http://www.nature.com/nsu/011004/011004-16.html

Language gene found

The first linking of a gene to language could speed our understanding of this most unique and most controversial of human abilities. 4 October 2001

JOHN WHITFIELD

Language problems run in the 'KE' family. Members of several generations speak "as if each sound is costing them their soul", one researcher has said. They struggle to control their lips and tongue, to form words, and to use and understand grammar. "To the naive listener, their speech is almost unintelligible," says geneticist Anthony Monaco, of the University of Oxford in England.

Researchers today unveil the single gene that, when it goes wrong, causes this speech breakdown. The gene - the first to be definitively linked to language - switches others on and off, and so could lead the way through a genetic network of language learning and use.

Finding one gene is like finding one part of a car. It looks useful, as though it's part of a larger mechanism. But we don't know what it does, what other parts it interacts with, or what the whole vehicle looks like. "It's an unbelievably complex system, and we've got one tiny glimpse," says Michael Tomasello, a psychologist at the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany.

We shouldn't have to wait long for more parts to turn up. Geneticists are on the trail of genes that control brain development and affect a range of mental disorders. The human genome sequence lets them do much of the groundwork on a computer, "saving what used to be months of work", says Robert Plomin, a behavioural geneticist at London's Institute of Psychiatry.

Forked tongue

The study of language divides researchers almost as starkly as languages themselves divide us. They disagree about whether language abilities are an innate feature of our biology or a product of our social interactions. Their opinions differ about whether the brain's language centres are specialized for these tasks alone, or are a part of our general mental machinery.

The controversy centres on theories first put forward by Noam Chomsky in 1959. That children learn to talk without instruction, and that adults construct an infinite number of new sentences from a finite number of words, convinced Chomsky that humans possess an inbuilt 'universal grammar' - a set of rules about the structure of language.

Forty years on, these ideas remain controversial. "You have to decide which side you're on - there's not much middle ground," says Bruce Tomblin, who studies the genetics of speech disorders at the University of Iowa in Iowa City. Tomasello, for example, believes that it is our ability to use abstract symbols that distinguishes humans from other animals, and is more likely to be genetically encoded in some way. Grammar, he says, "emerges historically - it's a sociological product, not genetic".

You don't need to believe in special language genes to believe, like Chomsky, in specialized, uniquely human language structures of the brain. "I don't think there are genes just for language, rather that genes build brain structures in such a way as to inform children what to expect," says Martin Nowak, who studies the evolution of language at the Institute for Advanced Study in Princeton. "It's impossible to learn language if we don't have a brain structure defined to expect it."

Family code

Family KE was first described in 1990. The way the disorder was shared out between the generations made it clear that just one gene was responsible, and the discovery was initially trumpeted as a 'gene for grammar'. When the breadth of the family's impairments became clear, there was a retreat from this claim - "I've heard it called the cold fusion of our field," says one psychologist.

The controversy still smoulders over whether the KE's symptoms have more to do with their inability to control their mouths, or some general brain problem, than with language centres. Supporters of a more purely linguistic interpretation of the family's difficulties point to the fact that the family's IQ, although below average, is within the normal range.

Monaco's team had been hunting the KE gene for several years. By 1998, they had pinned it down to an area of chromosome 7. Data from the Human Genome Project suggested that there were about 70 genes in this area. "We were marching down the chromosome," he says, using genetic markers to progressively narrow down the area that might contain the gene.

Two years ago, their march became a run. 'Patient CS', an unrelated boy with very similar difficulties to the KEs, turned up. Comparing the two allowed the researchers to stop their laborious rummage through chromosome 7 and zoom in on the gene. "It probably saved us a year or two," says Simon Fisher, another member of the Oxford team.

The same gene, called FOXP2, is damaged in the new patient and in the afflicted KEs. It belongs to a group that controls the activity of other genes by making a protein that sticks to DNA. The mutations in family KE and patient CS disrupt the DNA-binding area of the protein.

FOXP2 is "an important piece of the genetic puzzle of language", says psychologist Karin Stromswold of Rutgers University in New Jersey. But most language impairments are nowhere near as severe as those afflicting the KEs, and the patterns of inheritance in most families with language disorders are also more complex. The gene's "very messy" effects necessitate further studies of families with more limited impairments, cautions Stromswold.

Monaco's team is currently scanning the genomes of such families."I would be extremely surprised if the FOXP2 gene were a major determinant of more specific language impairments," he says.

FOXP2 is not unique to humans - it is switched on in the lungs and brain of mice. But subtle differences in its sequence or workings may illuminate why humans talk and animals don't, and how our ability evolved.

Ultimately, "we need to understand how genes give rise to brain structure, and how our brain structure gives rise to language", says Nowak. This job is just beginning: a full grasp of such processes is "50 to 100 years away", he says.

Shaking the tree

The network of language genes may be like a tree. Genes such as FOXP2 could be at the trunk - where sawing through them would knock out lots of aspects of language. Other genes might fine-tune aspects such as grammar further down the line; knocking these out would be analogous to lopping off a branch.

Psychologist Heather Van der Lely, of University College London, subscribes to this school of thought. She studies children whose speech and understanding of individual words are fine, but who, like normal adults learning a foreign language, are unable to master grammar. Such children muddle their tenses, saying 'yesterday I jump the fence', for example, and struggle to phrase questions.

"You have to explicitly teach them the rules of language," says Van der Lely. "They never have an intuitive knowledge - they always have to stop and work it out." These are the kind of 'pure' language deficits Stromswold wants gene-hunted. They lead van der Lely to believe in specialized grammar circuits in the brain, and genes to control their development.

Unsurprisingly, not everyone agrees. "It's hard for me to believe that we have genes devoted to influencing the brain in very specific ways that affect language and only language," says Tomblin. He thinks speech emerges from "general-purpose cognitive mechanisms, some of which may be more important for language than others. It's a less tidy view of things, but as I see the data, it looks more tenable."

Even apparently pure language disorders may be caused by complex interactions of many factors, warns Plomin. He believes there may be lots of different ways - genetic errors or environmental insults - to reach the same end language problem.

Sliding scale

People differ widely in their linguistic ability and behaviour - the age at which they begin speaking, for example, and the speed with which they master language. Plomin says that language development is probably controlled by "many, many genes, each with a small effect, working in many bits of the brain". Rather than language being something that you've got or you haven't, says Plomin, all these genes conspire to place people somewhere on the scale of linguistic ability.

Plomin is involved in a study of 16,000 pairs of British twins. It has found a strong heritable component to language disorders, but individual genes are hard to pin down: "I'm optimistic, but progress has been a lot slower than people thought it would be," Plomin says.

The genes and brains of unusually gifted linguists, people who can speak many different languages fluently, for example, might also reveal other genetic contributions to language learning. This approach has been neglected, Stromswold says, but a "surprising number" of professional linguists are the offspring of other linguists. "Linguists who marry linguists should trot on down to their local genetics centre," she adds.

It would be particularly interesting if their brains didn't work so well in other areas. "I'd look for linguists who can't balance a cheque-book," Stromswold says.