Should George Floyd's Past Record Be Admitted as Evidence in Derek Chauvin's Trial?

Explaining Minnesota's character evidence rule.

During pretrial proceedings Friday morning in the trial of Minneapolis police officer Derek Chauvin, Judge Peter Cahill indicated that he would admit some evidence of George Floyd’s prior arrest in May 2019. In that incident, Floyd resisted arrest to some extent and rapidly ingested drugs during or just prior to the arrest to conceal them. This caused his blood pressure to spike, putting him at risk of a heart attack. Floyd’s behavior in the prior incident is somewhat similar to his behavior in the fatal incident a year later. And it certainly paints him in an unflattering light. 

Some of the initial response to Judge Cahill’s ruling sounds a predictable note: George Floyd is not on trial. That statement, while true, illuminates nothing. It is common in criminal cases that a defendant’s guilt depends on the words or conduct of the victim. In self-defense cases, a defendant’s guilt depends on whether the victim attacked first. In rape cases, a defendant’s guilt depends on whether the victim consented. In fraud cases, a defendant’s guilt depends on whether the victims were in on it. In all of these situations, evidence of the victim’s behavior is critical, even though the victim is not a party to the proceeding—and is not “on trial.”

That is no less true in excessive force cases. Whether a police officer’s use of force is reasonable and justified depends in large measure on the arrestee’s behavior. And in this case, Floyd’s behavior bears on his cause of death. So necessarily, although George Floyd is not on trial, the jury will have to assess both Chauvin’s behavior and also Floyd’s behavior. 

Whether Floyd’s prior arrest should be part of that assessment is complicated. It depends on the character evidence rule.


The character evidence rule forbids evidence of character traits or general propensities. The most important implication of the rule is that evidence of a defendant’s prior crimes is generally not admissible at trial. The rationale for the rule is that jurors might give such evidence too much weight, or they might convict the defendant simply for being a bad person even if he is innocent in the incident in question. As some courts have said, each defendant stands before a jury with a clean slate. We try defendants for what they did, not who they are.

The rule has been part of the Anglo-American system for hundreds of years. It is codified as Rule 404 of the Rules of Evidence in Minnesota, as elsewhere. It is perhaps the most important rule of evidence in criminal practice. It is also perhaps the most frequently litigated rule of evidence in criminal practice because it is so complicated. It is complicated because it is subject to a variety of exceptions and qualifications.

Rule 404 says that evidence of a person’s prior crimes may not be admitted to show his character or propensity, but the same evidence may be admitted for other purposes. So for example, if a defendant is charged with assault, the prosecution may not admit his prior assault convictions to show that he is a generally violent person. But if the assaults involved the same victim, for example, the prior assaults could potentially be admitted to show their relationship and thus to show the defendant’s motive. 

In a bank robbery prosecution, the prosecution could not admit evidence that the defendant stole a car to show that the defendant is a no-good thief. But if the stolen car was used as a getaway car for the bank robbery, the prosecution could admit evidence of the car theft to show the defendant’s planning or preparation for the robbery.

While the character evidence rule most often comes up when the prosecution seeks to admit a defendant’s prior bad acts, the same general principles apply when the defense seeks to admit the victim’s prior bad acts—with additional exceptions and qualifications, of course. So this is the rule that governs whether Floyd’s prior arrest may be admitted. It cannot be admitted to show that he is generally a bad person with bad character traits, but it can potentially be admitted if the defense can show some other non-character purpose.

The inquiry is notoriously difficult. There are entire evidence law treatises devoted solely to this question. There are hundreds of appellate cases in Minnesota discussing the line between impermissible character and permissible non-character uses of other-acts evidence. You could spend a month reading them all, and you’d only come away more confused.

Part of the problem is that the character evidence rule rests on radically indeterminate conceptual foundations. We talk about character and propensity, but we don’t really know how to distinguish those concepts from all the other things that cause people to behave in certain ways—custom, culture, habits, biases, moral beliefs, mental illness, addiction, and so on. Rule 404 and court cases use the terms character and propensity interchangeably, but scholars who study evidence cannot even agree as to whether they are synonymous. (Say the words repeatedly, feel semantic satiation set in.) The concepts are familiar and yet impossible to pin down.

Conceptual indeterminacy aside, part of the problem is the legal system is in fact deeply ambivalent about the character evidence rule. We extol the virtues of the venerable rule—but when the rubber meets the road, we often find ways to ignore it.


For more than a decade I have taught evidence law to law students in Minnesota, usually in large lecture classes of 80 or more students. Teaching large classes involves some set pieces and gimmicks, and I have a set piece for character evidence that I used until recently. It worked in three parts. (I no longer use the first part, for reasons that will be obvious.)

Part 1 took place on the first day of class, before the students knew any evidence law. I gave them a summary of facts from a real case. I asked them to put themselves in the position of jurors: Read the facts carefully, then check guilty or not guilty at the bottom of the page. The case involved an allegation of sexual assault. The trick is that for half of the class, I gave one additional piece of evidence: The defendant had two prior convictions for sexual assault. Over the years, the half of class with the evidence of the prior crimes convicted at a much higher rate, in the neighborhood of 80 percent compared with about 40 percent in the control group. 

There was nothing scientific about my study, but it nonetheless confirmed the common-sense point that character evidence is powerful. The admission or exclusion of a single piece of evidence, especially character evidence, can be outcome-determinative.

Part 2 happens a few weeks later, on the first day of the character evidence unit. After some introductory readings, including some classic Supreme Court cases discussing character evidence, I ask the students whether the law is right to exclude character evidence. After all, we all make judgments based on character all the time in our everyday lives—when deciding whom to hire, or whom to marry. If we use character evidence all the time in regular life, why do we hide it from jurors? Why give them an incomplete picture when asking them to make life or death decisions? Partly to play devil’s advocate, I take the position that the character evidence rule should be abolished.

Most students react in horror to this idea. Allowing character evidence would give prosecutors an unfair advantage, they say. It would water down the presumption of innocence and the beyond-a-reasonable-doubt standard. It would exacerbate racial disparities in the criminal justice system. Most students are firmly committed to the belief that the exclusion of character evidence is a fundamentally important part of the American criminal justice system.

That commitment, however, wavers when I get to Part 3 of the set piece. By then, we’ve studied the character evidence doctrine in more depth, including all of its exceptions and gray areas and fictions. I return to the sexual assault case from the first day of class. I ask students to put themselves in the position of a trial judge: “Would you exclude this evidence of his prior crimes, and thus risk letting a serial rapist go free?”

Faced with that stark choice, students immediately run for the gray areas, searching for justifications to let the evidence in. Whether the justifications can plausibly co-exist with the commitment to the character evidence rule seems largely immaterial by then. The evidence has to come in, we all know it has to come in, and we will say whatever we need to say to make that happen. (Lawyers can justify anything.)

The point of my set piece in class is not that students are wrong or stupid or inconsistent. The point is that we are human and our abstract commitments to legal principles often cannot be maintained when the moral or political stakes get too high. 


What is true in the classroom is also true in the courtroom. Judicial opinions sometimes wax poetic about the character evidence rule and how it is a pillar of fairness in the system. Then, in the next line, they adopt highly strained arguments for admitting a defendant’s prior crimes anyway. 

In Minnesota, we’ve seen this play out largely through the “common plan” doctrine. That doctrine allows the admission of a defendant’s other crimes for the non-character purpose of showing a plan. Traditionally and in many jurisdictions, the other crime must be linked to the charged crime as part of a single overall scheme. Thus, in the bank robbery hypothetical above, the defendant’s car theft could be admitted to show his overall scheme, since they are constituent parts of a whole. But in Minnesota, courts have expanded the common plan doctrine far beyond its original bounds to include all prior similar offenses—even if they are not linked to the charged crime, even if they are not constituent parts of a whole. 

Similarity is an impossibly loose standard. To illustrate the point, I ask my students: “Am I similar to Barack Obama?” Another set piece, and it gets a laugh. The answer is obviously no: Obama is a transformational figure who was leader of the free world, someone who will be remembered in history books for centuries, and I am … not any of those things. Then again: We are both men, both lawyers, both Americans, both living at the same time, and for that matter we are both humans living on the planet Earth. Relative to all beings on the planet or all things in the universe, Obama and I are actually quite similar!

Without some preexisting baseline from which to measure similarity, there is no way to gain any traction on this question. Similarity inquiries are subject to endless framing games. They are, as a result, often standardless. And Minnesota’s common plan doctrine, which is a similarity inquiry with no objective baseline or frame, is largely standardless in practice.

Put differently, Minnesota’s common plan doctrine is a legal fiction. It allows courts to admit character evidence simply by slapping a different label on it.

A few years ago, the Minnesota Supreme Court informally requested the Advisory Committee for the Rules of Evidence to consider amendments to Rule 404 in order to create some clearer standards. I am chair of that committee. The committee proposed an amendment to Rule 404 that would have abrogated Minnesota’s loose common plan doctrine, returning the character evidence rule to something closer to its common-law roots. 

The political opposition was swift. It came not just from Minnesota’s prosecutors, but also and most forcefully from advocates for victims of sexual assault. Their opposition had little to do with the proper meaning of the character evidence rule—rather, their primary argument was that the court should not make any change that would make it harder to convict rapists. Not surprisingly, faced with that kind of bipartisan opposition, the Minnesota Supreme Court rejected the proposed amendment. Minnesota thus retains only a watered-down version of the character evidence rule.


Minnesota is not alone in this regard. The politics of crime are more or less the same across America, and the character evidence rule has been eroding across the country. Among other things, some jurisdictions have explicitly abolished the character evidence rule for rape cases, allowing a defendant’s prior crimes to show a propensity for sexual assault. Minnesota has not done that, but when it comes to our “common plan” doctrine, Minnesota is on the far end of the spectrum when it comes to admitting prior similar acts as supposedly non-character evidence. 

That loose doctrine generally makes it easier for prosecutors to admit unflattering evidence of defendants’ prior crimes. And now Derek Chauvin is exploiting that same law to admit unflattering evidence of George Floyd’s prior arrest. Chauvin’s attorney is arguing that the prior arrest is similar, with similar behavior, thus demonstrating Floyd’s common plan or practice or his modus operandi. If he swallowed drugs to evade detection and almost killed himself once before, perhaps he did it this time, too.

Is that inference a proper one for the jury to entertain? Is the prior incident similar enough that it shows a “common plan”? Is it character, or is it habit, or is it modus operandi—and what do those terms even mean? The character evidence rule, especially in Minnesota, does not provide a clear answer.


This case has been full of ironies. The felony murder doctrine, which can lead to excessively long prison terms, has long been a target of criticism from progressives who want to reform criminal justice. In this case, however, progressive attorney general Keith Ellison has deployed and endorsed an extreme version of felony murder in order to get Chauvin more time in prison. Erstwhile reformers have discovered the warm glow of retribution.

In most cases, during jury selection, criminal defendants seek to have more minorities on the jury, knowing that they are more skeptical of police and the criminal justice system, while prosecutors employ subtle tactics to remove them. In this case, however, it is the prosecution that wants more minority jurors, while the defense looks for more conservative white jurors.

In most cases, it is the prosecution that seeks to evade the character evidence rule, employing clever arguments and legal fictions to justify admission of a defendant’s prior crimes. Systematically, these prosecutorial tactics disproportionately disadvantage minority defendants, who are more likely to have prior offenses. In this case, however, it is the defense employing those same tactics to justify admission of a minority victim’s prior bad acts. 

Under current Minnesota law, Judge Cahill was correct to let in some of this evidence—or at least he wasn’t wrong. It is true that George Floyd isn’t on trial, but his behavior is relevant, and under our watered-down character rule, it is relatively easy to bring in his past similar acts.

The evidence of the prior arrest has the potential to be quite powerful. If you watch the video of the prior arrest, you might feel your mind changing just a little bit. Perhaps that is due to the video’s legitimate relevance. But perhaps it draws its power from somewhere else, playing on racist stereotypes lurking deep in our lizard brains. No amount of honest introspection can answer that question with certainty. Evidence law has no answer either. 

It is clear from watching jury selection that most of the jurors who have been selected so far are leaning in the prosecution’s favor from the outset. The evidence of Floyd’s prior arrest is the kind of evidence that could shake them from their pre-commitments. And if it shakes just one juror, that is enough for the defense.

Ted Sampsell-Jones is a professor at the Mitchell Hamline School of Law.