But Commitment Is Not Enough
Authority, weight, and the difference between rank and standing
There is a person who commits to everything. The launch will hit the date. The market will grow. The fix will hold. Each statement is a genuine commitment — named, dated, on the record. And each is close to worthless, because the person has no track record of being right about any of it and no standing in the domain they are pronouncing on.
There is another person, across the table, who knows the domain cold — twenty years, scars to show for it — and who will never commit to a call. It depends. There are considerations. One could argue either way. Their expertise is real and equally worthless to anyone who has to act, because it never crosses the commitment boundary into something you can hold them to — accountability without commitment.
The previous article established that every consequential decision should be captured as a committed fact, with its authority named (commitment requires authority). This article is about the uncomfortable second half of that sentence: naming the authority is not the same as the authority being worth anything. A commitment is only as good as the standing of whoever made it — and standing is a thing we can be precise about.
Commitment and authority are both necessary
The two hollow figures above are not opposites. They are the two ways the same thing fails.
A commitment from someone with no relevant standing is worthless. Deep standing that never commits is also worthless. Trust lives only where the two meet.
This is why trust takes so long to build and collapses so fast. Trust is not a feeling. It is the accumulated record of commitments kept by someone with the standing to make them. The Promise — from the opening of this series — erodes not because of one broken commitment but because commitments pile up that were never backed by a track record. The expert who won’t commit never starts building the record at all.
A truthful system therefore needs to track two things about any committed claim, not one: that a commitment was made, and how much that particular committer’s commitment weighs.
Authority weight
Call the second thing authority weight — the standing that a committer brings to a commitment in a specific domain. It is built from four things that can each be recorded and examined:
Track record. Has this authority committed in this domain before, and how did those commitments turn out? An authority whose past commitments were calibrated — confident when right, hesitant when wrong — has earned weight. One whose record is a string of confident errors has earned the opposite, no matter how senior.
Qualification. Does the authority have relevant expertise — training, experience, demonstrated skill — in this domain? Expertise is not transferable. A brilliant cardiologist has near-zero authority weight on a question of tax law, and pretending otherwise is how confident nonsense propagates.
Certification. Has the authority’s competence been verified by some process that others trust — a credential, an accreditation, a vetting? Certification is a shortcut: it lets a stranger assign weight without auditing the full track record themselves.
Authorisation. Does the authority have the right to commit in this domain at all? A person may be qualified and certified and still not be the designated writer for this decision. Authorisation is the organisational grant of commitment rights — and it is the place where authority weight collides with power.
None of these is the strength of the claim itself. A high-weight authority can be wrong; a low-weight one can be right. Authority weight does not tell you whether the commitment is correct. It tells you how much the fact of this authority committing should move your confidence — one input among several, not the verdict.
Positional authority is not epistemic authority
The most dangerous confusion in any organisation is between two kinds of authority that usually travel together and sometimes come violently apart.
Positional authority is conferred by rank. The manager, the director, the board. It grants the right to decide and the power to make decisions stick.
Epistemic authority is earned by being right. The analyst with the track record, the engineer who has seen this failure before, the nurse who has watched a thousand of these cases. It grants nothing official. It is just standing in the domain.
When the two align, decisions are both authorised and well-grounded. When they diverge, you get The Buried Finding. The junior analyst has the epistemic authority — the evidence, the relevant track record — and the manager has the positional authority — the power to decide what enters the record. The manager buries the finding. Positional authority overrides epistemic authority, and a true claim never reaches the floor of ground truth.
This is precisely why the immutability of facts is not a technical nicety but a moral safeguard. In a system where the analyst’s finding is recorded as a committed atomic fact the manager cannot delete, positional power can still ignore epistemic authority — but it cannot erase it. The record stands. The reckoning is merely delayed, and the trail of who ignored what is itself in the log.
Authority weight should follow epistemic authority, not positional authority. A system that assigns weight by rank rather than by standing will, reliably, bury its most valuable truths.
Authority is measurable — because commitments are recorded
Here the pieces lock together. The reason we can speak of authority weight as something concrete, rather than a vibe, is the commitment boundary from the previous article.
If every consequential decision is captured as a committed event — with its authority, its evidence snapshot, its moment — then every authority accumulates a visible, queryable record of their committed decisions and how those decisions turned out. Track record stops being reputation and becomes data. Calibration can be computed. Domains of genuine expertise can be distinguished from domains of confident overreach. Authorisation can be checked against who actually committed.
This is the foundation that the broader work — scoring, attribution, the design of who gets to write what — is built on. Standing is derived from a history of committed decisions and their outcomes, not asserted from a title. An organisation that records commitments faithfully can, over time, let authority weight migrate toward epistemic authority and away from mere position — slowly, on evidence. An organisation that records nothing is stuck assigning weight by rank forever, because rank is the only signal it has kept.
There are hard questions here that this series does not pretend to close: how to combine many authorities’ commitments, how to certify collectives rather than individuals, how to weight a confident minority against a hesitant majority. Those belong to the design of processes and harnesses that come later. The point for now is narrower and firm: authority weight is a real, recordable property, and a system that wants to approach truth has to track it.
One more input to one question
Step back and notice what we have accumulated. To know how much to trust a committed claim, it is not enough to know that someone committed. We also need to know the standing of who committed. And we will shortly need to know how strongly the underlying claim is supported, how far it sits from ground truth, and how fresh it is.
These are not competing theories of trust. They are separate inputs to a single question — how much may we rely on this? — and authority weight is one of them, sitting alongside commitment status and the facets we are about to meet. Treating it as one input among several, rather than the whole answer, is what keeps a system honest: it stops a high-ranking voice from settling a question by weight alone, and it stops a strong argument from being dismissed because of who made it.
We now have the substrate (commit and measure captures), the act that creates binding records (commitment at the commitment boundary), and the standing that gives a commitment its weight (authority). Part III closes with one more bind-time discipline: what happens when committed facts turn out wrong — When Facts Are Wrong. After that, we examine the claim itself — how strongly is it actually supported, on its own merits? That is where Part IV — The Facets of Trustworthiness begins: Every Statement Has a Tier.