Behavior By SENSMAN

Who Actually Makes the Decision? A Swarm's Quorum, and the Clerk in the Brain

A hive has a queen. She gives no orders. A head has a self. It may give none either.

Who actually makes the decision — scout signals accumulating past a quorum

A hive has a queen. She gives no orders.

Her work is laying eggs. Where to go and what to do are not hers to decide.

How this structure actually works has largely been measured.

What the queen emits is not an instruction

She is not silent. Pheromone from her mandibular glands suppresses ovary development in workers and holds the swarm together.

But that is a state signal, not an instruction. It contains no destination.

Nest site, foraging direction, swarming time. All of it rises from below.

The advertisement switches itself off

In a swarm, finding the nest site is the work of scouts — several hundred bees, a few percent of the whole.

A returning scout advertises the direction and distance of what she found with a waggle dance. Another bee reads it, flies out, inspects the site herself, and dances on return.

There is one device built into this.

A scout reduces the number of dance circuits by a fixed amount each time she repeats the advertisement.

A bee who found a good site dances long at first. Each repetition is shorter, and eventually it reaches zero. A bee who found a poor site starts short and is gone in two or three rounds.

Nobody issues a cancellation. The advertisement is built to expire.

A quorum of fifteen

So when is it decided?

Not by unanimity. Not by majority.

When the number of scouts gathered at a candidate site reaches roughly 15 to 20, the bees supporting it return to the swarm and begin the signals for take-off. The whole swarm lifts.

A quorum.

What matters is that this number has nothing to do with the distribution of opinion overall. The other several thousand bees have never seen any candidate. They do not participate; they follow the result.

So does the queen. She goes last.

And they stop each other’s dancing

This is the centre of the piece.

What happens when two sites are evenly matched? A swarm split in half dies twice over.

The solution was reported in Science in 2012.

A scout butts her head against a bee advertising a different site and delivers a short vibratory signal. That bee stops dancing.

A stop signal. And it is not delivered at random. It goes to bees advertising somewhere else.

The structure has a name: cross-inhibition.

HiveBrain
Optionscandidate sitescompeting alternatives
Support signalwaggle dancefiring of a neural population
Inhibitionstop signalmutual inhibitory connections
Triggerquorum of 15–20firing-rate threshold

That is the paper’s own point. The decision structure of a swarm and the models of competing options in the primate brain have mathematically the same form.

A hive has no centre. Whether a brain has one is an older question.

A clerk on the throne

People hold the self up as a commander. I judged, I chose.

Evidence that this feeling is manufactured afterwards has accumulated along several lines.

Split brains. In a patient whose corpus callosum has been cut, an instruction is presented only to the right hemisphere. The patient acts on it. Asked why, the left hemisphere — which handles speech — produces a plausible reason. For an instruction it never received.

Gazzaniga called this reason-building function the interpreter.

Choice blindness. The 2005 experiment is simpler. A subject picks one of two face photographs. By sleight of hand they are handed the one they rejected and asked why they chose it.

Most did not notice the switch. And they explained why they had chosen the face they had not chosen. Fluently.

The function that supplies reasons runs separately from the function that decides. And it does not stop when the first one is missing.

The self is less a king than a clerk. It records what has already happened, tidies the reasons, and stamps it at the end. My decision.

Libet is not settled

The experiment most often cited here is Libet’s.

In 1983 he reported that the readiness potential preceding a wrist movement began about 350 milliseconds before the subject’s reported conscious intention. A 2008 fMRI study reported predicting a choice up to 7–10 seconds in advance.

These results should not be read straight across as “there is no free will.”

  • The 2008 prediction accuracy was about 60 percent. A coin toss is 50. That is not a decision already made
  • In 2012 Schurger and colleagues showed the readiness potential may not be a decision signal. If spontaneous neural noise drifts until it crosses a threshold, the rise before a movement is accumulation, not intent

Written honestly: there is evidence pointing toward consciousness arriving late, and whether that rules out decision itself is not established.

Many articles overstate this. Overstatement is not needed; what remains is enough.

The hungry judge — famous and disputed

“Judges deny parole when they are hungry” is widely cited. A 2011 analysis of Israeli parole hearings found approval rates high right after a break and falling toward the end of each session.

Objections followed immediately that the ordering of cases was not random, and the matter is unresolved.

The example can be dropped. That bodily state enters judgement is well enough supported by other routes.

Choosing the materials is enough

Manipulation need not arrive as an order. Deciding what gets seen first, and what gets heard often, is enough.

That claim has an exact name.

The illusory truth effect. Repeated exposure to a statement makes it feel true. First reported in 1977 and replicated many times since.

And the 2015 study is decisive.

The effect appeared even in people who already knew the correct answer.

The paper’s title is the sentence itself: “Knowledge does not protect against illusory truth.”

Judgement is not broken. What enters judgement as material is changed. The process then runs normally and arrives somewhere else.

But this part needs care

Claims of the form “a headline seen at breakfast changes your choices that day” often lean on social priming research. Repeated replication failures there made it the emblem of the replication crisis.

So it is both safer and stronger to split the claims:

ClaimStatus
Repetition makes a statement feel trueestablished, even when you know better
We come to prefer what we see oftenestablished (mere exposure)
Easily recalled cases distort probability judgementestablished (availability heuristic)
Exposure to particular words changes behaviormany failed replications

The first three carry the whole argument on their own.

This article’s definition

A swarm decides without a centre. Support accumulates, competitors inhibit one another, and past a threshold everything moves. The queen follows last.

Evidence is accumulating that a head has something close to that shape. And the self arrives at the end of the process to handle the record.

Manipulation does not arrive as an order. Choosing the materials is enough.

Fixed here as this article’s definition. Before strengthening the will, the question is what the will is being made out of.

The screen looked at daily, the people kept close, the phrases driven into the ear by repetition. That becomes the floor of the next judgement.

People call the mind their own. The materials that made it have never once been entirely theirs.

In short

  • A swarm’s nest site is settled by several hundred scouts’ signals, not by the queen
  • Queen pheromone is a state signal, not a behavioural instruction
  • Scouts cut their dance each time they repeat it; poor sites expire on their own
  • The decision fires on a quorum of 15–20, not on unanimity
  • Scouts send a stop signal to bees advertising rival sites: cross-inhibition
  • That structure is mathematically the same as models of competing options in the brain
  • A split-brain patient’s left hemisphere invents reasons for instructions it never received
  • In choice-blindness experiments people fluently explain choices they did not make
  • Libet is evidence that consciousness is late; it does not rule out decision
  • Repetition makes statements feel true, even to people who know the answer

This piece is rebuilt for search readers from “A Clerk on the Throne” of the serial essay The Great Corona Swindle. The original’s point of view is kept, and each claim carries its primary source and how well established it is.

Frequently asked

How does a honeybee swarm choose a nest site?
Several hundred scouts disperse, find candidate sites and advertise them on return with the waggle dance. Other bees fly out, inspect the site themselves and dance in turn. When the number of scouts gathered at one candidate reaches a quorum of roughly 15 to 20, the whole swarm moves there. The queen takes no part and follows.
Does the queen bee not give orders?
She directs no behavior. Queen mandibular pheromone suppresses worker ovary development and holds the swarm together — a state signal, not an instruction. Nest site, foraging direction and swarming time are all settled by information rising from below.
Did Libet's experiment prove there is no free will?
No. In 1983 Libet reported that the readiness potential precedes the reported conscious intention by about 350 milliseconds. In 2012 Schurger and colleagues showed that this potential may not be a decision signal at all but spontaneous neural noise crossing a threshold. The interpretation is open.
Does hearing something repeatedly change judgement?
It is a well-established effect with a name: the illusory truth effect. Repeated exposure to a statement makes it feel true. A 2015 study found the effect even in people who already knew the correct answer. Knowing is not a defence.

References

#swarm#scout bees#quorum sensing#cross-inhibition#collective decision#Libet experiment#confabulation#choice blindness#illusory truth effect#free will

#분봉#정찰벌#정족수#교차억제#집단의사결정#리벳실험#작화#선택맹시#진실성착각효과#자유의지

Published September 28, 2026