Tom M. has a thoughtful post at Defective Equilibrium on coercion (a rejoinder to Paul Walker’s response to Tom’s post on private property and liberty). He provides a descriptive definition of coercion which stipulates that the action the coercer intends to bring about must make the coerced worse off and the coercer better off, suggesting that a normative definition of coercion would require a specification of the rights people have and thus a normative baseline. His descriptive definition:

A is coerced to Q iff:
1) B threatens A to perform Q, perhaps by imposing negative consequences for A if he/she performs R
2) In light of 1), A strictly loses when performing either Q or R
3) B gains from A performing Q.

An upshot of this definition is that taxation is not necessarily coercive in the descriptive sense:

Is taxation coercive? Under a descriptive framework it seems only coercive if the taxpayer loses from paying tax. If you are better off after paying your tax (and receiving the benefits of Government services, if any) you wouldn’t need to be coerced into paying your tax – the fact that you are threatened seems irrelevant. In a descriptive sense it seems then that the only people who are coerced into taxation are those people (if any) who lose out from the Government taxation.

I don’t like arguing over the definition of words, since every word is ambiguous to some extent and none has any True Meaning, but this just isn’t what most people seem to be talking about when they use the word ‘coercion’. I don’t see conditions 2 or 3 in Tom’s formulation as being implied in most political discourse, and it’s certainly not what I (and, I suspect, most libertarians) mean when talking about coercion. I can’t think of a single philosopher who takes Tom’s view (that is not to say that there are none, but that they are probably few).

Conditions 2 and 3 are, of course, likely to be empirically related to coercion: If A wanted to perform Q anyway, B wouldn’t normally need to bother coercing him, and if B didn’t gain from A performing Q, it’s unclear why he would want him to. It’s easy, though, to construct examples in which force is used even where conditions 2 and 3 aren’t met, especially once we introduce uncertainty and asymmetrical information.

A moral baseline is relevant, however, in specifying exactly what is to count as a threat. I largely agree with Nozick when he says that coercion requires a threat would need to shift A’s preference away from Q towards R, and that A would rather not receive the offer at all. Nozick, however, ignores actual force. I would say you’re coerced if you’re tied up in order to prevent you from performing an action.

My definition of coercion would be more like:

B coerces A iff:

1) B uses force to prevent A from performing Q, or threatens perform action S if A performs Q

2) A is harmed by S

Things are complicated slightly by the possibility of B withholding benefits rather than imposing costs. Suppose B is the major customer of A’s business, and wants A to marry B’s daughter. He threatens to withdraw his custom if A refuses. We might find this despicable, but those committed to property rights would say that B has every right to take his money elsewhere. This does, however, shift A’s preference towards marrying B’s daughter, and is an offer he would rather not receive. We could amend the definition of coercion to remove such instances, specifically excluding the removal of benefits as a form of threat. I’m inclined to bite the bullet and call this coercion, but insist that it’s not a rights violation.

This definition seems much more in line with our intuitive understanding of what coercion is, and surely proves more useful in political debate. On Tom’s definition we couldn’t talk about paternalism, if it really makes the individual in question better off, being coercive. The distinction between coercive and noncoercive paternalism is surely a useful one. If you’re a utilitarian who thinks that people often act against their best interests, you may want to say that coercion is often justified, but to deny that it is coercion seems a touch Orwellian.

I don’t see coercion as necessarily wrong. I agree with Tom that private property is a coercive institution, but justified on (broadly speaking) utilitarian grounds, and I think some level of taxation is justified, but maintain that taxation is inherently coercive. Coercion is, however, a prima facie wrong, and requires justification in each case, with the burden of proof lying with the coercer.

Free Will

Bryan Caplan has an interesting post on the philosophical problem of free will and behavioural genetics.

OK, now let’s get Bayesian.  If you could fully account for a person’s choices using genetics and measurable environmental variables, you’d count it as a confirmation of determinism, right?  Well, if you buy this argument, you also have to buy its mirror image: The harder it is to account for a person’s choices using genetics and measurable environmental variables, the stronger the case for free will.

I see predictive ability as utterly irrelevant to the question of whether or not we have free will.

The standard argument against free will goes something like:

1. Action X is freely chosen by agent A if and only if A could have done other than X

2. All events (including actions) are causally determined; therefore,

3. If agent A performs action X, he could not have done otherwise (from 1 and 2); therefore,

4. Action X is not freely chosen by agent A (from 1 and 3); and by generalization,

5. We do not have free will

In other words, if our actions are entirely caused by a combination of our genes and environment, things we have no control over, our actions are predetermined and therefore not freely chosen. All our choices are belong to the laws of physics.

Some people try to escape this conclusion using indeterminacy at the quantum level to argue against premise 2. This doesn’t work. Even if quantum indeterminacy filtered through to the macro level (doubtful, and if not we could just change premise 2 to ‘all events at the macro level are causally determined’ and reach the same conclusion), it would be pure random chance adding noise to causality. This surely doesn’t give us free will in any meaningful sense.

Imagine a simple robot programmed to kill any human it comes across. Does it have free will? Now imagine that instead of killing every human, it is programmed to flip a coin to decide the fate of those it comes across. Does it have free will now? Surely the answer has to be the same in both cases. Random chance doesn’t give us freedom.

I think free will and determinism are compatible. If we take humans for what they really are, i.e. meat machines conditioned to behave in certain ways by natural selection, the free will problem becomes tractable. There is no brute mental entity making choices in a vacuum, but it is us making choices nonetheless. We are physical things (we are also mental things, but every mental thing is a physical thing, differently described), and the causal determinism of the universe flows through us, as just another part of said universe, to produce our actions. Who we are and the choices we make may be entirely predetermined by the prior state of the universe, but this does nothing to alter the fact that we act based on preferences and reasons.

We could have acted differenetly had we been different people with different preferences and reasons, even if we could not have been different people. It doesn’t matter how we came to be the people we are, only that we are those people and that we make choices. That’s the only freedom we are are ever going to have and I, for one, am grateful for it.

Wise Words

From Bryan Caplan:

Weird hypotheticals are philosophers’ equivalent of controlled experiments.  When a scientist wants to test a physical theory, he sets up weird laboratory conditions that make it easy to find an exception to the theory.  Similarly, when a philosopher wants to test a moral theiry, he sets up weird examples that make it easy to find an exception to the theory.

Counterfactual resistance is the bane of my existence.