Kant’s first test asks whether the rule behind an act could be adopted by everyone. The lab models the rule as the defecting choice among ten players and computes what the defector gains when alone, and what the choice is worth once all have made it. When the gain disappears at universal adoption, the practice the act depends on has collapsed.
Core ideas
Games and the Social Contract supplied the many-player game. The test below reads two numbers off it, and the three ideas say what each number is for.
The test is applied to the rule an agent acts on, stated with the act, the circumstance and the aim, and not to a single deed. In the lab the rule is the defecting choice and the practice it exploits is the cooperating one.
A liar is paid only because other people believe promises. That is why the payoff of lying rises with the number of others who keep their word, and why it can fall to nothing when none does.
A maxim fails the first test when it defeats its own aim at universal adoption. Whether the world would be worse for everyone is a different question, and the lab’s other tiles answer it separately.
The material
One maxim, its payoff table, and the two numbers the test reads from it.
The consequentialist lessons ended with a forbidden list typed by hand: “The Trolley Problem as a Decision Matrix” struck an act from the table and did not say why it belonged there. The deontological theories are attempts to say why. This lesson and the two after it each give one such account a checkable structure: a maxim tested at universal adoption, a harm tested for whether the good depends on it, and an omission tested as a cause.
Kant’s first formulation of the categorical imperative says to act only on a maxim you can at the same time will to be a universal law. The usual gloss, “what if everyone did that?”, is where the trouble starts, because it sounds like a forecast of consequences. The lab lets the two readings be told apart, and this lesson takes the first of Kant’s two failures: the contradiction in conception.
A maxim is the rule an agent acts on, stated with the act, the circumstance and the aim. Here it is: when I need money, I will promise to repay it, knowing I will not, in order to be given it.
Model the practice of promising as a game with ten players. Each player either keeps promises (cooperates) or makes false ones (defects). Let k be the number of the other nine who keep their word. An honest promiser gets 2 whatever the others do. A liar gets 5k/(n − 1) with n = 10: nothing when no one else is honest, because no promise is believed, and 5 when everyone else is, because every promise is.
The test reads two numbers from this table. The first is the liar’s gain when alone: every one of the other nine is honest, so the liar gets 5 where an honest promiser gets 2, a gain of 3. The second is what the same choice is worth when everyone makes it: with no one honest the liar gets 0, against the 2 that honesty yields among honest people, a difference of −2. The lab prints the pair as one line. Its equilibrium and replicator tiles are the ones Games and the Social Contract read, and the test leaves them alone.
A maxim has a contradiction in conception when, adopted by everyone, the practice its act depends on could not exist and so the act could not achieve its aim. A promise is believed only where promises are usually kept; where everyone lies, there is nothing to exploit.
That is what the figure of 0 records, and it is why the test is not the forecast it sounds like. A forecast would ask whether the world of universal lying is worse. The test asks whether the liar can still do what the lie was for. The lab’s social-optimum tile makes the difference visible: total welfare among the ten is highest not when everyone is honest but with seven honest and three liars, because in this table an honest promiser loses nothing to a liar, and a liar among the honest collects 5 where honesty pays 2. That stipulation is generous to the liar. A table in which the deceived lose would move the optimum; it would not move the two figures the test reads, which are taken where all nine others are honest and where none is. A sum-of-welfare view therefore does not condemn the maxim as stated, and Kant’s test does.
Cutting in line pays only where others wait. Let a waiting person get 2 + 2k/9 and a jumper 1 + 5k/9, where k counts the others who wait. Alone, the jumper gains 2. When all jump, the jumper gets 1 where the orderly queue gave 4: the advantage was being ahead of people who waited, and nobody is left to be ahead of.
The tax preset shows where the test is less clean. Evading a tax that funds something everyone uses pays whatever the others do, and when all evade nothing is funded. Nothing in that is a contradiction in conception, because a world without the tax is perfectly conceivable. Kant’s second failure, the contradiction in the will, fits better: the evader cannot coherently will a world without the public good, since the evader wants it. The lab prints the same two numbers for both failures, and which one a number means is your reading, not its output.
The limit belongs here. Kant did not compute anything, and the lab does so only by treating the maxim’s point as a payoff you supply. It also cannot choose how general the maxim should be: “I will lie to this person on this day” passes any such test, and most of the test’s critics begin there.
Interactive
payoffs as exact expressions in k, and what follows from them
Method
Five steps. The third is where a story becomes a table, and every later step reads from that table.
Write the act, the circumstance and the aim: “when I need money I will make a promise I do not mean, in order to be given it”. A maxim without its aim cannot be tested for whether the aim survives.
Find what must be going on for the act to work. A false promise works because promises are usually kept. The cooperating choice is that practice.
Give the honest player’s payoff and the defector’s as functions of how many of the others keep the practice. Be explicit that these are numbers you chose.
The first is the defector’s gain when alone. The second is the defector’s payoff when no one keeps the practice, set against the honest payoff.
If the defector’s point vanishes, the maxim defeats itself in conception. If the world is merely worse, or the agent wants the practice, say which of the other failures it is.
Ten players, k of the other nine keeping their word. Read the two figures off the table.
The first figure is why a rule like this is tempting, and the second is why it cannot be a universal law. The lie gains only against a background of honesty that the lie, adopted generally, removes.
Given a maxim, you should be able to write it as a defecting choice with payoffs, compute the gain when alone and the payoff when everyone adopts it, and state whether the practice the act depends on survives.
Something to change about this lesson? Write it down, tick it off when it is handled, and export the list when you want to act on it. Kept in this browser only: nothing is sent anywhere, and nobody is notified.
No recommendations yet for this lesson.