Posted September 17, 2014 by Bernard J. Baars in biology of language

Is ROBO2 really a language gene? Hmmm…

dna-polimerThe popular science media are wondering whether a genetic language code has just been discovered (near the gene ROBO on chromosome 3) … based on a correlation between a human gene variation and the number of words spoken by two-year-old children in Europe.

The original article is entitled: “Common variation near ​ROBO2 is associated with expressive vocabulary in infancy.” Beate St. Pourcain et al (2014)  Nature Communications 5,  Article 4831 doi:10.1038/ncomms5831

In scientific writing the phrase “associated with” means “correlated with.”

But as every science student knows, the number of days in the year is nicely correlated with the number of nights — proving nothing whatsoever about causality. Daylight does not cause darkness, even though they are perfectly correlated.

Scientists desperately try to avoid confusing correlation with causality, and the Medical Research Council team in the UK that made this finding does not make this basic error. But popular media are driven by headline pressures using very simple, stereotyped templates, like “MAJOR DISCOVERY X JUST MADE BY SCIENCE!!!

Unfortunately, the article shows no causal relationship between a small genetic change and language learning in little kids.

Good students of science can learn by comparing the popular headline with the original article — very easy to find these days by searching Google Scholar or

Then you can be the scientific thinker. Is this headline really true?

Searching PubMed — endowed with 20 MILLION scientific abstracts, free of charge — we immediately see that the ROBO2 gene is found in three species — humans, rats, and mice.

Furthermore, an analog of ROBO2 is found in the flatworm C elegans, a famous study organism in biology.

Question: Shouldn’t a language gene be unique to humans, since only humans possess language in the full meaning of that word? If researchers were to find the same tiny piece of genetic code in C elegans, shouldn’t flatworms be talking to each other?

The fact is that the popular headline tells us nothing new.

Language researchers have understood for half a century that human language has a genetic basis, because:

1. Spoken language is limited to homo sapiens. It is extremely hard to teach language to other creatures, but in humans it’s hard to prevent babies from learning language. They do  it all by themselves. Nobody really knows how they do it.

(We don’t really teach natural language – we just create the natural conditions for language to be learned.)

2. Language-specific brain damage is only found in humans. The most famous examples are Broca’s aphasia (impaired speaking) and Wernicke’s aphasia (impaired comprehension and poor production).

3. Major features of language, like syntax and a giant vocabulary (around 100,000 words) is only known in humans.

4. Human babies learn a rich phonology (producing speech sounds) very early in life. Pronunciation of one’s native language tends to “freeze” by puberty, after which it is hard to learn to speak a very different language like a native. Language learning seems to have a “critical period” when it happens at amazing speed, not like later on. Critical periods indicate a genetic window of opportunity.

5. Our vocabulary and its remarkably rich set of meanings (semantics) has no known parallels among other species.

HOWEVER, we do share some very basic speech-like behaviors with other species — like whale song and bird song, making emotionally charged sounds, and major evolutionary functions of sound production.

Baby baboons screech just like human babies. They do it for similar evolutionary reasons.

The human vocal system is amazingly complex, with hundreds of specialized muscle groups.

We spontaneously learn a subtle and very large range of sounds, including the famous “click phonemes” in African groups like the Khoi San — the “bushmen” of the Kalahari desert. Some anthropologists suggest that click languages may be the earliest surviving languages in the world.


Bottom line: There is a great deal to be learned from sophisticated genetic studies. The famous FOXP2 gene was discovered in an London family with a variety of speech impairments. The media celebrated that major language discovery, until …. FOXP2 was found in alligators, too. 

Some days you might as well stay in bed.

Bernard J. Baars