Paternal Words, Maternal Sounds: Asymmetrical Gene–Language Co-evolution in Indo-European Languages

How languages are transmitted across generations has long been a central question in research on human migration and linguistic origins. One long-standing debate concerns whether language inheritance is more strongly associated with maternal or paternal ancestry. The mother tongue hypothesis proposes that languages are primarily transmitted along maternal lines, whereas the father tongue hypothesis argues that languages tend to follow paternal lineages because linguistic patterns often show stronger correlations with the Y chromosome than with mitochondrial DNA. For many years these two perspectives appeared to contradict each other. Research on Indo-European populations, however, suggests that the relationship between language and ancestry is more complex than either hypothesis alone implies.

To clarify this issue, researchers conducted a large-scale quantitative study involving 34 modern Indo-European populations, examining both genetic and linguistic data. The genetic component focused on haplogroups of the Y chromosome and mitochondrial DNA, representing paternal and maternal histories respectively. Rather than relying on traditional language family trees or classification schemes, the study analyzed two distinct linguistic dimensions: lexical systems and phonemic systems. These two components reflect different aspects of language evolution. Lexical data were drawn from a public dataset compiled by Michael Dunn, while phonemic information came from the PHOIBLE database, which catalogues the sound inventories of languages. Because vocabulary and sound systems often follow different evolutionary paths, separating them allows a more detailed investigation of how languages change over time.

Vocabulary tends to evolve alongside language divergence, as words are gradually replaced or lost during the branching of language families. Phonemic systems, however, may follow more complicated trajectories. Sound patterns can change through contact with neighboring languages, borrowing of phonemes, or spontaneous shifts within a speech community. As a result, phonemic structures often preserve traces of contact and interaction that lexical comparisons alone may fail to capture.

To compare genetic and linguistic patterns across populations, the researchers applied the Neighbour-Net algorithm, which reconstructs relationships among groups in the form of networks rather than simple branching trees. This approach can reveal signals of historical mixing, borrowing, or parallel evolution. The network generated from Y-chromosome data showed extensive reticulation, indicating that male populations across Indo-European regions experienced frequent historical contact and admixture. By contrast, the mitochondrial DNA network displayed a much clearer geographic division between eastern and western lineages, revealing a broad maternal structure separating Indo-Iranian and European populations.

When linguistic data were analyzed, lexical relationships formed a network that closely resembled a branching tree, consistent with traditional classifications of Indo-European languages. The phonemic network, however, showed numerous conflicting signals and lacked clearly defined monophyletic groups. This pattern suggests that phoneme systems are particularly susceptible to contact-driven change and may not preserve simple genealogical relationships among languages.

Neighbour-Net analysis results: (a) Y chromosome; (b) mitochondrial DNA; (c) lexicon; (d) phonemes（Image source：Zhang M et al. (2018)， CC BY 4.0 ）

The researchers then applied Mantel tests to measure correlations among genetic distances, linguistic distances, and geographic distances between populations. At first glance, both genetic and linguistic differences appeared strongly correlated. Yet such correlations can arise simply because geographically distant populations tend to differ both genetically and linguistically. To account for this, the analysis controlled for geographic distance using a partial Mantel test.

Results of the partial Mantel test. Paternal lineages are positively correlated with lexicon, and maternal lineages are positively correlated with phonemes. No correlation is observed for the remaining comparisons（Image source：Zhang M et al. (2018)， CC BY 4.0 ）

Once geographic effects were removed, a striking pattern emerged. A significant positive correlation remained between Y-chromosome distances and lexical distances. In other words, differences in paternal ancestry aligned closely with differences in vocabulary among Indo-European populations. At the same time, a separate significant correlation appeared between mitochondrial DNA distances and phonemic differences. Conversely, paternal ancestry showed no meaningful relationship with phonemic patterns, and maternal ancestry showed no relationship with lexical variation. Even paternal and maternal genetic distances themselves were not significantly correlated once geography was controlled.

These findings indicate that language inheritance does not simply follow either maternal or paternal lines as a whole. Instead, different components of language appear to be influenced by different ancestry patterns. Vocabulary shows stronger alignment with paternal genetic history, while phonemic systems correspond more closely with maternal ancestry.

The researchers proposed a historical explanation grounded in patterns of human migration and social structure. In many historical scenarios, male mobility was higher than female mobility. Warriors, traders, or colonists often migrated into new regions where local populations already existed. Because such male newcomers frequently held higher social prestige, the language they brought with them could become dominant within the new community. In such contexts, the vocabulary associated with the incoming population would spread widely, producing a link between paternal genetic lineages and lexical systems.

At the same time, local women who married immigrant men often remained rooted in their native communities. These women might adopt their husbands' language in daily communication and teach it to their children. However, when learning a second language, speakers frequently retain aspects of their original pronunciation. Linguists describe this phenomenon as language fossilization, in which elements of a learner's first language persist within their speech. Under these circumstances, women could adopt the vocabulary of their husbands' language while continuing to pronounce words with phonetic features influenced by their native language.

Children raised in such mixed households would therefore acquire a language whose vocabulary largely followed the paternal linguistic tradition, but whose phonetic characteristics reflected maternal pronunciation patterns. Over generations, this process could produce the asymmetric pattern observed in the data: lexical evolution associated with paternal ancestry and phonemic variation associated with maternal ancestry.

In this way, the study offers a reconciliation of the father tongue and mother tongue hypotheses. Rather than treating them as mutually exclusive explanations, the results suggest that both processes operate simultaneously but affect different layers of language structure. Vocabulary behaves more like the linguistic inheritance of fathers, while sound systems retain stronger traces of maternal influence.

It is important to emphasize that these conclusions arise from statistical patterns observed at a broad population level. They do not imply that every society or every individual family transmits language in the same manner. Cultural practices, social hierarchy, migration patterns, and marriage systems can vary widely across regions and historical periods. As a result, the balance between paternal and maternal influences on language may differ from one context to another.

Future research could deepen this understanding by examining language transmission at finer scales. Studies of early childhood language acquisition, family language environments, and social stratification could illuminate how linguistic features are passed from parents to children in everyday life. The analytical framework used in this research may also be extended beyond language to other cultural traits such as music, ritual practices, or political organization, offering broader insights into the intertwined evolution of human genes and culture.

Author: Shui-Ye You

Reference:

Zhang M et al. (2018). Reconciling the father tongue and mother tongue hypotheses in Indo-European populations. National Science Review.

（Paid content. Unauthorized reproduction or use is prohibited.）