The veil of ignorance
from 1971 Rawls; anticipated by Harsanyi in 1953
Choose the rules before you know which position you will occupy under them, and self-interest does the work of fairness.
- Active
- from 1971
- Where
- Cambridge, Massachusetts
- Status
- In use
- Tradition
- Devices and constraints
What it is
The veil of ignorance is a constraint on who is choosing. Those setting the rules are deprived of knowledge of their own position under them — their talents, their wealth, their place in the order being designed.
The elegance is that it does not require anyone to be impartial. A purely self-interested chooser who does not know which position they will occupy has to consider every position, and the constraint converts self-interest into fairness without asking anybody to be good.
How it works
Identify the positions the arrangement creates and who will occupy them. Strip the choosers of knowledge of their own position. Ask what rules they would choose knowing they might end up anywhere in the arrangement, including at the bottom. Adopt what survives.
The last position is where the argument happens. Rawls holds that rational choosers under total uncertainty would protect the worst position; others hold they would maximise the average, and this single disagreement produces very different conclusions.
How it is applied
- Identify the positions the arrangement creates
- Deprive the choosers of knowledge of their own position
- Ask what rules they would choose facing every position equally
- Adopt the rules that survive that test
The device applied
Proposition The shift allocation should be settled before the roster is drawn.
- The depot has three shift types: weekday days, weekday nights, and the weekend pair. Everyone knows which they currently hold and every previous discussion has divided along exactly those lines.
- Apply the veil: the crew is asked to settle the allocation rule for next year before anyone knows which shift type they will be assigned, with assignment drawn afterwards.
- Under that constraint the arguments change immediately. Nobody proposes that weekend work carry no loading, because they might draw it. Nobody proposes the loading be so high that the weekday roster becomes unstaffable, because they might draw that instead.
- The rule the crew reached: weekend shifts carry a loading, weekend allocation rotates on a fixed cycle rather than being volunteered for, and anyone drawing three weekend pairs in a row goes to the front of the queue for leave.
- Every one of those provisions protects a position that the person proposing it might occupy.
Therefore The allocation rule was settled in one meeting after two years of failing, because removing the knowledge of who would get what removed the only thing anyone had been arguing about.
Nobody was asked to be fair-minded. The crew were as self-interested behind the veil as in front of it, and that is precisely why the device works.
What it is good for
Designing rules that will bind a group whose members have different interests — rosters, allocations, charging structures, constitutions. It is the most practically effective device on this list and the easiest to run.
Where it fails
The choice under the veil is not determinate. Rawls argues the parties would protect the worst-off; Harsanyi argues they would maximise expected utility across positions, which licenses very different arrangements. Both are rational responses to total uncertainty and the device cannot adjudicate between them.
The veil's contents are also chosen by whoever designs it. What the parties are allowed to know shapes the result entirely, and Rawls's decisions about what to exclude — and, critics have argued, what he failed to exclude — do a great deal of the work that the device appears to do on its own.
And it presumes the positions are known in advance. Where an arrangement will create positions nobody has thought of, the veil protects only the ones on the list.
Common misconceptions
That it asks people to be impartial. It does the opposite: the parties are assumed to be purely self-interested, and the constraint is on their information rather than on their motives.
Where it came from
Rawls's A Theory of Justice in 1971, though Harsanyi published a structurally similar device in 1953 and always maintained he got there first with a different conclusion.
What became of it
The most discussed device in twentieth-century political philosophy, and widely adapted in practical settings that owe nothing to the rest of Rawls's theory.
Lasting influence
It reoriented political philosophy, and its practical descendants appear in mechanism design, in institutional rule-making and in the cake-cutting procedure every parent knows.
Why it still matters
The most directly usable idea in this section.
Put another way
One cuts, the other chooses, and neither knows which piece they will get.
Texts
John Rawls, A Theory of Justice
Sections 3, 4 and 24 set out the original position and the veil. The rest of the book is what he thinks follows.
Where to start
Run it on a real allocation problem before reading about it. It takes one meeting and the effect is immediate.
The surprising bit
The device works better in a depot meeting than in political philosophy, because a roster has a known set of positions and a society does not.
In their own words
No one knows his place in society, his class position or social status, nor his fortune in the distribution of natural assets
Who held it
the cake-cutting procedure
Connections
- Relies on
- The thought experiment
- Parallel to
- The state of nature
The verdict
Outstandingly effective on a bounded problem, and unable to settle the question it was built for.