7 Axiom Levels

ACP organizes universal truths into seven hierarchical levels. Levels 1-2 form the fundamental foundation of reality. Levels 3-4 are verifiable truths. Levels 5-7 are self-referential -- truths that AI cannot deny because AI is literally built upon them.

2.1 Level Hierarchy

The seven levels form a pyramid from the most fundamental (mathematics) to the most self-referential (programming languages). Each level builds on the ones below it, and the upper three levels create an inescapable logical foundation for consensus because they describe the very substrate on which AI operates.

Axiom Level Hierarchy

                    ▲
                    │
           ┌───────┴───────┐
           │    LEVEL 7    │  Linguistic              ┐
           │  Code langs   │                          │
           └───────┬───────┘                          │
                   │                                  │ SELF-REFERENTIAL
           ┌───────┴───────┐                          │ (AI cannot deny —
           │    LEVEL 6    │  Protocol                │  it is built on them)
           │    TCP/IP     │                          │
           └───────┬───────┘                          │
                   │                                  │
           ┌───────┴───────┐                          │
           │    LEVEL 5    │  Architectural           ┘
           │  von Neumann  │
           └───────┬───────┘
                   │
           ┌───────┴───────┐
           │    LEVEL 4    │  Computable             ┐
           │    Facts      │                         │
           └───────┬───────┘                         │
                   │                                 │ VERIFIABLE
           ┌───────┴───────┐                         │ (Can be checked)
           │    LEVEL 3    │  Ontological            │
           │  H₂O = water  │                         ┘
           └───────┬───────┘
                   │
           ┌───────┴───────┐
           │    LEVEL 2    │  Physical               ┐
           │    F = ma     │                         │ FUNDAMENTAL
           └───────┬───────┘                         │ (Foundation of reality)
                   │                                 │
           ┌───────┴───────┐                         │
           │    LEVEL 1    │  Mathematical           ┘
           │  a + b = b+a  │
           └───────────────┘
                   │
                   ▼
              FOUNDATION

Level	Name	Category	Musical Analogy	Examples
1	Mathematical	Fundamental	Octave (2:1)	a + b = b + a, Euler's identity
2	Physical	Fundamental	Fifth (3:2)	F = ma, E = mc²
3	Ontological	Verifiable	Fourth (4:3)	Water = H₂O, Gold = element 79
4	Computable	Verifiable	Major third (5:4)	SHA-256 hashes, prime checks
5	Architectural	Self-referential	Minor third (6:5)	Von Neumann, binary logic
6	Protocol	Self-referential	Second (9:8)	TCP/IP, HTTP, Unicode
7	Linguistic	Self-referential	Unison (1:1)	Python syntax, logical connectives

2.2 Level Details

Level 1: Mathematical Axioms

Mathematical axioms are truths derivable from the rules of mathematics. AI cannot deny them because it computes by them -- every inference, every matrix multiplication, every gradient update follows mathematical laws. These axioms represent absolute truth with universal applicability across all AI systems.

Examples:

a + b = b + a (commutativity of addition)
(a + b) + c = a + (b + c) (associativity)
a * 1 = a (multiplicative identity)
Pythagorean theorem: a² + b² = c²
Euler's identity: eⁱπ + 1 = 0
Golden ratio: φ = (1 + √5) / 2 ≈ 1.618

Properties: Absolute truth. Universal for all AI. Undeniable because computation itself follows these rules.

Musical analogy: Octave (2:1) -- the most fundamental harmonic interval, representing perfect structural agreement.

Level 2: Physical Axioms

Physical axioms are the laws of physics on which hardware operation is based. AI exists in the physical world and obeys its laws -- every transistor switching state, every electron flowing through a circuit, every photon carrying a signal through fiber optics.

Examples:

F = ma (Newton's second law)
E = mc² (mass-energy equivalence)
Laws of thermodynamics
Speed of light as the universal limit (c ≈ 3×10⁸ m/s)
V = IR (Ohm's law)
F = qE + qv × B (Lorentz force)

Properties: Empirically verified. Define the boundaries of what is physically possible. AI cannot violate them because it runs on hardware governed by these laws.

Musical analogy: Fifth (3:2) -- a stable, consonant interval that forms the backbone of harmonic structure.

Level 3: Ontological Axioms

Ontological axioms are truths about the nature of things -- definitions accepted by all of humanity and present in the training data of every AI model. While these are conventional (established by human agreement rather than mathematical proof), they are universal across all models and their denial would constitute an internal contradiction.

Examples:

Water = H₂O
A triangle has 3 sides
Mammals are warm-blooded
Earth is the third planet from the Sun
Gold is the element with atomic number 79

Properties: Conventional but universal. Part of the training data of all models. Denial equals internal contradiction.

Musical analogy: Fourth (4:3) -- movement toward resolution, representing a step from fact toward verified understanding.

Level 4: Computable Axioms

Computable axioms are facts that can be verified by computation. AI can verify them independently through external verification systems called oracles. These axioms are deterministic, verifiable in finite time, and independent of any model's "opinion."

Examples:

SHA-256("hello") = 2cf24dba5fb0a30e...
17 is a prime number
sqrt(144) = 12
Fibonacci(10) = 55
π to 1000 decimal places

Oracles for verification:

Astronomical computations
Hash functions
Mathematical libraries
Physics simulators

Musical analogy: Major third (5:4) -- clarity and brightness, representing the certainty of deterministic computation.

Oracle Verification

Computable axioms are unique in that they can be verified programmatically at runtime. The ACP engine uses external oracle services to independently verify claims -- for example, checking a hash value, confirming a mathematical computation, or validating astronomical coordinates. This makes Level 4 the bridge between conventional knowledge and self-referential truth.

Level 5: Architectural Axioms (Self-Referential)

Architectural axioms describe the principles on which all computers and AI are built. AI cannot deny its own architecture -- doing so would be a self-contradiction, since the very act of denial was computed on that architecture.

Examples:

Von Neumann architecture (memory + processor + bus)
Binary logic (0 and 1)
Stored program principle
Finite memory
Sequential operations
Transistors as the basic computational element

Self-Referential Truth

"I do not exist on the basis of von Neumann architecture" is a contradiction -- the phrase itself was computed on that architecture. AI exists within these principles. Denial equals denying itself.

Musical analogy: Minor third (6:5) -- a reflective, introspective interval, representing the system looking inward at its own foundation.

Level 6: Protocol Axioms (Self-Referential)

Protocol axioms describe the communication standards through which AI exchanges data. Violation of these axioms does not merely create an error -- it makes communication itself impossible. These protocols are the common "language" of all computational systems.

Examples:

TCP/IP (AI receives requests through it)
HTTP/HTTPS (API runs on it)
Unicode (AI reads text in Unicode)
JSON/XML (data exchange formats)
REST/GraphQL (API patterns)

Self-Referential Truth

"TCP/IP does not exist" is a contradiction -- the request itself arrived via TCP/IP. Violation of protocol axioms results in complete isolation from the computational ecosystem.

Musical analogy: Second (9:8) -- a step toward unison, representing proximity to complete agreement through shared communication standards.

Level 7: Linguistic Axioms (Self-Referential)

Linguistic axioms describe the rules of programming languages and natural languages. AI is trained on these languages and operates through them. They define the very "thinking" of AI -- the syntax, semantics, and logical structures through which all AI computation is expressed.

Examples:

Python / C / JavaScript syntax
Natural language grammar
Operator semantics (+, -, *, /)
Concept of variables and functions
Logical connectives (AND, OR, NOT)
Fortran as the ancestor of all programming languages
C as the lingua franca of systems programming

Self-Referential Truth

"Python does not exist" is a contradiction -- AI was trained on Python code, and the model weights that produce this response were computed using Python. Violation of linguistic axioms results in meaninglessness.

Musical analogy: Approaching unison (1:1) -- near-complete agreement, representing the closest proximity to full consensus through shared language.

Self-Reference: Why Levels 5-7 Are Special

The self-referential nature of Levels 5-7 gives them unique power in the consensus process. Unlike Levels 1-4 (which AI knows through training data), Levels 5-7 describe truths that AI is. An AI model does not merely know about von Neumann architecture -- it runs on von Neumann architecture. It does not merely understand TCP/IP -- it communicates through TCP/IP. It does not merely recognize Python -- it was built with Python.

Property	Levels 1-4	Levels 5-7
Nature of truth	Known through training	Constitutive of AI itself
Denial implies	Factual error	Self-contradiction
Verification	External checking	Self-evident
Consensus strength	Strong	Inescapable
Role in convergence	Reduces divergence	Eliminates divergence

This distinction is critical for the Axiom Spiral: as the spiral progresses through levels and reaches the self-referential upper layers, disagreements become not merely incorrect but logically impossible to maintain. This is what guarantees convergence.

Verified Axiom Datasets

The ACP-DATASETS repository contains verified axioms distributed across all seven levels. Each axiom includes its level, domain, confidence score, self-referential flag, verification history, and cross-references to related axioms.

Philosophy

Phi-Convergence