In April, the Supreme Court began to hear arguments about one of the central requirements of the Constitution. It’s right there, in Article I, Section 2, clause 3: For a government of the people to function, the people must be counted. The Founders wanted an “enumeration” to occur within three years of the first meeting of Congress, and then “within every subsequent Term of ten Years, in such Manner as they shall by Law direct.” A census, in other words.
That sounds straightforward, but, in 2018, the Trump Administration proposed adding a new question to the 2020 census. It’s a simple query, nine words long: “Is this person a citizen of the United States?” The Secretary of Commerce, Wilbur Ross, first suggested the change, at the urging of Steve Bannon, the former White House chief strategist, and it has since been met with widespread criticism. Six former directors of the Census Bureau, from both Republican and Democratic Administrations, have come out against the question. A federal judge in California deemed it unconstitutional; another, in New York, found that Ross’s proposal broke a “smorgasbord” of rules. Now the Supreme Court will weigh in.
The census has not contained a question about citizenship for seventy years. As the Constitution states, the census is supposed to enumerate all people in the United States, not just all citizens. Only seven per cent of the people living in the U.S. are non-citizens, but the number of people living in households with a non-citizen is higher—fourteen per cent. (Among Asian-Americans and Latin-Americans, that figure rises to forty-five and forty-six per cent, respectively.) The Census Bureau’s chief scientist, John Abowd, said that adding a citizenship question would deter at least six hundred and thirty thousand households from completing the form, and a government estimate predicted that 6.5 million people might not be counted as a result. That sort of distortion would have enormous consequences. The census determines the distribution of seats in Congress and the Electoral College. It’s also used to apportion nearly a trillion dollars in federal spending, to fund schools, hospitals, and roads. If a segment of the population—whether an entire state or an ethnic community—is undercounted, its resources are diminished accordingly.
The Supreme Court prides itself on being apolitical, but this case has already provoked deep partisan feelings. This is precisely because the census is, more than it might seem, a mirror of our politics. It has always reflected the promise and prejudices of our country, and its story drives to the heart of the Constitution—indeed, to its very first three words. Who are we, the people?
The Extraordinary Presence of Frida Kahlo
The Narrative Experiment That Is the Marvel Cinematic Universe
What to Do in N.Y.C. This Weekend: Jay-Z at Webster Hall, and Hadestown
Go-Go Music as a Rallying Call to Resist Gentrification
Dating While Muslim: The Uncomfortable Truths of “Ramy”
Julianna Barwick Is Using the New York Sky to Make Music
That question was important to the Founders, who understood that a precise census was essential for any government of the people to work. Figuring out who lived here had not been not a priority before independence. During British rule, when a new king inherited the throne, the broad reaches of the realm were assessed crudely, and the population was determined through guesswork. In 1714, with the accession of King George I, His Majesty’s counters estimated 375,000 white and 58,000 black subjects in the American colonies. Native Americans were certainly present, but they were not considered subjects, and they went uncounted.
During the American Revolution, the population shot up to about 2.75 million people, including half a million African-Americans. After independence, actuarial certainty remained elusive. It was not easy to track people who were perpetually moving west, away from the seat of government and the taxes they associated with it. Counting African-Americans was another challenge. In the earliest days of nationhood, under the Articles of Confederation, Southern states were delighted to count slave populations, because doing so would add to their electoral strength in Congress. But at other times they refused to count them—for example, when a federal tax was tied each state’s population. The compromise reached by the Constitution helped to preserve the Union, but at an egregious cost to democracy. African-Americans were to be counted as “three-fifths of all other persons,” an arrangement that inflated Southern clout and made it impossible to check the spread of slavery until the Civil War.
Still, the Constitution was a stronger document than the Articles of Confederation, partly because of the census it instituted. The act of counting rooted a rootless people. On the first Monday in August, 1790, U.S. marshals began to fan across the continent, enumerating everyone they could see. It was hard work. There were few rules, and some states needed extensions—especially those that had trouble accepting federal authority, such as Rhode Island and South Carolina. In the early days, the findings were not even published; they were simply posted, in each district, “at two of the most public places . . . to remain for the inspection of all concerned.”
That first census was completed in eighteen months. Nearly four million people were counted, which raised the membership of the House of Representatives from sixty-five to a hundred and five. A French statistician was astonished. He declared that the United States was an unprecedented phenomenon, calling the new nation “a people who instituted the statistics of their country on the very day when they founded their government, and who regulated by the same instrument the census of inhabitants, their civil and political rights, and the destinies of the nation.”
The very act of asking questions led to more questions, and the census grew fatter each decade. In 1849, the responsibility for the census was removed from its original home, the State Department, and sent into the Department of Interior, which, at the time, was home to such a grab-bag of programs that it was known as the Great Miscellany. The 1850 census was a particular triumph: for the first time, it spanned the continent and tracked the restless inhabitants of California, which had been recently acquired. It required more than three thousand employees to tabulate the results and transport them back to Washington.
By 1850, the population of the U.S. had reached a staggering 23,191,876—considerably more than England’s population (fifteen million). Size mattered to Americans, who could not help seeing growth as a sign of divine favor. Walt Whitman, in his poem “Year of Meteors,” found something lyrical in the dry information being culled across the country. “I would sing in my copious song your census returns of The States,” he wrote, not very poetically. He was far from alone; in the beleaguered early months of the Civil War, Lincoln took solace in predicting an eventual U.S. population of two hundred and fifty million.
With the North’s victory, the three-fifths rule was finally eliminated, which, paradoxically, gave more power to the South, since it could now count the entirety of its population. The Fourteenth Amendment stated that the allocation of House seats was to rely on “the whole number of persons in each State”—a reading that has been supported by the Supreme Court ever since. The enumerators went out again in 1870 and 1880, but the task of processing census data was increasingly becoming a headache for civil servants. Sometimes it took almost a full decade to tabulate the results—at which point, of course, it was time for a new census to begin. The questions just kept flowing. Here is one example, from the 1880 form, which shows the census burrowing into the land itself:
Level, alluvial or bottom land?
A sorting machine invented by Herman Hollerith.
Photograph by Hulton Archive / Getty
With the 1890 census looming, the Census Bureau was desperate for a faster way to count. In 1888, it held a contest: using samples of population statistics from St. Louis, it asked challengers to invent new ways of processing data. To general amazement, a twenty-eight-year-old former census employee named Herman Hollerith won easily, with an invention called the electric tabulating machine. One part of the contest asked the candidates to divide the data into categories. It took the second- and third-place competitors forty-four and a half hours and fifty-five and a half hours, respectively. Hollerith’s machine did the same task in five and a half hours.
Hollerith, the son of a German immigrant, grew up in Buffalo, tinkering with machines. He especially liked trains—he had already pioneered an “electro air brake” for them—and it was this passion that led to his idea for tabulation. He knew, from his census work, that it took too long for an enumerator to write out answers by hand, and he had noticed how quickly train conductors could punch passenger tickets as they moved through a crowded car. He reasoned that a device could record the census in a similar way.
After trying a few different models, Hollerith settled upon a long card, approximately six and a half inches by three and a quarter inches, as the best vehicle for the vast amount of information he was processing. He started calling them “punch cards,” because his device tabulated data by punching holes through pre-printed symbols on the cards. These holes represented every point of information on the census: country of national origin, age, marital status, general health, languages spoken. Suddenly, the most exquisitely sensitive points of identity could be instantly noted and calibrated. Race had not disappeared, of course; a world of meaning was spelled out in the upper-left section of the card, where a few fractions of an inch separated the symbols for white (“W”), black (“B”), mulatto (“Mu”), quadroon (“Qd”), and octoroon (“Oc”) Americans.
Hollerith’s tabulator was a triumph. The 1890 census was by far the largest yet designed; it asked more questions than ever before and deployed enough enumerators (46,804) to constitute a decent army. Those enumerators were supplied with twenty-five million population schedules, weighing more than three hundred tons. They sent back sixty-two million punch cards, all the same size and machine-readable. An experienced tabulation clerk could process eighty punch cards per minute. This new technology led to another huge census, the results of which added up to 21,410 pages. Thanks to Hollerith and his invention, it was the most detailed statistical portrait of the United States that had ever been created.
But the 1890 Census did not stop there. It also introduced large-scale data processing to a world that was hungry for it. Following his triumph, Hollerith pursued other ventures, including a census of Russia (the czars had not condescended to count their people since 1851), and an immense data project for the New York Central Railroad. His energies were also drawn to the private sector, and Hollerith founded the Tabulating Machine Company, which merged with other manufacturing companies to form the Computing-Tabulating-Recording Company. In 1924, it was renamed International Business Machines, or I.B.M.
A census worker records information on a punch card in 1940.
Photograph Courtesy U.S. Census Bureau
Hollerith had solved the processing problem, but he had not quite solved the issue of where to store all of the paper that was needed in the pre-digital era. The first ten censuses, from 1790 to 1880, had been bound into books, which cumulatively added up to 4,597 volumes. But the 1890 Census was simply too big to bind; it was five times larger than all the previous censuses combined. The population schedules alone would have added up to thirty thousand volumes. The chief clerk of the census predicted that it would require a seven-thousand-foot shelf to store them all.
Finding more than a mile of empty shelf space was difficult, even in a city that was already known for its bureaucracy. So the pile sat around, sparking no joy among the officials charged with its safekeeping. It sprawled across so many pages that copies could not be sent around the country, as earlier censuses had been. Only one copy existed, and it spilled over into different buildings—the Pension Bureau, the Patent Office, and a building called Marini’s Hall. In 1913, it was moved to a new Commerce Department building, where it was consigned to a basement storage area, adjacent to a boiler room. The papers were arranged on narrow pine shelves, in cardboard packages wrapped with twine. The shelves stood very close to each other, only twenty inches apart. The space was also very hot—more than ninety degrees when the boiler was running.
On the afternoon of January 10, 1921, an employee noticed smoke coming through openings around the pipes that ran from the boiler room into the file room. The building was evacuated, and firemen were called. The next morning, investigators found charred records floating in ankle-deep water. They believed that the census might have spontaneously combusted. The census director estimated that twenty-five per cent of the records were gone; half of what remained was damaged. The greatest effort in history to tell a nation’s story had gone up in a puff of smoke.
In the wake of the fire, the government began to reassess its responsibility to the past. In an op-ed, the Washington Post argued that new steps were needed to protect the Declaration of Independence and the Constitution, which were then kept in wooden cabinets at the old State Department. The idea of a National Archives also emerged in response to the disaster. The remains of the 1890 Census, meanwhile, were placed in an old brewery, where they lingered for another decade. Congress authorized their destruction on February 21, 1933, the day after President Herbert Hoover laid the cornerstone for the Archives.
Time is now running short before the 2020 census, our twenty-fourth, will begin. It’s possible that it could be a unifying experience. More likely, it will be bitterly contested, like every other aspect of nationhood. It will also be the most digital census in U.S. history, conducted mostly online and facilitated by enumerators guided by apps. Herman Hollerith would have surely embraced the speed and accuracy that these technologies provide. But they also incite certain fears—that a foreign power might try to hack the result, or that the system will fail for its own reasons. In 2016, a high-tech census crashed in Australia. Thanks to chronic underfunding of the Census Bureau, the census’s new tools have not been rigorously tested. For the fiscal year 2019, the Administration asked Congress to provide four hundred million dollars less to the Bureau than it received in 2009.
Still, we have no choice but to honor the precise instructions of the Founders. They knew that self-government begins with self-understanding. Accounting for every person in the nation was a matter of principle; no one would have blamed the Washington Administration if it had paused before the first count, in 1790, to deal with more pressing problems. But the nation went ahead with the enumeration, and a government of the people was established. For it to endure, we must once again ask who we are.