Anthony and Ben live in the same city. They grew up in different neighborhoods, move in different circles, and have never met. Due to an improbable coincidence their lives began to converge. A week apart, they committed identical murders – separately. Both men were addicted to drugs and, finding themselves broke and in desperate need of a fix, they decided to burglarize a dwelling. Armed with a pistol, they broke into a home and found some valuables. As they were leaving with the stolen property, they were confronted by a member of the household. Both men shot and killed the person who came upon them, and then fled. The authorities were notified, and a manhunt ensued. A few days after their crimes, they were each arrested. Unfortunately, for them, theirs is a death penalty state. The law of their state allows a defendant to be sentenced to death if convicted of capital murder – murder with one or more “aggravating factors,” including murder during the commission of another felony, such as burglary. After reviewing their cases, the District Attorney charged both men with capital murder. Several months later, Anthony’s case was tried before a judge and jury. Anthony was found guilty of capital murder. At the penalty phase of the trial, the jury voted not to impose the death penalty but, rather, to sentence Anthony to life imprisonment without parole. Not long afterward, Ben’s case came to trial. He too was convicted of capital murder. However, his jury unanimously voted to sentence him to death.
These hypothetical scenarios raise an issue that frequently arises in capital murder cases: unpredictability. Judges, law professors, and legal commentators have spoken of the arbitrariness of the death penalty, sometimes likening the decision to impose it in one case but not in a similar case to a lightning strike (see Dieter 2011). If we wanted to probe the reason for the disparate decisions, we would naturally turn first to the written law – in particular, to statutes enacted by legislators and decisions rendered by judges. But nothing in the written law would allow us to distinguish Anthony’s case from Ben’s. The burglary-murders were identical, yet only one case resulted in the defendant being dispatched to death row. Since the rules could not explain the divergent outcomes, what would? Press a lawyer for an answer and you will most likely be told “discretion.” Anthony’s jury exercised its discretion not to impose the death penalty; Ben’s jury exercised its discretion in the opposite direction. That is clearly unsatisfactory. To explain the decision with “discretion” is not to explain the decision. Discretion is simply a fancy word for “we don’t know” (see Baumgartner 1992).
But social scientists might not do much better. Ask why Ben was, and Anthony was not, dispatched to death row and the answer you will likely get is race. Ben must have been Black or must have killed a White person, or both. Anthony must have been White or killed a Black person, or both. Many social scientists know that a large body of research shows that race affects capital punishment. Some studies suggest that Black defendants are more likely to get the death penalty. Virtually all studies suggest that a defendant who kills a White victim is more likely to get the death penalty (for a review see Phillips and Marceau 2020). David Baldus and colleagues conducted the best-known study in Georgia in the 1970s. (We will have a lot more to say about this terrific source of data.) The Baldus study, which formed part of a constitutional argument against the death penalty in McCleskey v. Kemp, found a race-of-victim effect (Baldus, Woodworth, and Pulaski 1990).1 Less well known is that the race effect was only one of many effects that Baldus found. Baldus and his team presented to the Supreme Court a statistical model that contained 41 variables. The model showed that a defendant’s odds of receiving a death sentence were 4.3 times greater when the case included one or more White victims. But nine variables had an even stronger impact, including whether the victim was 12 years old or younger, whether there was a rape involved, and whether the victim had been physically tortured.2 The lesson is clear: race matters, but it is not the only thing that matters. To obtain a fuller understanding of who gets sentenced to death we must include race while also looking beyond it. That is what the geometrical theory of law does.
The Theory
The geometrical theory of law was born in a police car, according to its father, Donald Black (2002: 109). As a graduate student at the University of Michigan, Black rode around with police officers as they patrolled the streets of Detroit. When his major professor, Albert Reiss, secured a large grant to conduct an observation study of police officers in three cities (Boston, Chicago, and Washington DC), Black worked on the study, writing his doctoral dissertation from the observations. He went on to publish several influential papers and, eventually, a book on police behavior (Black 1980). His gaze soon began to wander beyond the police to law more generally. After completing his Ph.D., Black moved to Yale University where he began to dream big. In a manifesto published in the Yale Law Journal (“The Boundaries of Legal Sociology”), he argued that “the proper concern of legal sociology should be the development of a general theory of law” (1096). Four years later, Black delivered. In The Behavior of Law (1976), he proposed a theory designed to explain variation in law wherever and whenever it exists. The theory predicts and explains differences in the handling of cases, such as whether people call the police or contact a lawyer, whether the police arrest or a lawyer files suit, whether the case succeeds or fails, and, if it succeeds, the severity of the punishment imposed or the amount of compensation the defendant is required to pay. Black illustrated his theory by drawing on patterns of legal behavior not just in modern societies (e.g., modern America and Japan) but in earlier societies studied by historians (e.g., imperial China and colonial America) and struct...