Today's digital economy records transactions well and preserves accountability poorly. Authority is separated from execution; automated agents act across systems; receipts become logs controlled by the same platform that produced them; and responsibility disappears between organizational and technical boundaries.

DieWan is building an Accountability Economy: an economy in which a person, institution, or autonomous system can carry and exercise authority together with portable evidence of:

• Mandate — who authorized an action, what was authorized, and under which limits
• Action — what was attempted or completed
• Receipt — the evidence binding the action to its mandate
• Verification context — the exact rules, checkpoint, finality policy, proof provider, freshness bound, and trust mode under which a conclusion was reached
• Residual responsibility — what remains uncertain, disputed, reversible, or owed

The goal is not more surveillance, one universal ledger, or an algorithm that replaces judgment. It is better-structured proof: authority, action, verification, observation, and interpretation remain separate so that evidence can travel without responsibility disappearing.

The first constitutional primitive is now implemented: portable, recipient-verifiable non-equivocation for proof-carrying state transitions.

An isolated recipient can verify that a successor consumes a uniquely ordered source state and is bound to the delivered destination assignment—without trusting the sender's replay database. The V2 stack includes:

• canonical DAG and state-transition identities;
• disjoint consumed-state and repeatable-evidence references;
• immutable, non-downgradable exclusivity semantics;
• canonical proof-carrying consignments;
• typed closure proofs and finalized checkpoints;
• dimensioned assurance instead of authoritative booleans;
• atomic acceptance, conflict reasons, crash recovery, and reorganization downgrade/revocation;
• a published hostile conformance package and reason-code registry;
• reference/testnet source-closure adapters for Bitcoin signet, Ethereum Sepolia, Sui testnet, Aptos testnet, and Solana devnet.

These are reference and testnet capabilities, not a mainnet-production or legal-finality claim. Every conclusion remains bounded by its named verification context and trust assumptions.

flowchart LR
  PIT[Piteka<br/>authorize · execute · receipt] -->|canonical artifacts through SDK| PAR[Parwana<br/>meaning · closure · verification]
  PIT -->|signed evidence| TUP[Tuppira<br/>observe · index · lineage]
  TUP -->|read model| HEM[Hemion<br/>inspect · verify locally]
  HEM -->|local verification context| PAR
  CHAINS[Chain ordering / registries] -->|proof material| PAR
  TUP -.->|observes| CHAINS

Many systems make a platform, custodian, bridge, contract, or committee the final judge of validity. DieWan takes a different path:

1. Authority is represented in client-held, proof-carrying state.
2. A single-use condition constrains how it may be exercised.
3. The action consumes or updates that state.
4. Closure evidence binds the successor to an explicitly ordered source.
5. The recipient verifies the complete artifact before acceptance.
6. Observers may index what happened, but observation never upgrades itself into verification.

Proof travels. Custody does not. Verification stays local.

Portable non-equivocation is the foundation, not the final destination.

1. Accountable action — bind mandate, execution, receipt, and residual responsibility across humans, institutions, agents, and chains.
2. Accountable disagreement — preserve and reward reproducible incompatibilities without forcing premature consensus.
3. Emergent capability — describe a capability as a reproducible payoff attributable to a configuration and context, not as a narrative label.
4. Exchange, only if warranted — consider coordination or markets only after comparable objects, provenance, demand, and resistance to gaming have been demonstrated.

The ordering matters. Economic and interpretive envelopes may build on integrity; they may not reinterpret an invalid signature, forged anchor, or conflicting exclusive closure as contextually valid.

• Integrity is not truth. A signature proves integrity and attribution under a key, not the truth of every assertion.
• Mandate before action. Technical permission is not organizational authorization.
• Receipts preserve uncertainty. Unknown, unavailable, indeterminate, revoked, and invalid remain different states.
• Verification is contextual and reproducible. Same evidence, rules, and checkpoint should produce the same result.
• Observation is not authority. An indexer records what it saw and how fresh that view is.
• Native guarantees stay visible. Different chains and proof providers keep their real trust and finality assumptions.
• Privacy is part of accountability. Evidence should be purpose-limited, selectively disclosed, and tenant-isolated.
• No application owns protocol meaning. Products consume the versioned Parwana contract; they do not recreate it.
• Human responsibility remains. Automation and cryptography make evidence stronger; they do not make responsibility disappear.

DieWan does not currently prove human sincerity, cognitive independence, one-key-one-person, global real-time absence of conflict for an offline recipient, legal equivalence, or market value. Indexed absence does not prove non-occurrence. A testnet/reference adapter is not a production mainnet deployment. An obstruction assessment is not truth, and an emergent-capability record is not an exchange.

For every consequential digital action, ask:

Who authorized this? What exactly was authorized? What happened? What proves the match? Under which context was it verified? What remains uncertain, disputed, or owed—and who remains responsible?

If a system cannot answer those questions, technical validity alone is not enough.