It is a shame that grammar leaves no fossils behind. Few things have been more important to our evolutionary history than language. Because our ancestors could talk to each other, they became a powerfully cooperative species. In modern society we are so submerged in words--spoken, written, signed, and texted--that they seem inseparable from human identity. And yet we cannot excavate some fossil from an Ethiopian hillside, point to a bone, and declare, “This is where language began.”
Lacking hard evidence, scholars of the past speculated broadly about the origin of language. Some claimed that it started out as cries of pain, which gradually crystallized into distinct words. Others traced it back to music, to the imitation of animal grunts, or to birdsong. In 1866 the Linguistic Society of Paris got so exasperated by these unmoored musings that it banned all communication on the origin of language. Its English counterpart felt the same way. In 1873 the president of the Philological Society of London declared that linguists “shall do more by tracing the historical growth of one single work-a-day tongue, than by filling wastepaper baskets with reams of paper covered with speculations on the origin of all tongues.”
A century passed before linguists had a serious change of heart. The change came as they began to look at the deep structure of language itself. MIT linguist Noam Chomsky asserted that the way children acquire language is so effortless that it must have a biological foundation. Building on this idea, some of his colleagues argued that language is an adaptation shaped by natural selection, just like eyes and wings. If so, it should be possible to find clues about how human language evolved from grunts or gestures by observing the communication of our close primate relatives.
This line of thinking raised an exciting possibility: Perhaps language left a fossil record after all--not in buried bones, but in our DNA. Yet for years biologists could not find a single gene involved in language.
Ten years ago, that finally changed. In 2001 a team of British scientists announced the discovery of a gene, called FOXP2, that seems to be essential for language. FOXP2 came to light through the study of a family that had unusual difficulties with words. The KE family--so called in scientific papers for privacy reasons--lived in West London and included nine siblings, some of whom attended the same special speech and language school. Psychologists at the school discovered that four of the children struggled with language in a similar way. The meaning of sentences sometimes confused them: They might misinterpret “The girl is chased by the horse” to mean “The girl is chasing the horse.” They also had trouble speaking--dropping some sounds off the beginning of words, for example, so that they would say “able” when they meant “table.”
In 1987 the school headmistress referred the case to the Institute of Child Health at University College London. There, neurologists found that some of the children’s cousins had the same language troubles, as did some of the parents. Geneticists traced the condition to a grandmother and deduced that she probably carried a rare mutation that she had passed along. The mutation did not alter intelligence or psychological well-being; the KE family was normal in those regards. Its effects were limited to language--but within that narrow sphere, its effects were profound.
The family then came to the attention of geneticists at Oxford, who began a dogged search for the gene that caused these problems. They compared the DNA of family members, looking for distinctive markers shared only among the ones who had trouble with language. Among those with the language deficit, they found shared markers in a single region of chromosome 7. Years later, the scientists received a vital new clue when the same kind of language disorder was identified in an unrelated 5-year-old boy. He had experienced a particularly dramatic mutation, in which a piece of chromosome 5 had been swapped with a piece of chromosome 7. One end of the boy’s swapped DNA lodged itself in the same region that the Oxford team had identified in the West London family, right in the middle of the FOXP2 gene.
The Oxford researchers turned back from the boy to the KE family and, using the additional information, discovered that those members with language troubles shared a mutation in FOXP2 as well. Their mutation was far more subtle, however. Their trouble with language had been caused by the change of a single nucleotide of DNA--just one letter in the genetic sequence.
All land vertebrates carry a version of the FOXP2 gene, so some of the Oxford researchers then teamed up with colleagues from the Max Planck Institute for Evolutionary Anthropology in Germany to analyze what is unique about the variant in humans and to track how the gene had evolved in our ancestors. They determined that after the gene arose, more than 300 million years ago, it barely changed in most branches of vertebrate evolution to the present day. In the human branch, however, two amino acids in the protein produced by the FOXP2 gene changed notably over the course of just a few million years. The scientists concluded that FOXP2 experienced a fast pulse of natural selection in our lineage, a development possibly related to the emergence of language.
Several groups are now hard at work gleaning more details about the relationship between FOXP2 and language. Cognitive neuroscientist Frederique Liegeois of University College London is using fMRI scans to compare the brain activity of members of the KE family who have a mutated copy of FOXP2 with those who have a normal version. The most striking difference, Liegeois recently reported, arises when family members are asked to repeat a set of nonsense words, something most adults can do without trouble. Those with the mutation do badly at the task. They also have low levels of activity in several regions of the brain, especially the basal ganglia, a key hub for learning muscle movements. That makes sense, since one of the hardest aspects of speech is learning how to make the necessary rapid movements of the lips, tongue, and vocal cords.
Other scientists are probing the FOXP2 gene further by studying the protein it produces, known as FOXP2. The protein seems to be especially active while human embryos are developing. Simon Fisher--one of the original Oxford geneticists, now at the Max Planck Institute for Psycholinguistics in the Netherlands--has found that the gene switches on in neurons within certain regions of the brain, including the basal ganglia. The FOXP2 protein then latches onto other genes in developing neurons and switches them on or off as well. By orchestrating dozens of genes, FOXP2 appears to oversee the growth of new branches on the neurons, bringing about a level of complexity likely to facilitate language.
Humans are not the only species to benefit from FOXP2. Researchers have shown that the gene is associated with vocal learning in young songbirds, which produce higher levels of FOXP2 protein when they need to learn new songs. If their version of FOXP2 is impaired, they make singing mistakes. Other vocal-learning species, such as whales, bats, elephants, and seals, may also rely on the gene. To probe this connection, geneticist Wolfgang Enard of the Max Planck Institute for Evolutionary Anthropology engineered mice by replacing their FOXP2 gene with the human one. The mice did not start reciting poetry, but they did display some subtle changes. Instead of producing a high squeak, for example, the engineered mice produced lower sounds. Bigger changes took place within the animals’ brains. Enard found that in the basal ganglia and connected regions involved in learning, the human version of FOXP2 caused some neurons to develop longer branches than those found in normal mice. Around the same time, Fisher and his team engineered mice so that one copy of their FOXP2 gene carried the same mutation as that found in the KE family. In subsequent tests, the mice with the mutation did a worse job than normal mice at learning new motor skills.
These findings hint at what happened to FOXP2 in our ancestors. It may have started out hundreds of millions of years ago as a gene that helped regulate the learning of body movements, such as those involved in running, calling, and biting. Later mutations in the gene spurred more neural growth in certain areas of the brain, including the basal ganglia, creating the connections essential for learning and mastering complicated sounds and, eventually, full-blown language.
FOXP2 didn’t give us language all on its own. In our brains, it acts more like a foreman, handing out instructions to at least 84 target genes in the developing basal ganglia. Even this full crew of genes explains language only in part, because the ability to form words is just the beginning. Then comes the higher level of complexity: combining words according to rules of grammar to give them meaning.
Language is nearly endless in its forms. So the search for its behavioral fossils--genes associated with grammar and syntax--should keep scientists busy for decades to come.
Copyright 2011 Carl Zimmer