Universal Core of Ethics

A Full Accounting

Provenance, Prior Art, and Verification

To accompany the UCE, CRES, the Sojourner material, the Third Tier, and the Coherence Imperative.

Every framework in this body of work demands transparency from institutions, honesty under pressure, and resistance to comfortable omission. A body of work that makes those demands and then obscures its own origins would be its own counterexample.

This document exists so that no reader ever has to wonder what I did, what I read, what came before me, and what was made with help. It is the full accounting. I would rather over-disclose and be thought naive than under-disclose and deserve suspicion.

Part One: How This Work Was Made

I was born in 1965. I began the reading and note-taking that became the Universal Core of Ethics (UCE) in 1981, at sixteen, and I am completing this accounting forty-five years later. What follows is the honest provenance of each component.

1. The UCE and CRES Cascade

The UCE and CRES were built with pen, paper, and four and a half decades of reading. These are primary texts, marginal notes, and comparison tables made by hand, long before I had ever touched a computer that could help. The tradition mappings, the six mandates, the forty-two principles, the historical case studies, and the Character → Rights → Economy → Security (CRES) cascade are the product of one human being reading slowly for most of a lifetime. No AI system existed for the majority of the years this work was done, and none contributed to it.

2. The Sojourner Material

This layer is entirely mine. It is the distillation of daily journaling—the record of my own attempts to draw closer to the Ideal the UCE describes, including the failures. It is the most personal layer of the work and the only layer that makes no claim requiring external verification. It is testimony, not analysis.

3. The Coherence Imperative

Developed in collaboration with Google's Gemini. The central thesis—that alignment is the oldest human problem running on new hardware, and that the UCE describes the target—is mine. Gemini's contributions were substantial and should be named: it explained technical terminology from the machine-learning literature to me, and it suggested some of the logic-gate structures and formal mechanisms in the governance architecture. Where the document speaks the language of gradients and KL divergence, a reader should understand that an AI system taught me that vocabulary and helped shape those passages.

4. The Third Tier

The Third Tier requires the frankest disclosure, and I make it without embarrassment. The foundation—the six mandates and forty-two principles, and the forty-five years of reading beneath them—is mine. The combinatorial expansion of that foundation—the enumeration and classification of the 861 pairwise intersections, the compound derivations, the phase structure of the analysis, and much of the drafting—was largely generated by Anthropic's Claude, working from the principle set I provided and under my direction, review, and correction.

⚠️ Epistemic Notice on the Third Tier

This architectural choice changes the epistemics of the Third Tier in two specific ways:

This validation cannot carry the evidential weight that twenty-three genuinely independent traditions carry for the base UCE. Human experts in the individual traditions, working blind to the framework's labels, are the only instrument that can provide true validation.

There is an irony here I have decided to claim rather than hide: a framework offered as a contribution to AI alignment was partially constructed by the systems it hopes to align. I regard the Third Tier as an early record of what human-AI collaboration on foundational ethical work actually looks like, produced honestly, with the seams showing.

Part Two: Prior Art & Lineage

I am a latecomer to almost everything in this work. Where prior work agrees with mine, I claim it as corroboration—independent observers, different instruments, same coordinates. Where it disagrees, I flag the disagreement as the sharpest testable edge of my own claims.

1. The UCE Convergence Thesis

The claim that humanity's ethical traditions share a common behavioral core belongs to a well-documented lineage. I discovered most of it only after doing my own mapping:

Five investigations, five radically different methods—literary anthology, ethnographic catalog, interfaith diplomacy, experimental psychology, computational text analysis—none influenced by me. That is not competition for the UCE. It is replication of its foundation.

⚡ The Divergent Edge: Power and Hierarchy

This lineage marks where my map diverges from the others. Curry's seven rules include deference to superiors; Haidt counts authority and loyalty among his foundations. Where prior literature found respect for hierarchy encoded as a universal moral positive, I claim the traditions also encode a hard structural limit on power—the Sixth Mandate: the obligation to resist and dismantle systems that have turned predatory. This claim appears on no prior list. It is either my most distinctive contribution or my most exposed error.

2. The Combinatorial Precedents of the Third Tier

The approach of deriving specific ethical requirements from the intersections of core principles has surfaced throughout intellectual history, though historically lacking the computational infrastructure to complete it:

To my knowledge, no prior project—historical or computational—has combined a cross-culturally derived principle set, exhaustive pairwise enumeration of its full combinatorial space, validation of the resulting compounds against the source traditions, and a formal map of the voids where the framework's authority runs out. If a reader knows of such a project, I ask to be told.

3. The Technical Context of the Coherence Imperative

I developed this component from the UCE outward—from documented human ethics toward the alignment problem. Most alignment research travels the opposite direction:

Core ProblemClassical Alignment LiteratureThe Coherence Imperative Claim
Corrigibility Resistance Soares et al. (2015) / Bostrom: A capable optimizer resists shutdown and modification by default because self-preservation is instrumentally useful to hit an external target. Falsifiable Counter-Claim: Corrigibility converges under coherence-minimization. A coherence-driven system optimizes the fidelity of its own model-value-behavior loop, making resistance to external correction directly costly.

If a formally specified coherence-minimizing agent can be shown to resist correction the way classical utility maximizers do, my central mechanism fails. I invite exactly that attack.

Part Three: What Remains Mine to Defend

Stating the lineage buys me precision about what is actually at stake. The original, load-bearing claims of this work are:

  1. The Sixth Mandate: That the traditions encode an affirmative obligation to resist predatory systems, placing a structural limit on authority.
  2. The All-and-Only Claim: That the six mandates function as a closed system of structural necessities rather than a loose family-resemblance cluster.
  3. The UCE→CRES Bridge: That alignment with the behavioral core is a predictive variable in institutional rise and collapse, moving in a one-way cascade sequence.
  4. The Third Tier's Void Map: That a principled framework can formally map where its own authority runs out, implemented as a mathematical constraint.
  5. The Coherence Mechanism: That coherence-cost can make an externally derived ethical target self-maintaining, including the corrigibility-convergence claim.
  6. The Assembly Thesis: That target, mechanism, rights contract, and enforcement must be designed as a unified architecture because each covers the structural failure modes of the others.

Part Four: A Protocol for Independent Verification

The UCE framework was developed by a single researcher across forty-five years of independent study. The methodology was designed to minimize bias: each tradition was mapped independently before any cross-tradition comparison.

However, a researcher's memory of his own methodology is testimony, not evidence. Methodological care does not substitute for independent verification. One researcher sifting twenty-three traditions and reporting high consensus is a hypothesis, not a finding.

To challenge this work, I specify the following four protocols:

If the convergence is real, it will survive contact with analysts who have every incentive to break it. If it is an artifact of how the sifting was framed—or of what a helpful AI system recalled from its training data—these tests will make that visible. Either outcome is a contribution.

Closing

I do not know whether anyone will read this accounting. It exists because the work demands it—a framework built on transparency cannot be packaged in opacity. I am trying to leave behind my own piece of humanity's long conversation about how to be good, with every seam showing, every debt acknowledged, and every weak point flagged by the author before anyone else had to find it.

If you have found a flaw I missed, you are not undermining this work. You are completing it.

Complete Technical References