Mission Command: Decision Rights Under Uncertainty

Chapter V opens with the axioms. Command is "central to all military action"; inherent in it is lawful authority over subordinates, including the authority to assign missions — and accountability for their completion. Two design rules follow immediately. First, responsibility is non-delegable even when authority is delegated: commanders "may not absolve themselves of the responsibility" for missions whose execution they hand down. Second, authority is always bounded: its extent "is specified by the establishing authority, directives, and law." In Simon's (1947) terms, JP 1 writes the zone of acceptance down instead of leaving it implicit in custom — every grant of authority arrives with its limits attached. Unity of command adds a uniqueness constraint: all forces operate under a single commander with the requisite authority, and "two commanders may not exercise the same command relationship over the same force at any one time." Where unity of command is politically impossible — interagency work, coalitions — the doctrine downgrades honestly to unity of effort, "coordination through cooperation and common interests," which it treats as a complement, not a substitute.

The taxonomy itself is a nested set of authority bundles, each strictly contained in the one above it. Combatant command (command authority), or COCOM, is the full bundle: vested only in combatant commanders by Title 10, U.S. Code, Section 164, it covers organizing and employing commands and forces, assigning tasks, designating objectives, and "authoritative direction over all aspects of military operations, joint training, and logistics" — JP 1 enumerates twenty-four distinct elements of it, from prescribing the chain of command to convening general courts-martial. COCOM "cannot be delegated or transferred." Operational control (OPCON) is the same core — organize, employ, task, direct — minus directive authority for logistics, administration, discipline, internal organization, and unit training, and unlike COCOM it may be delegated at or below the combatant command level. Tactical control (TACON) shrinks further, to "the detailed direction and control of movements and maneuvers within the operational area"; it confers no organizational authority and no logistics. Support, the fourth relationship, is the interesting outlier — an authority JP 1 describes, remarkably, as "by design, a somewhat vague but very flexible arrangement," in which a common superior obligates one commander to aid, protect, complement, or sustain another.

Beneath the command relationships sit three authorities that explicitly are not command. Administrative control (ADCON) is direction over administration and support — organization of Service forces, resources, personnel management, logistics, readiness — the Military Departments' Title 10 responsibility. Coordinating authority grants the power "to require consultation between the agencies involved" but "not the authority to compel agreement": Mintzberg would file it under liaison roles and mutual adjustment, a coordination mechanism deliberately stripped of decision rights. Direct liaison authorized (DIRLAUTH) permits a subordinate to coordinate directly outside the chain — with the standing obligation to keep the granting commander informed. The taxonomy therefore separates three things most organizations blur into one reporting line: operational decision rights (who directs your actions), administrative decision rights (who feeds, trains, promotes, and disciplines you), and coordination rights (who may convene you without commanding you).

Chapter II specifies that the President and Secretary of Defense "exercise authority, direction, and control of the Armed Forces through two distinct branches of the chain" of command and control. The operational branch runs from the President through SecDef to the combatant commanders, for missions and forces assigned to their commands. The administrative branch runs from the President to SecDef to the Secretaries of the Military Departments, who — through the Service Chiefs — exercise ADCON: organizing, training, equipping, and supporting Service forces. A deployed Army brigade thus takes operational direction from a joint force commander while its personnel system, doctrine, equipment, and discipline flow through the Department of the Army. Organization theory would call this a two-dimensional matrix; JP 1's innovation is that the seam is governed by explicit rules rather than negotiation. Service component commanders are the designated interface: they answer to the combatant commander "on all matters for which the CCDR has been assigned authority" and to their Service Chief on "Service-specific matters such as personnel, administration, and unit training," with defined escalation paths — an unresolved logistics dispute goes from the CCDR through the Chairman of the Joint Chiefs of Staff to SecDef for resolution.

Movement between commands is likewise contractualized. The doctrine's rule is compact: "Forces, not command relationships, are transferred between commands," and every transfer must specify the command relationship the gaining commander will exercise and the losing commander will relinquish. Duration determines the default bundle — permanent or indefinite transfers mean assignment, with the gaining CCDR exercising COCOM; temporary transfers mean attachment, normally with OPCON. Only SecDef can direct a transfer, under procedures approved by the President. The result is a delegation cascade in which each echelon receives a defined bundle and passes a smaller one downward.

1. 01 President / Secretary of Defense
   Assign forces and missions. The Unified Command Plan establishes combatant commands; the Global Force Management Implementation Guidance assigns forces. Orders flow to CCDRs through the Chairman, who transmits but does not command.
2. 02 Combatant commander (COCOM)
   Prescribe the chain, delegate the rest. The CCDR prescribes the chain of command within the CCMD, organizes subordinate commands (subunified commands, JTFs, Service and functional components), and delegates OPCON or TACON — but never COCOM.
3. 03 Subordinate JFC (OPCON)
   Organize for the mission. A JTF commander exercises OPCON over assigned and normally attached forces, establishes command and support relationships between subordinate commanders, and may further delegate.
4. 04 Functional component (OPCON / TACON)
   Task what is made available. A functional component commander normally holds OPCON over own-Service forces and TACON over forces other components make available for tasking — the thinnest command bundle, matched to the narrowest scope.

Two features of this architecture deserve emphasis. First, the design embeds state-contingent authority: in a major emergency, a geographic combatant commander "may temporarily assume OPCON of all forces in the assigned AOR, including those of another command" — an authority that itself "will not be delegated," and that triggers immediate notification duties to the Chairman, affected commanders, and Service Chiefs. The org chart contains its own crisis mode, pre-authorized and bounded, rather than relying on improvisation when the contingency arrives. Second, even the vague support relationship is disciplined by paperwork: the establishing directive states the purpose, effect, and scope of support, the forces committed, its priority, the supporting commander's authority to modify the effort "in the event of exceptional opportunity or an emergency," and precisely how much authority the supported commander is granted. The supported commander gets general direction of the effort; the supporting commander retains choice of "forces, tactics, methods, procedures, and communications." That split — ends to the customer, means to the provider — is the taxonomy's smallest-scale rehearsal of mission command itself.

Note what the justification is not. It is not that subordinates deserve autonomy, nor that decentralization is intrinsically virtuous. It is a bounded-rationality claim: central control of detail is impossible at scale, so the only design question is whether the inevitable decentralization is organized or accidental. This is Aghion and Tirole's overload result stated as doctrine — a principal who cannot process everything will lose real authority over most decisions regardless of what the org chart says; the choice is whether to formalize that loss on favorable terms. JP 1 formalizes it. A joint force commander's general responsibilities are a delegation contract in four clauses: provide "a clear commander's intent and timely communication of specified tasks" with tasks "realistic yet leav[ing] subordinate commanders flexibility in their CONOPS and the ability to take the initiative as opportunities arise"; transfer forces to subordinates for their tasks; "provide all available information to subordinate JFCs and component commanders that affect their assigned missions"; and "delegate authority to subordinate JFCs and component commanders commensurate with their responsibilities." Authority commensurate with responsibility, plus an affirmative duty to share information downward — the two transfers, of rights and of information, travel together.

The subordinate's side of the contract is equally explicit. Subordinates accomplish assigned missions, and "authority normally is given to the subordinate commander to select the methodology for accomplishing the mission." But method-freedom carries notification duties: "Any departure from the plan by a subordinate commander should, if possible, be coordinated with other commanders prior to departure from the plan. In addition, the departure must be communicated as soon as practicable to the superior." Subordinates also owe advice upward and "timely information" on their situation and progress; commanders who merely share a boundary owe each other consideration of second-order effects, timely information about their own intentions, and mutual support "as required by the common aim." In incomplete-contract terms this is the renegotiation protocol: the contract anticipates its own violation, and prices deviation not as disobedience but as a reporting obligation.

Organizational form is itself treated as a delegated design variable. The JFC "has the authority to organize assigned or attached forces … to best accomplish the assigned mission based on his intent" — choosing among six standard options (subunified command, JTF, Service component, functional component, single-Service force, specific operational force) under four stated principles: simplicity, span of control, unit integrity, and interoperability. Span of control is explicitly multi-causal — "the number of subordinates, number of activities, range of weapon systems, force capabilities, size and complexity of the operational area, and method used to control operations (centralized or decentralized)" — a contingency statement Graicunas and Urwick would recognize. And the doctrine flags the delegation-eroding temptation directly: JFCs may establish functional components precisely to "integrate planning; reduce their span of control; and/or significantly improve combat efficiency, information flow, unity of effort", and "may elect to centralize selected functions within the joint force, but should strive to avoid reducing the versatility, responsiveness, and initiative of subordinate forces." Centralization is available, priced, and warned against in the same sentence.

The doctrinal placement matters: mission command appears in Chapter I as a property of the C2 joint function ("Mission command is the preferred method of exercising C2") and again in Chapter V as the second of ten C2 tenets — after "clearly defined authorities, roles, and relationships" and before information management, communication, timely decision making, coordination mechanisms, battle rhythm discipline, dependable support systems, situational awareness, and mutual trust. The ordering is an argument: decision rights must be clear before they can be pushed down, and everything after mission command in the list is infrastructure for making the push-down survivable. The same passage is candid about contingency — "The level of control used will depend on the nature of the operation or task, the risk or priority of its success, and the associated comfort level of the commander." Doctrine admits that the delegation dial is set partly by task risk (a contingency variable) and partly by commander psychology (which is not).

The instrument that makes this delegation safe is the mission-type order, and JP 1's definition is a textbook incomplete contract: it "direct[s] a subordinate to perform a certain task without specifying how to accomplish it," containing a mission statement of essential task and purpose — the what and the why — while "the senior leaves the details of execution to the subordinate, allowing the freedom and the obligation to take whatever steps are necessary to deal with the changing situation." Freedom and obligation: residual decision rights over method are not permission, they are assigned work. A JP 1 mission statement always carries five parts — who, what, when, where, and why — with how conspicuously absent. Commander's intent is the clause that governs the states of the world the order cannot enumerate.

Read through Aghion and Tirole, mission command is the deliberate version of what happens to overloaded principals accidentally. Their model predicts that when the principal cannot evaluate every proposal, informed agents acquire real authority whether or not it is granted — and that a principal who can credibly commit not to intervene buys more agent initiative and information-gathering. JP 1's design keeps formal authority concentrated (accountability, courts-martial jurisdiction, the non-delegable COCOM anchor) while committing, in doctrine, to non-intervention in method. The empirical warrant for that bet is the Bavelas–Leavitt lineage from Session 04: centralized networks win on simple tasks, decentralized structures on complex and changing ones — and JP 1 classifies its own environment in Chapter I as "uncertainty, complexity, rapid change, and persistent conflict." The doctrine even installs a judgment override on itself: joint doctrine "will be followed except when, in the judgment of the commander, exceptional circumstances dictate otherwise. … The judgment of the commander based upon the situation is always paramount." The rulebook's final rule allocates the residual decision right against the rulebook.

Sections 04–05 derived mission command's mechanics from JP 1's ten C2 tenets. That derivation is sound but incomplete, because the canonical list is not JP 1's. ADP 6-0 — the capstone Army publication, absent from this session's first construction — enumerates seven named principles of mission command, and three of them are things the JP 1 tenet treatment underweighted: competence (the tactical and technical proficiency without which delegation is reckless), shared understanding (promoted from a downstream requirement to a first-class precondition), and risk acceptance (named as a duty of the delegating commander, not merely a property of the situation). Where JP 1 lists mission command as the second of ten tenets and calls it "the preferred method of exercising C2" — one setting on a dial — ADP 6-0 states it categorically: "Mission command is the Army's approach to command and control" — not a method to select per task, but the standing posture from which control is relaxed or tightened. The difference is one of emphasis and evolution rather than flat contradiction (both still scale the level of control to the situation), but the emphasis matters: ADP 6-0 makes mission command the default and detailed control the exception.

The second correction is sharper and is primary evidence for this seminar's doctrine-as-memory thesis. ADP 6-0's Introduction is an admission of a doctrinal round trip. From 2012, the Army had replaced the term command and control with "mission command," using the one label for a warfighting function, a system, and a philosophy at once. By 2019 it had reversed course, rescinding "mission command warfighting function" and "mission command system," reintroducing command and control, and adopting the joint (JP 1) definition — explicitly to "align the Army with joint and multinational partners." The institution changed the names of its own decision-rights vocabulary twice in seven years, and wrote down why.

The logic map ADP 6-0 places at its front encodes the settled relationship the terminology fight was about: mission command is the approach nested inside command and control, which the command-and-control warfighting function then supports with people, processes, networks, and command posts. Mission command is neither identical to C2 nor a subordinate box — it is the philosophy that governs how C2 is exercised.

The Marine Corps states the same design as a warfighting philosophy rather than a set of principles, and in doing so supplies the incomplete-contract framing this session has used throughout. MCDP 1, Warfighting derives mission command not from a taxonomy of authorities but from the nature of maneuver warfare: because war is disorderly and the subordinate at the point of decision "will naturally better appreciate the true situation," command and control "must be decentralized." Its instrument is "mission tactics" — assigning a mission "without specifying how the mission must be accomplished" — and it names the arrangement in exactly the terms Simon and Aghion–Tirole would use.

MCDP 1 is the philosophy's purest statement precisely because it refuses to make initiative a procedure. It demands boldness and initiative "down to the lowest levels," insists that "errors by junior leaders stemming from overboldness are a necessary part of learning," and rules that the force "will not accept lack of orders as justification for inaction." That last clause is the mirror image of the whole delegation contract: where JP 1 prices deviation as a reporting obligation, MCDP 1 prices inaction as a fault. What sustains such a philosophy is not doctrine but people who can be trusted to exercise judgment — which is why the Marine Corps issues a companion learning doctrine, MCDP 7, Learning, that treats continuous learning as the institutional means of manufacturing ADP 6-0's competence principle and, through it, the trust the whole design rests on.

NDP 1 grounds mission command in the material history of the domain: "In the not-too-distant past, Navy and Coast Guard sea captains and Marine commanders had to operate for months at a time with minimal communication back to higher headquarters. This cultivated a rich heritage of mission command that is a particular necessity in naval warfare." It then sorts the theoretical field into three approaches to command: command by direction (the commander observes the battlespace and personally directs forces — the approach "applied since antiquity"), command by planning (attributed to Frederick the Great — "the commander tries to plan every move in advance, relying on highly trained forces and strict discipline"), and command by influence, "also known as 'mission command,' in which the commander outlines desired objectives for subordinates to accomplish and then relies on them to exercise on-scene initiative based on local situational awareness and lowered decision thresholds." The three are combined in practice — planning "provide[s] the framework that enable[s] subordinate commanders to understand the mission and higher commander's intent" — but the distinction between them is drawn on one axis, and it is Galbraith's axis from Session 04:

NDP 1 is unusually frank that this preference must be defended against its own technology. It quotes the historian Michael Palmer: "the natural tendency of most commanders has been and will always be, to control, to centralize; and yet the decentralization of command functions … generates greater fighting power." It notes that centralizing trends have arisen before and "will likely continue to appear, usually precipitated by advances in communication technology," and answers with Nelson and Arleigh Burke, who "shunned centralization … during periods when much-heralded communications technology seemed to offer the prospect of ever greater centralized control." The exceptions are named rather than smuggled: "the most prominent being command over nuclear weapons and restrictions on the delegation of certain information-related capabilities or special legal authorities" — precisely the simple-task, catastrophic-error domains where the Leavitt result favors the wheel. And the enforcement mechanism for initiative is a stated asymmetry in blame: "a subordinate who takes the initiative but makes an error in judgment is not to be faulted to the same degree as one who fails to act entirely. Propagating the initiative of the subordinate can only be achieved by encouraging the subordinates who demonstrate initiative." That sentence is Edmondson (1999) as doctrine — a codified guarantee that the interpersonal risk of acting is lower than the institutional risk of freezing, which is the condition under which delegated rights actually get exercised.

The publication's opening vignette is the design's existence proof. Off Samar on 25 October 1944, Commander Ernest E. Evans of USS Johnston — a destroyer screening escort carriers — sighted a Japanese force of four battleships, eight cruisers, and eleven destroyers and, "without waiting for orders," attacked: "All hands to general quarters. Prepare to attack major portion of the Japanese fleet." For more than three hours Taffy 3's destroyers and aircraft pressed the attack until the Japanese force, convinced it faced something far larger, turned away; Evans went down with his ship and received the Medal of Honor posthumously. No order to attack existed or could have arrived in time; what existed was intent, doctrine, and a norm that made acting the default. NDP 1's epigraph for the episode is Harry Yarnell's formula — "From skill and doctrine flows the initiative of the subordinate" — and Vice Admiral Robert Carney's 1948 retrospective on the Pacific war generalizes it: the separated commanders "were thinking in sufficiently like terms … so that even without authoritative coordinating command, they instinctively moved in directions which were basically sound." Real authority operated at the periphery because the center had spent years, not minutes, transferring it.

The primer's doctrine chain runs: ADRP 6-0 gives commanders three tasks — drive the operations process, develop teams, inform and influence audiences — and the first is "directly enabled" by knowledge management, whose Army definition is "the process of enabling knowledge flow to enhance shared understanding, learning, and decision making." Its objective sentence is a delegation-theoretic claim in logistics clothing: "Getting the right information, to the right people, in the right format, at the right time, in order to arrive at the right decision." The three observed failure modes are all failures of the shared-understanding substrate, not of the authority design: units at the training centers hold the same decision rights as everyone else, but their common operational picture "was not effective in contributing to shared understanding during movement or when troops were in contact with the enemy," their battle rhythms "do not follow a critical path" to commander decisions, and their information is not reliably accessible. Where Session 04's cognitive hierarchy was a design for individual understanding, the primer's version is explicitly collective — and it prices the failure: decision behavior degrades from anticipatory (low risk) to reactionary (high risk) as an organization slides down the ladder from shared understanding toward raw data.

The primer's positive prescription is that the commander must be the "knowledge leader": "Mission command strives to establish the mindset that situational understanding comes from a balance of bottom-up and top-down knowledge processes," and leaders "must drive KM as much as they drive the operations process" — beginning with the most essential guidance a senior leader can give, "to clearly identify the knowledge they require for decision making." In Aghion–Tirole terms the argument is exact: formal delegation without information transfer produces no real authority anywhere. A subordinate who cannot see the situation cannot exercise disciplined initiative; a commander who cannot see the force cannot supervise by exception; and so the delegation contract of Sections 04–05 has an unstated schedule attached — the common operational picture, the battle rhythm, the accessible record — without which both parties default. That schedule, and the machinery the Army built to maintain it (the KM officer, the working group, the five-step process, the maturity model), is Session 06's subject.

1. Is the non-delegable anchor a feature or a bug? COCOM cannot be delegated, and responsibility survives every delegation of authority. Aghion–Tirole suggest retained formal authority can chill agent initiative — the principal can always overrule. Does the military's combination (formal authority anchored, non-intervention promised by doctrine) capture the best of both, or does an always-available override invite exactly the selective micro-intervention that destroys the initiative delegation was meant to buy?
2. "The associated comfort level of the commander." JP 1 lists commander comfort alongside task nature and risk as a determinant of the level of control. Is this honest contingency design — trust is real capital, and its stock varies — or a doctrinal loophole that lets leaders practice detailed control while preaching mission command? Recall the failure mode from Session 01's frame: permission-seeking subordinates produced by leaders with constant digital visibility. Which reading does the NDP 1 centralization history support?
3. Deliberate vagueness as contract design. The support relationship is "by design, a somewhat vague but very flexible arrangement," disciplined only by an establishing directive. When is vagueness in an authority contract efficient — preserving adaptation under uncertainty — and when does it merely relocate conflict to the establishing authority who must adjudicate it? Compare dotted-line reporting in corporate matrices, which is usually vague by neglect rather than by design. Does writing the vagueness down change its behavior?
4. Can trust capital survive personnel turnover? JP 1 says mutual trust results from "demonstrated competence, and planning and training together"; NDP 1 derives initiative from "trust established through interaction with commanders." Military rotation cycles move people every two to three years. Is mission command's real trust interpersonal at all — or is it institutional, resting on the standardization and doctrine codebook of Session 03, so that any two strangers who share the codebook can transact on "trust" they never personally built? What would that imply for firms that want mission command without a doctrine system?
5. Does the digital COP re-centralize real authority? NDP 1 documents that every advance in communications technology precipitates a centralizing trend; the primer demands a live common operational picture as mission command's precondition. But if headquarters can see everything the subordinate sees, the information asymmetry that grounded the subordinate's real authority (Aghion–Tirole) evaporates. Is the shared picture a delegation-enabler, a delegation-destroyer, or both depending on who controls the interruption channel?
6. Does mission command go far enough? The edge debate. Alberts and Hayes (2003) argue that in a robustly networked force, command and control should be unbundled and pushed to "the edge," where nearly autonomous units self-synchronize on shared awareness and intent — dissolving the middle management that mission command still preserves. Mission command keeps delegation bounded (commander's intent as a limit), responsibility non-delegable, and the hierarchy anchored; power-to-the-edge would call that a failure of nerve, and mission command would call self-synchronization a recipe for desynchronized action without accountable command. Who is right, and does the answer turn on network reliability — Alberts and Hayes assume a robust network, while NDP 1 grounds the same heritage in commanders who had to operate "for months at a time with minimal communication"?
7. The error-forgiveness asymmetry. "An error in judgment is not to be faulted to the same degree as [a failure] to act entirely." Read as Edmondson: a codified psychological-safety guarantee, without which delegated rights go unused. Read as an economist: a moral-hazard subsidy for action over restraint, in a domain where action can be irreversible. Where should a hospital, a nuclear plant, or an SRE team set this asymmetry — and does the military's own answer (nuclear release is command-by-planning territory) concede the point?
8. Unity of effort — mechanism or euphemism? Where unity of command is politically impossible, JP 1 substitutes unity of effort, and coordinating authority can "require consultation … but does not have the authority to compel agreement." Is this a genuine coordination mechanism (Mintzberg's mutual adjustment, formalized) or an admission that no one decides? Test against the multinational case: JP 1 itself warns that parallel command structures without a single commander "should be avoided if possible."
9. The transfer question. Commander's intent works because Sessions 03–04 built its interpretive substrate: shared doctrine, common terms, pre-loaded common ground. Firms adopting "commander's intent" via OKRs or strategy memos typically copy the artifact without the substrate — no common professional education, no shared planning ritual, no error-forgiveness norm. Predict, from this session's theory, exactly how that fails. Is "empowerment theater" (rights formally delegated, never really exercisable) the modal outcome — and what is the minimum viable substrate?