Selection by lot
A Socratic walk-through of selection by lot — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why would a city hand its highest offices to whoever a drawing machine happened to pick?
In classical Athens, most public offices were not contested. They were drawn. Slabs of stone stood in the Agora with grids of slots cut into them; citizens pushed in their bronze identity tickets, black and white balls dropped through a tube beside the grid, and whichever rows the white balls fell to were the men who would serve. The Council of Five Hundred, the panels of jurors, the great majority of magistracies — allotted.
To modern ears this sounds like an abdication. Election at least tries to find the best person; a lottery does not even try. So the useful question is not "how could they be so careless?" but the harder one: what can randomness do that no rule, however carefully written, can do?
Reasoning it through
REASONING #Begin with something a lottery undeniably does, then work out whether it is worth the cost.
A random draw from a population produces, on average, a sample that looks like the population. Not in any one draw — a single Council of five hundred might come out lopsided — but the process has no systematic tilt, because it has no criterion at all. Now compare an election. Elections do not have no criterion; they have a hidden one. They select for whatever wins votes: reputation, oratory, leisure to campaign, wealth to spend, an existing network. Aristotle saw this plainly and put it flatly — election, he says, is oligarchic; the lot is democratic. He is not praising the lot's wisdom. He is naming its lack of a filter as the point.
But representativeness alone would not justify it. Push further and ask a different question: what happens to everything built on top of the selection?
Consider a would-be faction. Under election, a coalition is worth assembling, because votes accumulate and a bloc can deliver an office. Under sortition there is nothing for the coalition to act upon — you cannot persuade a machine, and a favour done to a hundred citizens buys you no more chance than a favour done to none. The whole apparatus of patronage loses its object. This is the deepest effect of randomness here: it does not merely pick fairly, it destroys the returns on manipulating the picking.
The same reasoning explains a detail that otherwise looks fussy. Jury panels were allotted on the very morning they sat, from the six thousand men enrolled for the year, and only then assigned to a court. Ask why. If the panel were named a week ahead, you would know exactly whom to bribe or threaten. Drawn at dawn, the set of people worth bribing is the entire citizen body, which is to say bribery has become unaffordable. Unpredictability is functioning as a security property, in the same way a cryptographic nonce does — it makes the attacker's preparation useless.
Now the obvious objection, which the Athenians shared. Some jobs need competence, and a lottery cannot supply it. Watch how they answered, because the answer is what makes the system work rather than a curiosity. Generals — the strategoi — were elected, annually, and could be re-elected; so were the senior financial officers. Expertise-critical roles were carved out explicitly. Everything else got the lot, but wrapped in three further devices: a pre-office scrutiny of eligibility, service in collegial boards of ten rather than alone, and a compulsory audit of conduct and accounts at the end of the term, with real penalties.
Notice what that combination implies about their theory. They did not believe any citizen would perform well. They believed that with vetting at the entrance, colleagues alongside, an audit at the exit, and a term of one year, the variance in individual quality mattered less than the removal of a permanent office-holding class. Randomness was safe because it was fenced, not because it was trusted.
The analogy
THE ANALOGY #Think of shuffling a deck before a card game. The shuffle does not improve anyone's hand — it cannot, it adds no cards. What it does is make the deal unforecastable, which is what stops a player from spending the evening arranging the cards instead of playing. The value of the randomness lies almost entirely in what it prevents.
a shuffled deal gives every player an equal chance and nothing else, whereas an allotted Council was expected actually to govern — so Athens had to add the scrutiny, the colleagues and the audit, which have no counterpart in a card game.
Clarifying the model
THE MODEL #Two misconceptions are worth clearing.
The first is that sortition means "anyone will do". It never did. Only enrolled adult male citizens entered the pool — a minority of the resident population, since women, slaves and metics were excluded — and candidates could be struck out at scrutiny. The lot ran inside a defined and, by our standards, narrow pool. Randomness distributes; it does not enfranchise.
The second is that the lot replaced election. It sat beside it, and the division of labour between them is the system's real insight. Where the office needed judgement that any citizen could supply, and where the danger was capture, they drew. Where the office needed a scarce skill, and where failure was immediate and visible, they voted — and then audited the winner anyway. That is not an ideological commitment to chance. It is a considered claim about which failure mode is worse in which role.
It is also worth saying plainly that this is not merely antiquarian. Randomly selected citizens' assemblies have been used in the last two decades in Ireland, France and elsewhere for exactly the first reason — to produce a body that has no reason to be captured and no career to protect. Whether they deliberate well is a live argument, not a settled result.
A picture of it
THE PICTURE #How to readFollow the citizen down the left edge through four encounters, in order. The draw is only the second of them — read the messages before and after it as the fence around the randomness, since the scrutiny screens the entrance and the audit judges the exit. The note names what the arrangement buys: because the second step is unforecastable, no effort spent on the first or third can be aimed at a particular person.
What became clearer
WHAT CLEARED #Randomness in Athens was not a substitute for judgement about people. It was an instrument aimed at the machinery around the appointment — it flattened the selection so that no coalition, patron or briber had anything to work on, and it produced a body statistically like the citizenry rather than like the class that wins elections. Its cost, unreliable individual competence, was paid down by vetting, collegiality, short terms and audit; and where that payment was not enough, they simply voted instead.
Where to go next
ONWARD #- How the dokimasia and euthyna procedures actually worked, and how often they caught anything.
- Why Venice and Florence later built elaborate hybrid lottery-and-election schemes, and what they were defending against.
- What modern citizens' assemblies inherit from this design, and where the analogy to Athens fails.
Key terms
TERMS #| Term | What it means |
|---|---|
| Sortition | the filling of public office by lot rather than by election. |
| Kleroterion | the stone allotment machine, with slots for citizens' bronze tickets and a tube of black and white balls that decided which rows were selected. |
| Dokimasia | the scrutiny of an incoming officeholder's eligibility, conducted before he could take up the post. |
| Euthyna | the compulsory audit of an outgoing officeholder's conduct and accounts. |
| Strategos | one of the ten annually elected generals; among the offices deliberately exempted from the lot. |
Every term the collection defines is gathered in the glossary.