It was Hollywood’s first scandal.

The year was 1921. Prohibition was young. The silent film era was at its peak, and the picture business boasted no greater star than Roscoe “Fatty” Arbuckle. 266 pounds, and billed by theaters as “worth his weight in laughs,” Arbuckle was larger than life. His paychecks from Paramount Pictures were unprecedentedly outrageous. Between 1913 and 1921, Arbuckle had been credited in nearly 150 films and shorts. Year after year, America crammed into theaters to watch his follies, roaring as he buried his titanic body beneath the sand at Coney Island, or donned a curly wig and deftly tap danced with a group of young women, or sprinted on top of a speeding train in a full suit, top hat wavering precariously in his hand.

Fatty Arbuckle was a clown. In one film, he stuffed a serving of spaghetti in his mouth so greedily that he didn’t even notice when his necktie mixed in with the noodles and nearly swallowed it. In another, he waddled too near to a tap dancer and had his hat kicked from his head. He got punched in the face trying to break up a fistfight. He ate paint. And he always danced with surprising grace, like a twirling Jenga tower in a perpetual state of near-collapse. A toppling Arbuckle would right himself, but only temporarily; his destiny was always to tumble dramatically to the ground and splay out across the floor in what looked to be several pieces.

Like a clown, Arbuckle always looks as though things are happening to him. He is a man with no agency. He falls, or gets hit with something, and the audience chuckles. But in truth, the clown is always in total control. He executes each false step and tumble with precision. Because while a clown pretends as though his life is filled with unknowns, he is the only one who knows exactly what will happen.

But there was little certainty in a San Francisco courthouse that September in 1921, where Fatty sat sweating before a judge. The tap dancing was over; Arbuckle was set to be charged with the manslaughter of a bit actress named Virginia Rappe. Ms. Rappe had attended Arbuckle’s soiree, a boozy Labor Day gathering at the St. Francis Hotel downtown. She died of a ruptured bladder four days later. It was up to the community to determine whether the rupture stemmed from an attack by Arbuckle. So, naturally, we brought in a jury.

It was a century before Hollywood’s deep-seated misogyny would be truly revealed to the world. That’s one of the most troubling legacies of the Arbuckle case; it highlights how long women have been fighting for America’s ears. And for far too long, the burden of proof has been misplaced.

We like to think that just because something is true that it will be easy to prove. But that is the exception, not the rule. And this case was not an exception. Rather, scavenging through the legend of Arbuckle’s case prompts a whirlwind of facts and fictions and biases. There were faked fingerprints. Bribed witnesses. Male jurors turned deaf ears to their female counterparts. Judges showed no moral center. In its day, the event was a tabloid’s fantasy; William Hearst, contemporary peddler of the San Francisco Examiner, claimed that Arbuckle’s scandal sold more papers that the sinking of the Lusitania. Arbuckle’s committal hearing alone drew a crowd so large and rambunctious that nearly 100 police officers were needed to keep the peace.

The clown had found his circus.

But after Arbuckle had been paraded through three bewildering trials, we had learned little about what happened that evening at the St. Francis Hotel. That makes sense: reconstructing the past is no simple task.

But we would never let that stop us from reaching a verdict.

That September, in court for his committal hearing, Arbuckle was unsure if his case would even be sent to trial. The tabloid papers had run wild with the idea of the obese Arbuckle crushing a young actress to death with his considerable weight, but in the courtroom, little evidence had been presented. Maybe that wouldn’t matter. There was intense public pressure on the courts. Newly minted San Francisco District Attorney Matthew Brady was prosecuting the case. Brady had political aspirations and was determined to get Arbuckle behind bars. He wanted show everyone that he could bring justice to Hollywood.

But despite his ambition, Brady’s case was strangely feeble. He had called several witnesses to the stand. There was Al Semnacher, Virginia Rappe’s manager, who’d been at the party. Semnacher had made widely-cited claims in the press that Arbuckle had bragged to him about performing a lewd act upon Rappe with a piece of ice. But in the courthouse, Semnacher’s actual testimony was clouded by an apparent lack of recollection, and his on-the-record opinion was that Arbuckle had been acting in a “gentlemanly” manner that evening in the hotel.

A St. Francis hotel maid had also been produced as a witness, but presiding judge Sylvain J. Lazarus felt her testimony was weak and questioned its relevance; the maid hadn’t even been in the room. Then there was the matter of the complaining witness: Maude Delmont, who was the impetus for Arbuckle’s arrest. Delmont had been at the party as well, and she was the only one who had actually seen something happen. She told the cops that Arbuckle had forced Rappe into a private room, telling her “I’ve waited for you five years, and now I’ve got you.” Delmont was the one who heard Rappe screaming from behind the locked door. She was the one who tried to rescue her. She was the one who had found Rappe sprawled on the hotel bed, writhing in pain.

Delmont was the prosecution’s star witness, the one who gave the case its life. But she was never even called to the stand.

Arbuckle’s attorney was livid at being robbed of the chance to cross-examine her. Even Judge Lazarus was shocked. “I regret exceedingly that the district attorney has not seen fit to place her on the stand,” he said, referring to Delmont. He threatened to throw out the case. He blistered Semnacher’s testimony, calling it worthless, and questioned the hotel maid. He tore the evidence to shreds, blasting the case against Arbuckle as “the merest outline; a very skeletonized description of what, if anything, occurred in the apartment.”

Hollywood’s reckoning would have to wait. Surely, Judge Lazarus was not going to put the uncertain clown to trial for manslaughter without a shred of evidence.

But remember the context. This was Prohibition. The world-famous Fatty Arbuckle had thrown a party in a suite of hotel rooms with booze, music, and dancing. Women had gotten drunk and stripped off their clothes. Decent people nationwide were appalled by the behavior. Sure, Arbuckle was on trial for manslaughter, not hedonism. But when Judge Lazarus announced his decision to send Arbuckle to trial, he claimed the case was about much more than the entertainer alone. Instead, he called it a “public lesson,” and reminded the court that they were not really trying Fatty Arbuckle the man, but rather their “present day morals.” Because of this, Lazarus felt pity for Arbuckle. He called the comedian a “poor, unfortunate man.” But that didn’t matter. He sent Arbuckle to trial regardless, without a lick of substantive evidence, because he was unimpressed by the precedent set by the clown’s social life.

Amid the continued media frenzy, Arbuckle’s first trial began in earnest that November. Once again District Attorney Brady came armed with witnesses. There was fellow partygoer Zey Prevon. The security man Jesse Norgard. They made their claims, but fared poorly under cross-examination. As for the defense, a nurse testified that Rappe had complained of internal pain in the months leading up to the party. Rappe suffered from a chronic bladder condition that was inflamed by alcohol, and she had been drinking heavily at the hotel that evening. The defense claimed that this was the reason for her ruptured bladder.

Then Arbuckle himself testified. He comported himself well and impressed the jury. He said he’d been trying to help the ailing Rappe that evening. He said he was looking out for her.

And complaining witness Maude Delmont was absent yet again. But by now it was clear why she was held back from the proceedings. Delmont had a criminal history of extortion; she had a known racket of accusing wealthy men of rape and blackmailing them into paying her. She’d even sent a damning telegram to various California attorneys, saying “WE HAVE ROSCOE ARBUCKLE IN A HOLE HERE CHANCE TO MAKE SOME MONEY OUT OF HIM”.

Delmont was not a credible witness. The D.A. needed her to make his case, but he knew better than to put her on the stand.

The jury deliberations would have been quick, except for the objections of one juror who was convinced Arbuckle was guilty. But that juror, Helen Hubbard, was hardly impartial. She should never have made it through the selection process; her husband worked extensively with District Attorney Brady’s office. She’d reportedly had her fingers in her ears during Arbuckle’s defense. When deliberations began, she immediately announced that she would vote guilty “until hell freezes over”, and refused to discuss the details of the case with the other jurors, saying she’d made up her mind in the courtroom. And even after forty-four hours of arguing with her cohorts, she had swung the opinion of only one other juror, Thomas Kilkenny. The jury voted 10-2 to acquit.

10-2 is a mistrial, but hardly an indication of widespread jury disagreement. The dissenting Thomas Kilkenny went silent after the trial, which was unusual given the intense public interest. Helen Hubbard pled her case in the media, but otherwise the jury had presented an all-but-overwhelming response to the case. Nearly everyone was certain; Arbuckle was innocent. It was just going to take one more trial to prove it.

But that’s not what happened.

Before we can understand what happened in the second trial of Fatty Arbuckle, we first need to explore a certain peculiarity of the human psyche. To elicit that peculiarity, let us skip forward to a contentious 21^st century debate. This is a divisive argument, one in which we would rather die clinging to our dogmas than change our minds. The mere realization that a loved one falls on the wrong side of the fence may be grounds to question your entire relationship with that person. The debate is that fierce.

I am talking about the Great Hawaiian Pizza Conflict.

You probably know Hawaiian pizza as an adventurous creation that combines the mundane pizza toppings of tomato sauce and cheese with ham and, famously, pineapple. Chances are, either you love Hawaiian pizza or you completely abhor it. There is no middle ground. It is a moral quandary just shy of being written into our legislatures; the President of Iceland has gone as far as to say that he wished he could ban pineapple as a pizza topping nationwide. Amidst the ensuing consternation, the pollsters over at YouGov released a UK survey that showed that while Brits were split evenly on the taste of Hawaiian pizza, a whopping 15% of the country would support legally banning it.

Gordon Ramsey has weighed in; he hates it. Justin Trudeau loves it. Don’t bother trying to keep up with the meme war; it’s exhausting. Thankfully, food blog Spoon University ended the debate once and for all in 2017, scientifically proving that pineapple does belong on pizza, only to watch Mel Magazine later end the debate again in 2020 by proving that it actually does not.

Frankly, I don’t touch the stuff.

But this is a joke, right? Surely we are not serious about legally banning a pizza topping that half of us think is great. The safe assumption is that this a faux debate. It’s tongue-in-cheek. Taste is subjective; my enjoyment of a particular food is not a testament to its righteousness. I just like it. We know that.

Yet we find the debate funny. Or entertaining. Or something. It’s hard to fathom why someone would even post their preference on a particular pizza topping to Twitter in the first place, much less how it could go viral. Why should that even spark a conversation? I get that the memes are funny. But I’m not sure I get why they’re funny.

Maybe the Great Hawaiian Pizza Conflict is really a recognition that the very idea of certainty is a joke. The debate feels like self-satire, an acknowledgement that we pretend to be sure of uncertain things. And we assert that conviction even more fervently inside the bowels of a heated debate. After all, disagreement is one of our favorite pastimes. Why should pizza be any different?

The truth is, we don’t come to an agreement on anything without incentive. What’s the closest thing to certainty that can be found in the world of American pizza? Pepperoni. It’s our favorite topping by a wide margin. If you don’t believe me, take a look at the pizza emoji on your phone. It defaults to pepperoni. As do we; roughly half of Dominos’ ordered pies are topped with those spicy little circles. But how are we so sure that it is the best—did Dominos present us with a wide variety of options? Did we choose pepperoni to be our darling?

Not quite. Think of the wee hours of Dominos’ audacious designs to take over the country’s pizza business; in order to spread across fifty states, Dominos needed to serve us the pizza that made the most sense for them. It was a negotiation between our taste buds and their shipping routes. And as the food gurus at Thrillist have reported, when Dominos scaled their delivery service in the 1960s, fresh chicken was the popular pizza topping. But that wouldn’t work for Dominos’ business model. They needed something that could be produced quickly and in great quantities, something that would survive a long voyage in the back of a truck, so they could get it to their shiny new franchises, no matter how far the company expanded. Pepperoni profiles even better for this process than other preserved meats because it takes less time to cure.

I suppose that makes it unfortunate that pepperoni is also one of the unhealthiest things you can eat. Those tiny patties are so tightly packed with salt, fat, and preservatives that adding them to an already indulgent slice of pizza is like putting a timebomb in your mouth. It drives up your chances of diabetes, heart disease, and some cancers. Maybe that is why no other country likes pepperoni toppings as much as we do; most don’t even care for it. Several countries, however, prefer the far leaner topping of fresh chicken.

Just like we did before Dominos told us there was a better way.

All consumer goods have upsides and downsides, but marketers promise their product’s flawlessness on no uncertain terms. Papa John’s guaranteed us “Better Ingredients, Better Pizza,” but they didn’t waste their money on high-quality toppings. They preferred to funnel those dollars into an ad campaign that said they had them.

Why wouldn’t they just buy the better toppings?

For the same reason that Dominos convinced us to like pepperoni—they knew we’d listen. Consumers have an incentive to believe; we permit deceptive pitches from corporations because there are simply too many consumer decisions in front of us. It is not reasonable to ask us to do the hard work of finding certainty as it relates to the best slice of pizza, the most diligent tax accountant, the safest car. But we still need our car to be the safest. This is why corporations base their marketing strategies around assuring us of their superiority; they don’t need to actually prove it. Their bet is that eventually we will succumb to their conviction. Because certainty can always be sold to those who have an incentive to buy it, and we don’t have the energy to question every retailer. If we did, the nation’s top pizza chains would probably have been born in New York, not the rural Midwest.

Our funneling towards pizza homogeneity doesn’t imply that the eventual pepperoni standard was chosen due to its supremacy. It just means that the fresh chicken wasn’t worth the fight. But that isn’t the case in other settings—for instance, a criminal trial. There the stakes are real, and we need an airtight process of achieving certainty. No longer can we defer to those who are pretending they know. Because when lives are in the balance, we can’t possibly rely on the same sloppy thinking that put thin circles of death on our favorite dish.

Right?

As we set to conclude Fatty Arbuckle’s case, let us first consider a thought experiment so simple that a grade-schooler could understand it.

You are a teacher at an elementary school. You have ten students in today’s class. You tell your students that you are holding up either one or two fingers behind your back. You ask them to guess which it is. Seven students guess that you are holding one finger, and three students guess that you are holding two.

What are the odds that the seven students who guessed that you are holding one finger are correct? 50%. The exact same odds of the three students who guessed two. Why is it not more likely that the seven students are correct? Because to give that majority group a probabilistic advantage, you’d have to have given them some sort of head start—a clue that the other students weren’t told, or two guesses, or smarter brains. Because if everyone has been given the same information, the majority and the minority have the same odds of getting the problem right. There is no advantage in the majority.

But our juries don’t believe that.

In the 1950s, the Chicago Jury Project peeked into the deliberation room and returned with some of the most famous and widely cited research on jury decision-making in American history. One of this project’s findings was that the first thing most juries do when they enter the deliberation room is vote. Who thinks the defendant was guilty? Who thinks not?

Here was the bombshell: 90% of the time, the majority on the initial vote represented the final verdict. It was just a matter of the many convincing the few to change their minds. All the majority had to do was sell the pepperoni. This may feel natural, but as with the classroom thought experiment, the majority voters didn’t hold any statistical edge over the minority. Both groups are presented the same information, and thus have the same probability of getting the correct result. Deferring to the majority, then, only makes sense under the assumption that on any given jury, there will be a majority of individuals who are smart enough to understand what happened, and a minority of remaining jurors who aren’t.

Do we think that’s true?

Helen Hubbard, the juror with her fingers in her ears during Fatty Arbuckle’s trial, certainly did not.

Keep in mind the enormous pressure on a juror. The defendant’s freedom is in your hands, as is the safety of the general public. It’s a complicated decision, and you can’t pass the buck. Trials are an enormous cost and time burden on everyone involved: the court system, the defendant, the jurors, judges, attorneys, witnesses, and in high-profile cases like Arbuckle’s, the public. This is why hung juries only occur in 5% of trials. Indecision is scorned. We plead with our jurors to come to some sort of certainty.

But absolute certainty is unreasonably difficult to attain. In fact, it is probably impossible.

So we shouldn’t be surprised when the Chicago Jury Project’s findings show us that jurors opt for the easiest path in their quest for certainty. Say a twelve-person jury walks into the deliberation room after a criminal trial. If their first vote comes up twelve to zero or zero to twelve, their work is done. But if that isn’t the case, there remain eleven possible iterations in which unanimity was not achieved prior to deliberation, as there may be between one and eleven jurors who think the defendant is guilty. One of those iterations is a six-six tie. In that case, there is not one path to ending the deliberations that appears easier than the other.

In the remaining ten iterations, the Chicago Jury Project tells us that if between five and eleven jurors think the defendant is not guilty, the jury nearly always acquits. If three or four jurors think the defendant is not guilty, the odds of acquittal are a toss-up. If only one or two jurors think the defendant is not guilty, conviction is nearly certain.

So in the ten iterations in which a jury has neither an immediately unanimous decision nor a totally even deadlock, they will choose the easiest road to resolving the trial 70% of the time (in the cases where one, two, seven, eight, nine, ten, or eleven jurors initially vote not guilty). 20% of the time, we don’t know whether they will take the easier road (in the cases where three or four jurors vote not guilty). Just one in ten iterations shows jurors accepting the hard road to certainty, and in that case, we start with a 7-5 initial split in favor of conviction. That is barely a majority in favor of guilt, and not strong enough grounds upon which to lock someone away.

So in nearly every case, the deliberation room is useless. If the first poll determines the result 90% of the time, we might as well have the vote in the jury box and accept the majority decision.

But a majority decision is not enough to resolve a criminal trial; just ask the Supreme Court. It must be unanimous. Why? Because our judicial punishments are severe. Convicted felons lose their most basic constitutional right: freedom; but even that penalty doesn’t satisfy our sadism. Take a step behind bars for a moment. Official studies regarding treatment of the incarcerated are uncommon, a telling fact in and of itself, but the Justice Department’s examinations of Alabama’s prisons in 2019 show us how little we care for the livelihoods of our felons. Homicide and sexual abuse are the norm inside those facilities, but the guards don’t seem to care. The report shows that “prisoners who are seriously injured or stabbed must find their way to security staff elsewhere in the facility or bang on the door to the dormitory to get the attention of correctional officers.” Others are “tied up for days by other prisoners while unnoticed by security staff.” For a graphic image of the violence seen by the Justice Department’s research team, look no further than the prisoner who “had bleach poured on him, was beaten with a mop handle, and was stabbed several times.”

Prison is probably the most violent place in the country. Yet no one on the correctional payroll feels responsibility to protect the people who are kept captive in those conditions. As a society, we could not subject our neighbors to such torture unless we were completely certain that they deserve it. This makes the fact that we will never be completely certain quite inconvenient.

If we came to terms with our uncertainty, we would need to totally rethink our penal system. 83% of federal criminal defendants are convicted. How would they feel sitting behind bars knowing that some jurors thought they were innocent?

Further, five in six released American prisoners are re-arrested within nine years. That indicates a broken theory of crime prevention. The very terminology of a “justice” system implies one based on retribution and revenge. But the problem with revenge is that if you dish it out unjustly, you yourself become criminal. A “judicial” system only works under the guise of certainty. Could we simply admit that to be a ruse, stop mistreating the convicted, and end the cycle of revenge?

Apparently, we cannot. Instead we have opted for obfuscation. We have chosen a jury process that masquerades as a unanimous vote but is in truth a simple majority decision. Majority decisions are fine. We use them frequently. But they are imperfect.

Thus, their power needs to be restricted.

So what did Fatty Arbuckle’s second jury think of his case?

Arbuckle’s second trial felt a lot like his first. Same prosecutor. Same judge. Same witnesses. If anything, District Attorney Brady’s case was even weaker. His witnesses kept proving problematic. Zey Prevon testified that Brady had forced her to lie in the first trial. Security officer Jesse Norgard was discovered to have been recently charged with the sexual assault of a young girl; he was seeking a reduced sentence from Brady. Fingerprint evidence from the first trial was called into question. And you can probably guess whether Maude Delmont was involved.

Arbuckle and his attorney were so convinced that he’d be acquitted that Fatty declined to testify.

And again, the jury returned from their deliberations nearly certain of what had happened. They voted 9-3.

Only this time, it was the other way around.

Nine jurors had voted to convict Arbuckle. This jury was nearly as certain as the first, but of the complete opposite verdict!

But is it really surprising that two juries would be presented the same evidence only to arrive at totally different conclusions? It is not hard to think of another example in which groups of Americans look at the same complicated picture and then fall into utter disagreement over it. We do it with pizza.

Everyone was certain, but no one was right.

Let’s revisit Arbuckle’s committal hearing, when Judge Lazarus sent him to trial even as he lambasted the prosecutor. Remember the only witness that had caught Lazarus’ eye? It was the hotel maid. She’d testified that she’d heard Virginia Rappe pleading “No. No. Oh my God.” That was real testimony. And seriously, how much more incriminating can a piece of testimony be in a rape case in which the victim is deceased?

But Lazarus wanted to write it off. Why? He said, “We must remember a possibility suggested in a line from Byron: ‘And saying she would ne’er consent, consented.’”

Lazarus had heard testimony from a witness who had claimed she’d heard Rappe pleading anxiously with her attacker prior to an alleged rape, but he didn’t really care because you just never know if a woman is saying no because she means it, or if she is just playing hard to get.

And he was the judge.

By the way, the “line from Byron” Lazarus mentions? That refers to a poem about Don Juan, the legendary mythical womanizer. If you don’t know him well, Don Juan’s legend first found the stage in “The Trickster of Seville and the Stone Guest,” which begins with him pretending to be another man in order to trick a woman into sex, and ends with him seducing yet another woman and killing her father when he tries to avenge his daughter’s honor. This was a common theme in Don Juan’s legend. He seduced women for sport and murdered those who got in his way.

Maybe not the keenest analogy at a rape and manslaughter trial.

Is it all coming into focus now? Arbuckle and his judges were all misogynists, and the stubborn men in the first jury had refused to listen to Helen Hubbard, one of the rare jurors who was willing to fight for truth even once a simpler path to certainty had been established.

That’s why the second jury voted 9-3 in favor of conviction. By then, all the facts were on the table. They were able to see what the first jury refused. Again, they were nearly certain. Surely the third panel would resoundingly direct a shackled Arbuckle to a life behind bars.

But of course they didn’t. Arbuckle’s third trial ended with a five-minute jury deliberation. After sixty seconds, they had already declared him innocent. They spent their remaining four minutes writing Arbuckle an apology. “Acquittal is not enough for Roscoe Arbuckle,” the jury said. “A grave injustice has been done.”

Hollywood’s first major scandal was finally closed by a jury that claimed complete certainty. At long last, we had arrived at pepperoni. The final jury’s confidence would be surprising, given the clear uncertainty seen in the earlier trials, until you remember that 95% of juries reach certainty even though the truth is almost always an unknown.

And regardless of what one of his 36 jurors would tell you, Arbuckle’s truth will always be unknown. His vehement acquittal did little to convince the American people of his innocence. His career never rebounded. He was banned from Hollywood. Heckled in public. The industry eventually reinstated him, but no one would hire him. He tried a comeback as a director under a pseudonym, but Arbuckle’s appeal had always been physical comedy. Clowns have no business behind the camera.

But in 1933, Arbuckle finally got his second break. It was late June; he had just finished filming a new series of shorts. They were good. On June 29^th, he signed a deal with Warner Bros. He was going to do feature films. With audio this time. Fatty was back. He told his manager, “This is the happiest day of my life … it’s a second honeymoon.”

He died in his sleep that night. He was 46.

Prohibition would end five months later.

Is Arbuckle’s death tragedy or cold justice? Like Hawaiian Pizza, it depends who you ask. History has been kinder to him. Many now believe that the judicial system treated him unfairly, and that the tabloids ran unnecessarily wild with his reputation. But are we more certain now than we were then?

Certainty begets power; it puts the gun in your hand and makes an enemy out of your opponent. It allows a citizen to permit the torture of a prisoner. When you claim certainty, any concession to your detractors becomes immoral. And with this mindset, why would we expect twelve random American jurors to all see the complicated evidence presented in a criminal trial the same way? We don’t see that cohesion anywhere else in this country. The only reason juries come to an agreement is because they want to go home. The longer the losers of the initial vote hold out, the longer it is until that can happen.

And when they get home and assure their loved ones that they have performed their civic duty, they are unlikely to mention that they did so by deferring to the loudest voice in the room. But they must’ve. Because what other means would they have to achieve the certainty their country is requesting of them?

Everyone is certain, but no one is right, and everyone goes home except the defendant.

Democracy at work.