This is the last of our series of posts about the prediction theory of law (PTL), and we may now sum up our negative assessment of the prediction theory as follows: although we concede the prediction theory cannot explain the law from Hart’s internal point of view, at the same time we must concede that other major theories of law also do a bad descriptive job of explaining how judges decide cases. (More simply put: the prediction theory sucks, but so too does every other theory of law!) Given this theoretical stalemate, what is to be done?
There are two possibilities. We could either try to formulate an entirely new and improved theory of law à la Hart, Dworkin, or Posner, or we could just “satisfice” and move on, i.e. try to figure out which theory of law is “less wrong.” In the remainder of this post, I will explain why I believe the prediction theory to be less wrong. Now, if all we want to do is figure out which theory of law is less wrong, we still need to state up front what our criteria of theory choice are and how those criteria are going to be weighted. Here, I will follow Brian Leiter’s lead by using the following three criteria: simplicity, scope or “consilience,” and coherence. (Prof Leiter identifies these criteria in his thoughtful paper “Explaining Theoretical Disagreement.” Although Leiter himself does not explicitly assign any specific weights to these criteria, for simplicity we will weigh them equally.) Next, let’s evaluate the prediction theory of law and its competitors in light of these criteria:
- Simplicity. Leiter’s first criterion is simplicity, i.e. we usually prefer simpler explanations to more complex ones. Here, however, it’s hard to say which theory of law is the simplest one, since all of these theories can be summed up in a single sentence or phrase, such as Dworkin’s normative notion of law as integrity, Hart’s influential concept of law as the union of primary and secondary rules, or Holmes’s simple idea of law as a prediction of what courts will do, so let’s move on.
- Scope. As Leiter correctly notes, we usually prefer more comprehensive or “consilient” explanations–i.e. theories that explain different kinds of things–to explanations that are too narrowly tailored to one kind of datum. Here, the prediction theory wins hands down. Why? Because the prediction theory is based on probability theory, and probability theory can be used to make predictions in a wide variety of domains. Competing theories of law, by contrast, are limited to the domain of law. (In fact, competing theories of law are even more limited than that–limited to the behavior of appellate courts.)
- Coherence. Lastly, Leiter notes that we usually prefer explanations that leave more of our other well-confirmed beliefs and theories intact to those that do not. Again, the prediction theory of law wins hands down. After all, as we have already noted, the prediction theory is based on probability theory, and the axioms of probability theory are already well-established and beyond dispute. (Although there are some fundamental disputes at the margins of probability theory, such as the controversy between frequentists and Bayesians, these theoretical disputes pale in comparison to the disputes about the nature of law among legal philosophers.)
There is a fourth reason why I prefer PTL to other theories of law: the prediction theory is a purely descriptive or non-normative theory of law. Dworkin and Posner’s theories of law, by contrast, are normative ones, since their theories ultimately rest on a single master value–in Dworkin’s case it is integrity, while in Posner’s, it is some ill-defined form of pragmatism or consequentialism. For his part, although Hart’s theory of law purports to be a descriptive one, normative values can slip in through the back door of Hart’s theory. In short, the prediction theory wins by default–or in boxing terms, by technical knockout (TKO). Where have I gone wrong?