Fable
How the three watchtowers decide whether to open the gate
Three beacon towers in the mountains jointly decide whether to light the main beacon. The east tower watches the sea entrance, the west tower watches the mountain road, and the north tower watches the city gate. A misjudgment by a single tower could make the whole city busy itself for nothing or miss an enemy movement.
At first, whoever shouted first lit the main beacon. Once the east tower mistook fishing lights for enemy ships, and the whole city was alarmed; another time the west tower could not see clearly in fog and nearly delayed the warning. The city guard decided that major signals required confirmation from multiple towers.
The new rule ran into trouble too. If all three towers had to agree, the whole city would be paralyzed when the north tower was snowed in; if two were enough, a single fog could fool both at once.
The defenders then established a consensus method: each station first reported observations and confidence, and messages were collected within a time limit; if two stations had consistent independent evidence, they could act, and if they disagreed, a fast rider was sent to verify; absent stations would record afterward, and it would be checked whether they had long been inaccurate.
The city warden first ruled that two towers must agree before lighting the fire. But on a foggy night, the east and north towers each saw only half a shadow, and the west tower couldn't hear clearly because of the wind. If you simply counted votes, you might still amplify the same misunderstanding, or miss the moment because one tower was late.
Later the three towers agreed on a more detailed method of deliberation: each reported the evidence it had seen, its credibility, and the time; when there was a conflict, the nearest rider post was asked to verify; in emergencies, temporary warnings were allowed but marked as unconfirmed. The central beacon no longer listened to whoever was loudest, but had multiple observation points piece together local observations into a decision that could be borne.
With this approach, the three platforms would still each misread things, but they rarely turned a single-point illusion directly into a citywide order. The value of joint decisions is not in having more voices, but in letting evidence from different positions cross-check one another. With this approach, the three platforms would still each misread things, but they rarely turned a single-point illusion directly into a citywide order. The value of joint decisions is not in having more voices, but in letting evidence from different positions cross-check one another. With this approach, the three platforms would still each misread things, but they rarely turned a single-point illusion directly into a citywide order. The value of joint decisions is not in having more voices, but in letting evidence from different positions cross-check one another.
The three beacon towers did not need to always see the same picture, yet they had to converge into an executable judgment in an agreed way. The difficulty of multi-party collaboration is preventing single-point misjudgment without letting the wait for consensus stall action.