[ Armenian Epigraphic Corpus ]
A digital framework for reading, encoding, and preserving Armenian inscriptions — before the stones that carry them are lost.
Ա · About
ArmEpiC is a long-term scholarly infrastructure for the systematic digital documentation, encoding, and interoperable publication of Armenian inscriptions. It combines epigraphic rigour with current digital humanities standards to build a preservation-quality, linked-open-data-ready corpus of Armenian inscribed heritage.
Initiated at the EPFL Digital Humanities Laboratory, ArmEpiC operates through a federated model of regional corpora — each with its own scholarly scope — unified under a shared semantic and encoding framework.
Every inscription carries full uncertainty notation, palaeographic detail, and editorial transparency, so a later scholar can disagree with one specific reading rather than with the whole edition. Many of these monuments are endangered; the encoded record is built to outlive access to the stone.
ՅԻՇԱՏԱԿ Է ԱՐՁԱՆՍ
ՏԷՐ ՍՏԵՓԱՆՈՍԻ
ՅԻՇԱՏԱԿ Է ԱՐՁԱ[Ն]Ս
ՏԷՐ ՍՏԵՓԱՆՈՍԻ +++++
ԹՎ(ԻՆ) ՈՂԳ
<div type="edition" xml:lang="xcl">
<ab>
<lb n="1"/>ՅԻՇԱՏԱԿ Է ԱՐՁԱ<supplied
reason="lost">Ն</supplied>Ս
<lb n="2"/>ՏԷՐ <persName ref="#stepanos"
>ՍՏԵՓԱՆՈՍԻ</persName>
<gap reason="illegible" quantity="5"
unit="character"/>
<lb n="3"/><expan><abbr>ԹՎ</abbr><ex>ԻՆ</ex>
</expan> <num value="693">ՈՂԳ</num>
</ab>
</div>“This stele is a memorial to Lord Step‘anos +++++, in the year 693.”
Armenian letter-numerals: Ո 600 + Ղ 90 + Գ 3 = 693 of the Armenian Era, which begins in 551 CE — so 1244 CE.
Բ · Corpora
ArmEpiC is a federated ecosystem: each regional corpus has its own scholarly scope, editorial team, and authority files, while sharing the common TEI/EpiDoc encoding schema, SKOS vocabulary framework, and URN identifier namespace.
The first regional corpus within ArmEpiC, covering Armenian inscriptions from the historic region of Artsakh (Nagorno-Karabakh). It documents inscriptions on monasteries, churches, khachkars, and civil structures, from the early medieval period through the 19th century — Gandzasar, Dadivank, and Amaras among them.
A record of Armenian inscriptions in Madras (Chennai), Kolkata, Hyderabad, and Mumbai — the epigraphic trace of the Armenian merchant diaspora in India from the seventeenth to the early twentieth century. Funded by the British Library Endangered Archives Programme under EAP1721, the project produced approximately 300 photographic records with full descriptive metadata.
The schema, the vocabularies, and the publication pipeline are already public and reusable. If you hold Armenian epigraphic material — a monastery, a cemetery, a parish, a museum collection — this niche is yours.
Propose a corpusԳ · Method
Five stages, each producing something a later scholar can check. The stages are numbered the way the inscriptions themselves are dated — with Armenian letter-numerals.
Record the stone where it stands.
Measured survey, condition notes, squeezes where the surface will take them, and a standard photographic set. Every record carries its own provenance: who read it, when, and from what.
Make the worn legible.
Reflectance Transformation Imaging and photogrammetry capture what a single photograph flattens. Raking light recovers strokes that have weathered below the threshold of ordinary vision.
Encode the reading, not just the text.
Editorial decisions — restorations, expansions, uncertain letters, lost lines — are marked in the file itself. A later scholar can then disagree with one specific choice rather than with the whole edition.
Name things once.
People, places, monuments, materials, and script types resolve against Wikidata, VIAF, GeoNames, and the Getty AAT — with a published Armenian vocabulary wherever no external authority covers the term.
Publish so it survives.
Each release is archived to Zenodo with a DOI, mirrored on GitHub, and rendered as a browsable site through GitHub Pages. The XML is the record; the website is only a view of it.
Դ · Data model
Persistent URN identifiers, CIDOC CRM event-based modelling, and SKOS vocabulary alignment. Every entity — inscription, monument, place, person, material, technique — is addressable through a stable URN.
urn:armepic:artsakh:ins:ART0001
Other entity types — mon monument · plc place · per person · mat material · bib bibliography
<?xml-model href="https://epidoc.stoa.org/schema/9.7/tei-epidoc.rng"?>
<TEI xmlns="http://www.tei-c.org/ns/1.0">
<objectType
ref="urn:armepic:objecttype:OBJ0006"
xml:lang="en">lintel</objectType>
<material
ref="urn:armepic:material:MAT0001"
xml:lang="en">tuff</material>
<rs type="technique"
ref="urn:armepic:technique:TEC001">
carved</rs>
<origDate
calendar="urn:armepic:calendar:armenian"
datingMethod="urn:armepic:calendar:gregorian"
when-custom="1271"
evidence="internal-date">
Հայոց 720 (1271)
</origDate>
<handNote
scriptRef="urn:armepic:script:SCR002">
<term xml:lang="en">Bolorgits Erkatagir</term>
<term xml:lang="hy">Բոլորգիծ երկաթագիր</term>
</handNote>
<!-- Honorific abbreviation -->
<expan
ana="urn:armepic:abbrtype:honorific">
<abbr>Սբ</abbr>
<ex>Սուրբ</ex>
</expan>
</TEI>Ե · Vocabularies
Every entity class is governed by a concept scheme with persistent URN identifiers, bilingual Armenian and English labels, scope notes, and hierarchical relations — aligned to EAGLE, the Getty AAT, and CIDOC CRM where those exist.
skos:ConceptScheme
skos:hasTopConcept Inscription type
skos:prefLabel “donation”@en · “նվիրատվական”@hy
skos:exactMatch EAGLE:typeOfInscription
Զ · Partners
Research host · Lausanne
Endangered Archives Programme · EAP1721
Academic partner · Yerevan
National Academy of Sciences · Yerevan
Civil-society partner
Three-dimensional documentation
Federated discovery and mapping
Է · Publications
Datasets and schema releases are cited the same way articles are — every release gets a DOI. The entries below are placeholders, pending first release.