Incentives Under Partial Verification

Status: Draft / Notes Non-normative Builds on: Application Semantics Over Bounded Verification

Motivation

In a bounded-verification system, incentives cannot assume:

• global visibility
• instant finality
• universal agreement
• synchronized knowledge

This note describes how incentives function when truth is local, partial, and time-bounded.

Core Constraint

Incentives must reward:

• correct behavior relative to a verifier’s frontier
• not absolute or global correctness

There is no omniscient judge.

Incentive Units

Rewards and penalties are tied to:

• verifiable actions
• provable statements
• observable behavior

Not intent. Not global outcomes.

Local Rationality

Participants optimize for:

• their own verification horizon
• expected future verification
• refusal risk

Rational behavior is contextual, not globally optimal.

Proof-Linked Rewards

A reward must be claimable only if:

• the claimant supplies a valid proof
• the proof verifies within the verifier’s frontier
• the proof remains consistent until acceptance

Unprovable claims are worthless.

Delayed Reward Model

Because verification may be delayed:

• rewards may be provisional
• settlement may lag execution
• acceptance may expire

Applications must model rewards as pending until proven.

Refusal and Penalty Asymmetry

Failure to prove is not guilt.

Therefore:

• penalties must require positive proof of misbehavior
• non-proof is not punishable
• refusal defaults to no reward, not punishment

This preserves safety under uncertainty.

Slashing Constraints

Slashing is only safe when:

• misbehavior is provable within bounded verification
• evidence is independently verifiable
• ambiguity is eliminated

If ambiguity remains, slashing must not occur.

Incentives for Availability

Proof providers are incentivized by:

• successful proof delivery
• timely responses
• consistency across requests

Non-delivery results in lost opportunity, not punishment.

Free-Rider Tolerance

Some participants may:

• consume verification without contributing
• appear inactive

This is acceptable.

Forcing contribution under uncertainty creates perverse incentives.

Economic Finality vs Verification Finality

Economic systems may:

• settle probabilistically
• hedge against refusal
• require confirmations across multiple frontiers

Finality is economic, not absolute.

Cross-Verifier Incentives

When interacting across verifiers:

• rewards require overlapping verified facts
• disagreement blocks settlement
• reconciliation is explicit

Incentives must tolerate disagreement.

Offline Incentive Behavior

Offline participants may:

• accumulate unclaimed rewards
• delay settlement
• rejoin with cached proofs

Systems must allow delayed participation without loss of safety.

MEV Considerations

Bounded verification limits MEV by:

• reducing global visibility
• fragmenting ordering assumptions
• localizing execution

MEV exists, but is scoped.

Design Principle

Incentives should encourage:

• proof production
• availability
• honesty under partial knowledge

Not speed. Not dominance. Not coordination.

Boundary Statement

No incentive mechanism can guarantee:

• global fairness
• universal truth
• perfect alignment

Bounded verification demands bounded expectations.

What Follows

Once incentives operate under partial verification, governance must also function without global truth.

The next note addresses this directly:

0x09 — Governance Without Global Consensus