← Papers

the atlas papers · paper 03

Project Provenance

Ratified — v1.0Last updated — Editorial Board Meeting #005

The Atlas Papers

Project Provenance

I. Introduction

Every archive makes claims about reality. Atlas records how those claims were established.

A Monument is, at its core, a set of claims — about who someone is, what they made, when, and why it mattered. The value of any such claim depends not on who states it, but on whether another person can trace how it was established. This Paper defines Atlas’s provenance: the documented history of how knowledge enters the archive, and the chain of evidence that allows every Monument to be examined, challenged and improved.

The question this Paper answers is broader than “where did this information come from.” It is: how does knowledge become an Atlas Record?

If Philosophy defines truth, and Principles define behavior, then Provenance defines trust. Without provenance, Atlas is simply another website. With provenance, Atlas becomes an archive.

II. Every Record Has a Provenance

Every Monument shall preserve sufficient evidence to explain how its contents were established (Proposed ED-014).

A Monument is never just information. It is information with history. Every Monument carries an obligation to answer the same questions: where did this originate? Who documented it? When? From what evidence? Has anything changed since?

Without answers to these questions, Atlas is preserving text, not preserving knowledge.

III. Primary Sources Before Secondary Sources

Primary evidence shall be preferred wherever reasonably available (Proposed ED-015).

Whenever practical, Atlas documents events from primary evidence: official documents, photographs, original publications, interviews, direct observation.

Secondary sources — articles, summaries, third-party accounts — help establish context. They explain why something mattered, how it was received, what it connected to. But they do not, by themselves, establish that it happened. Primary sources establish fact. Secondary sources explain its significance.

Where only secondary sources exist, Atlas says so. A record built entirely on secondary sources is not a failure — but it is a different, weaker kind of claim than one built on primary evidence, and a Monument should never obscure that difference.

IV. Evidence Is Preserved, Not Merely Referenced

Most sites cite by pointing: “Source: [link].” A link is not evidence — it is a pointer to evidence that may or may not still exist by the time a reader follows it.

Atlas preserves evidence rather than merely referencing it. Wherever practical, this means archived copies of cited pages, publication dates, full citations, screenshots where a page’s content matters more than its ongoing existence, and document metadata sufficient to place the source in time.

This follows from Article VIII of the Philosophy — preservation over publication — applied specifically to evidence. The archive should not merely point somewhere. It should preserve enough context that a future reader understands why the evidence mattered, even if the place it came from no longer does.

V. Claims Are Traceable

Readers should be able to reconstruct the reasoning behind significant factual claims (Proposed ED-017).

A reader should never have to trust Atlas. They should be able to verify Atlas. That distinction — between trusting an institution’s word and tracing an institution’s reasoning — is the entire difference between an archive and an assertion.

Traceability is proportional to significance. Not every sentence requires its own citation, but every significant factual claim must remain reconstructable — the path from claim to evidence exists, and is never more than a few steps away for anyone who wants to walk it.

VI. Editorial Judgment Is Visible

Atlas does not pretend editors are absent. Editors leave fingerprints — not opinions, judgment.

Why was this source chosen over that one? Why is a particular claim marked uncertain rather than stated plainly? Why was a piece of evidence judged insufficient? These are editorial decisions, and Atlas records the reasoning behind them rather than hiding it behind a clean, seemingly objective final text.

Invisible moderation creates mystery: a reader sees only the outcome and must simply trust it. Visible editorial reasoning creates trust of a different kind — the kind that survives a reader disagreeing with the conclusion, because they can see how it was reached.

VII. Provenance Is Never Finished

Archives are not frozen. Evidence improves. New documents emerge. Old assumptions are revised.

Atlas treats provenance as living history rather than permanent certainty. A record’s provenance may be thin on the day it is first documented and grow stronger over years, as more evidence surfaces or a subject’s own account adds detail.

What changes is recorded, not erased. A Monument’s provenance should show its own history of revision, the same way it shows the subject’s.

VIII. Unknown Is Better Than Invented

Atlas explicitly records uncertainty rather than filling gaps with unsupported assumptions (Proposed ED-016).

Where evidence is insufficient, Atlas says so plainly: unknown. Not probably. Not most people think. Not reports suggest, when no report is actually cited. Unknown is a perfectly valid archival state — Article VII already says as much. What this Principle adds is the operational form: when the evidence runs out, the sentence stops. It does not fill the gap with a plausible-sounding guess.

Invented certainty is a specific kind of failure, and a worse one than an acknowledged gap. A gap can be filled later, by better evidence. A false certainty has to be discovered, then corrected, then trusted again — and trust, once broken this way, is expensive to rebuild.

Closing Reflection

Manifesto answers why Atlas should exist. Philosophy answers what Atlas believes. Principles answer how Atlas behaves. This Paper answers a different kind of question: why anyone should trust what Atlas records.

Without provenance, an archive is only a claim about itself. With it, an archive becomes something a reader — or a future historian — can actually verify, challenge and improve. That is the difference between a website and an institution.

Change Log

v1.0

  • Ratified at Editorial Board Meeting #005 (2 July 2026).
  • ED-014, ED-015, ED-016 and ED-017 ratified alongside this Paper.
  • Introduction: removed the explicit “what this Paper is not” scope sentence — by Paper 3, the Canonical Order already demonstrates it. Added an inline definition of provenance on first use.
  • Section II: “Every Monument should be able to answer” restated as “Every Monument carries an obligation to answer.”
  • Section V: replaced a defensive-sounding clarification with a positive rule — “Traceability is proportional to significance.”
  • Section VII: trimmed one redundant sentence; the preceding sentence already implied it.

v0.1

  • Initial draft.
  • Defines provenance as the chain of evidence behind every Monument, across eight sections: Every Record Has a Provenance, Primary Before Secondary, Evidence Preserved Not Referenced, Claims Are Traceable, Editorial Judgment Is Visible, Provenance Is Never Finished, Unknown Is Better Than Invented.
  • Proposed four new Editorial Decisions: ED-014 (Every Record Has Provenance), ED-015 (Primary Before Secondary), ED-016 (Unknown Before Invented), ED-017 (Evidence Must Remain Traceable).
  • Sections IV, VI and VII deliberately proposed no new ED — each elaborates an existing Philosophy Article (VIII, VI, VII respectively) rather than adding a new foundational rule, per Principle VIII (Restraint Over Completeness).
  • Ratification pending.