sign lists (e.g. Labat, Borger MZL, Huehnergard).
- Museum and library cataloguing of tablets and inscribed objects; exhibition labels typically give period, language, provenience, and museum number.
- Digital corpora (CDLI photographs and catalogue; ORACC annotated texts; British Museum, Louvre, Yale, Penn, Vorderasiatisches Museum collection databases).
- Forensic and heritage practice: identification of looted or unprovenanced tablets; sign-by-sign authentication versus modern fakes.
- Rare contemporary revival/art: calligraphic or typographic use of Unicode cuneiform fonts (e.g. Noto Sans Cuneiform, CuneiformComposite) - not a living administrative script.
- Which of these real-world use hold for the sense of cuneiform this model covers, and on what evidence? provenance
Typical measurements
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- typical clay tablet face size (hand tablet) - roughly 30-80 mm on the short side, 40-120 mm on the long side for common administrative and letter tablets; royal and literary tablets and prisms much larger - mm
- wedge (sign) height on a typical tablet - about 2-6 - mm
- time span of use of the writing system - c. 3400 BCE to 1st century CE (last astronomical tablets) - year (historical)
- Unicode assigned Sumero-Akkadian cuneiform characters (BMP-supplementary blocks) - on the order of 900-1100 encoded signs across the three blocks, far fewer than all attested paleographic variants - code point
- CDLI catalogue scale - on the order of 300000-400000 registered cuneiform artifacts (growing) - artifact
- Which of these typical measurements hold for the sense of cuneiform this model covers, and on what evidence? provenance
Failure modes and hazards
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- Clay tablets are brittle: breakage, surface flaking, and loss of wedges destroy readings; unbaked tablets are especially vulnerable to water.
- Salt efflorescence and improper conservation (over-firing, consolidants, humidity cycling) can erase or obscure signs.
- Modern forgeries and composite 'Frankenstein' tablets mislead the market and the scholarly record.
- Illicit excavation and trafficking (especially post-1990s Iraq and Syria) strip provenience, which is often the only dating and archival context.
- Paleographic and homophonic ambiguity: one sign has many values (polyphony); different signs can write the same syllable (homophony) - misreading is a research hazard, not a physical one.
- Digital encoding does not capture ductus, damage, or un-encoded variants; naive font rendering is not an edition.
- Handling without support or with skin oils; display lighting and vibration in exhibitions.
- Which of these failure modes and hazards hold for the sense of cuneiform this model covers, and on what evidence? provenance
Regional variation
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- Southern Mesopotamia (Sumer/Babylonia): the core logo-syllabic tradition from proto-cuneiform through Neo-Babylonian/Late Babylonian; Sumerian then Akkadian as the main languages.
- Northern Mesopotamia / Assyria: Assyrian dialect orthography and ductus (Old, Middle, Neo-Assyrian hands) distinct from Babylonian in sign forms and some values.
- Anatolia (Hattusa and related sites): Hittite, Luwian (cuneiform, not hieroglyphic), Palaic, Hurrian written in a Mesopotamian-derived syllabary with local conventions.
- Elam (Susa and highlands): Linear Elamite is a different script; Elamite cuneiform is an adaptation of Mesopotamian cuneiform for Elamite.
- Northern Levant / Ugarit: a local alphabetic cuneiform of ~30 signs, not the logo-syllabic system.
- Urartu (eastern Anatolia / Armenia): Urartian cuneiform derived from Neo-Assyrian.
- Achaemenid Iran: Old Persian cuneiform is a largely independent semi-alphabet used on royal inscriptions (Behistun, Persepolis), alongside Elamite and Babylonian versions of the same texts.
- Modern cataloguing: English 'cuneiform', German Keilschrift, French cunéiforme, Arabic al-mismārī / al-kitāba al-mismāriyya; museum numbering schemes differ by country.
- Which of these regional variation hold for the sense of cuneiform this model covers, and on what evidence? provenance
Neighbouring kinds and how to tell them apart
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- Linear Elamite (and proto-Elamite) - Also an early Iranian-plateau script on clay, but signs are linear incised/drawn, not wedge-impressed; not mutually readable with Sumero-Akkadian cuneiform. Test: sign inventory and wedge vs. linear ductus.
- Anatolian (Luwian) hieroglyphs - Pictorial carved signs used for Luwian in the same region as Hittite cuneiform; not impressed wedges. Test: medium and sign forms (stone reliefs/seals vs. clay wedges).
- Egyptian hieroglyphs / hieratic - Pictorial and ink-brush scripts of the Nile; no wedge impression. Occasional co-occurrence in the Amarna correspondence is Akkadian cuneiform on clay vs. Egyptian on other media.
- Ugaritic alphabet vs. logo-syllabic cuneiform - Both use wedges on clay, but Ugaritic is a small abjad (~30 letters) with a different Unicode block (Ugar) and sign list. Test: inventory size and ISO 15924 code (Ugar vs Xsux).
- Old Persian cuneiform vs. Mesopotamian cuneiform - Old Persian is a separate, largely alphabetic invention (Xpeo) with simple signs and a closed inventory; not a late form of Akkadian writing. Test: script code, sign count, and language of the royal Achaemenid inscriptions.
- Cypriot syllabary / Linear A / Linear B - Aegean linear syllabaries incised on clay; visually linear, historically unrelated. Test: sign shapes (lines vs. wedges) and archaeological horizon.
- Modern wedge-like display fonts or 'cuneiform' decorative type - Typographic pastiche without a documented mapping to Unicode Xsux or to a period sign list. Test: whether signs correspond to a published sign list (MZL, Labat) and to encoded code points.
- Which of these neighbouring kinds and how to tell them apart hold for the sense of cuneiform this model covers, and on what evidence? provenance
Sources
- The Unicode Standard, Core Specification - Cuneiform, Cuneiform Numbers and Punctuation, Early Dynastic Cuneiform - Official encoding of Sumero-Akkadian cuneiform as three Unicode blocks (U+12000-U+123FF, U+12400-U+1247F, U+12480-U+1254F) and the distinction of that script from Ugaritic and Old Persian.
- ISO 15924 - Codes for the representation of names of scripts - Script identifiers Xsux (Sumero-Akkadian Cuneiform), Xpeo (Old Persian), Ugar (Ugaritic) as the standard names under which cuneiform varieties are catalogued in IT and bibliographic systems.
- Cuneiform Digital Library Initiative (CDLI) - The principal public catalogue of cuneiform artifacts (P-numbers), periodization, provenience, and the practical distinction of tablet vs. other object types used by museums and Assyriologists.
- Open Richly Annotated Cuneiform Corpus (ORACC) - Language- and project-level corpora (Sumerian, Akkadian, Hittite, etc.), lemmatization practice, and how the writing system is used as a research object rather than a single 'font'.
What the second pass must settle
- Does the registry intend cuneiform to encompass all traditions conventionally described by that term, including alphabetic traditions, or a narrower family?
- Which sign lists and editorial conventions should be accepted for each included tradition, and how should incompatible identifiers be reconciled?
- What evidence threshold should distinguish supported cuneiform identification from a merely wedge-like candidate or imitation?
- How should tradition-specific numerical and metrological conventions be divided between this model and neighbouring quantity or notation models?
- Which digital representations preserve the graphic distinctions needed for the intended agent tasks, and when must an image or three-dimensional witness remain essential?