The Truth Machine

by Geoffrey C. Bunn

  Accuracy statistics were also privileged in the scientific literature. In his 1917 paper on the “Systolic Blood Pressure Symptoms of Deception,” Marston claimed that measuring blood pressure constituted “a practically infallible test of the consciousness of an attitude of deception.”31 In one experimental study he concluded that he had correctly judged 94.2 percent of the curves he had examined.32 Harold Burtt reported that the systolic blood pressure method could correctly detect experimental crimes in “91 per cent of the cases as compared with 73 per cent” using the breathing method.33 Giving no experimental details, Keeler reported a response time-word association study that detected deception in sixty-two percent of cases. However, blood pressure and respiration techniques improved accuracy to ninety-three percent.34 Fred Inbau maintained that experimental cases were accurate “approximately eighty-five per cent” of the time, adding that in criminal cases, full confessions had been obtained in “approximately seventy-five per cent of those in which the record indicated deception.”35 In his history of lie detection, Paul Trovillo reported that the Scientific Crime Detection Laboratory examined 2,171 subjects between 1935 and 1938. Of this number twelve mistakes in diagnosis had been verified, wrote Trovillo, who calculated that the errors were of the order of “five-hundredths of one per cent.” And even if the errors were ten times this number, “they would still be a relatively small proportion of the total.”36

  Not everyone let the statistics go unexamined. In 1939 Walter Summers criticized Marston’s work “because of its impressionistic character, so that the apparent statistical result is valueless as a critique of the accuracy of the procedure.” Having quoted accuracy claims by Larson and Keeler, Summers concluded that a procedure that started with an experimental validity of only eighty-five percent was “an extremely hazardous thing to employ in the investigation of the guilt or innocence of any person.” “Even the 75% efficiency obtained in the numerous criminal cases leaves a very great probability of error… . The 75% efficiency by no means tells us the entire story, for it fails to relate the number of instances in which deception was actually practiced in a manner which eluded the examiner and the instrument.” Despite his forceful criticism, Summers did not dispense with the accuracy statistic altogether. The preliminary results obtained with his own psychogalvanometer—an instrument rejected by both Marston and Larson—had “showed an efficiency of better than 99%.”37 “Caution Keeler whenever you see him to cut out his talk about the infallibility,” Larson asked Vollmer in 1931, “because I know it is not infallible. Tell him to never have it used so that men are discharged because of the interpretation of the record, or legal action of any sort taken.”38

  The status of invention and the power of numbers were not the only ways advocates of instrument gained legitimacy. Another one was the depiction of the machine as a “black box.” Expository articles often included a photograph of the enigmatic instrument, a depiction that explained yet mystified at the same time. Here was a gadget fabricated from reassuringly complex components, all of which were encased within a scientific-looking “black box.” By describing the instrument thus, however, the question what exactly it was fabricated from remained unanswered. The first explanation conveniently rendered the second superfluous: that the instrument looked “scientific” was a sufficient testimony of its credibility. What exactly was within the black box was rarely explained. It was useful to have the lie detector described as “mysterious,” or as the Saturday Evening Post put it, as a “curious engine.”39 Henry Pringle began his 1936 Reader’s Digest article “How ‘Good’ is Any Lie?” with an account of the experience of submitting to a lie detector test: “We sat in a small room at the Scientific Crime Detection Laboratory of Northwestern University in Chicago. On a table behind me rested a small, box-like machine. This was the Keeler Polygraph, popularly known as the ‘lie-detector,’ and I was being subjected to an examination on my truthfulness. Professor Leonarde Keeler of the Northwestern University law school, was conducting the test.” Having praised the abilities of the intimidating “Keeler Polygraph,” Pringle wryly observed that the “vast majority of defendants … are entirely confident that they can outwit the little black box.”40 Keeler described his machine in the same manner. In 1948, he told Vollmer that a fellow operator had just completed his six-week polygraph course. “He is now carrying one of our black boxes to the Orient,” he wrote, “where he will be stationed for some time.”41

  Reporting on a sensational “eleventh-hour” lie test that “sealed the doom” of a convicted murderer, the Literary Digest described how Keeler “put an odd looking black boxlike machine, about two feet square, on the table” in the jail cell.42 In 1936, the New York Times described how a school principal wrecked his “experiment in psychology”—a homemade lie detector—that was attracting media attention.43 “The destruction of the black box equipped with dials and electric bulbs closed the incident.” In 1937, Scientific American pictured an instrument replete with dials, switches, a graphical recording device, and an assortment of tubes.44 The components looked scientific and complex, but the technical language obscured the fact that they were quite straightforward. Describing “Two Simple Ways To Make A Lie Detector,” Popular Science Monthly hit the nail squarely on the head when it told readers that “since the device is so simple, it is advisable to conceal it in a box so as to hide the mechanism and give it as mysterious and complicated an appearance as possible.” Although the homemade machine was little more than a toy, it nevertheless worked “on the same principle as the famous lie-detecting instruments used by criminologists in obtaining confessions of guilt from law violators.”45

  The semblance of science served not only to legitimate but also to threaten. The black box was not only mysterious but also frightening. Although the lie detector’s apparent role was to replace the third degree, it never managed to lose its intimidating character. This feature of the test was recognized in 1929 when the Literary Digest described the process when a suspect faces “a new kind of third degree, a strange machine that neatly separates falsehood from truth in the story he tells.”46 The Science News Letter might have been somewhat cynical, but its point was well made: “Chief usefulness of the gadget is as an aid to the police in scaring an ignorant or superstitious person into making a confession of crime. An empty black-box, if it looks mysterious, would serve the same purpose—and has been used for it.”47 In fact, an “empty black-box” had already been used to induce confessions. In July 1931, the Philadelphia Police obtained a confession of guilt from a boy with a lie detector made up of stolen radio parts. During the test, which was a curious mixture of modern science and ancient magic, the young suspect was seated in front of the contraption and told that if he told a lie “it would be reflected in a color register behind his back… . Wires were placed on his arms and legs, the police told him, to register his ‘blood flow’ in the device. All lights save a small red one were extinguished. Each time the police thought the boy’s answers to questions were wrong they told him the ‘lie detector’ showed green. Then pepper seeds were put on his tongue. He complained that they burned. ‘Those pills always burn the tongue of a liar,’ he was told.” After the youth began to cry he made a statement implicating himself and five others in a number of recent robberies.48 The “lie box,” as one suspect described it, was extremely intimidating by virtue of its apparent infallibility.49

  The rhetoric of the machine’s scientificity was completed by an image that would come to achieve iconic status. As the graphical output of the black box, “the chart,” achieved immense persuasive power as the essential record of truth or lies. The first illustration of such a chart appears to have been published in Balmer and MacHarg’s Luther Trant (1910) story “The Man Higher Up.” In the story “Professor Schmalz” uses a plethysmograph and a pneumograph to detect a person’s disliking of caviar: “The instruments show that at the unpleasant taste you breathe less freely—not so deep. Your finger, as under strong sensation or emotion,
grows smaller, and your pulse beats more rapidly.”50 In 1932, The Review of Reviews provided what would become the standard image. Accompanying a photograph of Keeler performing a lie detector test were pictures of two short strips of graph paper. The top graph, showing a gently undulating curve, was captioned “THE TRUTH.” Below it, another graph described a violently fluctuating line and was captioned “THE LIE.”51 A similar “Lie and Truth” image was chosen by The Literary Digest to illustrate an article, “Detecting Liars.”52 “One subject made both records,” read the caption. “The upper is the ‘lie’ record, the lower one is when he told the truth.” In a later piece, “Catching Criminals with Lie-Detector,” the magazine printed a photograph of “The blood-pressure record of an embezzler who decided to go straight.” The upper section, “taken before he confessed, shows characteristic tension, caused by lying. The lower section, after full confession, shows steady, normal pressure.”53 Such comparison graphs were very popular.

  Sometimes only a single strip of graph paper was shown. In such cases, the graph would invariably undulate smoothly until interrupted by a disturbance that signified the lie. An arrow might point the reader to the crucial moment. The legend on Scientific Monthly’s illustration was “The arrow on each record indicates the peak of tension in the subject’s blood pressure curve—the point at which the lie was told.”54 “A black arrow points to the jagged peak,” a 1937 Literary Digest illustration was captioned, “depicting a lie which sent Joseph Rappaport to the chair for murder.”55 The lie occurs at an apparently simple and discrete moment, and that moment could be detected, and its parameters calculated. Readers were led to believe that the graph paper could almost speak for itself, so obvious was the appearance of the lie. According to the rhetoric of the image, the chart did not need a human operator because the machinery seemed to work so well as to not require one: the truth was plain for all to see. Images of the chart obscured the operator, whose role it was to scrutinize the chart in order to render it intelligible. Reader’s Digest’s reporter was glad that the questions of his mock examination had not been too embarrassing, “for the wig-wag lines on the polygraph’s recording roll of graph paper indicated all too clearly when I had lied.”56 Responsibility for determination of truth was transferred from human to machine: “The needles flickered unsteadily, indicating, Professor Keeler said later, that Rappaport was lying… . Professor Keeler turned to casual questioning. The needles graphed a steady course.”57 Because the physiological detection of deception was essentially an interpretive human enterprise, an edifice had to be constructed around the instrument to deflect criticisms of its subjective and perhaps arbitrary nature. Although human scrutiny was necessary to interpret the curves on the chart, discourse about the lie detector worked to hide this fact.

  Yet, paradoxically, the lie detector’s advocates were anxious to stress that their expertise was indispensable for the success of the venture. Such a tension was part of the larger problem of the presence of the polygraph operator. If a machine could detect a lie automatically, why was an expert required? The operator was potentially a mere technician, a cog in the machine. Yet he was also the expert, the sage, and, ideally, also the instrument’s inventor. The charismatic authority of the expert mediated this fine balance between these two opposing poles.

  Not everyone found the rhetoric persuasive, however. In 1939, A. A. Lewis criticized the myth of the autonomous black box in an article for the Scientific Monthly. “Diogenes is back again with his lantern,” he wrote. “But this time the lantern looks too much like a laboratory to be regarded as a trick of burlesque employed to incite an insurrection against dishonesty”: “These detective instruments, like an X-ray machine, may turn culprit or criminal inside out, as by zigzagging curves and dial readings his deeper bodily changes are made visible, but it still remains to interpret the picture. The suspect’s guilt or innocence is not spelled out in unmistakable letters. The fact that these deeper, organic reactions are involuntary and can’t be made to belie culpability like a face ‘kept straight,’ does not guarantee that they can be subject to no other source of causation except actual misconduct.” By calling the lie detector operator a “laboratory magician,” Lewis hoped he would undermine the claim made by advocates that the instrument had scientific credibility.58 In fact, attributions of magic actually supported the agenda of lie detector advocates.

  The scientific “chart” so easily became a numinous “scroll.” Sacred truths were being revealed. An illustration of a “typical graph” in Scientific American encapsulated the conjunction of science and magic: “This graph depicts the sudden rise in blood pressure at the point of attempted deception. The subject was handed ten well-shuffled playing cards, with instructions to choose a card and then lie about his choice. Respiration at top, blood pressure below. Notice where he said ‘No’ to the three of diamonds. He later admitted that the three of diamonds had been his choice.”59 Observers were led to believe that lie detector could perform card tricks. Although the ostensible purpose of the “stim test” (as it would become known) was to “obtain controls,” the trick also functioned as a remarkable demonstration of the machine’s abilities. “Control readings” were normally obtained through the use of innocuous or irrelevant questions. But the importance of the card trick lay in its power of intimidation.

  Consider, for example, the point made by Father Summers, inventor of the psycho-galvanometer lie detector, to a reporter for Forum and Century magazine. He was shown ten playing cards and asked to make a mental note of one of them, “keeping my selection a secret… . I chose the deuce of hearts; the cards were shuffled and then displayed to me one at a time. ‘Is this your card?’ asked Father Summers, as each was turned up. I steadfastly replied, ‘No,’ keeping my eyes on the galvanometer dial to see what happened. When the deuce of hearts appeared, I said, ‘No,’ as coolly and disinterestedly as possible, but the indicator shot up like a jack-in-the-box. After two repetitions of the test, I broke down and ‘confessed’ that I was lying about the deuce of hearts.” “So you see,” said Father Summers, “if one perspires over a little thing like a playing card, what would happen if a real crime were being concealed.”60 The trick was already a component of the orthodox examination by the early thirties. “So delicate is the apparatus,” reported the New York Times in 1931, “that a subject can be asked to select a card from a deck and will react at once when the correct card is picked up.”61 In 1937, Keeler used the technique on Joseph Rappaport, the convicted murderer whose eleventh-hour jail cell lie test sealed his fate.62 A photograph accompanying a Newsweek story on the lie detector showed Keeler revealing a playing card to a female subject strapped to the lie detector. “Professor Keeler’s card trick works nine times out of ten,” read the caption.63 The entire procedure was a classic piece of misdirection. What looked like the work of the instrument was actually a ruse engineered by the operator using a marked playing card. Keeler, a keen amateur magician, had devised the sleight of hand.64

  The lie detector was considered magic because it embodied scientific progress: it discovered the truth, promoted justice, and was humane. “The machine has now been used in 60,000 cases,” the Saturday Evening Post enthused, “and its uncanny power of penetrating guilty secrets has been thoroughly established.” The lie detector was awe-inspiring because it was thought that it could achieve its impressive results on its own, without human assistance: “Automatically controlled pens will record the slightest deviation from the truth.” The attribution of consciousness was irresistible: “When a lawbreaker denies his crime during a lie-detector examination, the pens become feverishly animated. A guilty man, seeing that the machine is practically photographing his soul, usually cuts short the examination by confessing.”65 What was most magical about the lie detector was its “uncanny power” of agency. It was thought that the machine, not the human operator, did the work. The Saturday Evening Post articles were peppered with personifications and attributions of agency: “The lie detector acts
as a mechanical conscience”; “the machine had solved some baffling mysteries”; “The detector has a peculiar genius for geography”; “It can read a nervous, excitable person like a book; it can read a tough, hard-boiled character like a book”; “It airs a scandal in a sorority house, stops students from cribbing, bank presidents from tapping the till, and releases a guiltless man condemned to a lifetime in jail.”

  The instrument was often described as possessing human body parts, such as “an accusing finger.” In its review of electronic devices used for “the detection and prevention of crime,” Radio News asserted that the field was “richly aided by the sharp eye and keen ear of electronic devices of one sort or another.”66 “Three moving fingers of the Keeler Polygraph record these changes,” observed Reader’s Digest.67 And if it possessed a body, then it also possessed a mind. “The vast majority of defendants,” claimed the magazine, “are entirely confident that they can outwit the little black box.” A subject experiencing a lie test “made vigorous denials, but the polygraph betrayed him.” The instrument habitually took on a persona in these narratives, such as the street-wise detective. Reader’s Digest introduced its piece with the caption “The ‘lie-detector’ at work solving crimes.”68 “To the lie detector goes the credit for ‘cracking’ this strange enigma,” remarked Scientific American about one particularly puzzling case, “and searching out the murderer from among nearly 50 suspects, as well as locating the murder weapon.” It was a “machine that knows all the answers,” promising to show how “Leonarde Keeler’s astounding invention tracks down murderers, unmasks the liars and the larcenous, and can tell you just how honest you are—and intend to be.”69 A. A. Lewis described the instrument as a “mind reading device.”70 Although the lie detector’s scientific credentials apparently contradicted its supernatural abilities, both qualities were acceptable within a broader context that encouraged the responses of awe and intimidation.