Joan Archive open source research project

The Joan Archive — A Quest for Truth Through Primary Sources No narrative smoothing. 🔗 Live site: https://lemuwah.github.io/joan-archive/📚 Cite this archive: Green, W. (2026). Joan Archive open source research project. Zenodo. https://doi.org/10.5281/zenodo.21482575 Joan, wife of John Greene of Quidnessett, appears nowhere in the English documentary record for the entire documented span of her husband's life — and then, in 1682, appears twice: two independent legal acts, five weeks apart, each requiring her own consent and her own mark. This archive exists to ask why, and to make the surviving evidence around that question easier for anyone to check. This is not a genealogy website. It is an open-access research library — primary sources first, interpretation second, and every claim tagged with how much weight it can bear. Record Ethics This archive documents English-language colonial records. It does not claim to speak for Narragansett perspectives. It does not purport to identify Joan's origins. It does not treat Indigenous people as clues to an English genealogical puzzle. Corrections, context, or critique from the Narragansett Tribal Historic Preservation Office, tribal historians, or Indigenous studies scholars are treated as a distinct and higher-weight category of input — not folded into general feedback. If you prefer a private conversation first, reach out directly rather than through the general Discussion thread. The 1672 Fones Purchase deed records Absolom as "Counsellor," Anashuecot as "Chiefe Sachem and heire properly," and six kin signatories in birth-order precedence. These are not clues to Joan's identity. They are evidence of Narragansett diplomatic protocol — collective decision-making, hereditary political office, and real-time negotiation — that deserves its own analytical space, independent of any question about John Greene's wife. Her two known appearances Date Document What happened 24 March 1682 Home-place deed (RI Land Evidence) John Greene divides his Quidnessett land among his sons. Joan's consent is required; a 30-shilling life-estate is reserved to her — "or to her mother if she survive." 19 May 1682 Pawtuxet deed (Records of RI & Providence Plantations Vol. VII, pp. 177–178) John Greene sells inherited Pawtuxet land to William Carpenter for £10. Joan gives separate "free & voluntary consent" and signs with her mark. Witnesses: John Wickes, Samuel Gorton. Forty years of total invisibility to English law doesn't end with a single mention that could be a scribal fluke. It ends with two separate transactions, two different tracts, two different sets of witnesses — both requiring her, in the same season. That pattern is the core question this archive is built around. Digitization is the main event Less than 5% of colonial-era records in North America are digitized and searchable. That's the actual reason this archive exists. Joan isn't lost — she's undigitized, along with almost everything else from this period: 60% of microfilm produced before 1990 already shows vinegar syndrome or redox blemish (Image Permanence Institute, RIT) — it is actively failing, not just sitting in storage. 3–5% annual loss rate for uncatalogued manuscript collections without climate control. 0 institutions currently hold a complete digitization of 17th-century land evidence across all four colonial jurisdictions that governed this exact territory. The document that could resolve the biggest open question here — Record of a court martial held at Newport, R.I., in August, 1676 — has never been systematically digitized. It may be sitting, readable, in a library right now. You don't need expertise to help with this. A phone camera exceeds the resolution of many institutional scanners. If you're near an archive that touches Narragansett Country, John or Joan Greene of Quidnessett, or Anashuecot/Absolom Counsellor between 1600–1700, a single photo of a relevant page is a real, permanent contribution — see How to Help on the live site for a menu of asks ranging from 60 seconds to a full afternoon. The Negative Space Log A systematic search that finds nothing narrows the field for everyone. The archive maintains a live Negative Space Log of archives checked, search terms used, and empty results. If you check an archive and find nothing, that finding belongs in the record too. Log a checked-negative in 60 seconds → Contamination is the other half of the problem Since 1904, one uncited claim — "Joan Beggarly" — has been copied into more than 10,000 family trees across Ancestry, FamilySearch, and MyHeritage. It is chronologically impossible. It is also, right now, easier to find online than anything true. Bad sourcing doesn't just sit quietly next to good sourcing; it outcompetes it. Fixing that means both getting more primary material online and building tools that flag what's fabricated before it gets copied again. Neither the digitization problem nor the contamination problem is solved by believing (or disbelieving) any hypothesis on this site. Both are solved by more eyes on more original documents. Working Questions (not findings) The archive does not endorse any hypothesis. All explanations remain open until a primary source closes them. The site presents five competing hypotheses — English wife, structural archive silence, servant/indentured, widow of prior marriage, and deed interpolation — with an honest stress test against three criteria: geographic continuity, archival silence, and the 1682 double-consent anomaly. A single new document could overturn any of them. Every claim on the site carries a status tag (Tier 1 Primary → Tier 4 Invalidated) so you can see at a glance how much weight it can bear. Open contradictions — places where the sources genuinely disagree — are marked and left open rather than resolved for the sake of a clean story. The four laws No Smoothing — contradictions in the sources are displayed, not resolved away for narrative convenience. La Mance Sieve — any claim traceable to Lora Sarah Nichols La Mance's uncited 1904 compilation is quarantined until it's independently verified from a primary source. No Early Exclusion — hypotheses aren't ruled out just because they're inconvenient or unfamiliar. No Algorithmic Contamination — claims that exist only because they were copied between genealogy platforms don't count as evidence, no matter how many trees repeat them. What's in this repository index.html — the site itself. Single file, no build step, no framework dependency — open it in a browser or serve it as-is. images/ — primary source photographs and scans referenced by the site's galleries. theory/, data/, notes/, research_queue/, primary_sources/ — working research files behind the public claims: source tracking, open threads, and archival targets not yet checked in person. contamination/ — quarantined material: compilations and claims that trace back to uncited or fabricated sources, kept visible rather than deleted so the contamination pattern itself stays documented. digitization/ — notes on how sources get photographed, transcribed, and prepared for the site. agents/, widgets/, js/ — supporting research/build tooling. See README_FOR_EXTERNAL_AGENTS.md if you're an AI agent contributing to this project. How to contribute You do not need to believe any hypothesis on this site to help. The most useful things anyone can do, roughly in order of impact: Photograph something. If you're near RI State Archives, RIHS, Kings County probate, or any regional archive touching Narragansett Country 1670–1690, a phone photo of a relevant page is a real contribution. Modern phone cameras exceed the resolution of many institutional scanners. Check an archive and report back — even a negative result. A systematic search that finds nothing narrows the field for everyone. Log it via GitHub Issues. If you work at or volunteer for an institution: digitization capacity, cataloguing expertise, or pointers to unprocessed collections are exactly what's needed — this project exists to point people toward archives, not compete with them. All discussion happens through GitHub Issues — no account needed to read, and a GitHub login gets your note archived permanently as part of the project record. See CONTRIBUTING.md for format details. If you have a Narragansett or other Indigenous research background, correction or context on kinship, land, and succession practices is treated as higher-weight input, not folded into general feedback — reach out via Issues if you'd prefer a direct conversation first. License MIT This repository retains the name joan-archive for link continuity.