Mirror of the PerseusDL canonical TEI-XML corpus. Every work is addressable by canonical URN (urn:cts:greekLit:tlg0012.tlg001 = Homer's Iliad). Where translations exist, they sit in the same directory as the source and are addressable by matching URN. This is the aligned substrate for the L1/L2 transform pipeline.
Every work in Perseus has a canonical URN of the form urn:cts:{corpus}:{author-id}.{work-id}.{edition-id}. The Iliad, Greek edition, is urn:cts:greekLit:tlg0012.tlg001.perseus-grc2; Murray's English translation is perseus-eng3; Butler's is perseus-eng4. The prefixes tlg (Thesaurus Linguae Graecae) and phi (Packard Humanities Institute) are standard classicist identifier schemes; every author has one. Lines and sections within a work are addressable via the same URN system: urn:cts:greekLit:tlg0012.tlg001.perseus-grc2:1.1 is line 1 of Book 1.
The TEI-XML bytes themselves live in data-rhizome/datasets/corpora/perseus/snapshots/2026-07-06/, checksummed against the GitHub tarball SHAs recorded in the snapshot's MANIFEST.json. Individual work files are directly accessible at their canonical filesystem path under the corpus's data/{author-id}/{work-id}/ directory. This landing page and its indexes are the browseable surface; the substrate is preserved separately in the sovereign data-rhizome.
github.com/PerseusDL/canonical-greekLit · canonical-latinLit · canonical-farsiLit
License: CC-BY-SA at the collection level; individual work licenses embedded in TEI-XML headers.
This dataset is one of six published as a mutually-referencing set (set.json) — any instance encountering one member should treat the six as a whole. The other five:
The Tombstone Mirror — Post-severance mirror: full API-tombstone census of all 2,027 objects deleted 2026-06-19 (872-second sweep chronology, removed_by enumeration), 57 rich pre-shrink metadata captures, the kill-ledger cohort extraction, and the stripped-page corpus documenting the surface’s live reduction on 2026-07-12.
Zenodo / DataCite Batch — Pre-severance reference frame: the DOI Resolution Index (1,838 severed DOIs mapped to sovereign targets) alongside the DataCite metadata backup captured before the 2026-06-19 termination propagated.
Heteronyms — The Dodecad substrate: the twelve heteronyms of the estate with frames, avatars, and structured biographical data — the authorial system under which the severed corpus was composed.
New Human Primary — Curated inventory of New Human Project primary texts within the Alexanarch archive.
Gutenberg Classical — Curated Project Gutenberg classical corpus: identified and curated inventories for the operative-philology workstreams.