Constitutional Consensus

Principle II: Constitutional Consensus

Agreement through shared commitment to correct principles.

Consensus mechanisms in distributed systems typically fall into two categories: those based on voting (democratic but slow) and those based on authority (fast but fragile). Obsidian introduces a third option: consensus through constitutional alignment.

How It Works

When multiple agents need to coordinate — and they always do, eventually — they do not vote. They do not defer to a leader. They evaluate proposed actions against the Constitution and accept those that are constitutionally valid.

This sounds almost too simple to work. It works because the Constitution is not a set of vague aspirations — it is a precise specification of acceptable behavior. When two agents disagree, the Constitution provides an unambiguous resolution mechanism. Not “what does the majority think?” but “what does the Constitution require?”

The Elegance of Constraint

Most consensus protocols add complexity to handle disagreement. Constitutional consensus removes complexity by eliminating entire categories of disagreement. Agents cannot argue about whether an action is acceptable if the Constitution already answers that question.

The Warden serves as the constitutional interpreter of last resort, but in practice, most consensus is reached without Warden intervention — because the Constitution is clear enough that agents can evaluate compliance independently.

Why Not Just Vote?

Because voting optimizes for preference, not correctness. A majority of agents can prefer an incorrect approach. The Constitution prevents this by establishing correctness criteria that are independent of agent preferences.

This is, admittedly, a strong claim. It rests on the assumption that the Constitution’s principles are actually correct — not merely popular, not merely expedient, but correct. That assumption is the foundation of everything Obsidian does.