Axiology in AGISystem2 refers to dimensions 256-383: the 128 dimensions that encode values, preferences, and normative judgments. These are kept separate from ontological facts so the system can reason about values explicitly and transparently.

Why Separate Values?

Mixing facts and values creates problems:

• Hidden bias – Value judgments influence factual classifications
• Unexplainable decisions – Can't tell if a decision is based on facts or preferences
• Context brittleness – Values that make sense in one context don't in another

By putting values in their own dimensions, AGISystem2 can:

• Reason about facts with values masked out (veil of ignorance)
• Explain which values influenced a decision
• Swap value systems without relearning facts

Value Encoding

Axiological dimensions use the same Int8 range (-128 to +127) as ontological ones:

Value	Interpretation	Example
+127	Strongly positive	Very good, highly permitted
+64	Moderately positive	Generally good
0	Neutral / unknown	No value judgment
-64	Moderately negative	Generally bad
-128	Strongly negative	Very bad, forbidden

Deontic Logic Integration

The deontic axiological dimensions integrate with deontic reasoning:

# Assert something is forbidden
ASSERT Speeding PROHIBITED_BY TrafficLaw
# Sets deontic dimensions to negative values

# Query deontic status
ASK "Speeding PERMITTED?"
# → FALSE_CERTAIN (deontic dims strongly negative)

Implementation

// Axiology-only mask: only dims 256-383 active
const axiologyMask = new Uint8Array(384);
axiologyMask.fill(0, 0, 256);   // dims 0-255 = 0
axiologyMask.fill(1, 256, 384); // dims 256-383 = 1

// Pure value comparison (ignores facts)
const valueDist = mathEngine.maskedL1(conceptA, conceptB, axiologyMask);

Related Documentation

• Ontology – The factual dimension partition
• Bias – Fact/value separation for fairness
• Deontic Logic – Permitted/forbidden reasoning
• Veil of Ignorance – Reasoning without value influence
• BiasController – Managing value masks