No two people build a vowel the same way. A new study using real-time MRI scans of people speaking found that speakers with differently shaped vocal tracts move their tongues, lips and jaws in noticeably different ways to make the same sound, and the researchers argue those small differences may help explain how humanity ended up with thousands of distinct languages.

The work, published Oct. 13 in PLOS Biology, comes from Steven Moran and Tiena Danner of the University of Neuchâtel in Switzerland and Christoph Zollikofer and colleagues at the University of Zurich. The team proposes that natural variation in anatomy within any group of speakers plants the seeds of sound differences that, over many generations, can grow until groups no longer understand one another.

The finding reaches every reader who has ever noticed that a friend, a parent or a newscaster says a familiar word just a little differently. Linguists have long treated those individual quirks as background noise. This study suggests they may be part of the engine that drives language change.

Watching Vowels Form Inside the Head

The researchers used recordings from the USC 75-Speaker Speech MRI Database, an open collection built at the University of Southern California in Los Angeles that captures the inside of people's mouths and throats in real time as they talk. The team analyzed 32 speakers of American English, 16 men and 16 women, as they produced five vowels.

Real-time MRI works much like a video camera pointed at the inside of the head. It records the tongue, lips, jaw, soft palate, and throat as they move, frame by frame, without the need for sensors placed in the mouth. That makes it one of the few ways to see exactly how a speaker shapes each sound, rather than inferring the movements from audio alone.

The scans showed wide variation in vocal tract shape from person to person. Speakers compensated for those differences by using very different ranges of motion to reach the same target sound. Each person still produced a recognizable vowel, but the anatomy-driven differences also left an individual acoustic signature in their voice.

"Every person has a slightly different vocal tract and speakers compensate for these differences," the authors said in a PLOS release.

From Individual Quirks to New Languages

To test whether the pattern was unique to English, the team compared their results with speech data from Basque and Hindi, two languages from unrelated families. Variation within each of those languages turned out to be similar to the variation within American English, suggesting a baseline level of speaker-to-speaker variation exists in every language.

From there, the researchers built what they call a biocultural model. Within any community, people pronounce sounds slightly differently because of their anatomy. Those small, individual differences can spread as people learn from one another, and when groups separate, their sound systems gradually drift apart. Given enough time, the drift can produce languages that are no longer mutually understandable.

The authors say their results provide researchers with a null model, a baseline describing how much variation should exist naturally before any cultural or historical forces push a language in one direction. That baseline could help linguists tell ordinary anatomical variation apart from genuine sound changes in progress.

A Small Sample with Big Questions

The study has clear limits, and the authors acknowledge them. It is an observational study built on a small set of vocal tracts, and it compared only a few language families. The MRI data came from American English speakers, while the cross-language comparison relied on acoustic speech data rather than scans of Basque or Hindi speakers.

Just as important, the study does not show that anatomy alone creates new languages. History, migration, contact between groups, social identity, and many cultural factors all shape how languages change. The researchers propose that anatomical variation supplies some of the raw material, not that it determines the outcome.

The work was funded by the Swiss National Science Foundation, and the authors declared no competing interests, according to the release.

Variation Is the Rule, Not the Exception

For readers, the most practical takeaway is a change in perspective. Differences in how people pronounce the same word are normal and rooted partly in biology, not a sign that one speaker is right and another is wrong. That idea is relevant for speech therapy, language teaching and voice technology, where systems trained on a narrow range of voices can struggle with speakers who sound different.

The finding also fits a broader pattern in evolutionary science, in which small differences between individuals become the starting point for larger differences between populations. Biologists see the same principle in how species diverge, and the researchers argue language may follow a similar path.

Roughly 7,000 languages are spoken today by common estimates, and scientists still debate how so many arose. No single study will settle that question. This one adds a specific, measurable mechanism — the physical shape of the speaking body — to a list of explanations that has mostly focused on history and culture.

Next steps would involve larger samples, MRI data from speakers of many more languages, and long-term tracking of how pronunciation spreads within communities. Those studies could test whether the anatomical seeds the team describes actually grow into lasting sound changes. The bottom line is that the voice inside each person may be a small but real ingredient in the story of why humanity speaks so many tongues — a hypothesis now ready for wider testing.

What Readers Want to Know

What did the study find?

Using real-time MRI scans of 32 American English speakers, researchers found that people with differently shaped vocal tracts move their mouths and tongues differently to make the same vowels, leaving a personal acoustic signature.

How could that lead to new languages?

The researchers propose that small anatomical differences within a community create natural variation in pronunciation. When groups separate, those differences can build up over generations until the groups no longer understand each other.

Where did the MRI data come from?

From the USC 75-Speaker Speech MRI Database, an open dataset created at the University of Southern California in Los Angeles.

Does this mean anatomy causes languages to split?

No. The study proposes anatomy as one source of variation. History, migration, culture and social factors also shape how languages change.

What are the study's limitations?

The sample of vocal tracts was small, and only a few language families were compared. The MRI scans came only from English speakers.

Who conducted the research?

Researchers at the University of Neuchâtel and the University of Zurich in Switzerland. The study appeared in PLOS Biology.