language at home · political translation · the Trickster
The Trickster’s Republic
There are 55 American counties where most residents speak a language other than English at home. Together they swung 15.2 points toward Donald Trump in 2024. They do not speak the same language—and they did not move for the same reason.
There is an American political geography defined not by race, religion, region, or party, but by language.
In exactly 55 counties and county equivalents, most residents age five and older speak a language other than English at home. They do not speak the same language. They do not practice the same religion. They do not share a migration history, class structure, racial identity, relationship to the border, or local political system.
In 2020, those counties voted for Joe Biden by 35.9 points. Four years later, they voted for Kamala Harris by 20.7. The Republic of Fifty-Five Counties moved 15.2 points toward Donald Trump—more than twice the national swing—and 53 of its 55 members moved in the same direction.
The obvious question is why. The more interesting question may be why we assumed there had to be one answer.
The candidate can be common even when the meaning is not.
The 55 counties remain an artificial analytical Place: a strict ACS threshold joined to election returns. That artificiality is useful. It removes the temptation to define the story after seeing the vote and forces us to confront a geographically dispersed pattern created by one measurable fact.
Language at home is not a causal explanation for Trump’s gains. It is a lens that finds communities whose political information, trusted institutions, memories, and local vocabularies often sit at an angle to national English-language political discourse. Once those Places are assembled, their 2024 convergence is too large to ignore and too heterogeneous to compress into one demographic story.
The Trickster framework offers a way to live with that contradiction. Trump does not need Webb County, the Bronx, Miami-Dade, Apache County, Passaic County, and Bethel to want the same America. He only needs to remain legible inside each community’s own account of what has gone wrong.
The Trickster’s gift is not that he speaks every community’s language. It is that every community can translate him differently.
Method and evidence
Cohort. The universe is every U.S. county or county equivalent in which the 2020–2024 ACS 5-year estimate shows more residents age five and older speaking a language other than English at home than speaking only English at home. The strict rule yields 55 geographies. The existing URL retains “fifty-six” because the first published edition included Manassas City, Virginia, whose estimate falls just below 50%.
Election aggregation. County presidential results are Akashic’s canonical county returns. The cohort aggregate is an exact sum of those rows. The 2020 aggregate was D+35.890; the 2024 aggregate was D+20.698, a 15.192-point Republican swing. The national popular-vote margin moved 5.965 points Republican over the same interval.
Raw votes. Trump’s vote total across the cohort increased by 325,702 while the Democratic presidential total fell by 1,323,125; total presidential votes fell by 943,066. These changes do not identify individual switchers, nonvoters, new voters, or third-party defectors. The “path” labels in the table classify aggregate change only.
Precinct maps. The atlas uses Akashic’s verified 2024 precinct geometry and presidential result sidecars. It does not imply that precinct boundaries are stable across cycles; the 2020→2024 comparisons displayed beside the maps are county totals.
Interpretation. “Trickster,” “multivocality,” and “political translation” are interpretive frameworks, not measured causal variables. See Keir Martin and Jakob Krause-Jensen’s 2017 Anthropology Today essay “Trump: Transacting trickster”, Barry King’s 2018 analysis of Trump’s celebrity persona, and Associated Press reporting on 2024 outreach to small faith communities. The language measure comes from ACS table C16001.