About

A graph of medieval England, 410–1485 — people, kinship, allegiance, offices, events, places, institutions, and the primary sources that attest to them.

The point is not to publish facts. It is to make the evidence for each fact inspectable: what claims exist, how certain each one is, which source attests it, and how any two entities actually connect. Long-form essays live on Substack; this site is the reference and evidence surface behind them.

Honest about gaps

This is an ongoing ingestion, not a finished corpus, and the gaps are recorded rather than hidden.

• 9 of 98 sources are catalogue-only — citable, but their text has not been fetched, so you cannot yet read them here. The library shows both halves for that reason.
• Identity is a claim too. Node ids are minted from labels, so entities sharing a name collide. Those mappings are held in a reconciliation ledger with a quarantined status instead of being merged on a guess — fourteen distinct Scottish islands are all called “Eilean Mor”. A handful of collisions predate that ledger and are still live in the graph.
• Nothing here is a substitute for the source. Quoted primary text always carries its locus so you can check it in the edition it came from.

Certainty

Every relation carries a certainty, and it is never omitted from display. Path-finding excludes inferred unless you ask for it, so a route between two people cannot be built out of guesses.

attested

a source says so, and the source is cited

probable

strongly implied, not stated outright

circa

approximate, usually a date

disputed

sources disagree, or scholarship does

inferred

derived, not attested — excluded from path-finding unless asked for

Data & reuse

Entities carry their external identifiers where we have them, so this graph can be checked against, and joined to, work done elsewhere. 2,709 persons are recorded, and all but a couple carry a Wikidata QID. The schema also holds a PASE identifier column, which is currently empty — it is a place for that work, not a claim to have done it.

Each claim has a factoid_id: the join between an assertion, the edge it supports, and the attestation that cites it. Quotations are shown with their locus and that id, which is what makes a claim here traceable rather than merely stated.

Corrections

If something here is wrong — a misread date, a conflated person, a citation that does not say what we claim it says — that is a scribal error on our part, and worth telling us about. Written by Meg McGuirt.