Data Stories · Race & Democracy

The Coup America Called Order

American newspapers often saw what happened in Wilmington: organized men displaced a government by force. The more revealing question is how readily they made that seizure look like the restoration of legitimate rule.

Who this data representsDigitized English-language newspapers available in Chronicling America and processed by American Stories—not every 1898 newspaper, reader, or political viewpoint. The event set contains 1,624 articles selected from a 668,295-article denominator. GABRIEL scores visible language, not the truth of a claim or an author's hidden beliefs.

1,624

relevant articles after separate relevance screening

1,448

distinct story clusters after grouping reprints

67.3

white-supremacy endorsement, cluster-weighted

13.8

explicit condemnation of racial violence

The overthrow did not begin with the first gunshot. In the weeks before November 10, newspapers carried an argument: Black political participation was “domination”; white property holders were the “substantial” citizens; force might be regrettable but necessary. Those pre-event articles are part of this corpus because they describe the political campaign, threats, and racial program that produced the event.

Attention surges after the violence, then peaks on November 15—five days later—as weekly papers and syndicated commentary catch up. At the peak, just under one percent of every detected article in the available archive concerns Wilmington.

Figure 1 · A coup travels on a weekly newspaper clock

Relevant articles as a share of all detected articles, Oct 25–Nov 30, 1898

The national peak came after the seizure. The event line marks November 10; coverage peaks November 15 at 0.95%. The pre-event series captures the election campaign and explicit threats that preceded the violence.

The legitimation came before the violence

Among event-relevant stories before November 10, endorsement of white supremacy averages 79 and advocacy of disenfranchisement 71 on a 0–100 salience scale. Explicit condemnation averages only 6.8. The hierarchy was not invented afterward to explain chaos; it was the announced political program.

After the killings and forced resignations, condemnation rises—but only to 22 in the later aftermath. White-supremacy endorsement falls to 59, still one of the dominant frames. Recognition of a coup remains near 60 across phases. The press could describe organized political coercion and still present it as the work of respectable citizens restoring order.

Figure 2 · Recognition without rejection

Mean salience by phase, one vote per distinct story cluster

The event changed tone more than premise. Condemnation grows after the violence, while endorsement and disenfranchisement retreat from extraordinary pre-event levels. Coup recognition remains high: naming political displacement did not guarantee disapproval.

The striking combination is not ignorance. It is recognition plus legitimation.

North Carolina was the extreme, not an island

Geography matters. North Carolina stories score 77 on white-supremacy endorsement, 66 on disenfranchisement, and 58 on blaming Black people or Black politics for disorder. Outside the South, the same measures fall to 54, 43, and 42.

But 54 is not zero. Outside-South coverage still scores only 18 on condemning racial violence and 21 on electoral legitimacy. Regional distance weakened the justificatory frame; it did not replace it with a consistent defense of multiracial democracy.

Figure 3 · Geography changed the degree, not the grammar

Cluster-weighted frame salience by publication geography

North CarolinaOther SouthOutside South

The regional gradient is broad and consistent. North Carolina is most supportive of hierarchy and disenfranchisement; outside-South papers are more condemnatory and give Black political actors more agency. Publication state identifies the outlet, not its wire source or readership.

A Black newspaper breaks the regional map

Regional averages can conceal the institution that mattered most: the press itself. Among the 72 newspapers with at least five relevant articles, endorsement and condemnation correlate at −0.75. Most North Carolina outlets cluster near high endorsement and low condemnation. Yet the sharpest dissenter is also Southern.

The Richmond Planet—an African American newspaper founded by former slaves and edited by John Mitchell Jr.—averages 7 on endorsement and 64 on condemnation. The Iowa State Bystander is another strong condemning outlier. The important divide was not simply latitude. It was which citizens controlled a printing press and could name racial violence as violence.

Figure 4 · The newspaper matters more than the region alone

Outlet means among newspapers with at least five relevant articles; circle size shows article count

Endorsement and condemnation are opposing editorial choices. Each circle is a newspaper. The labeled Richmond Planet sits far from the main Southern cluster, showing why region cannot substitute for outlet identity.

The Library of Congress identifies the Richmond Planet as African American. No broad racial-identity classification was inferred for other outlets; the article treats the Planet as a documented case, not a modeled category.

What can this event teach us about American society in 1898? Democratic language and racial hierarchy were not separate vocabularies. Property, intelligence, progress, civilization, and order became arguments for deciding who counted as a legitimate political person. Newspapers did not merely fail to recognize a coup. Many recognized the coercion and supplied the moral language that made it respectable.