allen-proxmire/ED-generative: 27.0
Abstract
One primitive does the work: the arrow of time — the irreversible act of commitment. Most attempts at unification add structure. This one removes it. Event Density starts from a discrete relational substrate with no intrinsic metric, no background coordinates, and no dimensionless constants of its own, and asks what a single irreversible act can be made to carry. The answer, so far, is more than expected — and the claim the program actually rests on is this: The thing that selects a measurement outcome in quantum mechanics and the thing that supplies a preferred time in gravity are not two things. They are one primitive. The Problem of Time does not get bridged. It stops arising. What is here 242 papers, 506 documents, 16 arc ledgers. Weak-field gravity reproduced, MOND emerging natively as a horizon interference term rather than a modification, quantum mechanics reconstructed from relational lattices, gauge structure from indistinguishable channels, charge quantization from winding, rest mass from binding rather than a field. And a complete record of everything that failed. That second part is not modesty. It is the point. What makes this repository unusual Every load-bearing claim carries a tier: derived, grounded, inherited, postulated, measured, open. Not as decoration — as a contract. A claim tiered Grounded names the postulate it rests on. A claim tiered Inherited names where the number came from. A claim tiered Open says what would close it. The ledgers record corrections against the theory as carefully as results for it. A flagship prediction that went from confirmed to untested in one day when a systematics critique checked out. A repair carried for two days that turned out to invert the problem rather than fix it. A summary that overstated the arc's own open questions by two. Those are in here, dated, with the reasoning. Six automated checks run over the corpus and fail on drift: stale claims, broken provenance, contradictory definitions across documents, unverified rows, dead pointers. Each was built after a defect got through. How to read it If you have ten minutes — ED_UnifiedFramework_Report.pdf. The whole framework, box by box, with the limits stated in the same breath as the results. If you want the receipts — ED_ItemizedTheory_TieredClaims_v2.xlsx. Every claim, its tier, what it rests on, what would kill it. Sortable. Start with the Core Predictions sheet, then Ledger w Claims. If you want to attack it — physics-papers/predictions/22_Ways_to_Kill_Event_Density.md. Twenty-two ways, written by the author, each with a falsifier. The fastest route to a real objection. If you want one paper — physics-papers/gravity/Paper_a0z_MONDScaleTracksHubbleRate.md. A single prediction, registered before the data, with its kill condition stated in advance and its own confrontation reported honestly, including the parts that go against it. If you are an AI agent — read CLAUDE.md first. It tells you the one rule that matters: read the corpus before you build on it. It exists because a session once spent a day re-deriving results that were already there, then "falsified" its own reconstruction. What this is not Not peer-reviewed. Not complete. No distinctive prediction is confirmed. Two structural questions are open — why gauge multiplicities are {1,2,3}, and whether exact charge conservation forces anomaly freedom. One prerequisite is unproven: universal free fall. The dimensionless constants are inherited, by design and by argument. "Form-complete" here is a coverage claim — the substrate fixes the form of a law while the universe supplies its numbers. It is not closure, not unique derivation, and not empirical validation. The report says so in its first paragraph. The interesting question is not whether this is right. It is whether a program built from one primitive, that documents its own failures this carefully, is wrong in a way that can be found. It is set up so that it can be. Start with the twenty-two.
