Proof by Production
In 1844, twenty-eight weavers and skilled workers in Rochdale pooled a pound each to open a cooperative shop, then wrote down the rules they'd run it by: one member, one vote, open membership, dividends paid on purchases rather than capital. Those rules are still close to the textbook description of a cooperative. But what makes them worth reading now isn't that they were well drafted. It's that drafting them was only possible because two dozen-plus people had already done the thing the rules describe. They pooled money, trusted each other with it, and agreed on terms before any one of them could veto on behalf of the group. The document is evidence of its own claim before you've read a line of it.
Compare that to a different kind of document: a manifesto written by one person, alone, arguing that workers acting in concert can outlast owners acting alone, or that a community governed by consensus will outperform one run from the top. Nothing in the sentence is implausible. But the sentence, on its own, carries no evidence either way, because one person typing words about coordination demonstrates nothing about whether coordination among many people actually happens. The claim and the only thing that could support it are two separate objects, and the author controls just the first one.
Call it proof by production. A claim about collective action is only as convincing as the hardest problem that had to get solved to produce it. For a lone author, the hardest problem is finding the right words. For a cooperative, the hardest problem is getting two dozen people to agree on anything at all, including which words to use. When a document couldn't exist without first solving that second problem, reading it stops being a matter of trusting someone's description of coordination. It becomes checking a receipt for coordination that already happened, before you ever showed up.
This is also why open letters are such an uneven genre. Some get drafted by one or two people and then passed around for signatures once the text is locked, thousands of names stapled to a sentence none of the signers wrote. Others get negotiated line by line among the people who will sign it, and the text only exists in the form it does because enough of those people refused to sign anything weaker. The signature count at the bottom can look identical on both. The proof underneath it is not.
Order is doing the work here, not headcount. A petition signed after the fact by ten thousand people is evidence that ten thousand people read a sentence and clicked a button. That's real, but it's thin. Bylaws that couldn't have been typeset until a shared fund existed are evidence that people organized with each other before any of them got to see what they were organizing around. The second kind of proof is rarer, and it's rarer because it's harder to fake. You cannot retroactively insert a cooperation requirement into a document that already exists without it.
Anyone can write the sentence. Not anyone can produce the receipt.