Genes Are Not Languages

What genetics contributes to Bantu history, where its limits lie, and why ancestry, language and identity must not be treated as identical.

A DNA result can feel like a secret archive. Coloured charts promise to reveal where ancestors came from. Genetic studies can indeed transform our understanding of ancient migration.

But DNA cannot tell us what language a person spoke, what stories they believed, or what community they claimed. When studying Bantu-speaking histories, that distinction is essential: genes are not languages.


Three maps that will never match perfectly
Imagine three maps of the same region. The first shows related languages. The second shows pottery, crops, furnaces, and settlements. The third shows patterns of genetic ancestry.

They may overlap, but they measure different things.

A language can spread when people migrate. It can also spread when resident communities adopt it through trade, marriage, schooling, political influence, or daily usefulness. A pot can be traded without its maker moving. Genes can travel through one marriage even when neither community changes language.

Treating all three maps as identical turns a useful pattern into a false identity.


What recent DNA research found
In 2023, an international research team published one of the largest genomic studies of Bantu-speaking populations. It included 1,526 Bantu-speaking participants from 147 populations in 14 African countries, comparative samples, and genomes from 12 Late Iron Age individuals.

The results support substantial population movement beginning broadly 6,000–4,000 years ago in western Africa. Genetic diversity generally declined with distance from the western region in a pattern consistent with serial movements. Present-day Zambia and the Democratic Republic of Congo emerged as important crossroads in the researchers’ spatial models.

The same study also found significant gene flow from communities already living in the regions into which Bantu-speaking populations expanded. This matters. The genetic history is not a sealed line descending unchanged from one founding population. It includes encounters and mixture, with patterns varying by region.


A burial that complicated the story
Shum Laka in Cameroon lies within the probable homeland region of Bantu languages. Researchers recovered genome-wide data from four children buried there—two about 8,000 years ago and two about 3,000 years ago.

If geography, language, and ancestry were interchangeable, we might expect them to be direct representatives of the population later associated with Bantu-language expansion. Instead, their ancestry was closest to present-day western Central African hunter-gatherers and did not contribute substantially to most present-day Bantu-speaking populations in the models.

That does not make the linguistic homeland wrong. It shows that different populations occupied the same broad region over time. A site’s later language history cannot be projected automatically onto everyone who lived there earlier.


What DNA cannot recover
DNA can estimate biological relationships among sampled people and populations. It can identify mixture and model when groups may have separated. With ancient DNA, it can connect results to particular places and approximate times.

It cannot reveal a dead person’s spoken language directly. There is no “Kiswahili gene” or “Bantu gene.” Nor can a percentage assign cultural authenticity to a living person.

The evidence also has limits. Ancient DNA survives poorly in many tropical environments. Available burials are unevenly distributed and may not represent the wider population. Statistical models depend on comparison samples and assumptions. Modern communities have experienced thousands of years of movement since ancient events.

Researchers therefore compare genetics with linguistics, archaeology, history, and oral tradition. A genetic result is one witness, not the judge of every other source.


People are more than ancestry
The danger of genetic storytelling is that it can revive old racial thinking in scientific language. “Bantu” has sometimes been wrongly treated as a biological race. It is a language-family term applied to extraordinarily diverse communities.

People inherit DNA from parents, but they learn language through human relationships. They can become multilingual, stop speaking an ancestral language, or join a community through shared life. Identity is historical and social as well as genealogical.

Genetics makes the story richer when used carefully. It shows large movements and intimate encounters. It reminds us that communities met, formed families, and changed one another. Its greatest lesson is not that DNA defines people. It is that human history has always been connected.


Sources and further reading
• The 2023 Nature study presents its sampling, methods, migration models, and evidence of local gene flow.
• The Shum Laka ancient-genome study demonstrates why geography and later language history cannot identify ancient ancestry automatically.
• A peer-reviewed review examines Bantu-speaker migration and admixture with Khoe-San communities in southern Africa.
• “Moving Histories” explains why a language-family tree should not be mistaken for a complete history of people.