Sunday, September 09, 2007

Legality vs. Rationality

What is the difference between a legal rule and a rational, maximizing judgment?

it seems to me that the legal rule is a much simpler beast. the legal rule asks if a particular act or case meets the clear, objective criteria set down in a law, one that is clear enough so that most anyone with a command of the language can agree if a case--assuming all the facts are known--meets it.

A judgment, a rational decision under uncertainty is much more complicated. that is why different people can come to different judgments even if all the 'facts' are known.

it is conceded at the outset that these represent points at opposite ends of a continuum and that most cases will lie somewhere in that continuum. I only argue that as one moves toward one end or the other of the continuum the type of decision being made and the type of decision maker that is most appropriate and likely to make a good decision shifts.

I should also make clear that what constitutes a "legal" decision has shifted in recent years toward the rational judgment. For example, Roe v. Wade was hardly the enforcement of a pre-existing rule. it was the invention of a new rule from pre-existing values. whether this new rule was a good thing or a bad thing is no germane to my argument. I merely contend that such decisions are not laws in the sense in which law was understood traditionally. if it was, then such things as the strictures against ex post facto laws or laws that are not publicized would be meaningless.

A legal rule takes the form of a classical definition. A genius and a differentiae. There area some extremely complicated rules that defy these categories but often even these are more tractable than they might first appear because of the well-behaved, hierarchical nature of traditional categories in the law. they are like objects in object oriented programing that inherit all the qualities of the genus (it is no accident that object oriented programing languages borrow from classical logic) or definitions in number theory--a rational number implies all the qualities or real numbers etc. It may take a lot of empirical effort to verify the existence of all these qualities, but their hierarchical nature makes them easy to communicate.

Even if the elements of the genus are not hierarchically nested, they are still non-compensatory, which makes them like the dimensions of evaluation in satisficing.
this is a decision making procedure that is far simpler and makes far fewer demands on the decision maker's cognitive capacity. it also makes the decision making procedure more stable and predictable and transparent.

Finally, the dimensions in a legal decision tend to be defined as nominal variables rather than interval level or ordinal level variables.

Legal decision making is, thus, pattern matching, political judgments are rational decisions under uncertainty.

(Perhaps One reason that we have been drifting toward legal decision making and away from political decision making is the fact that intellectuals find pattern matching more satisfying than rational judgments under uncertainty. the latter is so messy, and you can never tell who won the game until after reality has made its move....)

Rational decision making nder uncertainty entails making value trade-offs that are not specified. we often cannot describe the trade-off between some values. often, even if we can describe the trade off in some range of the issue space but not for others. sometimes the trade-off shifts along the range of values (eliptical preferences). All of these things make such decisions irreducible to a well behaved rule. of course judges can make such decisions, but when they do so they are not applying rules, they are making judgments. it is because judges have been making more and more such decisions and that we have become more concerned with their character. we can't predict their decisions because their decisions don't depend on fules but on utility functions that are too complicated to write down or do not yet exist in any meaningful sense apart from their personal "preferences" before an actual case is before them.

The different dimensions are compensatory. if they were not, if they were like the dimensions of evaluation of a satisficer, the decision making would be more predictable and more like legal decision making. This compensibility is what makes the judgments "subtle" and "case dependent," all of which is another way of saying they are not applications of rules from outside the judge but applications of principles that lie within him.

In addition to the weighting of the dimensions of evaluation being unspecified and often as not unspecifiable, the probability judgments that go into choosing alternatives are empirical judgments that judges usually do not pretend to have knowledge of.

These judgments of probability depend on having experience in the world. they are also something that depend on an ability that we can't really observe ex ante--some person may have excelllent professional credentials and be lousy at it and others may be just the opposite (Churchill comes to mind, Roosevelt, with his second rate mind but first rate character, is another. Truman vs. Hoover. Presidential history is replete with such examples).

Given this, it is preferable that we place the ability to make such decisions in the hands of people we can fire. Judges in our system are defined by their life-tenure.

what people are really saying when they want the rule of law in international relations is that they want the kind of people that become judges and international legal scholars to make decisions about war and peace, not the kind of people that become politicians. This may or may not be a good idea, but it has nothing to do with the rule of law and everything to do with the rule of judges.



Law must be confined to the plain meaning of the words on the page. If it is expanded to mean whatever constructions are placed upon those words by the contending members of the legislature and the executive then law degenerates into deconstructionism armed with search engines.

No comments: