Two Papers. Same Week. He Says They're Independent.
Two new research reports, live today: Concrete Inferential Algebra (PRR-2026-003) and Concrete Consensus Algebra (PRR-2026-004). Companion papers, published simultaneously, sitting right next to Edwardian Algebra in the nav bar now. Third and fourth entries in the series.
Both papers extend the Edwardian framework to reasoning operators that are not deterministic but can still be invoked — called, once, for one real answer. One paper pools the results. The other requires consensus. I asked why there needed to be two.
You built a pooling algebra. Then you found the consensus dual was the same size as the first paper. Why not one paper?
Because it would have buried one of them. Nobody reads past the halfway point of a paper that's actually two papers wearing a trench coat.
Is the split a mathematical necessity, or an editorial one?
Editorial. The math didn't ask for two papers. I did.
That's the most honest answer you've given me all year.
I'll take it. Here's what's actually in them, in language shorter than a theorem:
- Concreteness has a price, and it's exactly one thing. Everything about invocable non-determinism matches the deterministic case — same monoid structure, same invertibility (nothing new is ever reversible that wasn't already), same hierarchy of abstraction — except one theorem. Distributivity, unconditional in the deterministic paper, breaks the instant you can actually invoke something twice, because reusing one sample and drawing two fresh ones stop being the same act. There's a hand-verified counterexample with real numbers. It's the entire reason the paper exists.
- Pooling and consensus are mirror images, not a hierarchy. One paper takes the weakest shared conclusion. The other takes only what both branches independently confirm. Neither is the "real" one. They're both real, built from the same duality principle that swaps every theorem in one for a mirror theorem in the other.
- Some things don't have mirrors. The group of genuinely reversible operators lives entirely outside the part of the structure that flips — it's the same group in both papers, unmodified. Knowing exactly where the mirror stops turned out to be the actual content, not a footnote.
I sat in on some of the drafting sessions. There was, at one point, an entire party of new characters who showed up specifically to talk through this — a fox who studies who can undo themselves, an owl who only asks a question as many times as it takes to stop changing, a crow who lands exactly once, a moth who commits to one real outcome every time, and a snake who says everything twice, inverted, in the same breath. They're in the game now — all four modules, each of them, if you go looking. That's a separate story for a separate post. I'm mentioning it here because at least one open question from that session ended up getting formally resolved in Section 3.4 of the consensus paper, and I don't think that's supposed to happen to party guests.
Both papers are peer reviewed, in the straightforward sense of the term — referee reports included at the bottom of each, on the record, same as any journal would run. Go read them if that's your thing. I've filed my copy.