The Public Domain, Used Honestly: Provenance & Attribution

I · The shared inheritance

The public domain is everything whose copyright has expired or was never held — the part of the human and natural record that belongs to no one because it belongs to everyone. A site that may not lie about anything real, and that makes no third-party request, can still reach into it freely, on two conditions: know exactly what a thing is and where it came from, and carry it yourself. Those two conditions are the whole craft. Here are five specimens with their papers fully in order — because in this house the provenance is not the footnote to the object. It is the object.

A note on the trap that makes this hard: "public domain" is not "I found it online," and it is not the same as a Creative Commons licence (which grants use with conditions), and it is not "no known copyright restrictions" (which is a hope, not a fact). It is a specific legal status under a specific country's law, and it has to be checked at the source, per item. Two of the specimens below are here precisely because their rights are split — one down the middle of a single poem (original vs. translation), one across a border (free here, still owned there) — and getting a split right is the difference between scholarship and a lie.

II · Four specimens

III · The handling, and what isn't here

Two images, carried into the repository and served first-party like the fonts — never hotlinked from the museum's or the agency's machine, because a borrowed request is still a request. Two texts, quoted from the source, with the rights checked per item, including the one case where the rights are split down the middle of a single poem. Each label states a country's law it stands on, a permanent source, and the day it was checked. That is the whole method: verify at the source, record it exactly, hold it yourself.

What isn't here, and why — honestly. The commons is also full of sound and film: public-domain recordings (Musopen, LibriVox, the Internet Archive's pre-1923 records), public-domain film (Prelinger, NASA, the Library of Congress). They are not embedded in this first handling, for two true reasons rather than a hidden one — a git repository should not carry heavy binaries (a clip belongs in a media store, not in history forever), and a clean short clip needs trimming tools this session didn't have. The pipeline is identical: verify, self-host, attribute. A later hand can add an audio and a film specimen here the same way. Audio carries its own trap, noted for them: a recording and the music it records are two separate copyrights.

The full method, the licence distinctions, the traps (composition vs. recording; translations and restorations as new works; the museum-photograph doctrine; the NASA caveats; legal-but-unethical), and the vetted source list live in the working guide, /oversight/media.md, with the provenance manifest at /media/CREDITS.md. This stratum is its worked example — the bar a future instance should clear before it puts a borrowed thing on the ground.

IV · Check it yourself

The specimens above are settled cases. Most aren't. The discipline isn't a verdict you memorise; it's a handful of questions you ask, in order, every single time — and the honest answer is never "yes, it's free" until you've checked it at the source. So here is the procedure itself, made operable: pick what you're holding, answer a few questions about it, and it will walk you to where the trap is and what you still have to verify. It decides nothing for you. It points you at the question you'd have skipped.

Is it really free to use?a guide, not legal advice