
Sri Lankan Historical Knowledge Graph
"A provenance-aware digital research infrastructure for exploring Sri Lanka's historical memory."
Connecting ancient chronicles, rock inscriptions, archaeological records, oral traditions, historical figures, locations, and contested events into an interconnected research environment.
Sri Lankan historical knowledge exists across Pali and Sinhala chronicles, thousands of Brahmi rock inscriptions, archaeological surveys, scholarly monographs, oral traditions, and localized folk memory.
Historically, these materials exist in silos. Traditional database models force flat, reconciled values (e.g., declaring a single absolute reign duration or character judgment), erasing the critical distinction between competing textual narratives.
When narratives diverge—such as conflicting chronicle perspectives on the character of King Elara or unverified battlefield locations—traditional systems either discard the divergence or present an arbitrary conclusion as definitive fact.
Lankavamsa is designed to connect these disparate materials while strictly preserving their source provenance, confidence ratings, and inherent historical uncertainties without unilateral reconciliation.
A multi-domain knowledge graph modeling history as an interconnected web of primary texts, physical evidence, oral memory, entities, and reified assertions.
Primary historical texts including the Mahāvamsa, Dīpavamsa, Cūlavamsa, Rājavaliya, and Pūjāvaliya.
Brahmi rock edicts, cave donor inscriptions, slab records, and copper plates cataloging tangible epigraphical testimony.
Material artifacts, structural stratigraphy, carbon-dating chronologies, and verified archaeological site excavations.
Localized folk memory, regional legends, ballads, and folk traditions preserved in cultural communities across Sri Lanka.
Monarchs, military commanders, religious elders, emissaries, lineages, and their recognized historical aliases.
Ancient kingdoms, settlements, fortresses, monasteries, battlefields, river crossings, and geographic coordinates.
Battles, military marches, coronations, treaties, diplomatic missions, and political transitions across historical eras.
Reified relational assertions with explicit citation of chapters, folios, confidence ratings, and review audits.
Lankavamsa prevents the conflation of raw historical testimony with modern scholarly interpretation.
What a historical work, lithic inscription, archaeological report, or oral recording contains in its original context.
A specific passage, folio quotation, photographic record, translated fragment, or physical measurement.
A structured assertion connecting a Subject entity to an Object entity (e.g. Dutugemunu DEFEATED Elara) attributed directly to evidence.
An explanation developed by synthesizing multiple pieces of corroborating or conflicting evidence across different accounts.
A tentative historical explanation or proposed reconstruction requiring further epigraphic and archaeological inquiry.
A historical claim in Lankavamsa is never an immutable truth; it is a structured proposition grounded in traceable evidence. When accounts disagree, both claims persist side-by-side with full bibliographic attribution.
The systematic workflow for exploring historical questions and tracing claims across the knowledge graph.
Formulate a research inquiry regarding a historical event, figure's lineage, geographical movement, or contested reign.
Locate relevant primary texts, stone inscriptions, archaeological surveys, and oral traditions referencing the entities.
Construct ontological graph relationships between figures, places, campaigns, and traditions with rule-governed validation.
Juxtapose corroborating, conflicting, and independent testimony across distinct textual and epigraphical traditions.
Audit any relational claim back to its exact primary passage, folio translation, chapter, or inscription location.
Develop and evaluate historiographical interpretations while explicitly preserving nuance and dissenting accounts.
The current V1 provides the foundational knowledge environment for constructing, auditing, and exploring the historical knowledge graph.
Biographical entities with alias indices, historical era bounds, and attached primary source citations.
Fortresses, monasteries, ancient settlements, and rivers linked to geographic locations and chronicle events.
Military battles, political treaties, and coronations cross-referenced with chronological approximations.
Folklore and regional legends preserved with distinct evidence tags separate from formal chronicle texts.
Comprehensive bibliography of primary works, epigraphic corpora, and scholarly publications with bidirectional citation indices.
Reified relational assertions categorized by evidence type, confidence rating, and audit status (Verified vs Disputed).
Global search across figures, places, events, and textual assertions with category filters and instant suggestions.
Multi-hop graph path exploration stepping through connected entities, claims, and supporting sources with persistent breadcrumb history.
2nd Century BC Historical Stress Test for Provenance-Aware Knowledge Graph Modeling
The Mahāvamsa, Dīpavamsa, and Rājavaliya provide varying degrees of detail regarding King Elara's reign in Anuradhapura and King Dutugemunu's mobilization from Rohana.
Correlating southern rock inscriptions and archaeological remains with chronicle claims of garrison strongholds (Mahiyangana, Kasatota, Vijithapura) along the Mahaweli river.
Chroniclers praised King Elara's legendary judicial fairness while simultaneously depicting his rule as foreign sovereignty—modeling both dimensions as attributable claims without distortion.
Explore entities, places, and asserted claims related to the Dutugemunu–Elara campaign inside the active V1 workspace.
A phased research program expanding from the foundational graph ontology to document-grounded evidence pipelines and computational historical investigation.
Core data models, reified claims, entity CRUD registries, Neo4j graph storage, and relationship constraints.
Direct processing of historical PDF texts into page-aware evidence passages, lithic inscription transcriptions, and passage-level graph linking.
Unified lexical and vector passage retrieval across Pali transliterations, Classical Sinhala texts, and English scholarly translations.
Automated visual mapping of historiographical consensus vs. isolated chronicle claims and archaeological refutations.