The quiet ones
The silence that looks like consent
Nobody said anything about seventeen of the thirty-eight removals. Four different things were going on in that silence and only one of them was agreement.
The quarterly reconciliation of the project ledger
week 3 · surfaced after 31 weeks
Shadowing new starters through their first week
week 6 · surfaced after 47 weeks
Reading supplier contracts before renewal
week 2 · surfaced after 22 weeks
Logging why a job was turned down
week 9 · surfaced after 64 weeks
Surfaced 4
Four removals were written down with the week each of them stopped. Every one of them produced an event inside the 78 weeks watched. Of these, four surfaced as a problem. The four removals nobody said a word about, all of which later surfaced as problems.
Seventeen of the thirty-eight removals passed without comment. In the room, that reads as a single thing: no objection, therefore consent, therefore proceed.
Reading the record afterwards, and from what people said when asked later, the seventeen silences were at least four different things.
Four kinds of nothing
Agreement. Some people genuinely thought the thing should go. Six or seven of the seventeen are of this kind on the available account, and in most of those cases the removal has indeed turned out fine.
Absence. Nobody who knew anything about the procedure was present. Four of the seventeen were removed in meetings that did not include anybody who performed them. Two of those four later surfaced as problems.
No standing. Somebody present had a view and did not think it was theirs to express — a newer member of staff, somebody from a different function, somebody who had already spoken twice that hour. This is the hardest kind to count and the record contains two acknowledged instances.
Nothing to say yet. The procedure was described in a sentence and the implications were not obvious in the ten seconds available. The person who would have objected worked out what the removal meant three weeks later, by which time the meeting was over and the thing was gone.
These are not degrees of agreement. They are four unrelated states that share a single observable.
The one that matters most
The second — absence — is worth isolating, because alone among the four it can be checked in the room for nothing.
Of the four removals made with nobody present who performed the procedure, two surfaced as problems: the standing slot and the reconciliation. Both are documented elsewhere on this site, and in both cases the person proposing could see the output of the procedure and not the work of producing it.
That is a two-in-four rate against one-in-thirty-four for everything else on the record, and while four is far too small a number to carry a ratio, the mechanism is not subtle and does not need a ratio. If nobody who does the thing is in the room, the silence in the room contains no information about the thing.
The check
One question, asked before a removal is agreed rather than after:
"Who here has done this?"
If the answer is nobody, the removal is not blocked. It is marked. It goes on the list with a note, and somebody asks a person who does it before the week is out.
On this record that would have cost four conversations and would have caught the standing slot immediately — the person who ran it could have said in thirty seconds what it was absorbing, and the record suggests they would have.
Why "any objections?" does not do this work
Because it is addressed to the room, and the room is the problem.
A general invitation to object routes the question to whoever feels entitled to answer it, which is the same group that has been talking. It cannot reach somebody who is absent, and it does not reliably reach somebody who thinks it is not their place.
A question addressed to a specific person about a specific thing behaves differently — not because people are braver when named, but because "who here has done this?" has a factual answer that somebody either can or cannot give. It is not asking for a view.
The limits of this
Three of the four kinds of silence are not fixed by any of this.
Agreement is fine and needs nothing. "Nothing to say yet" is not fixable in the room at all — the person needs three weeks, and the only remedy is a mechanism for a removal to be reopened without embarrassment weeks later, which most organisations do not have and this one did not.
"No standing" is a feature of how a place works and is not repaired by a question in a meeting.
So the realistic claim is modest: of seventeen silences, the check above addresses about four of them, and on this record those four contained two of the five expensive removals. That is a good return for a question that takes five seconds, and it is not a solution to anything.
What the seventeen are not
They are not seventeen mistakes. Thirteen of them produced no problem at all and several are among the clearest improvements on the list.
The claim is only that the organisation learned nothing from the silence, in any of the seventeen cases, including the thirteen where the silence happened to be correct. A silence that is right and a silence that is wrong are equally uninformative at the moment it occurs, and treating the first kind as confirmation is how the second kind survives.
How the four kinds were distinguished
A fair question, since all four look identical in the room, and the answer is weaker than one would like.
Nobody was surveyed at the time. The classification comes from conversations during the eighteen-month read-back, in which people were shown entries and asked whether they remembered them stopping.
That is a poor instrument. Recollection at eighteen months is unreliable, people reconstruct their own past positions, and anybody who turned out to be right has an incentive to remember having thought so.
What it is good for is a lower bound on one thing: the four removals agreed in rooms containing nobody who performed the procedure are a matter of record rather than recollection, because the attendance is documented. That category does not depend on anybody's memory, and it is the one the practical recommendation rests on.
The other three kinds of silence are offered as a plausible decomposition and nothing more.