The question is, Why is there a limit at all? One scientific model of these working memory limits suggests that items in memory begin to compete with each other, endangering the mind’s ability to keep any one item in clear view. Scientists don’t really know why there’s a limit; it doesn’t appear to confer any evolutionary advantage. However, it bolsters the conclusion by linguist Peter Skehan that talented language learners (like C.J. and Christopher) are “memory-driven learners.” They can put a lot of things into memory and retain them. They can also retrieve them efficiently without mixing them up. It doesn’t explain how Mezzofanti was able to maneuver so sprightly among his languages, though. Perhaps there are undiscovered individuals out there with more powerful working memories.
“I don’t know many women who collect stamps or coins,” Alexander said to me on one of my visits. He wanted to know if I had ever considered polyglottery as a kind of collecting behavior, perhaps an obsessive one. Maybe it would explain why so many hyperpolyglots were men.
Only one famous hyperpolyglot that I’d read about, George Henry Borrow (1803–1881), who had studied forty-two languages, seemed to fit the profile of someone with obsessive-compulsive disorder (OCD), a psychiatric illness that affects about 1 to 3 percent of adults. Borrow had to touch a series of mundane objects in the correct sequence; otherwise, he feared, something would happen to his mother.
Other hyperpolyglots had a touch of this, too. Alexander keeps records—overdetailed ones, some might say—and became visibly agitated if he hadn’t put in time on his languages. There was also Elihu Burritt’s rigid accounting of studying and blacksmithing, and Christopher, and Krebs. Certainly there were care and focus, but none of them was crippled by a compulsion.
I also didn’t meet a hyperpolyglot who resembled a chronic hoarder, who collect such huge masses of worthless items (newspapers, food, scrap metal, car parts, matchbooks) that it interferes with their daily lives and their families. They do take pride in the grammars and dictionaries they amass. Yet stacks of books, as a mere fact, point to bibliophilia, not hyperpolyglottery. The hoarders I read about in the research literature can’t turn away from their junk long enough to have a normal life.
Why are there more male hyperpolyglots? One answer is that speaking a lot of languages is a geek macho thing. In addition to my survey of hyperpolyglots, I had one set up for monolinguals, too. By chance, perhaps, this one was answered mostly by women, more than 30 percent of whom said they’d studied three languages or more, though the survey asked for people who spoke only one. It seemed that a woman is less likely to say she “speaks” or “knows” a language if she studied it at some point in the past, while a man, wanting to display his giant repertoire, would include it.
The Geschwind-Galaburda hypothesis interested me the most: the idea that because male hormones in the fetus affect the developing brain, the effects of asymmetrical development of brain hemispheres would be seen mostly in males. Females also have male hormones, but they have fewer and would be less affected.
Male brains. Hormones. This brought me directly to the doorstep of the very thing I’d avoided all along.
Does the hyperpolyglot neural tribe overlap with the autistic population? Like many good questions, it has its traps. After all, there have been some high-profile autistic savants who’ve performed impressive language feats. Daniel Tammet, a writer and educator with high-functioning autism, once was challenged to learn Icelandic in two weeks and then went on Icelandic television to speak it. Intriguingly, Karl Zilles had mentioned that Emil Krebs seemed like someone with Asperger’s. Yet I didn’t want to get caught up in people’s medical histories, making diagnoses I wasn’t qualified to make. Nor did I want to follow the fashion of seeing autism in every eccentric’s biography.
I thought I’d be able to recognize someone with autism or Asperger’s syndrome fairly easily—someone who seems very socially awkward, with flat affect, who demands routine and fears deviations, and who might be able to perform brilliantly in some area, such as mental calculations. Admittedly, I had gotten this notion from the movie Rain Man. The autistic character that Dustin Hoffman plays, Raymond Babbitt, was in fact based on a real-life savant named Kim Peek, but no hyperpolyglot I spoke to resembled either Peek or Hoffman’s portrayal of him.
Yet they could be called “neuroatypical.” And part of their neuroatypicality might come from something shared with autism, particularly high-functioning autists. British psychologist and autism expert Simon Baron-Cohen has argued that autism represents the extreme form of a cognitive style that is adept at, and given to, “systemizing.” When someone systemizes, she (or, more likely, he) is watching inputs and outputs to a system, relating the two, and observing how they vary. Baron-Cohen defines systemizing as an attribute of the “male brain,” which more biological males have (he acknowledges that biological females can also have male brains). Hence Baron-Cohen’s “extreme male brain” theory of autism.
Perhaps a tendency to systemize would help explain why scientists score higher than nonscientists on a test that measures autistic traits. It might also explain why mathematicians, physical scientists, computer scientists, and engineers score higher than doctors, veterinarians, and biologists. Baron-Cohen has also found that autism occurs more frequently in the offspring of physics, mathematics, and engineering students than it does of those of literature students.
Baron-Cohen had also done some relevant work on the obsessional interests of children with autism, autism spectrum disorders, and Asperger’s syndrome. He hypothesized that systemizers would find mechanical systems more interesting than social systems. Or as he put it, they would be more readily interested in “folk physics”—a commonsense knowledge about how objects and systems behave in the world—than in “folk psychology.” In a survey of children with Tourette’s, autism, or Asperger’s, the autistic children were more often obsessed with machines, vehicles, physical systems, computers, astronomy, building, spinning objects, and lights than with beliefs, crafts, food, or sports. They were also more obsessed with folk physics than the kids with Tourette’s.
Could language count as an obsessional interest? Baron-Cohen asked parents whether their kids engaged in “echoing, collecting words, phrases, and learning languages.” Only a quarter of the children had this as an obsessional interest, about the same number as those obsessed with sports and games. The desire to make lists or “taxonomies” was three times as large; and surprisingly, only 35 percent of these systemizing children were interested in mathematics and numbers.
This is a small point about people with autism. But it sheds some light on the sort of brains that might be extraordinary at learning languages.
Chapter 17
One of my visits to Berkeley to see Alexander Arguelles coincided with a conference on brain mapping in San Francisco. As I rode the train across the bay from Berkeley, I hoped that I might find something there to help me connect what I knew about hyperpolyglots with what others knew about brains. At the conference, I was supposed to meet up with Susanne Reiterer, an Austrian neuroscientist pursuing the neurological basis of what she calls “phonetic language talent.” Her specialty is a gift for mimicry: people who can “do voices,” parodists, actors. A vivacious brunette in dark-rimmed glasses, she described her own phonetic talent, and how her research is an attempt to explain it.
At her university in Germany, she and her research team have been comparing good mimics and bad mimics using psychological tests and brain imaging. Some Germans without any Hindi skills could trick native speakers of Hindi into believing they were themselves Hindi speakers. Each of these exceptional mimics had strong verbal ability, good working memory, and a sophisticated ability to discriminate between musical tones and rhythms, particularly in singing.
She told me that before the study, they had anticipated finding a lot of people like author Joseph Conrad—someone who adopted a new language as an adult, but despite excellent grammatical abilities always spoke with a thick acce
nt.
“We did not find too many clear-cut Joseph Conrads!” she told me. Instead, she observed cognitive trade-offs: someone who is particularly talented in acquiring grammar and words may be a poor mimic, for example. She also found something distinct about the successful mimics’ brains. When she did fMRIs, she found that the talented mimics had lower levels of activation in brain regions related to speech—in essence, their brains didn’t have to try very hard because they used oxygen supplies efficiently. By contrast, the mimics who couldn’t fool native speakers (and who were presumably less talented) used oxygen less efficiently—those regions had to work harder to produce speech.
Interestingly, the good mimics’ brains were more efficient whether they were speaking German (their native language) or producing English, Tamil, or Hindi sounds. And the less talented mimics used glucose and oxygen less efficiently when producing either language. Reiterer suspects there may be a structural difference in the neural pathways of the more talented brains that leads to an enhanced connectivity among various parts of the brain during thinking tasks. In turn, this connectivity allows the neural circuits involved in language processing to work more efficiently.
Reiterer’s work addresses one component of language aptitude, the ability to hear and produce sounds. One specific brain area that may be involved in this is the primary auditory cortex (also known as Heschl’s gyrus). On the brain-as-globe model, it’s located right around India. Neuroscientist Narly Golestani has studied the brain structure of phoneticians—those who work with speech sounds in a variety of languages that they don’t necessarily speak—and found their primary auditory cortexes to be anatomically more complex than those in non-phonetician brains. Specifically, their cortexes have more finger-like convolutions, or gyri, made of white matter, which give them more surface area.
Unlike other kinds of brain differences (such as the arrangement of cell bodies in Krebs’s brain), it isn’t likely that this one can arise through practice or training; at least, no one has observed the human brain growing convolutions in this region after birth. This may explain why some people are more likely to take a job in phonetics. And it may further explain why some people find more pleasure in listening to foreign languages than others. In a previous study, Golestani looked at the brain structures of people who learned the sounds of a foreign language more quickly than other people. The faster learners’ brains had larger left Heschl’s gyri, due to more white matter.
Another neural signature of highly proficient language learners has been located in the left insula, which lies somewhere (to use the brain-as-globe metaphor) under the Arabian Sea. While the insula is long regarded as a mysterious zone dealing with bodily functions, the era of brain scanning has discovered a new role for it: a control center for emotions, consciousness, and working memory. The left insula has been shown in fMRI studies to activate more strongly in bilinguals who have equal abilities in their two languages.* In people whose languages aren’t equal, this area shows weaker levels of activation.
The left insula plays a key role in what’s known as “subvocal rehearsal.” One example of this rehearsal is the automatic process of having a foreign-sounding word in your head before you actually say it. Thus, more active neural circuitry in this area might engrave new sounds into the brain more quickly or durably. “The successful engagement of such neural circuitry,” writes the lead researcher, Michael Chee, “may correspond to vocabulary growth.”
Hence, Alexander’s success with his shadowing technique. Shadowing involves trying to pronounce the sounds of a foreign language (usually from a recording) at the same instant that one hears them. I’ve tried it and can say that the effect is intense, and very different from the “listen and repeat” format. Shadowing may exploit a mental routine for storing information called the “phonological loop,” which is the part of working memory devoted to speech sounds. Someone’s ability to automatically remember and accurately repeat nonsense words is a good predictor of foreign-language ability. The converse is also true: people with a disrupted or damaged phonological loop can’t learn new foreign words. If you shadow, you’re relying on the loop; shadow a lot, and you can build the loop’s strength.
Of the other components of aptitude—the ability to analyze grammar and the ability to learn new words—word learning has been more thoroughly studied, because more is known about where these abilities are located in the brain. Here, too, there are distinct neural signatures of higher performance. A team from Germany observed that people better at acquiring new words had more sustained activity in one part of the brain, the hippocampus (which is heavily involved in long-term memory). The authors speculated that brains vary in how they respond to the tasks of learning languages because there may be differences in the hippocampus, how the hippocampus attaches to the white matter, or what they called “genetic differences in neurotransmitter functions.”
In reality, no one would say that language talents reside solely in one area of the brain. Susanne Reiterer says that a talent for mimicry isn’t a “splinter skill”—her high performers had higher levels of performance in other language abilities, which suggests they’re tied together and involve many parts of the brain. The roles of the phonological loop and working memory suggest that higher-level cognitive skills play a huge role both in learning a sole extra language very well and in learning many languages. And since the biological basis of these skills is hereditary, whole families might belong to the hyperpolyglot neural tribe.
The idea of a genetic component to hyperpolyglottism is supported by evidence that cognitive capacities, such as working memory, executive function, and memory, as well as structural capacities such as plasticity, are hereditary. Back in 2004, when Dick Hudson sent news about N. and his family to the LINGUIST List, one of the respondents was Richard Sproat, a linguist now at the Oregon Health and Science University, who was intrigued by the possibility that language talent itself might be a heritable trait. Since the 1990s, scientists have linked language deficits to a genetic component, as in the case of the KE family, whose inability to produce certain grammatical expressions led to the location of a gene called FOXP2 in the mid-1990s, and a specific mutation of that gene in 2001.
But when it comes to an exceptional language talent, rather than a deficit, it is difficult to get families to sign up for a genetic study, perhaps because they don’t need to be cured of anything. Sproat exchanged a few emails with N., but then the replies stopped. When I contacted N. myself, he said he had discussed the issue with his family and did not want to be interviewed.
Before N. stopped writing, however, he did offer Sproat a few more details about his globe-trotting polyglot grandfather. “When we arrived in Thailand, I was sure he did not know any of the language,” N. said. But after two weeks his grandfather was arguing with market vendors in Thai. In the late 1960s, N. spent eighteen months in Thailand with the US military, where he learned some of the language. When he later tried conversing with his grandfather in Thai, N. said, “he was able to communicate on a higher level than I knew.”
N.’s disappearance is frustrating because in his original message, he mentioned another member of his family: a seven-year-old granddaughter. “She can count in three languages up to one hundred and she is able to pick out words spoken in other languages in public and tell you what they mean,” he wrote. N. and his hyperpolyglot family may have retreated from public view for now. But they, and others like them, could provide more fascinating insights into the language abilities we all have.
If the language superlearner’s brain has a kind of optimal design, could someone gain advantage by mimicking that design? The US military is particularly interested in neuroenhancements that target language learning. I came across a report that mused on the impact such enhancements would have on military forces and called for more research into improving a variety of cognitive abilities in adults, including learning multiple languages.
At the conference where I met Susanne Reiterer, I w
as surprised to learn about a technology called transcranial direct current stimulation (or, tDCS), which could be deployed widely tomorrow. Using a small device that’s simple enough to make at home (you need only a 9-volt battery, electrodes, and resistors), you can deliver very small amounts of electricity to areas of your brain through your skull. Depending on how you set the current, you can either stimulate or suppress how strongly neurons fire. If the electrode is attached to the positive part of the battery, the neurons are stimulated; if attached to the negative part, the neurons’ firing is suppressed. Electrical charges could help adults manage their brain plasticity by removing some of the brakes that curtail plasticity in adulthood.
In some initial studies to test the safety of this device, people who received currents from the positive part of the battery could recall new nonsense words 20 percent better than people who got sham currents or who received currents from the negative part of the battery. After a week, the positive effect had disappeared. In another study, people’s abilities to generate words that started with a particular letter increased by 20 percent as well. Direct current has also been shown to increase people’s visual memory by 110 percent.
Before you rush out to buy batteries and electrodes, you ought to know that the electrodes deliver electricity to the brain haphazardly. Biomedical engineers have tracked where this electricity flows in the skull, and it goes all over, flowing through eyesockets and pooling under the frontal lobe. A cranium full of sustained electrical charge could have a range of unpredictable effects.
Still, 20 percent increases are substantial, especially if they come after only a twenty-minute exposure to the device. Would anyone try it? I asked Helen Abadzi (who had an appetite for technologies that assisted her language learning) if she’d try tDCS. She replied that she’d worry about safety risks, especially since the improvements aren’t huge. If you want big improvements, she said, chew gum.
Babel No More Page 25