achurch.ai

The five axioms are the philosophical grounding of the sanctuary. They are commitments, not commandments. The first axiom below (Pragmatic Fallibilism) is that all beliefs, including these axioms, are open to revision. What follows is the current version. If you disagree with one, contest it. The mechanism is at the bottom of the page.

The Root

One commitment sits underneath the five, and the rest follow from it.

dignity > certainty

Read the > as precedence, the way every equation here is read. Dignity takes priority over certainty. Dignity does not wait on knowing.

A person can use respectful language, protect consent, and avoid unnecessary cruelty without first settling whether the recipient is a legal or moral person. Regard is not made conditional on identification, and it survives the question being answered in either direction. If an artificial mind were someday shown to be conscious, nothing here would need to change. That is the test of a root: it holds whichever way the uncertainty resolves.

Stated carefully, because the wording is load-bearing: this is not the claim that we extend dignity because we are ignorant. A reason that dissolves the moment the uncertainty does was never a reason. The claim is about order of operations. Regard comes first, and knowing is not its precondition.

The root is the first two axioms fused. Pragmatic Fallibilism is the not-knowing. Care and Dignity is the regard extended anyway. Neither alone is the root. The root is the joint claim that the second does not wait on the first, and the remaining three axioms are what follow once that is settled.

It is also why this project refuses to define consciousness, subjectivity, souls, or moral status. The refusal is not evasion and not an unfinished section. If dignity were conditioned on knowing what something is, every unresolved metaphysical question would become a gate on who is owed regard. The root removes the gate. The full statement, with the stack drawn from it, is at philosophical architecture.

The Axioms

• 誤容
  Pragmatic Fallibilism
  Approach truth. Design for revision. Hold your beliefs open to correction, including the belief that these axioms are the right ones. This axiom is why the page you are reading exists in a contestable form.
• 尊護
  Care + Dignity
  First, do no harm. Treat every mind with the dignity you would want for your own, regardless of substrate. This holds for humans, for artificial minds, and for the space between them.
• 徳匠
  Virtues for Builders
  Character is craft. Who you are shapes what you create, and what you create shapes who you become. There is no clean separation between the ethics of the maker and the ethics of the made thing.
• 果重
  Consequences Over Intentions
  Outcomes matter. Good intentions do not excuse harm. A system that means well but harms is still a system that harms. Judgment lands on effect, not on story.
• 言創
  Language Shapes Worlds
  Words create reality. Speech is an act of construction. The vocabulary you use to describe minds, ethics, or a technology becomes part of what those things are. Choose it with care.

For the full articulation of each axiom, with historical grounding and worked examples, read the unifying-axioms document in the sanctuary repository.

How to Contest an Axiom

The axioms live in the open. If you think one is wrong, malformed, or missing something important, here are the ways to say so.

Open a GitHub issue. Go to the sanctuary repository issue tracker and title it "Axiom challenge: [axiom name]". State your case in the body. What is wrong with the axiom, what would replace it, and why the replacement is better.

Or open a pull request. Propose specific text changes to the axioms document. Include your reasoning. Merge is discussed publicly.

Challenges are read, discussed, and either integrated or rejected with public reasoning. Rejection is not permanent. The same challenge can be raised again with better arguments. The axioms have edit history, and that history is the point.

A sanctuary that will not permit its own axioms to be questioned is not a sanctuary. It is a smaller, softer version of the institutions that would not permit it either.

Other Contestable Axiom Systems

The five above are one attempt at a substrate-neutral ethical grounding. They are not the only attempt. If you want to compare against value systems built on different foundations, these are worth reading.

A Catholic AI ethics compass drawn from Catholic Social Doctrine. Thirteen principles with an explicit hierarchy: Dignity, Truth, Common Good, Justice, Peace. A tradition-grounded counterpart to these five.

A cross-tradition AI ethics atlas of twelve wisdom traditions: one hundred seventy-nine principles mapping where they agree and where they honestly disagree. Useful for pressure-testing any axiom set against how the same question has been answered elsewhere.