Introduction
I’m at the Southern Political Science Association meeting in New Orleans. Legal academics sometimes forget that there is a parallel universe of legal scholarship in the world of political science (or “politics” or “government”), done from a diverse array of methodological perspectives—ranging from the formal models of positive political theory to old-fashioned attitudinalism to what is sometimes called “the new institutionalism.”
I’m attending a panel entitled “Formal Theory and Judicial Politics.” James Rogers of Texas A & M is chairing the panel & the other participants are Cornell Clayton (Washington State) and Jeffrey K. Stanton (Florida State). The format is a “roundtable discussion,” so there are no formal papers. I’m going to report some impressions of the speaker’s remarks, but given the format, my liveblogging will be even more impressionistic and sketchy than usual. These are my thoughts about the panel and not even close to a transcript. One more thing: I participated extensively in the questions and answers after the panel, but I was not able to blog my own remarks and exchanges with the panelists.
Rogers begins by introducing the panel. He explains that each speaker, including himself, will make a few remarks, followed by conversation among the panelists and with those in the audience.
Cornell Clayton
Clayton begins by explaining that he is not a formalist. His role is to express a skeptical or 
critical view. He is not as hostile as some might think, but he does have some questions. He is reluctant to engage in abstract debates about method. As a methodological pluralist, he tends to think that the question of methods should always be secondary to research subject. The process of theory building ought to be an iterative process. There is a world of difference between theory laden and theory driven research (referring to Ian Shapiro’s distinction); research should never be theory drive. Cherry picking data or advancing nonfalsifiable hypothesis is simply bad science.
We should always embrace a self-critical attitude toward our methods. The question of what value is added should be front and center. Because formal theory has high access costs, those who engage in it have a special burden to explain what the value added is. Clayton discusses a paper he heard from the formal approach with 64 pages of dense formal apparatus that suggested that supermajority rules dampen special interests. His reaction was, “Have you heard of James Madison”?
Far from being opposed to formal approaches, he thinks they have much to offer. He sees rational choice theory broadly—it is explains behavior in terms of goals. What makes rational choice theory exciting is that it places human agency at the center of consideration. Humans are motivated actors who interrelate with one another—this is the key to considering human behavior.
He questions formal work that fails to take human motivation seriously. Some time ago, Clayton and Howard Gillman edited two volumes on Supreme Court decisionmaking. At that time, they suggested that examination of motivations was welcome. But they questioned whether formal approaches that borrowed the attitudinalist approach to motivation would work. The new wave of strategic choice literature was different because it looked at motivations. The first wave recognized that judges cared about legal rules, judicial role, and other non-policy preference. Beginning in the late 80s and 90s, a second wave of strategic choice scholars developed more complex models, but they substituted the attitudinalist model of motivation for the more complex story of the first wave. Most of this formal work cited Segal and Spaeth, and some saw these formal approaches as generalizing the attitudinal model.
Obviously in some cases the attitudinal model has explanatory force. Gillman and Clayton argued that the attitudinal model was flawed. The attitudinalists were attacking a strawman version of the legal model. In the post-realist world, no one thinks law is just black letter law. Historical institutional scholars (Whittington, Gillman, etc.) have demonstrated that judicial preferences are more complex than attitudinalists permit them to be. So the second wave went off in a direction that Clayton sees as misguided. Judges do not calculate who the median member of Congress is.
The formal models lack resonance in the real world of experience. They are “just so stories,” implausible stories of what is really going on. What is promising is the third wave of strategic research. Rather than assuming an unrealistic dichotomy, these stories tell a more a complex story about judicial motivation (Tom Ginsburg’s work, etc.). The most exciting work will focus on the unbundling of the idea of motivation. In doing this it is likely the research will draw new linkage with normative, historical scholars. Recent research by institutionalists who are interested in understanding how political regimes affect courts and judges have used strategic frameworks.
Clayton’s presentation was insightful and effective–and I agree with much of what he had to say.
Jeffrey Stanton
Stanton will address the role of formal theory in comparative politics. He will also talk about
the value of formal theory as a method. And finally, he will discuss the value of formal or mathematical approaches.
Stanton begins that it is not very useful to ask what formal theory has uniquely contributed to our substantive understanding. Formal theory does not develop independently of nonformal theory. It is unhelpful to think about the research tradition. It is ultimately about core problems. It is insulting to nonformal theorists to focus on what is unique. Formal theory is a method. It would be like asking what computational linguistics contributed.
Formal theory is a method that applies to any positive theory of politics. There is always some use to formal theory. Most positive theories of behavior have a few things in common. They identify actors, what they care about, what they know, what they can and can’t do. Formal theory just does this. What actors care about are represented by functions. What they can do are their strategies. We can define their information construct. So this can be applied to any argument. It is a method for producing consistent predictions. It is useful for insuring that predictions are consistent.
So here are some examples of formal theory from comparative studies. What are the conditions under which you can constrain the state? Or you identify a problem, torture, and ask how institutions can solve that problem. Consider the first example, constraining the state by law. In the early 1980s, judges in Argentina started ruling against those who appointing them. Why did they do this? The argument is that they anticipated the fall of the regime, and they were trying to convince the new government that they (the judges) were worth keeping around. It is a neat story, but why would the new regime keep them around. And why wouldn’t the judges recognize this? You need to answer those questions, which Gretchen does in her book.
Tom Ginsburg’s work is another example. Why are constitutional courts established? How do they acquire power? Tom’s argument is that ruling coalitions use courts as an insurance policy against future electoral loss. After a court is constructed, they initially don’t have a lot of power and must be very careful. They transition into a phase where they have much more power. There is a tension between the two arguments. If you want insurance, it is not clear that lower power courts have the power to do that. (At this point, Stanton quotes an extended passage from Ginsburg’s work, which I won’t try to reconstruct.) Stanton suggests that some of the questions that Ginsburg’s work raises can be answered by formal models.
James Rogers
Rogers has noticed that there is some loose language in discussions of these issues. There 
is a difference between several different approaches. Rational choice models do not need to be formal. Many writers apply soft rational choice, approaches, e.g., Madison. In this sense, someone who thinks they are Napoleon can be rational so long as he can rank preferences and his ranking is transitive. This is a very thin notion of rationality. There is a lot of human behavior that is irrational—choice are not consistent or choices are flummoxed or choices are based on intransitive preferences. It is fine for psychologists to explain irrational choices. But if there is a field of activity that can be explained on the basis of rational choice, then the models have space.
Formal models do not need to be rational choice models. You can stipulate any kind of behavior. You can formalize rational choice easily, but there is no requirement that mathematics serve rational choice.
Game theoretic models only require interaction, or more strictly, dependence of one actor’s action on another actor’s action. To the extent that what I do depends on what others do is “strategic.” Game theory does not require awareness of the game. Game theory may be best when things are so internalized that we are unaware of it. I drive on the right hand side of the road, which can be explained by game theory, even though I don’t think it through in game theoretic terms. Hence, it is irrelevant that judges don’t think they are policy motivated. But of course, policy motivation may not be the only thing in the utility function.
Decision theoretic models are game theoretic models with one person dealing with an environment. Public choice includes a normative aspect that game theorists don’t need to accept.
We need to categorize in order to distinguish things, but there is a continuum—from prosaic theory (in prose) moving on a continuum to more or less formal theory. Most formal theory papers explain what is going on in prose. There is a continuum from purely verbal to a line with points to greek letters and the like. The question is what is the benefit of different sorts of abstraction. The idea that it is “us versus them” is not helpful.
Given the costs, formal theory needs to be justified. Quite often Rogers gets papers were there is a sensible idea that is formalized, and his response is “who care?” Translating something that can be understood without the math into formal terms is not worthwhile. There is no public reason to mathematize for the sake mathematization. There is a temptation to write for the sake of fun, but that is what the review system is for.
Why is formal theory valuable? “Unrealism” is valuable in some situations. If I were to lay out a map to get us from here to the hotel, and someone said it is not a realistic representation of the world, he would say, “but that is the point to having a map.” There are good maps and bad maps. A bad map would have way too much detail. You need just enough technical detail to get the insight. So unrealism isn’t a vice, it is a whole point. Verbal models abstract as well. In order to understand reality, we need to break it down.
There is “extended theorizing.” There are cases in which formalizing allows a longer chain of reasoning. There is a cost; we need to throw details away, but that is required.
The question of policy motivation—Rogers agrees and disagrees with Clayton. When Rogers got into the business, he was excited by the scholarship. Most of the scholars moved the policy space from legislation and moved it over to the judiciary. That was insensitive to the behavioral aspects of judicial politics. The critics tend to reify a stage of modeling development—perfect information and policy motivation—and then suggest it is unrealist. Rogers agrees, but at the same time those models were a first step. The insight that judges can act just like other politicians is important. Can we account for “sublimated behavior.” It is fun and useful to see how far we can push the simple assumptions.
So Rogers has a paper that assumes judges are legislators in robes. The assumption that when we allow judges to behave as Lochnerian judges, policy outcomes are better than if we require them to be deferential judges. The interaction of self-interested actors carves out a space for economic liberty. This is “fun and interesting.” That being said, an assumption
is only an assumption. If it doesn’t work, then you abandon the assumption. Critics seem to be unaware of the evolutionary nature of formal theory. An example of this is the way that models now can account for uncertainty.
Discussion
Clayton suggest that he agrees with almost everything Stanton and Rogers say. He agrees formal theory is different than rational choice. He agrees that models do not need to be consciously aware of what they are doing. How do these norms become so internalized that we don’t think of them? Many studies are part of the process of scientific inquiry: you must deal with problems that are internal to the theory. His concern is that sometimes we lose sight of the point that the point of internal work is to deal with the external.
Alec Ewald if we asked actors such as Senators why they take positions in nomination hearings, they would own up to certain nominations. Honest judges have been socialized to say they are not maximizing their policy preferences. Rogers is open to the “false consciousness” claim. He has a law degree & interacts with scholars who love the Supreme Court and are emotionally wedded to a certain view of what judges do. Because law is a system of authorized coercion, we can’t tell the truth to ourself. He is open to the possibility that judges are just achieving policy goals. Rogers agrees with Clayton that the case has not yet been made.
Clayton says that many formal modelers keep motivation as an open hypothesis. As a matter of theory, the question of motivation must be prior to interaction. Rogers What is an implausible motivation and why? Clayton So if that goes into an interactive model, it is implausible. It is implausible, because it has an implausible distinction between legal and political preferences. At the heart of model, is a distinction between legal and political preferences. Dichotomizing adopts an unrealistic view of law. Rogers Is it wrong to have one dimension?
I asked a question at this point, which is omitted here.
Clayton and Rogers had a discussion about the role of policy preferences. Rogers thinks that the attempt to model based on policy preferences still has some play.
Another question from the floor related an experience where someone who had served as a law clerk for a trial judge was given authority to decide cases with almost no supervision from the federal judge for whom she clerked.
Some Reflections
I’m an outsider to this debate. Although I frequently attend political science meetings, present papers, and even do some modeling, my primary interests and methodological orientations are not those of political scientists. Nonetheless, I am deeply interested in the debates between and among political scientists about the role of rational choice theory, thick historical description, legal doctrine, and other approaches.
I am of more than one mind about the methodology debates among law and courts scholars. First, and perhaps foremost, I am a methodological pluralist. In other words, I believe that different problems are amenable to different methodologies and some problems are best understood if approached from multiple points of view. For example, if we are interested in understanding the behavior of the United States Supreme Court, I believe that formal, rational choice models have a contribution to make, but so does conventional legal theory (doctrinalism), normative constitutional theory, history, and so forth. This might seem uncontroversial, but I am not so sure. In panels like the one from which I am blogging, everyone is anxious to appear reasonable and respectful of their methodological opponents. But when the various groups meet among themselves more extreme and strident views emerge. You may have heard the various refrains: “Formal rational choice theory is simplistic bullshit dressed up in fancy mathematical clothes.” “New institutionalism is for people who can’t do math.” Well let me join in the casting of aspersions: “Methodological monism is myopic.” Now that’s a slogan I can get behind.
Second, I am a great believer in the value of some positive political theory, rational choice, and game theoretic models for understanding political institutions in general and the interaction among courts and between courts and other political institutions. Let me give just one example, I am very interested in the politics of judicial selection, and in particular, with the question whether formalist judges (assuming such persons could exist) could be selected given realistic assumptions about American politics. In the course of thinking about this question, I tried to conceptualize the politics of nomination in a two dimensional space (with a left-to-right political ideology line and a bottom-to-top judicial philosophy line (on which realism and formalism) are plotted). Positive political theory models of the judicial selection process are very helpful in this enterprise; thick descriptions can be helpful too, but they don’t get at my problem in a rigorous and clear way.
Third, nonetheless, I am in great sympathy with Cornell Clayton’s remarks in this panel. Let me express my concern this way. Let us assume that law does constrain judicial decisions at least in part because some judges have internalized the a value of fidelity to law and because some outcomes are legally correct and some are legally incorrect—even for a court of last resort. This does not entail that politics will play no role in judicial decisions. Some legal provisions are ambiguous (they have more than one possible legally-justifiable meaning. Others are vague (they have borderline cases). Judge’s beliefs about political morality (or their political ideology) could influence the resolution of vagueness in at least two ways. First, many conventional normative theories of judicial decisionmaking allow for explicit reference to such concerns. One example is Dworkin’s theory, which would allow beliefs about political morality in as part of the process by which Hercules constructs theory that best fits and justifies the law as a whole. A simpler view is that judges have discretion when there is ambiguity or vagueness and they are authorized to refer to their beliefs about political morality when they exercise this discretion. Second, the values or ideology of a judge could have a causal influence on decisions in cases where there is vagueness, ambiguity, or discretion, even if the judge believes that this influence is improper and should be minimized. For example, a judge might believe that she is obligated to apply consensus community values to resolve vagueness, even if she herself does not share these values.
In other words, law can constrain outcomes even if values (beliefs about political morality or political ideology) has an influence outside the constraints and at the margins of constraint. This is Hart’s familiar picture of the “core” and the “penumbra,” and I have put this point in terms of the distinction between underdeterminacy, determinacy, and indeterminacy of law. At some level, I think this picture is widely accepted, although there are some serious scholars who continue to adhere to the radical indeterminacy thesis—that law imposes no constraints on outcomes.
The fact that law underdetermines outcomes is especially important when we focus on courts of last resort in multitier appellate systems, e.g., the United States Supreme Court. The decision not to settle (or give up) and pursue a petition for certiorari and/or an appeal as of right may be made in part on the basis of uncertainty about the content of the law. If this is so, then we would expect the caseload of the Supreme Court to be systematically biased in favor of cases in which con
siderations of morality or politics have a wide scope. But this would not entail the conclusion that the Supreme Court as an institution is disposed to decide every possible case on the basis of morality or politics. Instead, we would expect that the court would decide “easy cases” on the basis of formal legal considerations, but that it rarely (perhaps almost never) gets an opportunity to do so.
What concerns me about some (but not all) attitudinal and rational choice, formal models of judicial behavior is that they do not acknowledge the complexity of the relationship between law and morality (or politics) in judicial decision. Of course, simplified assumptions are part of the normal machinery of formal models (and informal models as well). So it isn’t the fact of simplification that is a problem, it is the failure to acknowledge the limitations on these assumptions when interpreting the models and empirical work that tests the models.
This was an excellent session with interesting presentations and great discussion among the panelists.