The automobile’s impacts on communities, governance, and law enforcement

The localism that Alexis de Tocqueville ([Reference Tocqueville1835] 2000) detailed in Democracy in America still predominated at the turn of the twentieth century. Governance in the United States was largely characterized by an anti-statist, laissez faire voluntarism centered on the principle of self-regulation. Enforcement was loosely implemented at the community level by a confluence of local organizations, professions, churches, and other civic groups (Balogh Reference Balogh2009). This highly localized, associational mode of governance relied upon the voluntaristic participation of citizen subjects in collective enterprise and the (self-)regulation of behavior. This resulted in the limited formal role of criminal justice apparatuses in the governing of everyday life (Seo Reference Seo2019, 13).

However, associationism and localism were upended by the automobile. Invented in 1896, it enabled people to traverse distances at previously unimaginable speeds, rendering increasingly porous the tightly territorialized communities in which voluntaristic associational governance operated. In the words of then secretary of commerce, Herbert Hoover (Reference Hoover1924): “[L]ocality has been annihilated, distance has been folded up into a pocket piece” (Seo Reference Seo2019, 5).Footnote ² With a thirty-five-fold increase in automobiles on the road between 1910 and 1925 (from five hundred thousand to almost eighteen million), no longer was self-regulation an effective mode of governing behavior as the automobile turned otherwise upstanding citizens into casual lawbreakers who endangered others. With the resulting growth in mobility and suburban communities (that separated work and home), as well as the regularity of inter-community encounters enabled by the automobile, informal voluntaristic modes of community control gave way to complex regulatory systems and formalized law enforcement to uphold the safety and welfare of an increasingly abstract, physically endangered, and geographically diverse “public.”

During this time, the scope of police work shifted from issuing warrants, managing vagrants and loiterers, and providing privately contracted security services for the wealthy to, ironically, policing the upper- and business-class people who constituted the bulk of early drivers.Footnote ³ In this respect, “the automobile served as the justification and the modality for scaling up police law enforcement as a method for governing all of American society, drivers and vagrants alike” (Seo Reference Seo2019, 12). The role of the state in managing the welfare of these newly interconnected and abstract publics was expressed by the government in 1925 in the landmark US Supreme Court case Carroll v. United States, when it argued that “as soon as the motor car appeared upon our highways it was recognized as a dangerous instrumentality, and thereupon the public asserted its right to declare upon what conditions such machines might use the public streets” (Seo Reference Seo2019, 128).Footnote ⁴

The streets and highways now constituted the shape (and rights) of this newfound “public,” over which law enforcement claimed jurisdiction. As Lochlann Jain (Reference Jain2004, 61) puts it, “[e]nveloped in a rhetoric of freedom, the automobile concentrates the most astonishing degree of hyper-regulation” because “the technology of automobility has defined public space in virtually every U.S. community.” With the automobile, “freedom meets regulation and a potential for individuation rubs uneasily against an actualized homogeneity” (61). In this early period of automobility, the government struggled to institutionalize particular modes of governance over the roads. The factors in this negotiation were the individual freedom afforded by the rapid proliferation of the automobile, the way in which it redefined both public space and the “public” that used it, and the ways in which law enforcement and the courts attempted to regulate this commons and both the automobiles and bodies that used them.

Popular demands to address the newfound dangers of traffic led to the creation and expansion of professionalized police forces. Suddenly, police encounters became a common feature of modern life for the “law-abiding citizen” (aka Everyman) whom they were charged with policing (Seo Reference Seo2019, 15). Officers tasked with policing the roads were now more plentiful than ever, with “the number of police officers as a percentage of the population nearly doubling” in the sixteen smallest states, alone, between 1910 and 1930 (7). Moreover, because most localities did not begin adopting ticketing systems until the 1940s and 1950s, and the use of a summons emerged only shortly before, “a violation of the traffic code was treated like any other crime,” such that “offending drivers were arrested, taken into custody, and brought before a magistrate, which initiated full trial procedures, unless the defendant pleaded guilty” (91–92). As a result, court dockets were overwhelmed by an influx of traffic arrests. This, in turn, exerted institutional pressure to streamline cases (Frost Reference Frost1950).

A culture of mistrust: police corruption, government overreach, and privacy

The profound shifts resulting from the rise of the automobile, and the concurrent need for traffic enforcement, occurred during the same period when alcohol prohibition visibly corrupted police officers. Law enforcement, especially prohibition agents, became notorious for their use of bribes, selective enforcement, and, sometimes, direct participation in the trafficking and distribution of bootlegged alcohol (Oliver Reference Oliver2018). Public distrust of the police in enforcing alcohol prohibition had been growing for decades before the Volstead Act became law since some states had already instituted prohibition (for example, Maine prohibited alcohol in the 1850s).Footnote ⁵ Wesley Oliver (Reference Oliver2018) shows how seminal court cases responding to and addressing police overreach and misconduct in enforcing alcohol prohibition laws—for example, the exclusionary rule—began to develop at the state level well before national prohibition. With the advent of national prohibition, the public reputation of law enforcement only worsened:

Police forces’ invasive and often warrantless search-and-seizure practices, which sometimes included beatings and evidence planting, occurred during the same historical moment that the US Supreme Court upheld law enforcement’s broad power to search vehicles without a warrant in Carroll v. United States (Oliver Reference Oliver2018; Seo Reference Seo2019).

Compounding these privacy concerns and citizens’ mistrust, the American domestic security apparatus underwent a historically unprecedented expansion during the Great War. The 1917 Espionage Act criminalized opposition to the war and became the government’s most powerful tool for suppressing radicals of all stripes by allowing for the interception of mail.Footnote ⁶ It was only the first of several laws that criminalized dissent and made life especially hard for immigrants suspected of being radicals. The Sedition Act, passed in May 1918, imposed “a fine of not more than $10,000 or imprisonment for not more than twenty years, or both” upon anyone disposed to “utter, print, write, or publish any disloyal, profane, scurrilous, or abusive language about the form of government of the United States.”Footnote ⁷ In May 1918, Congress passed the Alien Naturalization Act, by which “any alien who, at any time after entering the United States, is found to have been at the time of entry, or to have become thereafter, a member of any anarchist organization” could be deported.Footnote ⁸

Finally, the Palmer Raids of 1919, orchestrated by J. Edgar Hoover, led to the arrest and deportation of between five thousand and ten thousand radical “enemy aliens.” These raids and deportations boosted Hoover’s early career at the Bureau of Intelligence such that, by 1921, he had amassed files on 450,000 people, a staggering figure considering the population of the United States was only 108.5 million (Ackerman Reference Ackerman2007). The American Civil Liberties Union was founded in direct response to the excesses of the Palmer Raids.

Although the Palmer Raids quickly became unpopular,Footnote ⁹ most people were not personally affected. However, the passage of the Eighteenth Amendment that instituted alcohol prohibition from 1920 to 1932 (with some states taking longer to repeal) was an obvious sign of government overreach into citizens’ everyday lives. Now, Americans across the political spectrum were vulnerable to being spied upon by prohibition agents, who turned to wiretapping saloons—the constitutionality of which was established by the US Supreme Court in 1928 when it upheld the conviction of bootlegger Roy Olmstead.Footnote ¹⁰ This degradation of civil liberties was supported by many prohibitionists at the time, including William Jennings Bryan, who used the example of drunk driving to argue that prohibition did not infringe on personal liberties—rather, the infringement of liberties came from the drunk driver who plowed into the teetotaler (Okrent Reference Okrent2010, 283).

The result of this flurry of legislation, new surveillance technologies, rapid expansion of law enforcement, and crackdowns on dissent was a period characterized by intense concerns about privacy and the coercive reach of the government into the lives of everyday Americans. It was in this culture of fear of government overreach that solutions to drunk driving would be negotiated.

The problems of drunk driving in the United States

By the mid-1930s, finding a way to keep drunk drivers off the road seemed an urgent project to government officials and researchers. The United States had a per capita rate of “autocides” (that is, driving fatalities) that was more than twice that of Great Britain. National Safety Council data indicate that “between 1913 and 1932, deaths from car accidents increased 500 percent; during the same period, the death rate for all other accidents dropped 42 percent” (Seo Reference Seo2019, 22). In 1932, alone, there was an estimated total of 29,500 deaths and 850,000 non-fatal injuries from automobile accidents in the United States, with rough estimates of the economic loss from these injuries and deaths (including wage loss, overhead insurance, and medical costs, but not including property damage) of six hundred million dollars (Cameron Reference Cameron1934).

It is difficult to find accurate estimates of drunk-driving accidents in this period.Footnote ¹¹ However, a three-year study conducted by Northwestern University researchers, published in the Journal of the American Medical Association in 1938, concluded that “alcohol is a ‘major cause’ of automobile accidents” involving personal injuries in the United States (Holcomb Reference Holcomb1938; see also St. Louis Post-Dispatch 1938, 64). As one judge noted during a drunkometer demonstration in Los Angeles, drunk driving was estimated to be a factor in at least 60 percent of car accidents.

Troublingly, many drivers were unaware that they were drunk or that their alcohol intake had any effect on their driving. As the judge said, the dangerous driver “is the man who has had a few drinks and is absolutely and thoroughly convinced of his own sobriety and of his unequaled driving perfection—the type who insists that he is a better driver when he has had a few drinks than when he is completely sober” (Los Angeles Times 1938, 3). Before the 1940s, drunk-driving statutes were often vague and highly varied. Prosecution relied solely upon the subjective judgment and testimony of witnesses, defendants, and publicly mistrusted police officers, which often failed to convict in courts (Lerner Reference Lerner2011). If the drunk driver could not be trusted to know their drunkenness, who (or, in this case, what) could be trusted to know the truth for reaching a verdict?

In short, the policing of drunk driving had a problem with the production and deployment of “truth,” or what Michel Foucault (Reference Foucault2008, Reference Foucault2012) termed “regimes of veridiction.” How could policymakers determine what levels of intoxication constitute a public threat? How could police officers—let alone, drivers—know the truth of a suspect’s level of intoxication? And how could juridical procedures deploy the truth of one’s intoxication in determining guilt and punishment? Solutions to drunk driving’s veridiction problems—and the public’s demands to address the dangers of drunk driving—would need to be worked out in a context characterized by widespread concerns about privacy, police corruption and mistrust (especially with respect to alcohol), and the invasive reach of the state in the everyday lives of Americans. How were these tensions resolved? The solution was to come in the peculiar form of chemical-filled glass tubes and rubber balloons.

The birth of the drunkometer

In 1931, Dr. Rolla N. Harger, a member of the biochemistry and toxicology faculty at the Indiana University School of Medicine, announced a revolutionary new intervention to make roads safer by scientifically determining the drunkenness of drivers. His new invention, which was quickly dubbed “the Drunkometer” by reporters, was comprised of a “portable arrangement of glass and rubber tubes, one containing a pink fluid” that comprised sulphuric acid, distilled water, potassium permanganate, and other chemicals.

It worked by first having a subject blow into a rubber balloon, then releasing that breath into a tube and bubbling it through the pink chemical solution. The resulting degree of color change in the pink solution would determine a blood alcohol level of 0.15 or higher. It would also, presumably, end guesswork and false prosecution on the part of the police, who until this point had relied on decidedly less scientific measures, like a subject’s ability to accurately recite tongue-twisters or walk a line. The original drunkometer prototype fit into “a case the size of an overnight bag,” and the police were required to keep a supply of balloons for suspects to blow into (Indianapolis News 1934b, 32).

In 1934, Harger had a working prototype, which he kept refining and using in public demonstrations, and, by 1937, he saw its first adoption by a police department in Indiana. While drunk-driving charges had once come down to a defendant’s ability to persuasively argue their own sobriety, now the drunkometer provided an instant metric, followed by a backup blood test (Aurora Journal 1937, 1; Wichita Falls Record News 1937, 9). As the issue of drunk driving gained national prominence, the drunkometer won acceptance, albeit piecemeal, gradual, and with much controversy.

The drunkometer was promoted by its maker as a big step forward for law enforcement, and it quickly became an important tool in the fight against drunk driving. In addition to identifying criminal impairment, many convictions also came about when, faced with the drunkometer results, drivers confessed. In the words of one journalist, “[t]he chemical reaction is even more reliable than the record of the widely-heralded lie detector, which has brought about so many confessions of guilt without actually being recognized by the courts as a source of competent testimony” (Marion Leader-Tribune 1938, 4). By December 1937, five months into their initial pilot program, Indiana police estimated that the drunkometer inspired 80 percent of drunk drivers to plead guilty.

The dramatic display of color-changing chemicals constituted a powerful form of “alethurgy,” a term Foucault (Reference Foucault2012, 7) used to describe ritual “manifestation[s] of truth as the set of possible verbal or nonverbal procedures by which one brings to light what is laid down as true as opposed to false, hidden, inexpressible, unforeseeable, or forgotten.”Footnote ¹² The procedures by which the “truth” of intoxication was brought to light—for officers and suspected drivers alike—not only provided colorfully clear evidence for law enforcement but also inspired many drivers to confess. But, as we document, the working out of juridical procedures to deploy this “truth” in the prosecution of drunk drivers would be hotly contested.

By 1938, Harger refined his machine to measure the presence of as little as an ounce of consumed liquor, lowering its cost to $87.50 and causing law enforcement to hail it as an inexpensive life-saving tool.Footnote ¹³ By then, the Woman’s Christian Temperance Union was teaming up with insurance organizations, police, and safety departments to sponsor a “transcontinental safety tour,” taking the road in a trailer filled with educational exhibitions—including the drunkometer—focused on the dangers posed by drunk drivers (Knoxville Journal 1938, 22; St. Louis Post-Dispatch 1938; Wisconsin State Journal 1938, 5; Daily O’Collegian 1940, 1). Harger began offering a drunkometer operators’ school for law enforcement agents at the Indiana School of Medicine, since a big obstacle to its adoption was a lack of trained operators. By 1944, demand for the course grew so much that it even enrolled a police judge from Ecuador and a prison official from Colombia (Franklin Evening Star 1943, 1; Waterloo Press 1943, 1; Star Press 1944, 7).

In 1946, Harger perfected a new drunkometer model that could be installed in patrol cars. This enabled more accurate readings by testing drivers at the scene instead of later at a police precinct. The drunkometer was pioneered by the Indianapolis police department. By the end of the year, other departments across the United States had adopted it, and many, like the Kansas City police, were using it in connection with another innovation: the blockade.Footnote ¹⁴

The next improvement came in 1947 when police could use the so-called “feather test” to measure the blood alcohol level of drivers who refused to blow into a balloon. On the premise that a person’s breath became “common property” once it left the body, police used a suction to capture a subject’s breath from the air once it had left the body. With its increasing adoption by law enforcement agencies around the country, alongside its newfound portability and adaptability to innovative policing tactics, the greatest challenges to the drunkometer lay in the courts of law and public opinion.

Determining measures of intoxication

The first obstacle to the drunkometer’s adoption was determining a scientific measure of intoxication and criminal impairment that could be legislated and applied to all suspect drivers. In the early twentieth century in the United States, scientific knowledge about intoxication and its effects was fairly limited. Drunk-driving laws and legal definitions of intoxication varied locally and were notably vague, which made prosecution difficult.

They were also particularly liberal in comparison to European laws. For instance, while two Michigan statutes enacted in 1925 made it an offense to operate a motor vehicle “upon any public highway of this state while under the influence of intoxicating liquor,” it did not define drunkenness (Thornton Reference Thornton1928). With the early assumption by US safety officials that six to eight ounces of liquor qualified as an unsafe amount for driving, the United States diverged sharply from its British counterparts who defined driving impairment with much smaller amounts of alcohol. For comparison, a 1935 article written by the British Medical Association (1935) asserted that as little as two and a half ounces of whisky caused a 13 percent slowdown in a driver’s reaction time.

The lack of a scientific measure of, and method for testing for, alcohol-induced impairment concerned the Department of Justice. The research division chief at the Bureau of Prohibition admitted in a July 22, 1931, letter to the renowned Swedish chemist, Eric Widmark, that “[s]uch tests [of drunkenness] in our various states vary from smelling the offenders [sic] breath to making him walk a chalk line, but no scientific test apparently is applied.” The letter continued: “We are particularly anxious to know what the alcoholic content of the blood must be before a person can be described as being under the influence of alcohol. Some median line must have been established on one side of which an offender is not under the influence of liquor and on the other side of which he may be said to be intoxicated. Just what that line is we should like to know” (quoted in Andreasson and Jones Reference Andreasson and Jones1996, 207).

Sweden was the epicenter of intoxication-related research in the 1920s and 1930s, led by Widmark’s lab. By the end of the 1930s, federally funded research in the United States, sponsored by the Works Progress administration, also began bringing new and more accurate information about the effects of alcohol to broader audiences (see, for example, Science News 1938). Though researchers still quibbled over whether a breath test could be as accurate as Widmark’s blood test, a general consensus began to emerge around the drunkometer’s value and the need for tighter controls on drinking and driving, beginning with clearly defined metrics of intoxication (Science News 1938, 205; Topsøe-Jensen Reference Topsøe-Jensen1941).

Legislating and policing drunkenness

In addition to inventing the drunkometer, giving exhibitions of the device across the United States, and providing expert testimony in drunk-driving trials, Harger was tapped in 1938 as one of five members of a joint subcommittee of the National Safety Council and the American Medical Association tasked with drafting a “model act to legalize the use of evidence from chemical tests of intoxication and to set limits of body alcohol concentration for motorists” (National Safety Council 2004, 8). The committee provided the following suggestions to states regarding blood alcohol levels that eventually constituted the Chemical Tests Section of the Uniform Vehicle Code adopted in 1944:

A growing body of research, led by the National Safety Council, provided the institutional knowledge base for the scientific rationalization of intoxication measures. Drunkenness was transforming from a subjective state into an objective biochemical condition that could be measured by the drunkometer and readily adapted by legislatures. A marriage between the scientific and juridical aspects of drunk driving was solidified in 1944 when the American Bar Association and the American Medical Association accepted the drunkometer as reliable.

By 1945, four states—Indiana, Maine, Oregon, and New York—had incorporated blood alcohol standards into state laws (Dayton Herald 1945, 16). By 1952, sixteen states had specific chemical test legislation, mandating a specific blood alcohol level beyond which a driver was considered intoxicated. That year, 369 cities in forty-two states used chemical tests for blood alcohol levels, representing 27 percent of all cities with populations over ten thousand (Seliger Reference Seliger1953). How, then, did this scientific truth about intoxication, in the name of public safety, play out in criminal cases? If the drunkometer was to provide a techno-scientific solution to the subjective discretion and testimony of publicly mistrusted officers, then its ability to accurately and reliably measure intoxication in a court-admissible form needed to be established.

As we document below, in addition to questions of the drunkometer’s reliability, courts also wrestled with the validity and evidentiary value of homogenized measures of alcohol-induced impairment. Can drunkenness be reliably measured, or universally applied to all imbibers, for determining a verdict? These questions would be the first issues taken up in the courts of law and public opinion across the United States, often with dramatic flair.

Validity, reliability, and admissibility of drunkometer evidence

Contestations over the validity, reliability, and admissibility of drunkometer evidence played out in both the courts of law and of public opinion. The former were brought forth by defendants as well as judges and juries. A key question was not only if the drunkometer’s results were valid and reliable but also if drunkometer results and/or officer testimony were the most appropriate sources for evidencing intoxication. Considering that the same amount of alcohol differently affects each driver’s abilities, should the veridiction of intoxication be derived from subjective testimony or objective test results (and homogeneous, universal measures)?

As we document, confusion over the admissibility and constitutionality of evidence, as well as the interpretation of results, resulted in a varied and fractured drunk-driving legal landscape in the 1930s and 1940s. In attempts to address this confusion, public demonstrations—inside and outside the courtroom—were used to convince the public, lawmakers, judges, and district attorneys of the device’s value. Ultimately, increasing public outrage over drunk driving superseded concerns about the drunkometer’s ability to reliably produce the relevant scientific “truth.”

The drunkometer’s first legal use in an August 1937 Indianapolis drunk-driving municipal court trial of a married couple provided mixed results. When the defense attorney, who told the judge that he had had three cognacs and a beer during lunch, offered himself as a subject for the drunkometer, results indicated that he was sober. As an expert witness, Harger explained that a person’s body burns two ounces of alcohol every three hours, so by the late afternoon those four drinks would barely register. Nonetheless, the judge dismissed the charge of driving under the influence yet fined the husband and wife each five dollars on the charge of drunkenness (Indianapolis Star 1937a, 5; 1937b).

In addition to the reliability and validity of drunkometer results, judges also had to weigh their evidentiary value relative to the established precedent of officers’ assessment and testimony. For example, a judge in South Bend, Indiana, acquitted a man with a BAC of 0.187 percent after his lawyer argued that the drunkometer in question may not have been in good repair at the time and could have delivered a faulty reading. The defendant had admitted to drinking two beers and a “small drink of wine” before driving fifty miles per hour in a twenty miles per hour zone. But, as the judge pointed out, the test could show different results with different individuals such that one person might seem intoxicated with a BAC of 0.05 percent, whereas another person with a 0.15 percent reading might not actually be intoxicated. Ultimately, the district attorney agreed with the judge that the “best evidence of intoxication is testimony whether the person charged with that offense staggered and whether there was the odor of liquor on his breath” (South Bend Tribune 1943, 13).

In a similar case in Indianapolis—the city most identified with the drunkometer—a municipal court judge dismissed a manslaughter charge against a man with a 0.173 percent BAC drunkometer result, who had killed a beloved organist. This decision was based on police testimony that he appeared sober and that the drunkometer required a skilled operator. Notably, Harger had testified in this case that he himself had trained a police lieutenant to operate the device (Indianapolis News 1946, 39).

Jurists and judges continued to raise questions about the device’s reliability and universal applicability, with one judge from Salt Lake City pointing out that people metabolized the same amount of alcohol very differently (Evansville Post 1939; Indianapolis News 1939, 6; Salt Lake Tribune 1940). Moreover, relying solely upon drunkometer evidence also raised procedural concerns. For example, in Palo Alto, a judge rejected it on the grounds that “a machine, or a book of instruction for one, cannot be cross examined. I think very few courts would convict upon the evidence of a mere machine” (Toronto Star 1946, 17; Peninsula Times Tribune 1947). Others were similarly suspicious, like an Indianapolis judge who dismissed drunk-driving charges against a man whose drunkometer reading indicated that he was drunk at the time of his car accident, saying he “was dubious as to their actual worth” (Indianapolis Star 1938b, 1). Methodological issues also occasionally cropped up, which only compounded these concerns, as when a driver from Sun Prairie, Wisconsin—who overturned his car, injured a passenger, and registered 0.169 percent BAC—was released without charges after lab workers advised that the initial drunkometer test failed to register a number (Wisconsin State Journal 1947, 9). In other words, the drunkometer did not always immediately map onto established juridical practice.

In addition to weighing constitutional and jurisprudential concerns, judges also considered a range of legal strategies that defendants used to challenge the drunkometer. Though these strategies varied in content, they often centered on challenging either the validity and reliability or the constitutionality of the drunkometer (the latter will be addressed in the following section), with varying success. For example, one particularly creative defense lawyer successfully floated the theory that his client’s drunkometer reading of 0.159 percent BAC could have been caused by gargling with whisky right before the test (Evening Republican 1943, 1). In a 1941 case in Salt Lake City where a city physician had administered a drunkometer test to a driving under the influence suspect that found him to be intoxicated, the defendant was acquitted after a municipal court judge refused to admit the physician’s testimony because “the prosecution was unable to lay the proper foundation and unable to show that the device is an accurate scientific instrument capable of doing what it claimed for it” (Freeport Journal-Standard 1941, 14).

Conversely, in a Dayton, Ohio, case, a driver claimed to have consumed only one beer when he was arrested after his car plowed into a traffic beacon. Arresting officers testified that he staggered and swayed, and his BAC measured 0.23 percent. The judge “flatly repudiated” the defendant’s claims, along with similar claims made by three other men. All four were convicted, but the judge delayed sentencing in case the defendants’ attorney wanted to file an appeal, suggesting that, even a decade after the device’s adoption, there were still legal uncertainties surrounding it (Dayton Herald 1947b, 10).

Confusion also remained about how much weight should be given to drunkometer evidence and how to interpret its findings using legislative definitions of intoxication. Juries would occasionally reject drunkometer evidence even when the physical evidence suggested drunkenness, as in the case of a man who plowed into a parked police car and blew 0.137 percent. In the end, the judge dismissed his previous guilty verdict on the grounds that the prosecution had not adequately proved that he was intoxicated since only a result of 0.15 was considered legally intoxicated (Indianapolis Star 1943, 20).

Even after a new traffic code in Indiana took effect that legalized drunkometer use, some juries refused to convict on drunkometer evidence alone. In one case, the jury refused to convict despite the drunkometer tests showing a 0.208 percent blood alcohol level. In another, a judge in an Indiana municipal court refused to accept drunkometer evidence because of “public confusion” surrounding the tests and interpretation of the results (Chicago Tribune 1939, 11; Daily Herald (Jasper) 1950, 4).

Even though the test was new and its results fairly easy to understand, public skepticism about both its value and reliability was significant. In the court of public opinion, troves of newspaper columns comically questioned the usefulness of a machine to determine what to most people seemed obvious. For example, “[a] great many citizens know when they are drunk, without asking a ‘drunkometer’ to prove it,” wrote one editorialist, while also expressing the thought that “some bright young man” would be inspired to “add an attachment that will stop the car at that stage and tuck him into bed.” Another quipped that “there are other tests of sobriety which have been recognized long before there were such wise men as biochemists or toxicologists. One who makes love to a lamp post at 2 a.m. is drunk, regardless of what a drunkometer might say” (Reidsville Review 1931, 3; Indianapolis News 1934a; Owensboro Messenger 1934, 4). Other editorialists condemned its adoption for reasons of expense.Footnote ¹⁵

To address the skepticism of the public, courts, and lawmakers, dramatic public demonstrations often served to increase (and, occasionally, undermine) confidence in the drunkometer—many of which were sponsored by the Woman’s Christian Temperance Union (WCTU). Most importantly, public demonstrations across the country hammered home the idea that an imbiber may not be the best judge of their own inebriation—namely, that the intoxicated subject, like Barbee Hook, could not be relied upon to know the “truth” of their impairment. In a public demonstration for a Minneapolis police precinct, a test subject pronounced himself “sober as a judge” after downing ten ounces of straight whiskey in two gulps. His drunkometer reading showed otherwise, and the tester used this as evidence that the value of the drunkometer lay in its objectivity, in contrast to the unreliability of the drinker (Minneapolis Daily Tribune 1947, 1). Conversely, in a similar demonstration in Rock Island, Illinois, a police department gave three policemen variously five ounces, eight ounces, and sixteen ounces of one-hundred-proof whiskey over a two-hour period. The drunkometer registered the alcohol levels of the latter two at exactly the same level, a result explained by the tester as one driver weighing 275 pounds and the other 172 pounds (The Dispatch 1944, 10).

While these examples show how public demonstrations of the drunkometer varied in their audiences and outcomes, they never failed to garner media attention. Nearly half of the newspaper articles in the 1930s and 1940s resulting from our search covered local demonstrations of, and police trainings for, the drunkometer. On the one hand, national headlines were made when the machine misperformed, such as before an Oklahoma City legislative committee when it failed to detect inebriation in someone who had just downed four drinks (Arizona Republic 1937, 18; Escanaba Daily Press 1937, 9; Indianapolis News 1938b, 6; Indianapolis Star 1938a, 19). On the other hand, demonstrations also served the purpose of introducing and familiarizing public officials with the drunkometer, as when fifty judges and district attorney’s aides in Los Angeles watched an unidentified man drink measured amounts of alcohol at a safety council meeting in order to better understand drunkometer readings (Wilmington Daily Press Journal 1944). Perhaps most important of all, the public demonstrations shifted public understanding about how intoxication could be most effectively known. Despite the drunkometer’s occasional failures, the imbiber emerged as the least trustworthy source for discerning their intoxication.

Alongside public demonstrations, newspapers increasingly helped shape public opinion about the drunkometer and bolstered its perceived value in the domain of justice by frequently highlighting cases of innocent drivers cleared by drunkometer results.Footnote ¹⁶ An important aspect of the drunkometer’s public appeal was that it “safeguarded the innocent” by providing clear evidence that a sober subject was not drunk but suffered from one of the “65 pathological afflictions including a bump on the head that can make you appear drunk. Insulin shock even makes you smell like it” (Wing Reference Wing1946, 4). This eventually became one of its chief selling points: “A man with a skull fracture, a person with high blood pressure, a diabetic, a compressed air worker—yes, even a man with a bad sunburn—could be, to all outward appearances, drunk as a mash-loving billy goat” (Miami News 1946, 21). As this journalist concluded, “[C]ertainty: The drunkometer will stop imprisonment of sick people” (21).

This was demonstrated in the positive coverage of an Indianapolis case in which police arrested a man involved in an automobile accident, suspecting him of drunkenness. He was released when his drunkometer test came back negative for alcohol, and the police concluded that his erratic behavior was caused by illness (Journal-Gazette and Times-Courier 1938, 6). Between the drunkometer’s inception in the 1930s and its increasing use in the 1940s, newspaper columnists went from roasting the drunkometer to hailing it as a tool that could exonerate the innocent from the punitive reach of misguided police officers.

Despite the occasional embarrassing challenges to the drunkometer, popular demands for punishing drunk drivers proved a formidable force—for which the drunkometer offered a simple solution to a complex problem. For example, although the Ohio state legislature had recently failed to pass a law to set a definite standard for blood alcohol levels and enable courts to rely solely on drunkometer evidence, increased prosecution of drunk drivers still proved popular. A poll conducted by the Dayton Daily News (1947, 19) in the wake of four prosecutions found that “[t]here was very little sympathy shown for the drunken driver by Daytonians.” Three out of ten of them called for jail sentences, and an equal number favored jail sentences coupled with fines. Some respondents pointed to the postwar population boom as an additional reason to come down hard on these drivers, one telling the pollster that, “[w]ith the community so over-populated these days, the drivers have a civic responsibility and anyone who breaches this bond by driving when he’s drunk should suffer the extreme penalty” (19). Sober driving was viewed by respondents as an important part of the new social contract in the modern era of automobility (Dayton Herald 1947a, 7; 1947b, 10).

Prosecutors were also eager for more courts to accept prosecutions based on drunkometer evidence, which had already changed the drunk-driving landscape of some municipalities. For example, in Dayton, Ohio, the first month of the drunkometer’s use in 1947 saw an increase in drunk-driving arrests from four to forty-three. Soon, local newspapers were regularly publishing the names, blood alcohol levels, and punishments of drunk drivers. Gone were the days when a judge might fine a drunk driver five dollars. Now fines, often one hundred dollars or more (equivalent to over fourteen hundred in 2024 dollars), were combined with public shaming to great effect (Dayton Herald 1947a, 14).

Pressures by both prosecutors and the public to crack down on drunk driving—and protect the innocent from false accusation—ultimately overrode concerns about accuracy and validity. This emerged alongside increasing acceptance of drunkometer results as not only admissible but also sufficient scientific evidence for prosecuting (and streamlining) drunk-driving cases. By 1945, the state supreme courts in Iowa, Ohio, Wisconsin, Massachusetts, and Indiana affirmed the reliability of the test. Although academic challenges to the scientific and methodological basis of the drunkometer occasionally cropped up—which Harger repeatedly refuted (see, for example, Harger Reference Harger1944, Reference Harger1949; Burgee Reference Burgee1957)—questions regarding the constitutionality of extracting biochemical evidence from the body still needed to be resolved.

Limits of constitutional protections for the biochemical body

Even if concerns about the reliability of results could be overlooked, the drunkometer presented fundamental questions about the limits of constitutional protections against unreasonable search and seizure (Fourth Amendment) and self-incrimination (Fifth Amendment). Were the biochemical excretions of the body—in this case, breath—within the legitimate jurisdiction of law enforcement? If something as intimate as one’s breath was subject to seizure without consent, what would even constitute an “unreasonable search and seizure” against which the Fourth Amendment was written to protect? And if the Fifth Amendment was supposed to protect against coercion to provide self-incriminatory testimony, at what point should the line between breath and speech be drawn?

The 1925 Carroll case established the constitutional basis for broad police powers to search vehicles based on “reasonable suspicion” of violating a law, including those who appeared intoxicated. Yet a notable number of judges and prosecutors across the country initially refused drunkometer evidence. Their common ground: finding its use coercive and, therefore, unconstitutional.

Illinois, which did not establish a legal BAC limit of 0.15 percent until 1956, offers an illustrative example (Giannoulias Reference Giannoulias2023). In a January 1940 formal opinion, the state’s attorney general, John Cassidy, argued that, unless voluntarily submitted, compulsory “drunk tests” violated Fourth and Fifth Amendment protections, declaring that “[c]onstitutional guarantees against unreasonable search and compulsory self-incrimination must not be weakened by imperceptible practice of courts or by well intentioned but mistakenly over zealous executive officers.” Despite conceding that “the law should keep pace with scientific developments in the field of crime detection and admit as competent evidence the results of recognized and well established scientific tests,” Cassidy argued that “whether a person is drunk or sober is a fact open to the observation of all and its comprehension requires no particular scientific knowledge or training” (Decatur Herald 1940, 1). In other words, not only does the drunkometer violate constitutional protections against search and self-incrimination, techno-scientific evidence of intoxication is unnecessary—drunkenness is a fact res ipsa loquitur.

A year and a half later, the state passed a law allowing drunkometer results as admissible evidence but only for proving sobriety, not drunkenness, thereby preserving the prosecutorial boundaries between the courts and the suspect’s body (Chicago Tribune 1931, 2).Footnote ¹⁷ Later that month, the governor vetoed a bill that would have compelled courts to admit as evidence “the results of chemical analysis of body substances” in drunk driving cases after another state attorney general, George Barrett, again ruled it unconstitutional (Daily Independent 1941, 1). Then, in 1943, State Senator William Knox again introduced a bill—endorsed by the Chicago City Club’s public health committee—that would allow drunkometer results as evidence of intoxication in drunk-driving cases, which eventually was defeated in the state’s House of Representatives in 1945 (Chicago Tribune 1921, 21; Daily Journal-Gazette and Commercial-Star 1943, 8; Belleville News-Democrat 1945, 6).

Regardless, cities like Evanston and Rock Island purchased drunkometers, trained their officers, and regularly used them to police drunk-driving suspects, with local courts frequently accepting their findings (so long as suspects consented to the tests). For the precedent at that time, physician testimony in court cases was not considered “qualified to give a final and absolute opinion on drunkenness,” and arresting officers were deemed most qualified to determine intoxication based on “how a man talks, how he walks and how his breath smells” (Daily Herald (Chicago) 1951, 46). With this precedent, the limits of consent left judges with little choice but to acquit drivers of drunk-driving charges when they refused drunkometer tests, as was the case when a judge in nearby St. Louis ruled that “[a]nyone who had enough sense to refuse to take the test must have been sober” (Belleville News-Democrat 1950, 5).

It was not until 1951 that an Illinois Appellate Court ruled that drunkometer results were admissible evidence. The judges rejected the defendant’s argument that, after colliding with another vehicle, he was pressured into taking the test and was not properly advised of his right to refuse. With a 0.31 percent BAC, and his intoxication corroborated by officer testimony, the judges upheld his guilty verdict for driving while intoxicated (Chicago Tribune 1936, 36).

Immediately after the ruling, police departments in Chicago and across the state began purchasing and training their officers to use the drunkometer (Chicago Tribune 1951, 6). As one columnist in southwestern Illinois argued, it was now everyone’s civic obligation to demand that their local governments purchase and begin using drunkometers to criminally charge what he described as the great “menace to safety on streets and highways” who were responsible for “nearly half of the traffic fatalities and injuries at night and on weekends.” Because “the word of the arresting officer that the driver was drunk usually is deemed insufficient evidence,” the drunkometer provided a solution by leaving “no doubt about it, and the evidence stands up in court.” The stakes for the public, according to the columnist, were grave: “Either you choose to continue in the ineffectual rut that exposes yourself to irresponsible drunks running automobiles, or you’re for wiping out the menace and are willing to take the steps needed to do that” (Belleville News-Democrat 1952, 4).

As Illinois Attorney General Cassidy alluded to in 1940, a key question was at what point does one’s breath (or blood or urine) become public domain, if at all. Ever zealous to defend the drunkometer, Harger responded to a Kansas City municipal judge’s ruling that a man’s breath was his personal property by contending that “after breath leaves the body it is public property and that police have a legal right to sample it” (Kansas City Star 1939). Although Fourth Amendment battles over bodily fluids in both federal and state supreme courts initially resulted in a mixed legal landscape during this period, by the early 1950s, twenty-three states had taken legislative action to uphold the use of biochemical tests of intoxication (Anonymous 1957).

Resolving this debate about who “owned”—or in what domain resided—a body’s biochemical excretions would not only shape the constitutionality of the drunkometer but also that of the urinary, blood, hair, saliva, and other drug tests that would explode in popularity in the final quarter of the twentieth century. This, in turn, fundamentally reconstituted the relationship between authoritative institutions and the citizens they govern. In that respect, Harger’s assertion proved prescient.

Similar questions about Fifth Amendment protections against self-incrimination were challenged and resolved in several high-profile court cases. The Indiana Supreme Court’s 1938 ruling on Florence Davis’s case (discussed in the introduction) rejected Fifth Amendment arguments about coercion and self-incrimination with the drunkometer (Indianapolis News 1938c, 1; Indianapolis Star 1938b, 1). Then, in a 1947 case, the Indiana Supreme Court ruled that drunkometer evidence was admissible in court since the defendant, who agreed to the test, was appealing his drunk-driving verdict on the grounds that the drunkometer’s use violated his constitutional guarantee against being compelled to testify against himself. Even if reaching a verdict in this specific case was easy, the ruling had lasting consequences: “A former attorney general advised the state police department that a ‘reasonable’ amount of force is not improper, and in a local case a special judge decided that after a man’s breath leaves his body it becomes public property” (Indianapolis News 1943, 6).

By 1941, the Iowa Supreme Court disagreed with the very premise that breath could be considered testimony. Their ruling made clear that “no person gives testimony against himself unless he is under oath” (George Reference George1941, 1). And in a Detroit traffic court in 1945, where the attorney for a man convicted of injuring two in a drunk-driving accident claimed that “his client was in no condition to give consent to the test”—in other words, he was too drunk to consent—the judge ruled that being administered a drunkometer test did not violate consent (Detroit Free Press 1945, 11).

That same year, the National Safety Council’s Committee on Alcohol and Other Drugs released a report, based on a review of legal studies, that concluded: “Most legal authorities agreed that the prohibition of ‘self-incrimination’ referred to testimony of the defendant and not to the right of enforcement officials to examine the body of the defendant, including examination of breath and body fluids” (National Safety Council 2004, 11). This was followed by a 1952 report from the National Safety Council’s Subcommittee on Tests for Intoxication that approved the following provision on “informed consent” for state laws in its Uniform Code (which was immediately adopted by several states, including New York in 1953):

Though the adoption of these recommendations by states were piecemeal, consent to search their bodily fluids would eventually be given by all drivers who obtained a driver’s license. The biochemical body could now be made to tell the truth of its intoxication on behalf of the suspect subject who had already waived any protections against the seizure of their bodily fluids, regardless of their awareness of having done so. With this legal precedent and the aid of biochemical surveillance technologies, the jurisdictional extension of authoritative institutions over the most intimate aspects of the body was solidifying.

Gradual acceptance of the drunkometer and biochemical surveillance

As a hotly contested solution to the problem of prosecuting drunk-driving cases, the drunkometer was used by law enforcement and accepted in courts in most major jurisdictions by the early 1950s, fundamentally changing the drunk-driving and policing landscape along the way. For example, following a dramatic rise the previous year in drunk-driving accidents, by the end of 1945, use of the drunkometer in Detroit had brought the heaviest docket in the history of the traffic court. In November and December alone, drunkometer tests resulted in 252 drunk-driving cases, compared with only 205 in the preceding ten months before it was used (Battle Creek Inquirer 1943a, 5; 1943b 3; Cessnock Eagle and South Maitland Recorder 1945, 3; Leaf Chronicle 1945, 2).

The use of the device, and drivers’ awareness of its use, had a dramatic impact: in November 1946, the Detroit Free Press (1946b, 5) reported that drunk driving accidents in the city had declined by 47 percent (see also Republican and Herald 1946, 4). Moreover, courts in Detroit even began taking measures against jurists who refused drunkometer evidence, with one judge exclaiming, “it is just a waste of time to try cases under such circumstances” (Detroit Free Press 1946a, 13). Following Detroit’s lead, police departments around the country began purchasing drunkometers, and the first drunkometer training school west of Chicago opened in Pasadena in 1947 (Metropolitan Pasadena Star-News 1947, 25).

In addition to its growing juridical acceptance and institutional adoption came an increase in educational programs aimed at curbing drunk driving. These campaigns successfully effaced constitutional concerns when reframing the public debate about drunk driving as one about the rights of pedestrians and unimpaired drivers to be safe—and drunk driving as an incorrigible and morally objectionable source of needless death.

Take, for example, one publicly distributed pamphlet: “We are horrified at the Korean War casualties, but we seemingly shrug our shoulders at the much heavier casualties on our highways and streets.” The authors argued that, by separately allowing citizens both to drink and to drive, the law failed to protect the safety and welfare of law-abiding citizens: “For the most part, in this country, the law looks after you pretty well. But too often, the law can’t protect you from the drinking driver. He is legally permitted to drink, he owns a driving license, and he is, therefore, seldom prevented from driving. Yet the drinking driver continues to slaughter pedestrians right and left” (Seliger Reference Seliger1953, 10).

Even with the increasingly scientific discourse of alcoholism during this period—in clear contrast to the approach taken by the WCTU—the message was not always neutral or scientific when it came to drunk driving: “[We] draw upon our experience with hundreds of alcoholics and criminals to point out that alcohol is both directly and indirectly responsible for the crime, poverty and sickness in the world,” with the authors arguing that “religious faith helps any person—young or old—to cleaner, more wholesome living and is especially effective in helping young people to withstand the social pressure that may induce them to start drinking” (Seliger Reference Seliger1953, 23–26).

The mainstream revival of moralized temperance discourse on alcohol as a vice merged with increasing juridical and legislative uniformity regarding the mandatory collection of bodily fluids as a legitimate domain for using science to protect public welfare.Footnote ¹⁸ The dangers posed by untrustworthy drunk drivers to the innocent public of sober drivers and pedestrians would override protections against authoritative institutional reach into the most intimate aspects of the body.Footnote ¹⁹ The constitutionality of roadside breathalyzer tests—the drunkometer’s more mobile and efficient successor—would not be upheld by the US Supreme Court until 2016 in Birchfield v. North Dakota.Footnote ²⁰ Nonetheless, the practice of mandatory testing of a suspect’s bodily fluids became the national norm by 1954, when “a total of 45 states and 464 jurisdictions of 10,000 population or more were reported to be using chemical tests to determine blood alcohol concentrations,” and twenty-three states had passed chemical test statutes (National Safety Council 2004, 14; see also Anonymous 1957).

No longer would courts need to rely upon law enforcement testimony to protect the public from drunk drivers; the body would be made to tell the truth of its intoxication in courtrooms across the country. Along with it, a new mode of governance emerged that reconfigured the contours of citizenship in the modern era of automobility.

From the drunkometer to contemporary drug testing

Though the dangers of drunk driving are well documented and perhaps less contestable as a social problem (compare Gusfield Reference Gusfield1981), the political tactics involved in governing chemical citizenship can be adapted to a seemingly endless array of “threats” in the name of upholding the safety and welfare of purportedly threatened “publics” (however abstract). This quality renders politically strategic the power to construct and disseminate substance use-related threats, which is perhaps most obvious with the wide (and often contradictory and highly racialized) range of “drug scares” that have permeated the US political landscape and public imagination for over a century (see, for example, Reinarman and Levine Reference Reinarman and Gene Levine1989; Reinarman Reference Reinarman, Adler and Adler2006).

Notably, what began as specific debates over the use of the drunkometer in court initially included other testing methods. Alternative methods in the first half of the twentieth century included collecting and analyzing spinal fluid,Footnote ²⁸ urinalysis (which was accepted in many courts in the 1950s) (Courier-Journal 1950; York Dispatch 1950; Akron Beacon Journal 1951; Rochester Democrat and Chronicle 1951; Dayton Daily News 1952a, 1952b; Cincinnati Enquirer 1956), other competing breath analysis technologies (National Traffic Law Center 2025), and blood tests (which were originally used as a back-up for the drunkometer’s findings) (Indianapolis News 1936, 19; Indianapolis Star 1937a, 5; 1937b; Argus-Leader 1945, 3; Wisconsin State Journal 1946, 10). Whereas spinal fluid and blood tests were consistently seen by courts as being too invasive (that is, contrary to the Fourth Amendment) during this period for a variety of constitutional and historical reasons, breath and urine tests were seen as legitimate.

These rulings—which were regularly affirmed in state and US Supreme Court cases all the way to the 2016 US Supreme Court ruling in Birchfield—gradually produced a legal and institutional boundary around the body: fluids that were emitted “outside” the body (for example, breath, urine, and, eventually, hair follicles and nails) were considered public domain, and those that required technical extraction from beneath the flesh (for example, spinal fluid and blood) were considered private and required a warrant. In the end, the drunkometer—small, portable, and easy to use—won out, but these other testing methods presaged future debates and methods for surveilling and testing substance use.

This legal flesh boundary, for example, was partially dissolved by the Schmerber v. California ruling in 1966 that determined “forced blood draws do not violate Fifth Amendment protections against self-incrimination,” based on the exigent circumstances doctrine, since “a car accident or other emergency has the potential to severely limit the amount of time in which police could secure evidence (like BAC) that is destroyed over time—metabolized by the body” (Johnson Reference Johnson2025, 3).Footnote ²⁹ Yet police use of blood draws remained rare until their recent proliferation since Arizona established a police phlebotomy program in 1995, which has since been replicated in at least sixteen other states (Johnson Reference Johnson2025). Still, breath-based alcohol tests remain standard practice for policing drunk driving cases because “they are quick, promptly displaying blood alcohol concentrations (BAC) and do not require a warrant” (3).

Importantly, while drunk-driving legislation and the drunkometer homogenized intoxication with blood alcohol content measures that were universally applied to all suspected drivers, drug tests take this further by homogenizing results to the mere presence of particular metabolites in the body, since drug tests are unable to measure intoxication (Walsh and Trumble Reference Walsh, Trumble, Robert and Louis1991; Willette Reference Willette, Robert and Louis1991; Holtorf Reference Holtorf1998). With drug testing, mere consumption, not intoxication, now constitutes a sufficient basis for identifying “threats” and governing access to the rights and benefits of citizenship.Footnote ³⁰ Moreover, just as lawmakers maneuvered around consent by integrating it into driver license agreements, so too would similar strategies emerge with “informed consent” to drug testing in probation and parole (as a requirement for supervision), health care (as a requirement of “patient contracts” to access controlled substances), athletics (as league requirements for competition), employment (as a condition for hiring and/or maintaining employment), education (often as a condition for participation in extracurricular activities as well as in the form of the so-called “community covenants” prevalent in private schools), and public benefits (as a requirement for application and ongoing receipt, depending on the state and benefit).

A body of research has documented the often severe consequences of “failing” a drug test in a wide range of institutional fields that include, but also extend well beyond, the criminal-legal system, demonstrating the normalization of biochemical surveillance in policing substance use in virtually every aspect of social life. Among others, this includes corrections and community supervision (Simon Reference Simon1991; Phelps and Ruhland Reference Phelps and Ruhland2022; Piehowski and Phelps Reference Piehowski and Phelps2023; Schiraldi Reference Schiraldi2023); drug courts and mandated treatment (Paik Reference Paik2006; Mackinem and Higgins Reference Mackinem and Higgins2007; Gowan and Whetstone Reference Gowan and Whetstone2012; Tiger Reference Tiger2013; McKim Reference McKim2017; Stitt, forthcoming); health care (Collen Reference Collen2009; Chiarello Reference Chiarello2024); employment (Gilliom Reference Gilliom1994; Tunnell Reference Tunnell2004; Milloy Reference Milloy, Lauer and Lipartito2021); family and parental rights (Boyd Reference Boyd2019; McCabe Reference McCabe2022; Nguemeni Tiako, Jordan, and Sweeney Reference Tiako, Jordan and Sweeney2022); public benefits (Budd Reference Budd2011; Whitesell Reference Whitesell2024); athletics (Sluggett Reference Sluggett2011; Bennett Reference Bennett2012); and schools (Finley and Finley Reference Finley and Finley2005). The current overdose crisis, for which drug testing—much like the drunkometer and drunk driving—presents a seemingly commonsense tool to help address a pressing social problem, has only reinforced this mode of intimate bodily surveillance while expanding the range of participating institutions (Chiarello Reference Chiarello2024).

Whether the drunk driver on the roads, the drug-addicted criminal, the crack-using welfare recipient, the drug-using mother, or the steroid-using athlete, each stereotyped element is configured around the production of an innocent public, a chemical-based threat, and a technology capable of detecting the suspect chemical: innocent pedestrians and drivers + drunk drivers + drunkometer/breathalyzer; innocent taxpayers + welfare-receiving crack users + benefits-dependent urinalysis; innocent children + drug-using mothers + maternity-ward testing; innocent citizens + drug-addicted criminals + weekly probation officer-administered drug tests; innocent competitors + steroid-using athletes + pre-competition drug screenings. Through field-specific configurations of these elements, the relevant institutional authority is capable of surveilling the body at the most intimate level in the name of protecting its innocent constituents, extracting the chemical truth from it without need for collecting testimony, considering circumstances, and weighing context.

In fact, this is essential to, and integral to the reproduction of, the underlying moralized subject in question: the untrustworthy drug user (or drunkard, cheater, and so on). Chemical citizenship, as a mode of governance, both relies upon and reinforces the notion that those who use “drugs” (or drink excessively) are not rational—both because of their intoxication and their choice to become intoxicated—and therefore are dangerous and cannot be trusted to know or tell the truth, let alone manage the responsibilities and privileges of citizenship (O’Malley and Valverde Reference O’Malley and Valverde2004). If the subject cannot be trusted to speak the truth, the body can nonetheless be trusted to reveal the truth through techno-scientific procedures that reach into its veins, urinary tracts, hair follicles, and breath. The body can be made to speak for, and even defy, the subject.