The African Myth: Everything We Thought We Knew About Black American Admixture is Wrong

A probability model exposed the fault line in America’s favorite origin story

Photo by The New York Public Library on Unsplash

Not a member? Read it here.

THERE’S A REASON certain narratives stay alive in the academic world, repeated in lecture halls and literature reviews like gospel. They feel true. They align with our intuitions. They fit the story we’ve been telling ourselves as a collective for decades. In the field of human population genetics, one of these narratives is the widely accepted belief that Black Americans carry between 70–75% African ancestry, due to them being the direct descendants of the enslaved population in the United States.

But what happens when you actually run the numbers?

The Descendants of Slave Problem

For years, we’ve operated under a comfortable assumption that Black Americans, a post-colonial group created from the mixing of African slaves with Indigenous and European peoples in early colonial America, are predominantly of African descent. It makes intuitive sense, right? African slaves arrived, they mixed with some of the people in the New World, and together they formed this distinct population. And while historians have quietly admitted that only between 350,000–450,000 African slaves were brought to what would become the United States, most scholars and historians continuously assert that Black Americans carry forward that African heritage in meaningful ways.

Except… they don’t.

After running a probability-based ancestry model of continuous African immigration and mixing from the years 1620–1808 (when the importation of Africans was banned in the United States), the results are startling. Given the numbers from census counts, tax rolls, shipping logs, and estate records of Africans imported, as well as the overall slave and White populations, in 2025, it is likely that Black American admixture is composed of only about 8.9% African ancestry. Even with generous assumptions about social and geographic clustering and dilution effects from population mixing, Black Americans remain overwhelmingly not African.

What The Textbooks Get Wrong

The error hasn’t been in the data. For years, historians and anthropologists have been providing analysis of primary source documents related to colonial America. We’ve uncovered that only about 400,000 slaves, not millions, were actually brought to the United States. We’ve learned about the personal and economic benefits of forced intimacy between owners and slaves. We’ve traced the laws and statutes outlining the massive reclassification and erasure of many American Indian tribes.

No, the error was never in the data. It was in our mental model. We looked at the Black American population and saw African slaves arriving year after year. We saw approximately 4,000,000 slaves freed by the Emancipation Proclamation at the end of 1865 and assumed those numbers translated into substantial African ancestry that has endured undiluted for another 160 years beyond slavery, despite the integration between earlier colonial groups and with the millions of people from different ethnic groups that have emigrated to the U.S.

But most importantly, we forgot about the colonists.

Free European colonists made up 84-95% of the population and dominated every reproductive pairing probability in colonial America. Not just by raw population presence, but by access, status, legal authority. And they were deeply motivated by financial gain to influence domestic reproduction of slaves. According to this historically supported mathematical model, when an African slave paired up, there was a 95% chance their partner was a European colonist. Not another African slave. Not even a non-African slave or indentured worker, even though they were geographically and often socially clustered with indigenous individuals and contracted servants from overseas.

So yes, African × African pairings produce 100% African offspring. But based on the documented population numbers provided by historical records, those pairings could have occurred 0.0566% of the time in the Early colonial period. The overwhelming majority of African ancestry was immediately diluted to 50% (African × European colonist pairings), then 25% in the next generation, then 12.5%, and so on.

The Narrative We Tell vs. The Math We Ignore

Academia has built an entire framework around Black American identity that assumes meaningful African contribution. Researchers talk about Black Americans as an “African-derived” population. Scholars turn a blind eye to consistent numbers found in primary source documents that fail to support the narrative that millions of people could descend from the small amount of continent-born Africans brought to the U.S.

And yet, we never question the dissonance that arrives when we’re asked to simultaneously believe that these same people were supposedly working, serving, and living as foundational builders of mixed-population colonial America, yet were somehow also geographically and socially isolated enough to form a stable, endogamous population that would retain 70–75% African ancestry for centuries, despite originating from a 20-person founder population that never exceeded 10% of the larger colonial population.

The math tells a different story: Black Americans are 91.9% not African.

Even if you account for clustering (non-African slaves/workers and Africans living in the same conditions, governed by the same restrictions, deliberately choosing each other as partners more often than random chance would predict), you only get to 10.3% African ancestry in the population by the end of the African slave importation period.

Rethinking Black American Identity

The findings from this model have real implications for how we understand population mixing, ancestry, and identity formation, especially for Black Americans.

If Black Americans aren’t significantly African, then what’s the basis for saying they are? Why do we continue to analyze African + non-African slave populations as a combined population when their shared African ancestry is statistically negligible? But more importantly, why is there a concerted effort still in 2025 to still link Black Americans to Africa, while other post-colonial admixed groups such as Puerto Ricans are considered to be of the Americas?

The answer, I suspect, is that we’ve conflated social categories with genetic ones. Over time, the groups that were bound together by slavery and the word “Negro” in colonial America evolved to share cultural ties, historical connections, geographic proximity. But genetically?

Black Americans aren’t African.

They never were.

Does that mean Black American identity is meaningless? Of course not. Social identity, cultural heritage, and shared history are things that matter deeply and exist independent of genetic ancestry proportions. But if we’re going to do serious quantitative work on population mixing, and if we’re going to build models, make predictions, inform policy, we need to stop letting comfortable narratives override mathematical reality.

Where Do We Go From Here?

Our society is at an existential crossroads. We can continue teaching the old models, the ones that feel right, the ones that align with our intuitions about how history and mixing should look.

Or we can face the math.

We can acknowledge that minority populations contribute ancestry proportionally to their size and mixing rates, not proportionally to how much we focus on them in our research. We can stop grouping Africans and Black Americans together as if 5% shared ancestry creates some kind of meaningful genetic category. We can finally admit that there has been an intentional mislabeling of a group that is likely indigenous to the Americas, put on our big boy pants and face the consequences of creating a society based on lies and misrepresentation.

The choice is ours. But we can’t say we weren’t warned. The numbers have been sitting there the whole time, waiting for someone to actually calculate them.

Turns out, African Americans aren’t really African at all.

The mathematical model underlying this analysis used a probability-based approach to available historical data on colonial populations. The approach used pairing probabilities, clustering coefficients (c = 4.0), and immigration dilution factors (d = 0.5) across three 62-year periods to estimate the percentage of African ancestry present in the U.S colonial population. You can read the full analysis here.