Confessions of a Utopophobe

Here, I explain what’s wrong with “ideal theory” in political philosophy.*

[ *Based on: “Confessions of a Utopophobe”, Social Philosophy and Policy 33 (2016): 214-34. ]

I. Ideal Theory

In political philosophy, there is some debate about the relative merits of “ideal” versus “non-ideal theory”. Ideal theory is described in two different ways (sometimes confused with each other):

1. When we’re doing ideal theory, we choose political principles based (partly) on the consequences those principles would have if everyone perfectly complied with them. (I call this “strict compliance theory”.)

2. When doing ideal theory, we try to describe a perfect society. (I call this “perfection theory”.)

Sometimes, the distinction is applied just to theorizing about justice, where the ideal theorist is said to choose principles of justice based on the assumption of perfect compliance, or to be describing the perfectly just society.

Examples

John Rawls

In his original position thought experiment, Rawls imagines the parties in the O.P. choosing the principles to govern their society, on the assumption that whatever principles are chosen will be perfectly followed. He also thinks the result tells us what would happen in a perfectly just society.

So, when the parties decide that the state should distribute wealth in the way that most benefits the poorest people, they imagine that the government will know how to do that, they will want to do it, there won’t be special interest groups trying to manipulate the wealth redistribution system for their own benefit, people won’t cheat the welfare system by lying about their income, etc. They get to assume all that, not because it’s true, but because Rawls stipulates that we’re doing ideal theory.

Joseph Carens

Carens came up with a way to make socialism work: Everyone gets taxed such that their after-tax incomes are equal. In addition, however, suppose that everyone just voluntarily works their hardest to try to maximize their pre-tax income (despite getting no monetary reward for this), out of concern for the good of society.

Don’t say that’s unrealistic, because we’re doing ideal theory now. That is what should happen, even if it’s highly unlikely that it will happen.

G.A. Cohen

Cohen imagines a friendly, socialist camping trip, in which everyone voluntarily shares all resources, and everyone cooperates for the common good. No one expects to get extra rewards for giving benefits to the group, because they’re all friends. This is how society should work.

If you complain that this is unrealistic, you’re confusing “should” with “would”.

Jason Brennan

Brennan responds to Cohen by imagining a friendly, capitalist society in which everyone voluntarily trades with each other, everyone respects each other’s rights, and the poor are fine because people voluntarily contribute to charity to help anyone in need.

If you complain that this isn’t how real capitalist societies work, you are again confusing “should” with “would”.

II. Why Ideal Theory?

Why do some philosophers consider ideal theory interesting or useful?

A. “Ideal Theory Sets the Goal”

Some say that we have to do ideal theory first, so that we know what goals we are aiming at. Once we know that, we will do some non-ideal theory to figure out how to achieve those aims.

Rawls:

“until the ideal is identified … nonideal theory lacks an objective, an aim, by reference to which its queries can be answered.”

Stemplowska & Swift:

“[W]ithout knowing our long term goal, a course of action that might appear to advance justice . . . might nonetheless make less likely, or perhaps even impossible, achievement of the long-term goal.”

Reply:

You don’t have to describe the perfect society in order to pursue a useful goal. You can work to reduce particular injustices without knowing what a Perfect Society or Perfect Justice is like. There is no reason to think that doing this would prevent you from attaining societal perfection.

Analogy: Your car has a flat tire. You can try to change the tire without first articulating the nature of the Perfect Car. Indeed, discussing what the Perfect Car would be like would be a useless distraction. Nor is there any reason to think that changing the tire would somehow prevent you from attaining vehicular perfection later.

B. “Ideal Theory Defines Justice”

Maybe I have begged the question by assuming that we can identify particular injustices without a theory of justice. Maybe “injustice” just denotes a departure from perfect justice, and thus we need to know what perfect justice is before we can identify any particular injustices.

A. John Simmons:

“To dive into nonideal theory without an ideal theory in hand is simply to dive blind, to allow irrational free rein to the mere conviction of injustice and to eagerness for change of any sort.”

Reply:

We don’t need a theory of perfection in order to identify problems in any other area:

• If your car has a flat tire, you can know that this is a problem without knowing what a perfect car would be.

• I know that Ted Bundy was a bad person, even though I don’t know what a perfect person is or what the complete ultimate theory of virtue is.

• I know that Game of Thrones is a good TV show, despite having no theory of the aesthetic ideal.

There’s no obvious reason why justice should be different.

C. “Ideal Theory Enables Comparisons”

Maybe we need ideal theory to know which injustices are worse than which. John Simmons again:

“[T]he priority to be given to grievous (over less grievous) injustices is to be understood in terms of the lexical ordering of the principles of ideal theory violated by the injustices at issue.”

Reply:

We don’t need a theory of perfection to compare problems, either. E.g.,

• I know the car’s flat tire is a higher priority than the dust on the body, despite having no theory of the perfect car.

• I know Ted Bundy’s crimes were worse than my jaywalking, despite having no theory of the ideal person.

• And I know that wrongly imprisoning 500,000 people is a worse injustice than mispronouncing Kamala Harris’ name, despite having no theory of ideal justice.

D. “Today’s Utopia Is Tomorrow’s Reality”

Humanity has seen dramatic progress over the course of history. For most of history, if someone had described a society like our current society, they would have been dismissed as utopian. So it’s likely that future societies will be much better still. So it makes sense for us to aim high in planning for the future. Maybe we will one day have something close to a perfect society!

Reply:

The first part is true: the future will probably be vastly better than the present. But that doesn’t mean that we can usefully theorize about a distant future society that is nearly perfect.

Compare: Suppose some political philosopher back in 1500 A.D. wrote some treatises about what society should be like in the 21st century, including recommendations for how society should be run, what institutions and policies we should have, etc. Without knowing any details, how useful would you think the 1500 A.D. philosopher’s plans would be to us?

I think the reasonable answer is, “Completely useless.” Maybe they’d be of historical interest. But no one would dream of trying to use them.

That is how people in 500 years will look at anything we write today about what society should be like in the distant future. Except that our treatises will be even more useless, because society will probably change even more in the next 500 years, because the pace of change has been increasing.

Compare: Perhaps one day we will have flying cars with antigravity technology. But thinking about this possibility is of no use when your actual car has a flat tire. Similarly, it is not useful to think about a possible future society in which socialism would work.*

[ *Why isn’t thinking about anarcho-capitalism similarly useless? Briefly, I think an-cap is much less utopian than socialism, because it does not require any changes to human nature to work. ]

E. Strict Compliance Theory

The above discussion concerned perfection theory. What about Strict Compliance Theory, the idea that we should choose political principles based on their consequences in a world of perfect compliance? This is an importantly different idea. Rawls asserts it, but I don’t see any argument for it.

Consider two examples:

1. The Recusal Rule: If a judge has a personal relationship with a defendant, he has to recuse himself from the case.

   We wouldn’t have this rule if everyone always complied with all principles of justice; we wouldn’t need it, since judges would be perfectly impartial and just, even when their friends and family members were involved. Yet obviously, we should have this rule.

2. The Drug War: Many people criticize the drug war for its failure, and for the huge costs that result from noncompliance. E.g., we have hundreds of thousands of people sitting in prison for drug crimes, we have more other crimes, there is increased police corruption, etc.

   On strict compliance theory, we’d ignore these costs since they wouldn’t happen if everyone complied with the laws. Yet obviously, we shouldn’t ignore them.

What matters is the effects of a policy in a world with realistic levels of compliance, not its effects in a world of perfect people. There’s no obvious reason why the principles of justice should work differently.

Approximation

In science, people sometimes use idealizations; e.g., ignoring air resistance in computing the trajectory of a projectile. This is because the idealization is an approximation to reality (e.g., because it is known that the effect of air resistance is small). It is not done in cases where the idealization would not approximate reality.

In the political case, is the “idealization” of a world of perfect compliance at least an approximation to reality?

No, it isn’t. Of particular interest to Rawls, the assumptions that government agents have perfect knowledge of and desire to implement the distribution that maximizes the welfare of the poor, and that no one will try to game the system for personal advantage, are not approximations to reality. They are utterly false, and that makes a huge difference to what we should do.

III. Utopian Illusions

Ideal theorists are prone to three big mistakes.

A. Agentless Norms

Norms require agents: It cannot be true that A should be done unless there is some particular agent who should do it.

Ideal theorists often disregard this, e.g., by talking about what “society” should do. (Society isn’t an agent, so there’s nothing that it “should” do.)

E.g., Carens says that we should have a system wherein everyone maximizes their pretax income, while the state enforces equal after-tax incomes. But there is no agent who can bring this about. The state can’t bring it about, since they can’t make all the citizens pure altruists. No citizen can bring it about, since they can’t make their compatriots altruists. And “society” can’t do it, because society isn’t an agent.

B. Crazy Standards

There is such a thing as too much altruism. E.g., suppose a healthy patient in the hospital wants to donate his heart, lungs, and both kidneys to other patients. We would not say, “That’s wonderful, let’s schedule the surgery!” We would conclude that he was crazy.*

[ *See the episode of House, M.D., “Charity Case”. ]

That’s similar to (though more extreme than) Carens’ (or the socialists’) idea of the people who just selflessly work for the good of society. That isn’t consistent with a normally functioning human motivational system.

C. Excess Abstraction

Frequently, ideal theorists such as Rawls try to arrive at political conclusions by starting from general, abstract principles, like “it’s unjust that people have unequal resources due to morally arbitrary differences between them.” This is a highly unreliable way of forming conclusions, because philosophers are almost always wrong about abstract, general principles.

Instead, a better approach is to start from widely-shared, strong ethical intuitions about concrete cases, then draw analogies from these cases to more controversial cases. (Examples: (a) The Starving Marvin analogy for immigration; the killer’s accomplice example for gun control, discussed in my other papers.)