NZPols

Tuesday, December 16, 2003

PROCRASTINATION II: My Philotype(TM) is Baconian (apparently):

You have a fascination with machinery and mechanism, and suspect that the whole universe is a vast organic machine. You sometimes wonder where the "off" switch is, and what would happen if you flipped it. Your commitment to empiricism sometimes leads you to get lost in the wonder of the details, and your tendency to lapse into mathematical discourse and inappropriate uses of the word "orthogonal" make other people look at you funny. You write thick books at a moment's notice full of ingenious ideas that are widely reviled by the intellectual leaders of your day.

¶ 9:27 am

Monday, December 15, 2003

PROCRASTINATION: WHICH PHILOSOPHER ARE YOU? That's right, it's the latest in the endless stream of "who am I?" quizzes, for those with so little self-identity that they feel a constant need to "be" other people/cartoon chararcters/household appliances etc. As I am wont to do, I took the test more than once.

Surprisingly both sets of answers were apparently 100% compatible with Jean-Paul Sartre. (With a less surprising dose of John Stuart Mill thrown in for good measure.)

I can't tell whether this means it's a steaming load of turd, or that it's so good it could see through my cunning attempts to fool it by switching answers.

¶ 7:52 pm

Tuesday, December 02, 2003

THINKING OF THE CHILDREN: NRT then accuses me of "resorting to this tired old line" to justify social engineering. Apart from the fact that being a "tired old line" didn't (last time I checked) constitute a valid logical criticism, I think NRT may be pursposefully dodging my question. We don't actually disagree on the substantive issue of whether the government should get involved in social engineering here (though unlike NRT I'm open to being convinced). What I take issue with is this apparent claim that rights act as side constraints upon action.

Encouraging" marriage is as illegitimate as "encouraging" people to go to church. These things are simply too important to people's life-plans, their sense of self, for the government to interfere with them - no matter what the benefit. [Emphasis mine]

No balancing of rights allowed. No weighing of benefits. Who cares if the whole world is destroyed, as long as the government doesn't give tax breaks to married couples. Seemed a bit extreme, which is why I aksed whether it was really what NRT thought.

NB: If required, the reasons why I think that rights-as-side-constraints don't work can be found in the second half of my previous post.

¶ 8:32 pm

PRESUMPTIVE RULE CONSEQUENTIALISM AND THE PROBLEM OF JUSTICE:

From this post, it seems that NRT and I agree on the following:

(a) a consistent consequentialism will allow (indeed, often require) a presumptive rule consequentialist approach; and

(b) presumptive rule consequentialism must face the "justice" problem.

What we seem to disagree on are:

(c) whether there is a distinction between presumptive rule consequentailism and act consequentialism.

(d) whether the "justice" problem is actually a problem.

On the first point, NRT writes:

it's difficult to see how presumptive rule consequentialism differs from act consequentialism in any significant way

I don't really know how to explain this any more clearly. One has a presumptive rule, the other doesn't. One bases decisions on the rule unless a sufficient level of certainty is reached; one bases decisions on the balance of probabilities in every instance. One leads to better results than the other.

Perhaps an illustration will help. I'm 60% certain that dobbing in the inncoent will lead to better consequences than telling the truth. Act consequentailist: dobs in the innocent. Presumptive rule consequentialist: tells the truth.

I apologise if I'm labouring the point, but I thought this was really obvious the first time around.

On the second, the claim is that:

The retreat to rule consequentialism is driven primarily by clashes between consequentialism and justice. The classic example here is that of McCloskey, who points out that consequentialists (well, he talks about utilitarians - hedonic consequentialists - but the argument is applicable to general consequentialism unless an infinite disvalue is placed upon certain actions) must bear false witness against an innocent person if it that would stop a riot. Or use collective punishment to prevent guerilla activity in wartime. Or impose harsh punishments on the guilty to meet public demand for retribution. Or support laws designed to defend an insecure democracy by making it a crime to "arouse the suspicions of the government".

In other words, any arbitrary injustice is allowed provided it makes people overall better-off by whatever metric is used. This is widely regarded as a knock-down argument for act-consequentialism.

Depending on how you construct the examples, I don't think the conflict between consequentialism and "justice" is necessary. (And I have to say I'm not particularly fond of the term "justice" in this context because it seems to assume the fasilty of consequentialism to begin with. In fact, I think that this sort of circular assumption is the only way to turn McKloskey's argument into a "knock-down" one.)

But let's assume we push the examples such that they guarantee a conflict. In that case I would bite the bullet. Take the choice to bear false witness. To guarantee the conflict, assume that the choice is between sending one innocent to the gallows and watching a rampaging mob kill 10 innocents. In this case, I think "justice" loses. I have no reason to prefer the life of one innocent to the lives of 10, except for the delusion that the blood of the 10 is not on my hands, because "somebody else did it". If I know that my actions will result in those deaths I cannot disclaim responsibility, and the attempt to do so seems to me to be little more than a convenient psychological defence mechanism, dressed up as morality. The way this appears to turn what many people consider "moral" into nothing more than a form of self-serving egoism will be somewhat disturbing to some people. Indeed, it was disturbing for me too at first. But that doesn't mean it's wrong.

I do think that it is important to draw a distinction between the good (fewer people dead), and individuals' obligation to pursue the good (by bearing false witness). Because it is psychologically difficult to deal with being responsible for the death of innocents, it may not be morally required to bear false witness, if telling the truth allows one to maintain the illusion of not being responsible. However, the claim that it is morally illegitimate to save innocent lives, cannot, in my view, be sustained. So I'll stick to my presumptive rule consequentialism thank you.

NB: The "retreat" to rule consequentialism isn't driven by the "justice" problem at all. If you accept the "justice" problem is a problem then you don't accept consequentialism and if you run to rule consequentialism as a result of it, you've abandoned consequentialism altogther. (Evolutionary biology might be able to explain the emergence of moral intutions about "justice" - which might be thought to undercut the force of the justice problem - but that's entirely spearate.)

¶ 8:30 pm

HAVE TOYS, WILL THROW: Okay, No Right Turn, you win. I give up. Your arguments have battered me into submission, and I can take no more.

However, before conceding defeat, I feel obliged to let you know that this has little to do with their strength. If that were all it was about, I be happy to continue our battle of wits to its eventual resolution. But the fact of the matter is, I just can't be bothered with the endless cutting and pasting of your comments into separate posts over here, every time I want to make what should be a two line rebuttal of your (or, more often, a two line clarification of my) position.

It's your choice whether you get comments or not. But if you want to respond to my arguments, why not use mine. I know they can only handle 1000 characters at a time, but if that's what’s bothering you, I can change.

Or not.

¶ 8:30 pm

Monday, December 01, 2003

A QUESTION FOR NO RIGHT TURN: No Right Turn says:

as a liberal I think there's a wide sphere around the individual which is off-limits to government intervention. Libertarians would justify this on the grounds of natural rights; Rawlsians from the original position; Kantians on the grounds of respect for autonomy; and rule consequentialists on the basis that people are generally happier that way. What's in that sphere will vary depending on your specific basis, but I think all four versions agree that what you believe and who you live with are well and truely [sic] covered. "Encouraging" marriage is as illegitimate as "encouraging" people to go to church. These things are simply too important to people's life-plans, their sense of self, for the government to interfere with them - no matter what the benefit.

No matter what the benefit? Recall that this is in the context of preventing violence against children, something which is usually considered to be a pretty serious rights violation itself. In apparently refusing to balance these rights, you appear to be endorsing a very Nozickean "rights as side constraints" type view of the world. Is this REALLY what you mean? (If so, I think your view is incoherent, but I'll save that argument until you respond.)

¶ 11:52 pm

GOOD LOGIC, BAD LOGIC: Earlier today, No Right Turn nicely pointed out the logical flaw in NZPundit's tirade about Helen Clark.

However, on a separate matter, I think NRT's logic is somewhat lacking. In a tangent to an argument about the role of the nuclear family NRT appears to imply that a presumptive rule consequentialism of the sort I advocated in a previous post was somehow illegitimate: either you're a rule consequentialist or you're not; no exceptions.

A quick rejoinder should suffice. Perhaps I'll get the chance to develop this in more depth at some point in the future.

Assume that act consequentialism in respect of a particular issue has an average payoff of:

AC = 0.6x100 + 0.4x(-100)

= 20 utils (or whatever).

(This assumes a 60% chance of guessing right, with a 100 util payoff and a 40% chance of guessing wrong, with a -100 util penalty.)

Assume also that a rule consequentialist approach has an expected payoff of:

RC = 0.8x100 + 0.2x(-100)

= 60 utils.

(I.e that the rule will gove the right decision in 80% of cases, and the wrong decision in 20%.)

A consistent consequentialist, faced with a choice between act and rule versions, should choose the rule, resulting in a net gain of 40 utils. However, there remains a problem, in that there will still be cases where the rule clearly results in bad decisions. If we know this with a sufficient degree of certainty in a particular case, surely we should depart from the rule? Although this might appear to be a return to the failed strategy of act consequentialism, it is not, provided that we only depart from the rule when a certain level of certainty is reached (under act consequentalism decision would be based on the balance of probablities).

In fact, in the long run, we would get better consequences if we departed from the rule in any instance where we were more than 80% sure it would give the wrong result.

Assume that n = the proportion of cases in which we are more than 80% certain that the rule is wrong. Then the average payoff of such a strategy is:

PRC = [(1-n)*(0.8*100 + 0.2*(-100))] + [n*(y*100 + (1-y)*(-100))], where y > 0.8

= [(1-n)*(0.8*100 + 0.2*(-100))] + [m*(0.8*100 + 0.2*(-100))], where m > n

= [(1-n+m) x (0.8x100 + 0.2x(-100))]

= (1-n+m)*60 utils > RC

The maths is wee bit bit messy, but the point is that you can usually improve on rule consequentialism in the long run by allowing departures from the rule in limited circumstances. In fact, far from being incoherent or illegitimate, this would seem to be a requirement of a consistent consequentialist approach.

¶ 11:28 pm