Authors: Ken Alder
Agnes de Mille, Keeler’s old college friend, had a rare peek at how Keeler pulled this off. While visiting Chicago with her dance troupe, she watched with bank managers through a one-way mirror as Keeler examined an applicant for a teller’s job. The young man appeared nervous. At one point, when Keeler pressed him for information about his past, he asked to be released and began to disentangle himself from the apparatus. "At this point," de Mille recalled, "Leonarde got rough. I have never heard him talk to anyone with such abruptness. He told the boy, in effect, that unless he took the test, he could not have the job." Cowed, the young man answered in a subdued tone. Suddenly, he admitted having served time for theft. Inside the observation room, de Mille heard the bankers gasp. But Keeler did not register disapproval. He completed the interrogation, unhooked the subject, and handed him a cigarette lit in his own mouth. Then came the payoff. The trouble was his mother-in-law; she had ruined his marriage. For a half hour, the young man poured out his soul. Keeler listened sympathetically, shook the young man’s hand, and sent him out the door. Then he turned to the bankers and told them the young man was a safe bet. "He’s perfectly honest as far as I can tell," he said, "but he did serve a jail term."
De Mille was shocked, though less by the young man’s fate than by her sweet Nard’s ruthlessness. Still, she could not help but be impressed—as one choreographer to another—with Keeler’s skill at striking a powerful emotional chord in the body of his singular audience—and then releasing the tension. As a further demonstration, he strapped one of de Mille’s male dancers into his machine and announced, "Now I’m going to ask you an embarrassing question"—presumably about his sexuality. The needles jerked so dramatically that Keeler didn’t even ask the question. The nervous laughter was a tribute to the test’s insidious power.
Keeler exploited this same technique to extract confessions in criminal investigations, at a clip of 75 percent among "guilty" suspects. In his courses at Northwestern, Keeler showed the police and prosecutors how easily it could be done. First, maneuver subjects into taking the test by telling them that only guilty people feared it; this was not difficult with suspects aware that their chance for release depended on cooperating. Then convince the subject that the lie detector worked, and repeat the questions until a full confession was forthcoming. It wasn’t science, but it did the trick.
After all, the courts may have ruled out polygraph results at trial, but they still accepted confessions obtained during an exam—even when the confession was oral, even when there was only one witness, and even when the operator falsely informed the subject he had failed the test. Such shenanigans had long been a legitimate tool of police interrogation. Keeler’s colleague at Northwestern, Fred Inbau, wrote the book on how to pull it off, in terms any police officer could understand. His slender volume
Lie Detection and Criminal Investigation
became the bible of police interrogators, teaching officers the byways of psychological coercion, tricks deemed so likely to induce false confessions that they earned Inbau a sharp rebuke from Chief Justice Earl Warren in the
Miranda
decision of 1966.
How successful were Keeler’s methods? In 1933, while demonstrating his so-called jury box in front of 200 members of the Kansas Bar Association, he was on the verge of getting a young man to confess to a car heist—until a reporter from the
Topeka State Journal,
onstage a few feet away, hissed loudly, "Don’t be a darn fool." More to the point, Keeler’s technique could be used by others.
At the end of the 1930s Keeler privately surveyed police units using his polygraph technique. The thirteen respondents were East Cleveland, Toledo, Indianapolis, Kansas City, Buffalo, Honolulu, Madison, St. Louis, Cincinnati, and Wichita, plus the state police of Michigan, Indiana, and North Dakota. By their reckoning, of the total of nearly 9,000 subjects tested, about 97 percent had "voluntarily" agreed to take the exam, with only 1 percent refusing, and 2 percent confessing before even being strapped down. Of the one-third of subjects labeled "deceptive," an impressive average of 60 percent were persuaded to confess. Of the steadfast remainder, roughly half were later convicted and half had their cases dismissed. Contrariwise, of the two-thirds found to be not deceptive, only 0.3 percent went on to be found guilty at trial.
This survey, long buried in Keeler’s private files, may offer the best picture we will ever have of how the police deploy the polygraph when they think no outsider is watching. Not that we ought to take these statistics at face value. If anything, they show how the lie detector had become a self-fulfilling prophecy, a way for police and prosecutors to triage suspects: quickly resolving cases through confessions, marking out suspects to prosecute aggressively, and setting aside those to consider for release. The director of police in Kansas City admitted that "if he had to choose between the polygraph and the rest of the detective department, he would choose the polygraph." An officer with the Michigan state police estimated that his polygraph alone, by disposing of cases, had saved $25,000 in court costs in1938.
Of course, this efficiency came at a price. An examiner in Madison, Wisconsin, boasted that on being threatened with the machine four subjects had confessed to a crime—or five, he noted in the margin, if you counted the one who committed suicide. The police counted no more than 1 percent of these confessions as false. Americans like to think that we have left the age of inquisitorial justice behind, but the confession, for all its troubling ambiguities, remains our "queen of proof."
But the one dramatic exception among Keeler’s respondents belies the polygraphers’ confidence in the worth of these confessions. The operator with the Indiana state police achieved a confession rate of only 6 percent, one-tenth the rate of the others. Why the huge difference? The operator in Indiana was the only one trained not by Leonarde Keeler but by John Larson. Because suspects often made faulty admissions, the operator wrote to Keeler, "we rely on the Polygraph simply as an aid in investigation and not a means of bluffing an individual into a confession."
A noble sentiment—but Keeler’s bluffing better suited the sort of justice then becoming the American norm. The sad truth was that an ever-expanding caseload had long pressured both police and prosecutors to resolve cases short of trial—while judges winked. And it was not until the end of Prohibition that the distinctly American practice of "copping a plea" suddenly became scandalous. Data released at the time indicated that in Cook County in 1926, only 4 percent of those charged in preliminary felony hearings faced a jury trial, and 76 percent of those indicted for a felony had pleaded guilty to a lesser charge. Today, the subterranean process of plea bargaining still resolves some 90 percent of criminal convictions and is the predominant method of determining what counts as the "truth" in American criminal law. Such bargains, as one critic noted at the time, constitute false confessions or "legal fictions…, psychologically more akin to a game of poker than a process of justice." In this process, the lie detector plays a crucial role in determining which suspects prosecutors will pursue and which will be given less scrutiny. So the device remains central to the process of American justice even though it is banned from the courtroom. In the game of justice poker, the polygraph bluff plays a significant role.
This bluffing explains a central paradox of lie detection. Despite tremendous advances in physiology, the hardware of the polygraph has hardly changed since the 1920s. At unguarded moments, polygraphers have admitted as much. "The lie detector has not been improved since John Larson invented the thing in Berkeley," noted one expert in the 1950s. "It has been beautified. It has a better looking box and costs a lot more. Otherwise it is the same machine." Given the nature of the ruse, the interior workings of the machinery are almost beside the point.
Indeed, police examiners regularly wring confessions by putting suspects on sham devices with contemporary flash. In the 1930s a high school principal in Newark, New Jersey, got students to confess to a mock lie box; and a policeman in New York City used a towel and a ticking alarm clock to get a youth to admit filching $10 from his parents. For those interested in something with more action,
Popular Science
laid out "Two Simple Ways to Make a Lie Detector," with advice on how to amaze your friends with the card trick. In the 1980s cops were extracting confessions by putting a suspect’s hand on a photocopy machine filled with paper printed with the word "L
IE
!" In our era of cognitive science, cops have taken to placing the suspect’s head in a colander with wires attached. Naiveté of this sort among criminals may elicit a chuckle—it got a big laugh in the U.S. Supreme Court—but the joke is double-edged. Though some suspects may be duped into confessing to a mechanical placebo, cops and prosecutors are also locked in a self-fulfilling prophecy: deciding the fate of suspects on the basis of a doubtful test.
This is a general problem. Psychological studies suggest that when subjects are convinced that an apparatus can measure their true feelings—even when it is an empty box—they admit having socially unacceptable opinions. It is the strategy recommended in every interrogator’s handbook; to get a subject to confess, lower the stakes by convincing him that you already know what he has done and that you will not personally judge him. This is why the Roman Catholic Church grants confessors and penitents privacy and a degree of anonymity; and besides, parishioners are told, God already knows and forgives all your sins. Likewise, in confessing to the lie detector, the suspect is led to believe he is confessing not to a person, but to science, which knows and understands all. The judge and jury, of course, may not be so accommodating.
This is not to say that all—or even most—Americans have ever put full faith in the lie detector. Despite the best efforts of Keeler and his confederates, a whiff of hokum has always trailed after the device since its early days in Berkeley. But whatever a person’s degree of skepticism, there always remains a residual skepticism about skepticism—the sort of self-doubt thatP. T. Barnum knew how to exploit so well. There is always a lingering suspicion that the damn machine just might possibly work. Wasn’t it plausible, after all, that guilty knowledge made cheeks blush and hearts pound? Who wouldn’t feel anxious while being strapped into the chair? And then, when the operator guessed your card the first time out….
The one major technical innovation in the polygraph since the 1930s actually confirms the power of this ruse. In the 1990s new computer algorithms were developed that could analyze the subject’s physiological responses with mechanical neutrality. But because the algorithms might preclude operators from accusing subjects of lying (whatever the machine said), the nation’s top examiners at the Department of Defense Polygraph Institute report that most operators usually turn the computer off.
In sum, Keeler and his followers operated his lie detector according to the same logic as judicial torture. This explains why the police, despite their initial resistance, ultimately welcomed the technique. So not only did Keeler pack the lie detector into a "box" that almost anyone, even a minimally trained police officer, could operate; but because of the way he conceived of its operation, it actually enhanced the discretionary power of the examiner, who was less interested in the polygraph record per se than in using it to screen suspects, intimidate detainees, and extract confessions. In this way, Keeler spurred the replacement of the violent third-degree methods of the old-style municipal police with the psychological third-degree methods of the modern professional police.
From Larson’s vantage point, however, Keeler’s project was a betrayal. One year after their triumphant return from Appleton, when Keeler pushed again to introduce his lie detector into court, their split became public. In 1931, a case confirmed everything Larson despised about Keeler’s methods even as it conjured up demons that Keeler found hard to contain.
The case actually turned on an exoneration. In March 1931 Virgil Kirkland, a high school football star of Gary, Indiana, was tried a second time for the rape and murder of his "sweetheart," Arlene Draves. In a written confession, Kirkland had admitted that after getting Draves drunk at a party on wine laced with formaldehyde, he and four friends had driven her around town for hours and repeatedly "ravished" her in the backseat before they realized that she might be dead and deposited her at the house of a local physician. Kirkland’s defense was that she had died from a drunken fall, not a blow from Kirkland’s fist.
For months the case made headlines in Chicago, spiced by tidbits like the report of the medical expert who exhumed the victim’s body to perform a second autopsy and concluded (somehow) that Draves had not been a virgin at the time of her rape. When the first jury returned a verdict of first-degree murder, Judge Grant Crumpacker offered Kirkland a second trial, as he personally did not believe a blow from a fist constituted premeditated murder. On the first day of the new trial the defense attorney tried a new strategy: he announced in the presence of the jury that Kirkland had taken a lie detector test which had proved him "one hundred percent innocent."
This strategy was the handiwork of Dr. Orlando Scott, a surgeon in Chicago who had become a courtroom alienist and had twice been cleared of perjury charges for filing false medical reports. It was Scott who had led the exhumation and autopsy. It was Scott who arranged to have Keeler test Kirkland at midnight in his jail cell and saw to it that the results were reported in the local newspaper as the jury was being seated. And it was Scott who orchestrated the testimony of the members of Northwestern University’s crime lab in support of the lie detector, with Keeler expressing his willingness to test Kirkland in open court if that would help.