In the autumn of 1912, the white residents of Forsyth County, Georgia, expelled its Black community. The trigger, or the excuse, was the killing of a young white woman. A Black man was lynched in response, followed by the judicial murder of two more Black defendants and a campaign of white violence and intimidation that chased away more than a thousand Black residents. Blood at the Root, Patrick Phillips’s history of the ethnic cleansing, shows that the participants even included white children, who could dispossess a Black family by threat alone.
Such events do not fit neatly into the categories that conventional political science long used to understand politics in the United States. As debates over how to understand Jan. 6, 2021 demonstrate, until recently most practicing American social scientists reflexively dismissed such events as incidental, regrettable exceptions to a democratic, rule-based system.
Many political scientists like political behavior to fall into neat boxes, whether those be typologies cleanly defining terms or spreadsheets in which every row contains a discrete observation. They recognize that there’s always phenomena that won’t fit, cleanly, but those can be the basis of future research—or relegated to the “error term,” the leftover bin for the facts that theory doesn’t explain.
When the implicit definition of democracy is democracy with American characteristics, the exceptions don’t even register as exceptions—until some event so far out of the comfort zone of mostly white, upper middle-class academics forces us to confront them as if they were brand new.
Well into my own graduate training, mainstream researchers confidently assumed that U.S. democracy being peaceful and law-based was not only true but essentially beyond debate. The Center for Systemic Peace’s widely used Polity scores, for instance, give the United States between a +8 and +10 from 1809 to 2016—a stable, indeed maximally scoring, democracy. That period includes the Civil War, when the losing side launched a violent conflict rather than accept the election results.
Yet political violence and repression was common during that time. The United States Political Violence (USPV) database records numerous riots around elections during the mid-19th century. In April 1855, for example, hundreds of nativists “invaded” a German area of Cincinnati, Ohio, and destroyed more than a thousand ballots. Subsequent fighting led to two deaths. In August of that year, nativist Protestants attacked German and Irish neighborhoods in Louisville, Kentucky, killing at least 20. In Baltimore, election violence became routine in the 1850s, with 30 dead and 350 wounded in the 1856 election alone.
The bloodiest efforts came in the repression of Black people. The USPV lists nearly 70 incidents of political riots and assassinations in the decade after the formal cessation of hostilities, mostly in the South but including riots in Philadelphia and Indianapolis. Consider Louisiana, where white Democrats massacred dozens of mostly Black Republicans protesting for their suffrage in 1866. In 1868, Klan members killed 200 Black people in Louisiana in an effort at voter suppression. In 1873, as many as 150 people, almost all Black, were killed in Colfax, Louisiana, in violence stemming from the 1872 Louisiana elections.
In 1874, the White League, an armed group of over a thousand members led by former Confederate officers, overthrew the government of Louisiana, killing at least 13 police officers in the process. They charged the government with corruption, but their real objection was that it was racially progressive. (Three days later, the government was restored by the U.S. Army.)
Widespread political violence around elections only really ended when the federal government conceded that the South would be run by whites. Even then, anti-government violence took place.
In 1898, a group of armed white people overthrew the recently elected local government of Wilmington, North Carolina, forcing the mayor, council, and police chief to resign at gunpoint and replacing them with a new, unelected government. (Along the way, they burned down a local newspaper printing office and murdered 60 others.) If this was not a coup, then it’s hard to imagine what the term means in any but the most specialized and limited applications.
Flattering coding rules used to produce datasets make it too easy to dismiss any aberration when a look at the historical record keeps turning up aberration after injustice after atrocity. Historians, scholars of Black history, and political scientists specializing in race and ethnic politics have long been sharply critical of the idea that that concepts like democracy, sovereignty, or the rule of law can be as bluntly coded as standard datasets long did.
American democracy did not penetrate to state level until the 1960s. Nearly a quarter of the states denied voting rights to Blacks—who made up a majority in some of those states before the Great Migration—from the late 19th until the mid-20th century. (As Isabel Wilkerson’s The Warmth of Other Suns demonstrates, Northern and Western states were only somewhat more welcoming than the American apartheid states.) The University of Michigan professor Robert Mickey defines those Jim Crow-era Southern states as “stable, one-party authoritarian enclaves” that “curtailed electorates, harassed and repressed opposition parties, and created and regulated racially separate—and significantly unfree—civic spheres.”
A federal union with authoritarian states cannot but be at least partly authoritarian itself. White-dominated electorates in the so-called Solid South voted reliably and almost exclusively for Democrats in presidential and congressional elections. That nearly replicated the structural biases of institutions during the slavery period, when the counting of enslaved people as three-fifths of a free person gave the South extra seats in the House and the Electoral College. Despite the abolition of slavery, the imposition of Jim Crow meant that neither Congress nor the presidency were elected by fully democratic, or even representative, means.
Other concepts assumed to be simple in the U.S. context turn out to be messier on closer examination. Consider Max Weber’s workhorse definition of the state: the monopoly of the legitimate use of physical force within a given territory. By that definition, large swathes of the United States approached failed-state status for much of the 19th and early 20th centuries.
Between 1883 and 1941, 4,467 people, about three-quarters of whom were Black, were lynched in the United States. In the West, sociologists Charles Seguin and David Rigby found that in a confirmation of Ida B. Wells’s original argument, lynching in Western states resulted from the absence of government, claimed victims from many races, and declined as state institutions developed.
In the South, though, a different pattern held. The state contended for power with local actors and did not always win. Instead of becoming less frequent as state power grew, lynching became more common around the end of the 19th century. As historians Stewart Tolnay and E.M. Beck wrote in 1995, “A black man, woman, or child was murdered nearly once a week, every week, between 1882 and 1930 by a hate-driven white mob.”
Lynchings were frequently carried out in public, taking on a celebratory cast (Tolnay and Beck’s book is titled A Festival of Violence). The corpses of the murdered were mutilated, and photographs of the victims and killers turned into postcards. The state, such performances announced, possessed no monopoly on the legitimate use of force that the ruling class, supported by the white crowds who turned out to approve of the lynchings, could not reclaim.
Many condemned these abuses, but, just as today, there were voices even at elite institutions counseling patience and unity. An 1898 Yale Law Journal article defended lynching as a natural outcome of Reconstruction having given the ballot to former slaves too early, and urged “education,” not federal intervention, as the cure. Woodrow Wilson, a leading historian and political scientist long before he became president of the United States, defended the Ku Klux Klan and white terrorism as “aroused by the mere instinct of self-preservation.” Such efforts eventually paid off in helping to efface such atrocities from textbooks even as monuments—and popular culture—embraced Lost Cause nostalgia for the Confederacy.
Revisiting the United States as a partial or flawed democracy means confronting much of the history that celebrants of the liberal world order claim as a series of triumphs for democracy. None of the major combatants in the First World War were fully democratic, given that none of the democracies offered full suffrage to women (although a few U.S. states did). The U.S. Senate that ratified the Charter of the United Nations in 1945 was a body institutionally committed to securing the rights of those authoritarian U.S. states, not individual freedom or national democracy. Rewriting this history requires rethinking whether and how this world order became “liberal.”
Reassessing these definitions also matters for the social-science theories that build upon and employ them. The “United States” coded in statistical models deployed in the seemingly debates about the democratic peace was not really all that democratic. That might explain why defenders of the proposition had to invent epicycles—the fixes that premodern European astronomers employed to make their theories fit astronomical observations—to explain away inconvenient facts, like how the United States overthrew democratic regimes when Cold War imperatives demanded it.
To be sure, social scientists have lately become more skeptical of the conventional verities of progress. The V-Dem Institute in Sweden has compiled indices of democracy that are more sensitive to conditions like racial segregation. They show the United States as substantially less democratic than other countries, like the United Kingdom or Sweden, for most of the 20th century. Political scientists investigate topics that once attracted little attention, like the relationship between American political violence and social transformation, how national economic integration led to the decline of lynching, or how the “carceral state” (more than 2 million people are held in U.S. prisons or jails) degrades U.S. democracy today.
Yet old habits are hard to break. Generations of social scientists’ and historians’ choices to marginalize the entirety of the American experience left the broader conversation ill-prepared to understand the current moment. In the immediate aftermath of the storming of the Capitol, reporters and others turned to the British seizure of the Capitol building in 1814 as the closest analogue. Turning to foreign invasion rather than domestic precedents, however, says a lot. It suggests that people do not know the domestic precedents even exist, and it reinforces the notion that American political violence is “unthinkable.” (Even describing the 1814 incident as “foreign” is complex. The burning of Washington in 1814 was carried out by a British force that included marines previously enslaved by Americans—and motivated by hatred of the slavery system.)
Reality is never as easy to tame as neatly formatted spreadsheets imply. Rather than insisting on narrow definitions of concepts, it’s time to think more openly—and less defensively—about the totality of U.S. political history and behavior at home and abroad. Rethinking what seemed obvious, and taking steps to focus on the root of our problems, fits the needs of a country that’s seen its long string of peaceful transitions of power broken.
