Tuesday, February 11, 2003

Rules, Cases, Analogies

God of the Machine has rules. He talks about the innocent casuistry I propose - case-based analogical reasoning - as though it were covertly rule based, and he calls this "little-rulism". He says,

Little-rulism is an accurate description of how most people actually do reason morally, most of the time.... (I speak from long and painful experience.)

Well, I speak from long and happy experience in saying the same thing. He says that "big-rulers don't have this problem" of engaging in messy analogical reasoning, but I suggest that if you want to do a job right, you have to get your hands dirty. Applying one master rule to all moral deliberations is clean, but no one has ever produced a rule that could perform the task truth-conducively. Any rule you can name generates the wrong judgment in some cases if it is taken strictly.

This is why critical common-sense-ists - people who conduct messy moral deliberations by reasoning by analogy - are not even little-rulists. They don't use rules. Rules hold in every case. The little rules don't, so they must be treated merely as rough generalizations; call them "principles".

So, we have the case of killing innocent Iraqis. Aaron says, "[F]or any little rule you formulate, a case will fit it either exactly or not at all." Yes, but it fits more than one "little rule" (principle): don't kill innocents; it is permissible to defend yourself. So, when Aaron says,"First, why not, as a little-ruler, stop at the first rule? We have found a perfect fit," the answer is that the case will fit one rule better than any other conflicting rule, since one and the same action cannot be both right and wrong. And when Aaron says, "Second, if we continue to pursue the question, when do we stop? How many little rules must we examine?" the answer is that we stop when, upon careful reflection, we cannot find a rule which fits better than the leading candidate rule. Abortion, or euthanasia of the brain-dead: "Well, it's killing a human, and that's wrong.... But, wait, this human is not a person; it has no brain activity. So, it is not wrong....I can't see that any other rule overrides."

And, "Finally, since all applicable rules will fit the case perfectly, how do we adjudicate among them, lacking 'more' or 'less' analogous as a standard?" Aaron claims without argument that critical common-sense-ism is really rule-based and not based on analogical reasoning. I maintain that it is not rule based but based on analogical reasoning. Still, the question is a good one. How do you know which of two conflicting rules to apply? Well, they aren't "rules" but principles; each is a gesture at a set of analogous cases about which we would make the same judgment. Whenever the case at issue has been shown to have a property listed by a rule - rule 1 - as relevant, rule 1 is applicable: a candidate for the place of "overriding rule". But you will often find a rule 2 which is also applicable and overriding in all other cases in which rule 1 is applicable; it therefore replaces rule 1 as the leading candidate for the case at issue. You look for a rule 3 to displace rule 2 similarly, and proceed until exhaustive search shows that a rule to displace the leading candidate rule cannot be found. (You could be wrong. Twenty or 500 years later, such a rule could be discovered; the case of refuting the morality of slavery in the U.S. South is an example, as is the women's suffrage movement.) In addition, you might instead have found that rule 2 overrides in some cases in which rule 1 applies but is defeated by rule 1 in other cases. Obviously, a third rule is at work, as well. So, you look for a rule 3 which overrides rules 1 and 2 in other cases, and you conclude that it overrides in the case at issue, as well. This is what it means to say that moral reasoning is the endeavor to determine whether the case at issue is "most analogous to" the cases in which rule 1 overrides, or to the cases in which rule 2 overrides, or the cases in which rule 3...., etc. Finding the overriding rule, the final winner, is finding the set of cases to which the case at issue is "most analogous". (Again, I've used the word "rule" when I shouldn't have; these "rules" don't override in every case in which they apply; they're only principles.)

Critical common-sense-ism is like Aaron's orderly hierarchy of rules, except that, contrary to what Aaron thinks, there are too many rules for a reliable finite decision making proceedure to be written. Practical reasoning isn't rule-based because the cases we encounter are indefinitely many in kind, shading from the rule 1 shade of gray to the rule 2 shade in so many ways that the hierarchy cannot be given as a formula, let alone formulated on the basis of one rule, as Aaron, Kant, Hobbes, and the utilitarians try to do. Even the practical reasoning involved in finding your way from your house to a location downtown is so complex that rules won't work to describe it. (This is why AI is not going to be rule based, but neural-net based. Nets are weighting connections, not rules.) Scientific reasoning bears the same complexity. Math, deductive logic, and perhaps jurisprudence are the only kinds of genuinely rule-based decision making procedures, as far as I can tell.