Authorization is everywhere. Applications contain it. Exchanges contain it. Identity systems contain it. Settlement systems contain it.
Yet authorization itself is rarely treated as a first-class network primitive. Permission is stored inside applications, institutions, and platforms. Coordination occurs because those systems retain permission on behalf of participants. Authorization remains embedded within larger systems rather than existing independently of them.
AON does not introduce authorization. It externalizes it. Authorization becomes independently addressable: created, propagated, referenced, verified, and consumed without becoming embedded within the coordinating layer.
Traditional systems attach execution to authority. An exchange executes because it retains permission. A processor executes because it retains permission. An application executes because it retains permission. Execution is an institutional privilege.
AON separates execution from authority. Participants publish authorization objects describing what may occur, under which conditions, and until when. Once an executable authorization graph exists; authorization, conditions, and proofs all present. Execution does not need to be assigned to a trusted operator.
Any participant may discover the graph, verify it, and execute the authorized action. The executor does not approve. The executor does not decide. The executor does not create legitimacy. The executor exercises authorization that has already become valid.
Execution becomes a permissionless role rather than an institutional privilege.
Independent systems frequently need to coordinate. Payments must settle. Assets must transfer. Identities must be verified. Obligations must be fulfilled. Historically, this coordination has required participants to surrender control to a coordinating institution.
Authorization remains attached to the participant. Independent systems coordinate through authorization objects, conditions, proofs, and receipts rather than through institutions that retain permission on behalf of users. Applications remain sovereign. Namespaces remain sovereign. Participants remain sovereign.
Coordination emerges through discoverable authorization relationships rather than transferred control.
Blockchains coordinate shared state. Ownership, balances, and state transitions become valid because participants agree upon a canonical state. Consensus is required because the network itself becomes a source of truth.
AON operates under different assumptions. The network does not maintain ownership. The network does not maintain balances. The network does not determine truth. Truth remains local to namespaces. State remains local to the systems that maintain it.
Consensus becomes unnecessary not because agreement is undesirable, but because the protocol does not coordinate the category of state that requires agreement. AON coordinates authorization relationships rather than shared state.
Historically, coordination and authority have been treated as synonymous. To coordinate with another system, you transfer control to a coordinating institution. That institution retains permission. Authority accumulates. This is not a flaw in any particular system. It is the structural consequence of how coordination has worked.
AON explores the possibility that coordination and authority are separable. If authorization can remain attached to the action being coordinated rather than becoming embedded in the coordinating layer, independent systems may coordinate without surrendering control. Custody becomes optional. Execution becomes competitive. Authority has no mechanism to accumulate because permission never persists beyond the action it was granted for.