WP 12: Orthodox

Reflections on Motion Picture Evidence

Brian L. Frye

Film will only become an art when its materials are as inexpensive as pencil and paper.
—Jean Cocteau,


Courts have long admitted motion pictures as evidence. But until recently, making motion pictures was expensive and cumbersome. Today, making motion pictures is cheap and easy. And as a result, people make so many of them. As Cocteau predicted, the democratization of motion pictures has enabled people to create new forms of motion picture art. But it has also enabled people to create new forms of motion picture evidence. This article offers a brief history of motion picture evidence in the United States, and reflects on the use of motion picture evidence by the Supreme Court.

A Potted History of Motion Picture Evidence

Motion picture evidence is any motion picture introduced in a legal proceeding in order to establish a fact or illustrate an argument. The idea of motion picture evidence probably coincided with the invention of motion pictures. Among the first motion pictures were the “actuality films” created by Auguste and Louis Lumière in 1895, each of which briefly documented an actual event. Filmmakers soon realized that motion pictures provided a new kind of evidence. In 1913, Edwin Thanhouser and Lawrence Marston made The Evidence of the Film, a 15 minute melodrama that imagined the use of a motion picture film as evidence in a trial.1 But in practice, the standard 35mm nitrate film was too expensive, unwieldy, and dangerous for practical use in evidence gathering or presentation in court.2 Cameras were too large to conceal, and film was explosively flammable.

Moreover, courts resisted the introduction of motion picture evidence. In 1920, a California trial court refused to admit into evidence a motion picture re-enactment of a murder offered by the defense, on the ground “that ‘juries are naturally prone to accept them as absolutely correct,’ and that, therefore, they were admissible only under certain circumstances.”3 And in 1922, when a New York trial court allowed the introduction of motion picture evidence in a tort action, the appellate court reversed, holding that motion picture evidence was not admissible:

Aside from the fact that moving pictures present a fertile field for exaggeration of any emotion or action, and the absence of evidence as to how this particular motion picture film was prepared, we think the picture admitted in evidence brought before the jury irrelevant matter, hearsay and incompetent evidence, and tended to make a farce of the trial.4

In 1923, Kodak introduced 16mm safety film, the first commercially successful amateur motion picture film format. The cameras were small and reasonably priced, and the film was both relatively inexpensive and non-flammable. Suddenly, almost anyone could make movies, including judges.5 Most of those people made home movies. But some of them made evidence films, or films intended for use as motion picture evidence. Law enforcement saw the potential of evidence films in investigating and proving crimes, and lawyers saw similar potential in civil cases.6

Soon, genres of motion picture evidence began to emerge.7 One of the first was “surveillance films,” which provided covert documentation of criminal defendants or parties to civil lawsuits. For example, police secretly filmed not only criminal activity, but also strikes and political rallies, for use as evidence.8 And after the introduction of motion picture sound, police secretly filmed interrogations in order to document confessions.9 Similarly, private investigators secretly filmed targets in order to create motion picture evidence for use in civil cases.10 For example, insurance investigators made “malingering” films, in which they filmed tort plaintiffs in order to disprove claims of disability: “The advent of the miniature camera and small motion-picture camera, particularly, has marked the end of easy money for insurance fakers.”11 And private investigators made “philandering” films, which provided proof of infidelity.

Tort plaintiffs eventually responded to “malingering” films with “day-in-the-life” films, which documented how their injuries affected their lives.12 While these day-in-the-life films were occasionally introduced as evidence at trial, often they were used as “video settlement brochures,” to encourage defendants to settle:

The lights in the mahogany paneled conference room dim. There, projected on a 72-inch screen, is a day-in-the-life video of a car accident survivor who lost her leg in the crash. The attorneys in the room see the ugly reality of her stump, a sight she and her family must endure for the rest of their lives. She painfully and painstakingly changes the dressing on her oozing wound.

No words can conjure up the revulsion in the room created by the stark video. The attorneys for the insurance company know that nothing they might say would diminish the intense sympathy for the victim this video would generate in a jury. They might wince at your settlement demand, but more than likely, they will agree to settle.13

In addition, both plaintiffs and defendants created “demonstrative” films, intended to help juries understand their theory of a case by providing a motion picture illustration. For example, lawyers commissioned motion picture re-creations of car accidents or animated illustrations of how a patented device works.14 Demonstrative films are not introduced as substantive evidence of a fact, but rather as demonstrative evidence, to illustrate or explain testimony or evidence. 15

The introduction of inexpensive consumer video in the 1970s enabled the widespread creation of “testimonial” films. 16 Police began to videotape interrogations and lawyers began to videotape depositions, for use in court. Some courts even experimented with “Pre-recorded Videotaped Trials” (PRVTs), in which the jury watched a movie of the testimony and arguments, rather than live testimony. Not only did PRVTs reduce the duration of trials by excluding extraneous information, but also they enabled judges to rule on the admissibility of testimony and evidence outside the presence of the jury.17

Contemporary Motion Picture Evidence

Initially, evidence films were obscurities, created only by lawyers and government officials, and shown only to a jury, if ever. But as creating motion pictures gradually became easier and less expensive, more and more people could and did create evidence films, often unintentionally.

The Zapruder Film

The paradigmatic evidence film is surely the 26-second 8mm home movie that Abraham Zapruder filmed of President John F. Kennedy’s assassination on November 22, 1963. It is probably the most thoroughly examined motion picture ever. The Warren Commission investigation and report on the assassination, which found that Lee Harvey Oswald shot Kennedy and acted alone, relied heavily on the Zapruder film, as did several subsequent official investigations.18 Among other things, the Warren Commission used the Zapruder film to help determine the trajectory of the bullets that struck Kennedy.19 Not to mention the innumerable conspiracy theorists who have relentlessly scrutinized every frame of the film, searching for evidence to prove their pet theory of who actually killed Kennedy and why. Notably, the Warren Commission made 35mm enlargements of each frame of the Zapruder film, in order to facilitate close analysis.20

The Rodney King Video

The notoriety of the Zapruder film is rivaled only by George Holliday’s video of Rodney King being beaten by officers of the Los Angeles Police Department. At 12:30 a.m. on March 3, 1991, King was driving on the Foothill Freeway in Los Angeles with two passengers, when officers Tim and Melanie Singer of the California Highway Patrol clocked him speeding and initiated pursuit. King was intoxicated and refused to stop. After a high-speed chase involving several police cars and a police helicopter, the police eventually cornered King and forced him to stop. Soon afterward, five LAPD officers arrived on the scene: Stacey Koon, Laurence Powell, Timothy Wind, Theodore Briseno, and Rolando Solano.

Officer Tim Singer ordered King and his passengers to leave the car and lie face down on the ground. The passengers complied, and were beaten before being arrested. King eventually emerged from the car, laid face down on the ground, and grabbed his buttocks. Officer Melanie Singer believed King was reaching for a weapon and ordered him to remain on the ground.

At that point, LAPD Sergeant told the Singers that the LAPD was taking over. Koon ordered the other four LAPD officers to manually subdue and arrest King, but King resisted and attempted to stand. Koon then tasered King twice. At that point, King ran toward Officer Powell and collided with him. Powell hit King with his baton, knocking him to the ground, then hit him several more times. When Powell stopped hitting him, King again rose to his knees. On Koon’s orders, the LAPD officers began beating King, hitting him with their batons at least 33 times and kicking him at least 6 times. King was seriously injured by the beating, suffering multiple fractures and brain damage.

Shortly after Koons tasered King, George Holliday began videotaping the incident from the balcony of his nearby apartment. Holliday’s video begins with King running toward Powell and documents King’s beating and arrest. Two days later, Holliday told the LAPD about his video, but was ignored, so he gave it to KTLA-TV instead. KTLA broadcast Holliday’s video, excising the blurry beginning where King runs toward Powell, and soon every television station in the country did the same.

Holliday’s video galvanized the nation. It presented seemingly irrefutable documentary evidence of brutal and unjustifiable violence used by the LAPD against unarmed suspects. President Bush described the video as “sickening.” On March 14, 1991, the Los Angeles district attorney criminally charged officers Koon, Powell, Briseno, and Wind with use of excessive force. And in April 1991, Los Angeles Mayor Tom Bradley created the Independent Commission on the Los Angeles Police Department, to conduct “a full and fair examination of the structure and operation of the LAPD.” Among other things, the Commission explicitly recognized that there would have been no investigation without the video.21

The criminal charges were tried in nearby Ventura County before a jury that did not include any African-American members. Ironically, the prosecution and the defense both relied on Holliday’s video as their best evidence. While the prosecution’s strategy was to let the video “speak for itself,” the defense’s strategy was to explain what the video meant.22 The prosecution presented the video as simple and incontrovertible evidence of guilt. But the defense used the video to illustrate its theory of why the LAPD officers had acted reasonably under the circumstances, arguing that the officers had reasonably believed that King was dangerous and used targeted force to subdue him, consistent with their training. Specifically, the defense used slow motion and freeze-frame to examine and explain the meaning of key events depicted in the video.23 On April 29, 1992, the jury acquitted Koon, Briseno, and Wind, but hung on Powell. Massive riots immediately erupted in Los Angeles and other cities across the country.

On August 4, 1992, the United States Department of Justice indicted the four officers on federal civil rights charges for using unreasonable force and failing to stop an unlawful assault. In the federal trial, both the prosecution and the defense used the video to illustrate their respective theories of the case. Rather than merely presenting the video, the defense also interpreted it. On April 11, 1992, the jury convicted Koons and Powell, but acquitted Wind and Briseno. The district court sentenced Koons and Powell to 30 months in prison, heavily mitigating their sentences on the ground that King had resisted arrest, justifying certain uses of force. Among other things, the district court asked what kinds of evidence the video did and did not provide. Specifically, it observed:

Although the videotape creates a vivid impression of a violent encounter, careful analysis shows that it is sometimes an ambiguous record of the crucial events. A meaningful understanding of the events it depicts required the explanation of witnesses who are experts in law enforcement.

. . .

Although the videotape is reliable evidence of the physical events that unfolded in the course of Mr. King’s arrest, it provides less direct evidence concerning two factors central to Officer Powell’s illegal behavior. To determine that at any given point Officer Powell’s behavior crossed the line into illegality, the Court must find that Officer Powell used objectively unreasonable force, and that he intended to use excessive force.

. . .

Accordingly, in evaluating the videotape, the Court must take into account the totality of the circumstances, including facts not displayed on the tape. Mr. King had not been searched. Mr. King did not respond to the electrical charge of Sergeant Koon’s taser. Sergeant Koon testified that he observed Mr. King sweating profusely. At least during the initial stages of the arrest process, Sergeant Koon may reasonably have suspected that Mr. King was under the influence of PCP because of Mr. King’s erratic and recalcitrant behavior.24

The court of appeals reversed the district court’s sentencing decision, holding that the district court lacked discretion to mitigate under the sentencing guidelines, but the Supreme Court reversed the court of appeals and reinstated the original sentences, holding that the district court did have discretion.25

Reactions to the Rodney King Verdicts

After the state trial, many people were surprised and incensed that none of the officers were convicted, especially given Holliday’s damning video. Some chalked up the outcome to racism, observing, inter alia, that the jury had no African-American members. Others argued that the defense’s use of slow motion and freeze-frame to analyze the video had improperly distorted reality. But as film scholar Bill Nichols observed, this objection is incoherent:

Of course, slow motion is a “distortion.” So is the raw footage. (It puts three-dimensional events onto a two-dimensional surface, produces foreshortening, and introduced problems of graininess, focus, and sound.) Any post facto account, be it from an eyewitness, participant, prosecutor, or historian, distorts. History, which is what we must reconstruct here, is always a matter of story telling: our reconstruction of events must impose meaning and order on them, assign motivations, assess causes, and propose moral judgments (in this case, guilt or innocence).26

Likewise, according to Richard Sherwin, images lack intrinsic meaning, and acquire meaning only in context:

In the end, we can say that the prosecutor’s naïve realism undid him. Contrary to his naïve assumption, visual images do not make meaning all on their own. Visual meaning is highly malleable. As photojournalists know, captions can turn a photo’s intended meaning on its head. If you do not provide a context of meaning, if you do not wrap a sequence of images in a narrative of your own, you will leave open the possibility that their meaning will be captured by the narrative of another.27

But as film scholar Noel Carroll has observed, there is nothing epistemically unique about the rhetoric of motion pictures:

But clearly selectivity, even if it is an inevitable feature of film, is not a unique feature of film. Every mode of inquiry and its attendant channels of publicity – from physics through history to journalism – is selective. So nothing special is discovered by revealing the selectivity of the motion picture apparatus. Moreover, insofar as we do not regard physics or history as exiled from objectivity just because they select, then there should be no impulse to suspect the nonfiction film’s credentials, on a priori grounds, merely because it is selective. We can’t have chemistry or economics without selectivity. Indeed, it is their selectivity that makes them possible. Why should we expect things to stand differently with nonfiction films?28

While it is trivially true that images lack intrinsic meaning, the same is true of any representation, indeed any human endeavor. But contextless perception is impossible. The people who watched the video on their televisions at home perceived it in the context of their lived experience. The prosecutor relied on that context, believing that it supported the narrative he wanted to convey. The problem was that the defense provided a new and different context that undermined the prosecutor’s narrative.

As Naomi Mezey has observed, “the image cannot speak for itself.” On the contrary, images are “open to interpretive dispute and ought to be critically examined like text or testimony.” In light of that observation she asks:

How might prosecutors have used visual literacy to present the video or respond to the counter-narrative? One approach would have been to mimic the defense technique by unearthing details that supported their case or showing why the technique itself was misleading. Another possibility would have been to speak for the video by explaining why people were right to see it as excessive force by police.29

Of course, in the federal trial, the prosecutors did adopt their own version of the interpretive techniques used by the defense in the state trial, to good effect. But in addition, the state prosecutor arguably did attempt to explain why people were right to see the video as conclusively proving the use of excessive force. At least, the prosecutor’s strategy was to rely on the jury’s intuition that beating an unarmed man could not possibly be justified. But the defense used the same video to provide a detailed explanation of why each element of the beating was indeed justified under the circumstances.

The underlying problem is that the question presented to the jury was not a question of fact, but a question of law. The jury was not asked to determine the facts of what happened, which were not meaningfully in dispute. It was asked to determine whether the officers’ actions were reasonable under the circumstances, which is inherently a question of law.30 The video alone could not answer that question, because it is a question of judgment, not a question about what the video depicted.

Ultimately, the jury was not asked to determine what the video showed. It was asked to determine what the video meant. It could have held that, as a matter of law, beating an unarmed man is unreasonable. But it did not. For better or worse, it applied the law as argued by the defense and instructed by the judge, and asked whether each act was unreasonable under the circumstances, from the perspective of the police officers. When the state prosecutor failed to offer a counter-narrative on that question, the jury found for the defense. But when the federal prosecutor did offer a counter-narrative, the jury convicted two of the defendants.31

Cellphone Video

Today, motion picture evidence is ubiquitous. Police cars are equipped with dash-cams, police officers wear body-cams, and surveillance cameras are omnipresent. But perhaps more importantly, almost everyone routinely carries a motion picture camera in their pocket or purse, built into their cellphones. At the merest suggestion of official misconduct, a phalanx of cellphones emerges to document it.32 As a consequence, video evidence of police violence has proliferated.

Unsurprisingly, the interpretive dilemma posed by Holliday’s video has persisted. Many people watch these videos and understandably conclude that the police actions depicted are objectively unreasonable. But courts typically see the same videos as probative evidence of questions of fact relevant to determining the ultimate legal question of reasonableness, and frequently reach different conclusions. Which presents the question, how should courts evaluate the relevant and meaning of different kinds of motion picture evidence?

In the last decade, the United States Supreme Court has occasionally posted video evidence on its official website. Currently, the Supreme Court website hosts video evidence relating to three cases: Scott v. Harris, 550 U.S. 372 (2007); Kelly v. California, 555 U.S. 1020 (2008); and Brumfield v. Cain, 135 S.Ct. 2269 (2015).33 I will consider each in turn, describing both the factual background of each case and the motion picture evidence in question, then examining the Supreme Court’s interpretation and use of that motion picture evidence. I will also describe certain motion picture collages that I created using the motion picture evidence and oral argument recordings, intended to illustrate – and hopefully illuminate – how the Supreme Court has used and interpreted motion picture evidence.

Scott v. Harris

Ironically, the first video evidence posted on the Supreme Court website concerns another African-American driver pursued by the police. At 10:42 p.m. on March 29, 2001, Deputy Clinton Reynolds of the Coweta County, Georgia Sheriff’s Office clocked Victor Harris’s car traveling at 73 miles per hour in a 55 miles-per-hour zone. Reynolds flashed the lights on his cruiser, but Harris did not stop, so Reynolds initiated pursuit. Harris drove at speeds between 70 and 90 miles per hour, which exceeded the posted speed limit, passed other vehicles in places where passing was not permitted, and ran two red lights. However, he used his blinkers when passing and turning.

During the pursuit, Reynolds radioed dispatch and reported that he was pursuing a fleeing vehicle. He did not state that he initiated the pursuit because the vehicle was speeding. Deputy Timothy Scott of the Coweta County, Georgia Sheriff’s Office heard Reynolds’s broadcast and joined the pursuit as Harris approached Fayette County, Georgia.

When Harris entered Peachtree City in Fayette County, he slowed down and turned into the parking lot of a shopping complex, where two Peachtree City Police cruisers were already stationed. Scott drove to the opposite side of the parking lot and tried to prevent Harris from leaving the parking lot by blocking the exit. Harris drove around Scott’s cruiser, but collided with it, damaging the cruiser. Harris then entered the highway and drove south at high speed. Reynolds and Scott continued to pursue him.

Scott requested permission to execute a Precision Intervention Technique (“PIT”) in order to stop Harris. A PIT is a method of stopping a vehicle by colliding with it at a specific point. Scott had not been trained to execute a PIT. Sergeant Mark Fenninger of the Coweta County, Georgia Sheriff’s Office authorized Scott to execute a PIT. However, Scott decided that he could not execute a PIT because Harris was driving too fast. Instead, Scott rammed his cruiser into Harris’s vehicle, causing Harris to lose control of his vehicle and crash. As a result of the crash, Harris was seriously injured and rendered quadriplegic.34

On October 16, 2001, Harris filed a complaint in federal district court against Scott, Reynolds, Fenninger, Sheriff Michael Yeager, and Coweta County. Harris alleged federal claims under 42 U.S.C. § 1983 for violation of his rights under the Fourth and Fourteenth Amendments of the United States Constitution, as well as state claims for violation of his rights under the Georgia Constitution, and common law tort claims.35

The defendants filed a motion for summary judgment. In his response to the motion, Harris conceded that summary judgment was appropriate on his some of his claims. The district court granted summary judgment and dismissed some of Harris’s other claims. But it denied summary judgment and did not dismiss Harris’s Fourth Amendment claims against Defendants Scott and Fenninger for the use of excessive force and the authorization of the use of excessive force; municipal liability claim for failure to train against Coweta County; and negligence claim against Coweta County based on the acts of Defendants Scott and Fenninger.36

Scott and Fenninger appealed the district court’s denial of their motion for summary judgment of Harris’s Fourth Amendment claims against them based on qualified immunity. The court of appeals affirmed the district court’s denial of Scott’s motion for summary judgment, but reversed the district court’s denial of Henninger’s motion for summary judgment. The court held that Scott was not entitled to summary judgment based on qualified immunity because, viewing the facts in the light most favorable to Harris, a reasonable juror could find that Scott’s use of deadly force was objectively unreasonable under the circumstances. The court specifically rejected Scott’s argument that Harris’s driving “must, as a matter of law, be considered sufficiently reckless to give Scott probable cause to believe that he posed a substantial threat of imminent physical harm to motorists and pedestrians,” concluding, “This is a disputed issue to be resolved by a jury.” However, the court reversed the district court’s denial of Henninger’s motion for summary judgment because Henninger did not authorize Scott to ram Harris’s vehicle.37

Scott filed a petition for certiorari in the United States Supreme Court, which was granted. The Court held oral argument on February 26, 2007, and issued an opinion written by Justice Scalia reversing the circuit court’s denial of Scott’s motion for summary judgment on April 30, 2007. Specifically, the Court held, “The car chase that respondent initiated in this case posed a substantial and immediate risk of serious physical injury to others; no reasonable jury could conclude otherwise. Scott’s attempt to terminate the chase by forcing respondent off the road was reasonable, and Scott is entitled to summary judgment.” It based its holding on two videos of the pursuit, recorded by cameras installed in Reynolds and Scott’s cruisers, stating that no reasonable juror could have believed that Harris was not “driving in such fashion as to endanger human life.”38 The majority opinion also attached the videos of the pursuit.39

Justices Ginsburg and Breyer concurred, arguing that Scott’s actions were reasonable under the circumstances. Both Ginsburg and Breyer specifically objected to the adoption of a per se rule that it is reasonable for a law enforcement officer to use deadly force against a motorist who is endangering the public. Justice Stevens dissented, arguing that a reasonable juror could find that Scott’s actions were unreasonable under the circumstances.40

Reactions to Scott vs. Harris

Many scholars and commentators criticized the Court’s decision in Scott v. Harris, typically arguing that it had improperly reversed the circuit court because Harris presented a question of disputed fact that should have been decided by a jury: were Scott’s actions reasonable under the circumstances?41

Perhaps most notably, Dan M. Kahan, David A. Hoffman, and Donald Braman published an article in the Harvard Law Review arguing that the Court’s reasoning was incorrect, based on an empirical study showing that different people could and did have different opinions about the reasonableness of Scott’s actions. 42 Their study asked 1,350 people from a diversity of backgrounds a series of questions intended to determine whether they thought Scott’s actions were reasonable under the circumstances. The results showed that while the majority of the survey participants agreed that Scott’s actions were reasonable under the circumstances, some of the survey participants disagreed, and that people from certain social groups were more or less likely to believe that Scott’s actions were reasonable.

The article argued that the survey showed that the Court’s conclusion in Scott v. Harris was based on “cognitive illiberalism,” a form of unconscious bias that prevented most of the Justices from perceiving the videos from the perspective of diverse social groups. And it argues that this “cognitive illiberalism” may decrease the “democratic legitimacy” of the Court’s decisions. While the Court saw Scott’s actions as obviously reasonable, it failed to realize that others might disagree. Indeed, as the article emphasizes, Justice Stevens’s dissent rejects the majority’s conclusion that Scott’s actions were reasonable.

The article was quite influential, and led many scholars to conclude that the Court’s decision was incorrect. The problem is that the Kahan study may not have asked the question that the Court actually answered. Unfortunately, Justice Scalia’s opinion in Scott v. Harris is arguably ambiguous about the basis for its outcome. The study assumes that the Court reversed the circuit court on the ground that, based on the videos, no reasonable juror could find that Scott’s actions were unreasonable under the circumstances. And it provides convincing evidence that at least some reasonable jurors could find that Scott’s actions were unreasonable under the circumstances.

But the Court may actually have reversed the circuit court on the ground that, based on the videos, no reasonable juror could find that Harris’s actions were not dangerous, and that police officers who use deadly force to stop dangerous drivers are entitled to qualified immunity.43 This reading of Scott v. Harris is consistent with the Court’s ultimate holding:

The car chase that respondent initiated in this case posed a substantial and immediate risk of serious physical injury to others; no reasonable jury could conclude otherwise. Scott’s attempt to terminate the chase by forcing respondent off the road was reasonable, and Scott is entitled to summary judgment.44

And it is bolstered by the concurring opinions of Justices Ginsburg and Breyer, both of whom explicitly stated their opposition to the adoption of a per se rule that police officers who use deadly force against dangerous drivers are entitled to qualified immunity.45

In other words, the Court seems to have adopted a per se rule that police officers who use deadly force against dangerous drivers are entitled to qualified immunity. And it concluded, based on the videos, that Harris was objectively driving dangerously, because he was speeding, passing when not permitted, and running red lights. One can certainly disagree with the wisdom of adopting such a per se rule. Perhaps the law should require police to use deadly force more judiciously by empowering jurors to evaluate the reasonableness of their actions under the circumstances. And one can certainly disagree with the Court’s assumption that speeding, passing when not permitted, and running red lights qualify as “dangerous” driving. But if they do, it is hard to see how a reasonable juror could find that Harris was not “driving dangerously,” because the videos conclusively establish that he was speeding, passing when not permitted, and running red lights. While the videos cannot determine whether Scott’s actions were reasonable, they can determine certain facts about what happened.

Indeed, the authors of the Kahan study later acknowledged that they did not ask the question posed by the Court – whether the videos showed that Harris was driving dangerously – but rather asked the question the Court “should” have asked – whether the videos showed that Harris was driving sufficiently dangerously to justify Scott’s actions:

Because Justice Scalia, on the basis of the video and the background facts of the case, concluded that Harris, the fleeing driver, was entirely at fault, we asked our subjects to indicate agreement or disagreement with the statement that the decision to pursue Harris wasn’t worth the risk to the public, and also to apportion fault for risk between Harris and the police. We also asked our subjects to state their level of agreement or disagreement with the proposition (the outcome of the case, essentially) that use of deadly force to terminate the chase was warranted in light of the risk that Harris’s driving posed to the public and the police. As Orin notes, we found, again, that a majority took positions consistent with that of the Scott majority, although on these issues there was even sharper dissent among demographically and culturally defined groups.

It’s true, as Orin notes and as we discuss in the paper, that these issues wouldn’t be submitted to a jury under the apparent, bright-line rule that Justice Scalia announced: “A police officer’s attempt to terminate a dangerous high-speed chase that threatens the lives of innocent bystanders does not violate the Fourth Amendment, even when it places the fleeing motorist at risk of serious injury or death.” 127 S. Ct. at 1779. But Justice Scalia’s justification for such a rule was that there was only one reasonable way to assess the balance of risks in a case like Scott. We wanted to find out whether ordinary people—whose judgments presumably are part of the reasonableness calculus—would agree. Our finding that in fact members of diverse subcommunities balance the risks differently (from one another and from the Court majority) raises the normative question whether the Court should have formulated a test that bars a jury from considering this matter.46

The Kahan study convincingly showed that, based on the videos, reasonable jurors could disagree about whether Scott’s actions were reasonable. But it did not necessarily show that reasonable jurors could disagree about whether Harris was driving dangerously, at least insofar as “driving dangerously” is given an objective definition consistent with Harris’s actions as documented in the videos.

Perhaps the Court adopted the wrong legal rule. Perhaps juries should determine holistically whether a police officer’s decision to use deadly force against a dangerous driver was reasonable. And perhaps allowing them to do so would increase the political legitimacy of liability determinations.

But it seems odd to conclude that the Court’s conclusion in Scott v. Harris was the product of unconscious bias, if it intentionally adopted a per se rule that police officers who use deadly force against dangerous drivers are entitled to qualified immunity. While the Court may have adopted the wrong rule, it did so quite consciously.

And its application of that rule seems clearly correct. The videos conclusively establish that Harris’s driving was objectively “dangerous,” based on the Court’s definition, and the Kahan study provides no evidence to the contrary. It only shows that reasonable people could disagree about whether Harris’s driving was sufficiently dangerous to justify Scott’s actions. Notably, Justice Stevens’s dissent effectively conceded that Harris’s driving was dangerous, but argued that it was insufficiently dangerous to justify Scott’s actions.

In a similar vein, Naomi Mezey argued that “the Scott majority displayed a pronounced visual illiteracy” by “assuming that the video showed them precisely what they were looking for.”47 Specifically, she noted that the Court failed to consider the perspective from which the videos were filmed, which may have increased the perception of danger. And it did not consider relevant facts not documented in the video relating to the relative dangerousness of the car chase. “For all the video shows, it does not provide much evidence about ‘the circumstances’ from which courts are to determine reasonableness.”48

Mezey is quite correct that the Court did not consider many facts relevant to the relative dangerousness of Harris’s actions. And the Court also did not consider the perspective from which the video was filmed. But perhaps the Court did not think those questions were relevant to its legal conclusion? If the legal question before the Court was whether Harris was “driving dangerously,” and “driving dangerously” is defined as, inter alia, speeding, passing when not permitted, and running red lights, then perhaps the Court considered those questions simply irrelevant. The video documented certain facts. If those facts were sufficient to show that Harris was “driving dangerously” as a matter of law, then the Court was correct to grant summary judgment, because there was nothing for a jury to decide.

Of course, that is not to say that the Court’s adopted the right legal rule. Perhaps the question of “dangerousness” should always be a holistic question decided by the jury. And perhaps the Court adopted an overly mechanical standard for “dangerous driving.” But these are normative questions about the proper standard of care to apply to police and citizens, and structural questions about the respective roles of the court and the jury, not questions of fact about what the video did or didn’t show.

A Reasonable Man

My movie A Reasonable Man combines the videos at issue in Scott v. Harris with excerpts from the oral argument before the Supreme Court.49 It encourages viewers to consider the Court’s description of the videos in relation to the videos themselves. Specifically, it asks viewers to consider how the Supreme Court interpreted the videos and how the videos affected the Court’s decision. Should the Court have adopted a bright line rule that it is reasonable for a police officer to use deadly force to stop a fleeing motorist who is driving dangerously? Did the Court correctly find that the videos provided conclusive evidence that Harris was driving dangerously? Or should the Court have permitted a jury to make those determinations?

Victim Impact Evidence

Historically, the United States and the overwhelming majority of states permitted courts to impose capital punishment on criminal defendants for a variety of crimes, including murder and rape. In Furman v. Georgia (1972), the Supreme Court narrowly held that capital punishment violated the Eighth Amendment prohibition on “cruel and unusual punishment,” in a fractured opinion.50 In Gregg v. Georgia (1976), the Court reversed itself, holding that capital punishment for first degree murder did not violate the Eighth Amendment, so long as capital trials are bifurcated into separate guilt and penalty phases.51

In the guilt phase of a capital trial, the jury determines whether the defendant is guilty of a capital crime, and in the penalty phase, it determines whether capital punishment is justified, under the circumstances. During the penalty phase, the jury may consider, inter alia, both mitigating and aggravating evidence relating to the defendant’s culpability.

The penalty phase was intended to enable defense lawyers to present mitigating evidence relating to the defendant’s character or circumstances, in order to convince the jury not to impose capital punishment. But prosecutors and victims’ rights advocates soon asked courts for permission to present “victim impact evidence,” or evidence relating to the effect of the crime on the victim, the victim’s family, and the community.

In Booth v. Maryland (1987), the Court held that the introduction of victim impact evidence during the penalty phase of a capital trial violated the Eighth Amendment, because it was irrelevant to the sentencing decision. 52 And in South Carolina v. Gathers (1991), it reaffirmed Booth, holding that the introduction of any evidence relating to the victim’s character violated the Eighth Amendment. 53

But in Payne v. Tennessee (1991), the Court reversed itself and held that the introduction of victim impact evidence did not violate the Eighth Amendment.54 According to the Court, “In the majority of cases, and in this case, victim impact evidence serves entirely legitimate purposes. In the event that evidence is introduced that is so unduly prejudicial that it renders the trial fundamentally unfair, the Due Process Clause of the Fourteenth Amendment provides a mechanism for relief.”55 And it specifically observed that allowing the admission of mitigating evidence, but not victim impact evidence, was unfair to the state, because it prevented the introduction of evidence relevant to the jury’s determination of the appropriate punishment.56

After Payne, the Court offered no further guidance on how to evaluate the admissibility of victim impact evidence. Different states adopted different rules relating to victim impact evidence, and many states began to permit the introduction of many different kinds of victim impact evidence, including victim impact videos.57

Opponents of victim impact evidence argue that it should be excluded because its emotional appeal reduces the legitimacy of the criminal justice system by distorting the jury’s decisionmaking process and creating an unavoidable risk of improper discrimination among victims.58 Specifically, the admissibility of victim impact evidence may encourage juries to punish criminals who harm certain kinds of people more harshly than criminals who harm other kinds of people. For example, victim impact evidence relating to the murder of a child may have more emotional impact on a jury than victim impact evidence relating to the murder of a mentally ill adult.

By contrast, proponents of victim impact evidence argue that it increases the legitimacy of criminal justice system by enabling the jury to better understand the social harm caused by a crime.59 The criminal law is not an abstraction. It inevitably addresses the experiences of both defendants and victims. And victim impact evidence enables the jury to better understand the actual consequences of a defendant’s criminal acts.

Kelly v. California

In 1993, Sara Nokomis Weir was 19 years old. She lived in Burbank and worked at Warner Brothers Studios. In her spare time, she went to the gym, where she met Douglas Oliver Kelly. Eventually, she hired Kelly as her personal trainer, and they became friends.

Kelly lived with his girlfriend Michelle and her 10-year-old son Eric in Michelle’s North Hollywood apartment. On August 30, 1993, Michelle discovered that another woman had visited the apartment. When Kelly returned to the apartment, she locked him out, but he kicked open the door and assaulted her. The next day, Michelle called the police. Kelly was arrested and spent the night in prison. Michelle obtained a restraining order and moved in with her sister. Her landlord changed the locks on her apartment, but Kelly climbed onto the balcony and broke into Michelle’s apartment.

Weir spent the Labor Day weekend with her adoptive mother, Martha Farwell. On September 7, Weir called in sick to work, explaining that a friend had committed suicide. She was never heard from again. Robert Coty managed an apartment building across the street from Michelle’s apartment. Shortly after Labor Day, Coty saw Kelly “dominating” a naked woman in Michelle’s apartment.

On September 15, Michelle and Eric returned to their apartment. Eric discovered Weir’s decomposing body under his bed. She was naked and wrapped in a blanket. She had been stabbed 29 times with a pair of scissors. A plastic bag covered her head and was taped around her neck. Eric’s baseball helmet was stuck on her head, over the bag. Kelly’s fingerprints were everywhere. Kelly disappeared. But on November 24, he was arrested at the United States-Mexico border, attempting to re-enter the United States. He had in his possession forged checks belonging to Weir. Later, Weir’s car was found in Mexico.

Kelly was tried in California state court for first-degree murder, with special circumstances of robbery and rape. During the guilt phase of Kelly’s trial, the court admitted evidence that he had defrauded or raped several different women. During the penalty phase, the court admitted victim impact evidence, including a victim impact video prepared by Farwell. The video consisted primarily of photographs and videos of Weir’s life. It was narrated by Farwell, and accompanied by music written and performed by Enya, Weir’s favorite musician.

Kelly did not introduce any evidence during the guilt or penalty phases of his trial.60

The jury found Kelly guilty and sentenced him to death. On appeal, Kelly argued, inter alia, that the trial court abused its discretion by admitting the victim impact video, because it was too long and too emotional, especially because it included irrelevant elements like the Enya soundtrack. The California Supreme Court affirmed the trial court judgment, concluding that the trial court properly admitted the victim impact video. It concluded:

Most of the videotape was factual, relevant, and not unduly emotional, and the trial court had discretion to admit it. To the extent it contained aspects that were themselves emotional without being factual—the background music and the final portion, perhaps—we are confident that permitting the jury to view and hear those portions along with the rest of the mostly factual and relevant videotape was harmless in light of the trial as a whole. These days, background music in videotapes is very common; the soft music here would not have had a significant impact on the jury.61

Kelly filed a petition for certiorari in the United States Supreme Court, arguing that the admission of the victim impact video was so unduly prejudicial that it violated his rights under the Due Process Clause of the Fourteenth Amendment. The Court denied Kelly’s petition. Justices Breyer and Stevens voted to grant the petition, and Justice Stevens issued a statement objecting the Court’s denial of the petition.62

Justice Stevens provided a brief history of victim impact evidence, observing that the Court had initially prohibited the introduction of victim impact evidence, but later reversed itself, and began permitting it, so long as it was not “unduly prejudicial.” He further observed:

Victim impact evidence is powerful in any form. But in each of these cases, the evidence was especially prejudicial. Although the video shown to each jury was emotionally evocative, it was not probative of the culpability or character of the offender or the circumstances of the offense. Nor was the evidence particularly probative of the impact of the crimes on the victims’ family members: The pictures and video footage shown to the juries portrayed events that occurred long before the respective crimes were committed and that bore no direct relation to the effect of crime on the victims’ family members.

Equally troubling is the form in which the evidence was presented. As these cases demonstrate, when victim impact evidence is enhanced with music, photographs, or video footage, the risk of unfair prejudice quickly becomes overwhelming. While the video tributes at issue in these cases contained moving portrayals of the lives of the victims, their primary, if not sole, effect was to rouse jurors’ sympathy for the victims and increase jurors’ antipathy for the capital defendants. The videos added nothing relevant to the jury’s deliberations and invited a verdict based on sentiment, rather than reasoned judgment.63

Justice Stevens also attached the victim impact video at issue to his statement.64 Kelly remains on death row at San Quentin.

Reactions to Kelly v. California

Unsurprisingly, opponents of victim impact evidence argued that the victim impact video was “unduly emotional,” and supporters of victim impact evidence disagreed. The dispute hinged on whether the use of background music makes a victim impact video “unduly emotional.” Many felt it did. As Bennett Capers observed, music affects decisions by provoking inchoate and involuntary emotional responses:

Music does more than heighten the emotion experienced by listeners. Even “background” music is not really in the background. It is an independent component that not only affects the body; it is a way of knowing that can induce collective action. In a way we have yet to fully comprehend, music, through its emotive power, can tell a listener how the story should end. Whether a life should end. And it does this without language. No court stenographer can record what the music said. Even an audio recording marked as an exhibit and preserved as part of the record on appeal is likely inadequate since any listening will fail to capture music’s collective effect, that in-the-moment effect. This effect, even without language, has its say. This effect, even without language, adds to the mix of thoughts jurors have. This effect, even without language, can say vote yes. Vote death.65

But the doctrinal disagreement is essentially tactical.66 The underlying theoretical dispute goes to the purpose of the sentencing phase. Opponents of victim impact evidence see the sentencing phase of a capital trial as an opportunity for the defendant to introduce mitigating evidence intended to convince the jury not to impose capital punishment. By contrast, supporters of victim impact evidence see the sentencing phase as also an opportunity for the prosecutor to introduce evidence of the impact of the crime on the victim and the community. In both cases, the relevant evidence is necessarily emotional. It must appeal to the jury’s feelings about the appropriate punishment under the circumstances. The question is, which feelings should matter? Should the sentencing consider only the relative culpability of the defendant in the abstract? Or should it consider the social impact of the defendant’s crime as well?

The oddity of the doctrinal standard is that it asks courts to determine whether a victim impact video is “unduly emotional,” when the entire point of a victim impact video is to appeal to the jury’s emotions. In other words, the doctrinal standard implies that ineffective victim impact videos are probably admissible, but effective victim impact videos are not. That cannot be right. If the kinds of emotions provoked by victim impact videos are relevant to the jury’s decision, then they should be admissible, irrespective of how much emotion they provoke. But if those emotions are not relevant, then the videos should not be admissible, even if they are ineffective.

Sara Nokomis Weir

My movie Sara Nokomis Weir combines the victim impact video at issue in Kelly v. California with excerpts from the oral argument before the California Supreme Court.67 It asks viewers to consider the court’s description of the video in relation to the video itself. And it encourages viewers to consider how the video affected the court’s decision. What is the purpose of asking the jury to impose the sentence? Is this the kind of evidence the jury should consider in deciding what sentence to impose? Is the court asking the right questions?

As film scholar Jaimie Baron has observed:

[I]t imposes a “reparatory gaze,” that seeks to redress a possible wrong. Although Frye’s film does not take an explicit stance vis-à-vis the legality or ethics of the victim impact video, I suggest that by pointing out the inability for the court to adequately discuss the video, the film implies that the legal system may require revision. Frye’s film thus attempts to reestablish an ethical gaze over and above the one solicited by the victim impact video, one which asks us whether this is the kind of “evidence” we wish to allow a jury to see when a human life is at stake.68

Brumfield v. Cain

Shortly after midnight on January 7, 1993, Kevan Brumfield murdered Corporal Betty Smothers of the Baton Rouge Police Department. At the time, Smothers was off-duty, working her second job as a uniformed security officer for a local grocery store. Smothers drove the assistant manager of the grocery store, Kimen Lee, to a local bank in her police cruiser, in order to make the store’s nightly deposit. When Smothers and Lee arrived at the bank, Brumfield and his accomplice, Henri Broadway, opened fire. Brumfield shot Smothers five times in the forearm, chest, and head with a .380 caliber handgun. Lee also received multiple gunshot wounds, but managed to drive the cruiser to a local convenience store for help. Emergency responders took Smothers and Lee to the hospital, where Smothers was pronounced dead on arrival. Lee survived and described Broadway to the police.69

On January 11, 1993, the police arrested Brumfield for Smothers’s murder. After several hours of interrogation, he eventually provided a videotaped confession. In 1995, Brumfield was tried in Louisiana court, convicted of first-degree murder, and sentenced to death. During the trial, Brumfield introduced some evidence of intellectual disability. In 1998, Brumfield appealed to the Louisiana Supreme Court, which affirmed his conviction, and the United States Supreme Court denied his petition for certiorari.70

In 2000, Brumfield filed a petition for post-conviction relief in Louisiana state court, alleging that he was ineligible for capital punishment based on insanity. In 2002, the United States Supreme Court held in Atkins v. Virginia that the execution of “mentally retarded” criminal defendants violates the Eighth Amendment prohibition of “cruel and unusual punishments.”71 Brumfield amended his petition to claim intellectual disability and request an Atkins hearing, based on:

1) his IQ score, obtained prior to trial, of 75; 2) his slow progress in school; 3) his premature birth; 4) his treatment at multiple psychiatric hospitals; 5) various medications he was prescribed; and 6) testimony that he exhibited slower responses than “normal babies,” suffered from seizures, and was hospitalized for months after his birth.72

On October 23, 2003, the court denied Brumfield’s request for an Atkins hearing to determine intellectual disability. Brumfield filed a writ of habeas corpus in the Louisiana Supreme Court, which was also denied.

On November 4, 2004, Brumfield filed a writ of habeas corpus in federal district court, alleging that Louisiana improperly denied his request for an Atkins hearing. In 2010, the federal court held an Atkins hearing, and in 2012, it enjoined the State of Louisiana from executing Brumfield on the ground that he was “mentally retarded.”73 The State appealed, and the circuit court reversed, on the ground that the district court improperly failed to defer to the Louisiana state court’s decision to deny Brumfield’s petition for an Atkins hearing.74

Brumfield filed a petition for a writ of certiorari in the United States Supreme Court, which was granted. The case was argued before the Court on March 30, 2015. As per his usual practice, Justice Thomas was silent during the oral argument. On June 18, 2015, the Court vacated the circuit court’s decision and remanded for further proceedings. The majority opinion was written by Justice Sotomayor and held that the state court’s refusal to grant Brumfield’s request for an Atkins hearing was unreasonable because he had introduced enough evidence to create a “reasonable doubt” of intellectual disability.75

Justice Thomas wrote a dissenting opinion, in which he argued that Brumfield failed to provide clear and convincing evidence of intellectual disability and that the majority improperly failed to defer to the state court’s determination of fact that Brumfield was not intellectually disabled.76 He also attached the video of Brumfield’s confession.77

In addition, Justice Thomas reflected on Brumfield’s actions in relation to the actions of Smothers’s children. Specifically, he noted that when Smothers died, her teenage son Warrick Dunn acted as the parent of his younger siblings. Dunn eventually joined the National Football League, playing for the Tampa Bay Buccaneers and Atlanta Falcons, and continued to care for his siblings. He is also active in philanthropy, founding charitable organizations in his mother’s name to provide free grief counseling services to children in the Baton Rouge area, among other things.78

On Reasonableness

My movie “On Reasonableness” combines the video of Brumfield’s confession with excerpts from the United States Supreme Court oral argument in Brumfield v. Cain.79 It allows viewers to compare the statements of fact about Brumfield’s intellectual capacity made by the lawyers representing Brumfield and the State of Louisiana to the motion picture evidence of Brumfield’s behavior when he confessed to murdering Smothers. It also encourages viewers to reflect on whether that comparison affects their perception of the relative strength of the legal arguments presented by Brumfield and the State of Louisiana, or the majority and dissenting opinions of the Supreme Court.

Finally, it asks viewers to consider why Justice Thomas, who typically refrains from speaking during oral arguments, chose to speak in this case through the attachment of motion picture evidence to his dissent. Why did he want people to watch the video of Brumfield’s confession? Why did he think it was relevant to Brumfield’s intellectual disability claim? And what does it actually show? Does this video provide any relevant information about Brumfield’s intellectual capacity? If so, how can and should it be used?


The history of motion picture evidence is closely tied to the history of amateur cinema. As the difficulty and expense of making motion pictures decreased, the quantity of motion picture evidence increased. Today, we are awash in motion picture evidence. While courts have long admitted and relied on motion picture evidence, its public visibility has increased, as people have created and courts have received more motion picture evidence of different kinds. Even the Supreme Court has begun to share motion picture evidence as exhibits to its opinions, in order to illustrate its reasoning. Reflecting on when and why the Court has used and shared motion picture evidence may illuminate the different ways in which motion picture evidence informs and shapes legal reasoning and rhetoric.


Brian L. Frye is an Associate Professor of Law at the University of Kentucky College of Law. His scholarship focuses on intellectual property, charity law, and art law. He is also an essayist and artist of no particular distinction.