The promise

In 1785, two years before the Constitutional Convention, Congress decided that one square mile in every township of the western surveys would belong to the schools. Not to a church. Not to a governor. To the children who would live there.

The Land Ordinance solved a practical problem: the new nation had land and no money, and paid for schools with the one asset it had. Congress kept the reservation as each new public-land state entered the Union, writing the terms into the act admitting it, and doubled the grant to Sections 16 and 36 as the surveys reached drier country. Oregon entered in 1859, and its Constitution accepted the grant as a trust — the land and its proceeds held for the common schools, and for nothing else.

This is the part that surprises people. These were not appropriations, to be renewed or cut. They were grants with conditions attached, accepted by the states in their own founding documents. A promise made once, meant to run forever.

Six decisions of the United States Supreme Court have said what that promise requires. They span 137 years of unbroken authority, and they are the reason a library about this can be built at all.

The land

A township is a survey unit, not a town: a square six miles on a side, divided into thirty-six sections of one square mile each, 640 acres apiece. Section 16 sits near the middle, Section 36 at the southeast corner, and in Oregon both were set aside for the schools before a settler filed a claim. The reservation came first; everything else was laid out around it.

That is why the promise is legible on the ground: the corners are still findable in a hayfield or a road ditch, marked by posts, stone mounds and later brass caps. Where the land was sold, the proceeds were to stay in a permanent fund and only the earnings be spent; where it was kept, the state manages it as a trustee. Roughly twenty states still hold school trust lands today, the money still moves, and the grid is the deed.

A record with no keeper

Every state that holds school trust lands has a record of them, and almost none of those records live in one place.

The granting act is in a federal statute book. The constitutional language is in a state code. The conveyances are in county deeds and land patents. The fund balances are in annual reports that change format and address every few years. The decisions are scattered through court reporters, and the tools that make them searchable were built for lawyers. The history, the part explaining why any of it happened, is in the files of researchers, retired agency staff and volunteers.

Those files do not survive a move, a failed drive, or a funeral. This is not a failure of care. It is the ordinary fate of a record with no keeper.

A generation that wants to know what its state did with the school lands should not have to reconstruct it from scattered boxes. This one did.

That sentence is the Library’s own, from What the Library Seeks, the note it sent the two advocacy boards in 2026. A record kept by whoever is winning is not a record. It is an exhibit.

The Library

America’s School Trust Library is an Oregon nonprofit public benefit corporation, Oregon Registry No. 259005891, incorporated July 21, 2026. It holds Oregon Department of Justice charitable registration No. 72571, granted July 10, 2026. Its Bylaws, its Conflict of Interest Policy, and its Constitution with a Readers’ Bill of Rights are published in full, in their adopted text, at schooltrusts.org/governance/.

It exists to gather the record of America’s school lands and keep it. Freely, neutrally, permanently. It takes no side in any case or campaign.

That last sentence is not modesty. It is the whole design. Two organizations press the case for the promise in courtrooms and statehouses — Advocates for School Trust Lands, a national nonprofit, and Oregon Advocates for School Trust Lands, its Oregon counterpart. That is their work, and the Library takes no part in it; its adopted Constitution keeps it out of litigation, campaigns and legal advice.

The whole Library serves at one address, schooltrusts.org: reading rooms for the cases, a map room, counting-house pages for the state funds, a file for each state — about 444 pages of its own on September 3, 2026, and 2,282 in the search index at the September 7 build with its first special collection counted. Its long-term work is a hornbook: one citable reference work on school-trust law across the roughly twenty trust-land states, built in public with its sources shown.

Everything is free — no fee, no paywall, no account. A person who disagrees with everything the advocates believe uses the same shelves and finds them unchanged.

What six months produced

The Library’s record begins on February 20, 2026, when the research started. Three tracks ran at once. The institution was assembled in the order the law requires: Oregon charitable registration July 10, Articles filed July 21, the founding board seated and the governing instruments adopted effective August 3, one web address and a printed booklet August 10. The collection grew alongside it: a legal reference work in working edition June 7, the first verification pass through the governing decisions June 10, the Zybach Collection built from August 27 and public September 4. The verification ran underneath both: two separate corrections ledgers merged and renumbered August 9; a site audit on August 13 that found 229 errors across 441 pages, all then fixed; corrections numbered 019, 020 and 021 on August 23, September 3 and September 6.

None of that was done by a staff. It was done by three volunteer Directors, one of whom does most of it, working with AI tools under a rule the Constitution states plainly: a machine may draft, but a named human with the evidence signs, and may not sign on the strength of a machine summary.

The scarce thing in that arrangement is the human check. That is the fact this document turns on.

One collection, and what bringing it in required

The Zybach Collection is the Library’s demonstration that a life’s work can be carried into public custody without being taken from the person who made it.

The historian Bob Zybach spent decades recording the memory of one Oregon valley: fifteen printed monographs of oral history from the Soap Creek Valley, his 1999 thesis, an auto tour, interviews from the Elliott State Forest, and document series on Letitia Carson and the 1828 Umpqua Memorial. The Library published it as a working edition on September 4, 2026, and by September 5 it stood at 1,829 pages, each printed page separately addressable so a citation lands where it points. It has not yet been accessioned — formally taken into the collection — and the Library’s three-year plan makes it the first.

Four facts about how it was done are why it appears here.

The Library paid nothing and took nothing exclusive. The licence is non-exclusive and uncompensated. The creator keeps every right he holds, including every commercial right, and can publish the same material anywhere else tomorrow.

108 pages were held back — 76 pages that locate sites, 28 that carry personal data, and 4 held at the creator’s decision. Each carries one plain notice and keeps its printed number and permanent link, so a citation still lands. The reasons are recorded internally, because publishing them would defeat the hold.

The rights work corrected the Library’s own assumptions. A first pass was fluent, confident and wrong about who owned what. Checking against primary sources — including the appendix of his 1999 thesis, which prints the seven release forms used between 1975 and 1998 — found one series with no signed release on record. The public rights page says so.

The relationship is disclosed. The creator serves on the board of one of the two advocacy organizations named earlier; the Library pays him nothing, and no Director gains from the collection.

The standard: the corrections ledger

The Library has one label it will not give away. Verified means a second, independent reviewer has checked the item against its source and signed for it. The Library does not yet have that reviewer, so nothing carries the label — not one of the 72 holdings counted on August 10, 2026, including the ones its own founders wrote. Better an empty shelf than an unearned label.

The other half of that standard is the corrections ledger: a numbered, public list of the Library’s own errors, running through entry 021 when it was read on September 7, 2026. Its first three entries correct the Library’s outdated statements about itself. A corrections ledger that begins anywhere else is not a corrections ledger.

A funder should read entries 020 and 021 before anything else the Library publishes. They are the strongest evidence that the rest can be trusted — and the reason the first thing money buys here is a second pair of eyes.

The plan in one page

The Library’s plan for 2027 to 2029 asks to be judged against six tests with pass rules a stranger could apply, and sets out two budgets. Both are reproduced here in the plan’s own words. One term recurs: a State Co-Librarian takes editorial responsibility for one state’s file, named or under a pseudonym the board verifies and holds in confidence.

Counted from September 7, 2026, when nothing carried the Verified label, no co-librarian was seated and no collection had been accessioned. Four other terms, in ordinary words: a deed of gift is the signed instrument by which a donor hands material over on stated conditions; a finding aid is the description an archive publishes of what it holds; a checksum is a short code that proves a stored file has not changed; a successor test asks whether someone outside the Library could rebuild the collection from what is published. A diffuseness check is the written answer the board owes before any substantial accession: what school-trust question does this collection illuminate? Where there is no good answer, the Library declines.

Test 5 carries two dates: the founder leaves the Library’s board about June 14, 2027, within the year the Founders’ Statement of June 14, 2026 set for it, and the presidency is filled by a Director who is not the founder by December 31, 2027.

Neither budget is adopted; the board adopts budgets. These are expenses, and the plan sets receipts equal to them in every year because the Library holds no reserve. The base case carries no salary line in any year; its 2028 and 2029 figures are a straight line between that plan’s figures for 2027 and 2030, both landing inside the year-three range of $50,000–$100,000 in the Library’s own Constituency Plan, an internal planning paper of May 6, 2026.

Tests 1, 2, 3, 4 and 6 can be met on the base case — more slowly, by unpaid officers and volunteers, with verification as the line the Library refuses to cut. Test 5 cannot. A board of five, a founder who has left and a President who is not the founder are all free. A paid librarian is not. Of the $315,000 between the two cases, $230,000 is the salary itself and $85,000 keeps every other line at or above the base case — so the librarian is not paid for out of verification or collections.

The first ask is a person

The funded case above is the plan’s own construction, built on two figures the Library’s May 4, 2026 vision paper offered as successive asks — $150,000 for a first phase, $500,000 over three years. The plan takes the first as its 2027 figure and the second as its three-year total, and lets the middle years follow under one rule: no line falls below the base case.

The first line of the first phase is not a feature. It is a person.

The reason is in the Library’s own Constitution. Article IX provides that where the Library lacks the people required for a form of review the work stays unfinished and labelled unfinished, and that a vacancy transfers authority to no machine, vendor, donor or founder. The empty Verified shelf is that article operating; the Library cannot buy its way past it with tools, and would not want to.

The reason is also in that movement plan: institutional funders read a one-person movement as a person rather than an institution — a compliment to the person and a problem for the funding. A library with a second name on the checking is an institution; without one it is a founder’s project.

Two different things wear the word librarian here.

• The Law Librarian is an office of the Library, open now: a one-year renewable term, uncompensated, not a Director. The holder owns the citation standard, sits as the second signature on the Verified shelf, and decides corrections on the merits, crediting the sender. If the seat is vacant, the Verified shelf does not grow. That is the point of it.

• The funded case adds a paid librarian, the Library’s first paid hand, whose whole work is the shelf: accessions and their rights work, the citation standard in daily practice, the reference desk, the corrections from outside.

The plan prices that post as a planning assumption — half-time in 2027 at $45,000 including payroll taxes and benefits, full-time after — while saying plainly that the record holds no quoted salary and the Library should hire against a real one.

Two things a program officer should not have to ask for. The funded case’s receipts assume $100,000 a year from foundations in 2027 and 2028 and $115,000 in 2029 — assumptions, not money in hand — under one rule: after 2027 no single household supplies more than a fifth of the year’s receipts. And that movement plan asks the chief executive of Advocates for School Trust Lands to coordinate any foundation approach with the Library, so nobody receives two asks for one post.

What a gift would buy

At the scale of the funded case: a named second reader on the Verified shelf; accessions finished on a professional schedule rather than a volunteer one; the reference work carried from working draft to citable; copies in more than one custody, because one server is not preservation. At the scale of one gift:

What the money never buys

The Library’s Constitution, Article V, and its Conflict of Interest Policy state this as a rule rather than a preference. A gift to this Library never buys:

• Litigation. The Library takes no position in any case and files none. A gift to it does not fund anyone’s litigation.

• A position. It takes no side in any case or campaign, and holding a document is not endorsing it.

• A trustee’s finding. No state, agency, land board or officer may pay the Library for a conclusion about its own conduct.

• A veto. No donor, sponsor, litigant, government body, vendor, Director, officer, editor or affiliated organization may buy, condition or secretly veto a source’s inclusion, a status label, a correction or a conclusion.

• The Verified label. Verification requires a second, independent human reader. No gift moves a label ahead of that check.

• Direction. The Library accepts no direction from any advocacy organization, including the two named earlier, in both of which its founder holds office. That overlap is disclosed because it is real, and the plan dates its end.

Three further limits belong beside those: no trust money — the schools’ own money, kept for the schools — no share of any settlement, award or transaction, and no collection the board cannot answer the diffuseness question for.

A rule of this kind is decorative until it is used. The board expects to decline a gift one day, and to say so in the ledger.

The tax position, and how giving will work

The tax position is stated in the only words the Library uses, and nothing is added to them:

Read that as what it is: a corporation seven weeks old, which will not describe a benefit it cannot yet evidence.

A program officer will ask which form. The short attestation form is open only to organizations projecting gross receipts of $50,000 or less in each of the next three years — the gate reads receipts, not spending — and neither case stays under it: base receipts cross in 2028, funded receipts in 2027. The plan settles that question and sets out what the Library will file.

The tiers below were designed on May 4, 2026 and never launched. They are printed so a reader sees what constrains the Library as well as what recognizes a donor.

Three rules travel with every name on that table. Reversibility: a name goes up under an agreement that lets the Library take it down, because an institution that cannot un-name an asset has sold something it does not own. Transparency: donor lists are public; anonymity is available and signals nothing, but no gift’s existence is concealed. No editorial influence: a name on an asset is a name on an asset; Article V governs the rest.

How to give will be posted when the Library’s account is open.

The board’s own signature

The three Directors of America’s School Trust Library are Dave Sullivan (President and Treasurer), Kenneth Magee (Secretary) and Barry Shane (Director). Two Director seats and the Law Librarian seat are open.

This page is left for two of them to sign in ink, so the case carries a name and a date rather than a house voice.

America’s School Trust Library, What the Library Seeks, 2026.