In the course of my research, much of which relates to selection of legal and ethical decision rules, I’ve often pondered the meaning of the word “rule” itself. What does it actually mean to have a rule, or to be guided by one? The more I think about it, the more I think that a rule means a prescription for action that relies on an intermediate degree of abstraction. When rules become either too abstract or too situation-specific, they begin to lose their rule-like character.
Take the rules of etiquette. One rule is to say “Thank you” when someone gives you a gift or performs a service for you; another is to say “Please” when you wish someone to do something for you; and so on. Now, imagine if we replaced these rules with a single, highly abstract directive to just “Be polite.” That directive would not provide much useful guidance. Lacking more information, the decision-maker would have to decide for each and every interaction what would be a “polite” thing to do.
On the other hand, what if we had a different prescription for every possible interaction? One directive for when someone passes the butter, another for when someone passes the salt, another for when someone holds open a door, another when someone holds the elevator, etc. The decision-maker’s problem would be very similar to that which he encountered under the “Be polite” rule. Any time he encountered a novel situation – and arguably, every situation is novel in at least some infinitesimal degree – he would have to decide the correct action, without much assurance that it’s correct. (Would it be okay for him to use the passing-the-butter response the first time someone passed him I-Can’t-Believe-It’s-Not-Butter?)
The actual rules of etiquette have an intermediate degree of abstraction, neither so broad as to include all situations, nor so narrow as to differ for each and every specific situation. They identify abstractly-defined types or kinds of situations.
The same analysis applies to legal rules. A rule of tort liability that said “Do the right thing” or “Be careful” would not be terribly helpful, unless the decision-maker had some idea of what other people think is right in more narrowly defined types of situation. But if the categories were too narrow, depending too much on the particular characteristics of the particular situation, the decision-maker’s dilemma would be the same. A more useful rule of tort liability tells the decision-maker how to act in situations with an intermediate degree of abstraction; e.g., “Always yield to cars already on the freeway when merging.”
I surmise that my definition of rule-ness could be applied to the rules of virtually every area of human interaction: the rules of language, the rules of ethics, the rules of games, etc.
One implication of my position is that many so-called rules of law, such as the ubiquitous “balancing rules” and “reasonable man” tests, are not very rule-like at all. They are standards, which have a higher degree of abstraction than rules. Standards can be useful in choosing rules, and in dealing with novel situations that arise in the “cracks” between rules. But they can also create considerable uncertainty if allowed to substitute for, rather than supplement, rules with an intermediate degree of abstraction. A similar problem afflicts case-by-case decision-making, which errs on the side of too little abstraction instead of too much.
I realize that I’m shifting between a descriptive notion of rules (what we mean by the term “rule”) and a normative notion (what rules should be like). But I happen to think the entities we call rules usually do have this characteristic of intermediate abstraction, despite some exceptions. If we sometimes apply the word rule to entities that are highly abstract or highly specific, it is because abstraction and specificity are separated by a gradient, not a sharp line.
Comments are closed.