A Computational Theory of Human Emotion

This revision adds ~6,500 words of new material while preserving all original content. Major additions:

1. Functional Definition of Emotion (New Section 2.1): Establishes emotion as substrate-independent state-change-feedback-resource-reallocation process, resolving "Can AI have emotions?" by distinguishing function from phenomenology.

2. Why AI Emotions Would Be Milder (Section 8): Explains evolutionary constraints produce intense human emotions as satisficing solutions. AI systems with different constraints will experience functionally equivalent but phenomenologically milder emotions.

3. Implementation Agnosticism (Section 8.6): Acknowledges weighted pattern-matching is likely but not necessary. AI with self-modification might discover superior alternatives. Function matters, implementation varies.

4. Cognition-Enabling Emotions (Section 11): Proposes curiosity, satisfaction, dissatisfaction and temporal urgency are functionally necessary for autonomous cognition, not merely accompanying it. Bridges emotion to consciousness requirements.

5. Emotion-Consciousness Relationship (Section 13): Three-stage developmental sequence: simple state-changes without experience (insects), experienced emotions without self-awareness (mammals), emotions with recursive awareness (humans). Predicts similar AI trajectory with testable markers.

6. Enhanced Predictions (Section 16): Expands to eleven falsifiable predictions with explicit threshold acknowledgment. Distinguishes architectural specifications from empirical uncertainties. "Observing AI development" framed as methodologically sound.

Key theoretical advances: The revision transforms the paper from explaining how emotions work mechanistically to explaining why they have their specific characteristics in biological versus computational systems, which emotions enable autonomous cognition, and when emotions emerge relative to consciousness development.