Four Classes of Promise the Architecture Is Currently Carrying, and the Instrument That Records Them Before They Are Invoked.
Every architecture carries a liability that appears on no balance sheet: the full set of promises it is currently obliged to keep.
Some of those promises are written. Most are not. A few were made by people who are no longer present. Several were made by nobody at all, in the sense that no one ever said them out loud, and they are binding anyway because the architecture has behaved as though they were true for long enough that a counterparty is entitled to rely on it.
The operator economy tracks this liability with a contracts folder and a good memory. Both instruments fail for the same reason. They record only the class of promise that was formally executed, which is the smallest and least dangerous of the four, and they leave the other three to accumulate unobserved until one of them is invoked at an inconvenient moment.
This essay introduces the framework. The Commitment Ledger. Four classes of promise. Explicit. Implied. Inherited. Self. Each class has a different origin, a different mechanism of accumulation, a different cost of breach, and a different discipline of retirement.
The ledger is not a metaphor. It is a document, kept, reviewed on a cadence, and inherited by the next custodian along with everything else. The architect who keeps one operates with a clarity about their own obligations that most operators never achieve, and the clarity is directly convertible into the structural condition Monday named. Leverage that has been promised away is not leverage.
This is the second framework of Volume IX. It pairs with Monday’s Leverage Architecture and stands underneath Friday’s examination of margin, because the largest single cause of quiet margin compression is a promise the architecture is still honouring and no longer priced.
Why the contract framing fails
Three reasons the contract framing fails the architect.
First reason. It covers a minority of the liability. In any architecture of reasonable age, the executed contracts represent perhaps a third of the promises the counterparties believe they hold. The remainder lives in patterns of behaviour, in sentences spoken in meetings, in arrangements that predate the current ownership, and in the architect’s own private undertakings. A register that covers a third of a liability is not a register. It is a partial inventory that produces false confidence.
Second reason. It records formation and never records retirement. Contracts are signed and filed. They are rarely revisited to ask whether the obligation is still appropriate, still priced correctly, or still wanted. Promises therefore accumulate monotonically. An architecture at year fifteen is carrying obligations that were sensible at year four and have been quietly subsidised ever since.
Third reason. It treats the promise to self as a separate category of thing, belonging to personal discipline rather than to structure. This is the most consequential of the three errors. The undertakings an architect makes to themselves are structural instruments and not private sentiment (Volume VI’s Vow). An architect whose self promises are routinely revised without record will, in time, hold the other three classes with the same elasticity, because the underlying capacity is the same capacity.
The ledger corrects all three. It covers the full liability, it forces an act of retirement, and it holds all four classes in one instrument, on one cadence, in one hand.
Class One. Explicit.
Explicit promises are the ones that were stated, in words, to a named party, and can be produced.
They include the executed contract and everything else in that register: the delivery date given on a call, the price held open in an email, the public statement about what the architecture will do, the commitment made to a team member about a role, the assurance given to a capital counterparty about a reporting cadence.
The class is the easiest to track and the most expensive to break, which is a fortunate combination. It is also the class where architects most often assume they are complete and are not, because the formal contracts are filed and the verbal explicit promises are not.
Four properties to audit at the explicit class.
First. Inventory completeness. The architect can produce the full list, including the promises made verbally in the last quarter. The test is whether the list is longer than the contracts folder. If it is not, the inventory is incomplete rather than clean.
Second. Precision. The promise is specific enough that both parties would agree on whether it had been kept. Vague explicit promises are the most dangerous entries in the ledger because both parties leave the room satisfied and holding different documents.
Third. Capacity backing. Every explicit promise has capacity allocated against it. A promise with no allocated capacity is a forecast the counterparty has been told is a commitment.
Fourth. Expiry. The promise has a defined end, or the architect has consciously accepted that it does not. Open ended explicit promises are the ones that become subsidies.
The dominant failure mode at the explicit class is the unrecorded verbal commitment. The architect says something reasonable in a meeting, does not write it down, and forgets it inside a fortnight. The counterparty writes it down and remembers it for three years, usually producing it at the moment it is least convenient.
Class Two. Implied.
Implied promises are the ones the architecture made by behaviour rather than by statement.
No one said the architect would respond within the hour. The architect responded within the hour, for two years, and a counterparty is now entitled to be surprised when the response takes a day. No one said the discount would apply again. It applied three times. No one said the founder would attend the quarterly meeting. The founder attended fourteen of them.
Pattern is promise. This is the mechanism the operator economy consistently underestimates, and it is the reason architectures accumulate obligations far faster than they accumulate contracts. A courtesy repeated becomes an expectation. An expectation held becomes a standard. A standard breached becomes a grievance, and the grievance is genuine, because the counterparty did not invent the expectation. The architecture taught it to them.
Four properties to audit at the implied class.
First. Source pattern. The architect can name the specific behaviour that created the expectation. If the behaviour cannot be named, the promise cannot be withdrawn cleanly.
Second. Counterparty reading. The architect knows whether the counterparty actually holds the expectation. Some patterns are noticed and some are not. Auditing the ones nobody noticed is wasted effort.
Third. Cost of continuance. What honouring the implied promise costs annually, in capacity and in margin. This number is almost never calculated and is almost always larger than expected.
Fourth. Withdrawability. Whether the promise can be ended, at what notice, and with what conversation. Implied promises are withdrawable, but only at a boundary and only by an explicit act.
The dominant failure mode at the implied class is the accidental standard. The architecture creates a durable obligation through repeated generosity, never prices it, never records it, and eventually resents it. The resentment is misdirected. The counterparty did nothing except accept what was repeatedly offered.
Class Three. Inherited.
Inherited promises are the ones the architect did not make.
They arrive with an acquisition, with a succession, with a departing executive who committed the architecture to something before leaving, with a founder’s arrangement that predates the current structure, and sometimes with an industry convention that the architecture has never examined and simply assumed applied to it.
The architect who takes over an architecture inherits obligations along with assets, and the act of custodianship includes a deliberate reckoning with what was inherited (Volume III’s Custodian). This class is where that reckoning is performed. It is performed poorly almost everywhere, because inherited promises are invisible until invoked and the person who could explain them has usually gone.
Four properties to audit at the inherited class.
First. Provenance. Who made the promise, to whom, when, and under what conditions. Provenance determines whether the promise is even coherent under present circumstances.
Second. Bindingness. Whether the obligation is legal, moral, or reputational. All three are real. They carry different costs of breach and different instruments of retirement, and conflating them produces both unnecessary honouring and careless breach.
Third. Currency. Whether the promise still fits the architecture it now sits inside. Many inherited promises were correct for a smaller, differently shaped organisation and are structurally wrong for the present one.
Fourth. Disposition. Whether the architect has decided to keep it, renegotiate it, or retire it, and whether that decision has been communicated. An undecided inherited promise is the most dangerous entry in the entire ledger.
The dominant failure mode at the inherited class is silent assumption. The architect neither honours nor retires the promise. It sits, undisturbed, for years, until a counterparty invokes it in a negotiation where the architect has no information about its origin and no position from which to dispute it.
Class Four. Self.
Self promises are the undertakings the architect has made to themselves.
The daily architecture they committed to hold. The category of work they said they would stop taking. The hour they said would be protected. The standard they said would not be traded for a quarter. The number below which they said they would not price.
The Vow (Volume VI) develops the substance of the self undertaking at length. What the ledger adds is the record. A vow without a record is not a vow. It is an intention that will be revised without anyone, including the architect, noticing the revision occurred.
This class governs the other three. An architect who breaks promises to themselves weekly is developing a general capacity for elastic commitment, and the capacity does not stay in one class. The consistency layer (Volume V’s Trust Architecture) is built or eroded here first, in private, long before any counterparty observes anything.
Four properties to audit at the self class.
First. Specificity. The promise is stated in terms that make breach observable. A commitment to work more strategically cannot be breached. A commitment to hold Tuesday mornings unscheduled can.
Second. Witness. Someone, or some record, can observe the promise. The Circle exists in part to supply this property, because a room of peers at compatible altitude is the only witness structure most architects will accept.
Third. Cost of breach. What actually happens when the promise is broken. In most architectures the answer is nothing, which is why the promise is broken.
Fourth. Continuity under stress. Whether the promise holds in a thin quarter. Self promises that hold only in comfortable conditions are descriptions of preference rather than commitments.
The dominant failure mode at the self class is the silent revision. The architect does not break the promise. The architect amends it, quietly, to match what they did. No breach is recorded, no cost is paid, and no learning occurs, and after several years of this the architect’s stated architecture and operated architecture have separated entirely.
The integration of the four classes
The four classes form one liability and behave differently under load.
Explicit promises are cheap to track and expensive to break. Implied promises are expensive to track and cheap to break individually, which is precisely why they are dangerous. Each single breach costs little and the aggregate pattern costs everything, because the pattern is what counterparties read. Inherited promises cost nothing until they are invoked, at which point they cost whatever the invoking party decides they cost, because the architect has no provenance with which to argue. Self promises cost nothing externally and determine the architect’s capacity to hold the other three.
The ordering runs upward from the self class. An architect who holds their own undertakings develops the internal instrument that makes the explicit class trivial to honour. An architect who does not will find that the explicit class requires continuous effort, because there is no underlying disposition doing the work.
The gate is supplied by an instrument already in the catalogue (Volume VII’s Decision Architecture). Every new promise is a decision, and its categories determine how much deliberation the promise warrants before it enters the ledger. A promise made at the wrong deliberation level is the most common origin of an entry the architect later cannot honour.
The connection to Monday’s framework is direct. Each of the four classes converts, over time, into transferred leverage. An implied promise of availability transfers time. An inherited promise of pricing transfers alternative. An explicit promise made without capacity transfers position, because the architect will be negotiating from a place of having already failed.
The ledger itself
The instrument is a document, and it is deliberately plain.
Four sections, one per class. Each entry records the promise, the party, the origin, the annual cost of continuance, the expiry or the absence of one, and the disposition. Explicit entries are added at the moment they are made, including the verbal ones, which is the single habit that most improves the instrument.
The implied section is the one that requires deliberate construction, because nothing prompts its entries. It is built by asking, of each material counterparty, what they would be surprised by if the architecture stopped doing it. The surprises are the entries.
The inherited section is built once, thoroughly, at the point of custodianship, and reviewed annually. It is the section most likely to be empty in architectures where it should be longest.
The self section is short and is the section the architect is most tempted to omit. It should be written in the same document as the other three, in the same hand, because the point of the ledger is that the four classes are one liability.
The ledger is kept in a form the next custodian can read without explanation. This is not administrative tidiness. It is the difference between transmitting an architecture and transmitting an unexplained set of obligations that the successor will discover one invocation at a time.
The audit cadence
The Commitment Ledger runs on two clocks.
Quarterly, the architect reviews the explicit and implied sections. These accumulate fast. A quarter is roughly the interval at which unrecorded verbal commitments become unrecoverable and at which a repeated courtesy hardens into an expectation.
Annually, the architect reviews the inherited and self sections. The annual questions are direct. Which inherited promises remain undisposed. Which self promises were silently revised rather than kept or broken. What is the total annual cost of continuance across the implied section, and would the architect agree to that figure if it were presented as a new proposal.
The last question is the instrument’s sharpest edge. Most architectures, presented with the true annual cost of their implied promises as a fresh proposition, would decline it without hesitation. They continue paying it because it was never presented, only accumulated.
A note on the frameworks already in the catalogue
The Commitment Ledger sits alongside three frameworks the Circle has published and completes something each of them leaves open.
Consistency has been described as a layer built across years (Volume V’s Trust Architecture). The ledger is the instrument that makes consistency inspectable rather than aspirational, because consistency is precisely the ratio of promises kept to promises made. How a promise enters is governed elsewhere (Volume VII’s Decision Architecture). What happens to promises at transmission is governed elsewhere again (Volume III’s Custodian). The fourth class has a doctrine of its own (Volume VI’s Vow).
Together they describe a single discipline. The architect knows what they have promised, to whom, at what cost, with what expiry, and has decided about each one rather than inherited it by momentum.
Doctrine close
The architecture’s largest unrecorded liability is the set of promises it is currently obliged to keep.
Explicit. Implied. Inherited. Self. Four classes, four origins, four costs of breach, four disciplines of retirement, held in one document on one cadence. The self class governs the other three. The implied class accumulates fastest. The inherited class is invisible until it is invoked at the worst available moment.
The Circle is a room of architects who keep the ledger, who retire promises by deliberate act rather than by neglect, and who can state at any moment what their architecture owes and to whom.
Structure. Access. Positioning.
The Architect Circle is application only. Two hundred and fifty seats globally. createyourownmarket.com.au.