The First Word: The Search for the Origins of Language

by Christine Kenneally

  10. You have a human brain

  On July 28, 2005, Lacy Nissley was scheduled for neurosurgery at Johns Hopkins Hospital in Baltimore. Before she was born, the neurons in Lacy’s right hemisphere migrated to the wrong place in her brain. The hemisphere became enlarged and started to cause seizures that were only poorly controlled by medication. As time went on, Lacy’s seizures got worse. Nothing could be done to make her right hemisphere work well, and while it was attached to the rest of her brain, it corrupted the way the left hemisphere worked. The only chance Lacy had to live a normal life was for her to undergo a hemispherectomy. In this radical operation, Lacy’s neurosurgeon would remove her right hemisphere, essentially taking out half of her brain.

  Four hours into the operation, Lacy’s neurosurgeon, Dr. George Jallo, his resident Dr. Violette Renard, and the OR nurse Sean Stelfox stood in a small, still crescent around Lacy’s head. Earlier, Jallo had removed the frontal lobe. He then used micro-scissors to cut around the parietal lobe, and now he and Renard were slowly working their way around each side, making tiny little pinches into the cut with electric cauterizing forceps. Occasionally, Jallo used a flat metal spatula to lift the lobe up and back so he could push the bipolar forceps farther in. As the cut became deeper and wider, the tissue on either side browned and blackened, and the lobe, which was initially stationary, started to move back and forth as more of it was detached from the rest of the brain.

  Deep at the bottom of the parietal wedge lay the white matter of Lacy’s brain. Everything else was colored or discolored, but the long cables that connect neurons to one another gleamed toothpaste white. They came apart like string cheese. Stelfox bent toward Jallo clutching a small plastic bowl with both hands. Using normal forceps, Jallo picked out the lobe—it was the size of an infant’s fist—and dropped it into the container. Stelfox held it aloft. “The parietal lobe.”

  Four hours later, the right hemisphere was gone.1 From the top of Lacy’s head, her cranium looked like a wide, uneven bowl, revealing the white-pink base of the skull from the inside and the larger, deeper cavity that had held the frontal and parietal lobes. In the middle was a shallow mound where Jallo left a layer of axons to protect the ventricle. The white matter was now gray-black. Jallo and Renard lightly touched their forceps to it, and the cauterizers fizzed, and occasionally popped and spluttered, sealing the brain against micro-hemorrhages. Just below the mound were the basal ganglia, small dark squiggles in the emptiness. Over and over Stelfox poured in saline, and Jallo and Renard drew it out again.

  Jallo filled the right side of Lacy’s head with saline, and over the next few days it would be replaced by the brain’s constant drip of cerebral spinal fluid. He then reattached her skull using four tiny dissolvable plates made of sugar. Overall, the hemispherectomy took nine hours, and at the very end Renard bandaged Lacy’s head and gently turned her onto her right side, sticking on tape that said “This side up.”

  Lacy was released from the hospital a week later.2 Around one hundred children have undergone a similar procedure at Johns Hopkins, and with extensive therapy to help them relearn how to walk, talk, and think, the overwhelming majority of them have flourished.

  Hemispherectomies are a drastic but necessary operation for a small group of people, most of them children. Faraneh Vargha-Khadem, a professor of Developmental Cognitive Neuroscience at the University College London Institute of Child Health, has followed up on a large number of children who have undergone hemispherectomy. Her best-known case was Alex, a young boy whose left hemisphere was removed when he was eight and a half years old. Alex was virtually mute before the surgery, and his comprehension of words had developed only to the level of a four-year-old. But around ten months after his left hemisphere was taken out and his antiseizure medication was withdrawn, he began to speak first in single words and later in phrases and then in sentences. Even in the normally dry tones of science journals, you can perceive the researchers’ surprise. “To our knowledge,” they wrote, “no previously reported child has acquired a first spoken language that is clearly articulated, well-structured and appropriate after the age of six years.”

  How can a brain do such a thing? At this point in human evolution, there are so many neurons in our brains that the potential number of connections between them is thought to be around 500 trillion. We’ve had these enormous brains for about 200,000 years, and it took us almost all this time (190,000 years) to start opening our skulls and interfering with them. It took another 9,900 years to really start working out how the brain functions. Since 1990 the neuroscience of language has run a course similar to that of animal cognition and language evolution in that it has undergone revolutionary changes. Our picture of language in the brain since then has been transformed almost beyond recognition.

  Nothing in the traditional view of how the brain and language function could account for Lacy and Alex. A skeptic might argue that Lacy can talk because her right hemisphere was removed—scientists used to believe that language was located almost entirely on the left side of the brain. But if that were the case, Alex would be forever mute. Indeed, in the last few decades, a number of children have demonstrated that they are able to talk after removal of the left hemisphere. Most of them suffer some kind of deficit, but their language is more than good enough to enable them to get by in the world.

  In the past the only way to deduce the workings of the brain was through the successes and mistakes of primitive neurosurgery and “experiments of nature,” cases where unfortunate individuals suffered brain damage from some kind of accident. Observers were able to determine the damage postmortem and then plot in a crude way how it had affected behavior and thinking while the victim was alive.

  Phineas Gage is the best-known case study in accidental neuroscience. Gage was a railroad laborer, and in 1848 the inadvertent sparking of some gunpowder sent a bolt of iron shooting through his brain. He survived, but his personality changed completely. He became surly and difficult and struggled with decision making and planning. Gage’s state before and after his injury revealed a great deal about the role of the frontal lobes in the workings of the brain.

  Today magnetic resonance imaging and positron-emission tomography allow scientists to peer inside a normal living brain and see how it works in real time. Electroencephalograms, another useful technology, measure the electrical waves that are naturally emitted by the brain. These brain waves change in response to different input, which in a language experiment might include normal and ungrammatical sentences. More recently, neuroscientists have developed a way to keep neurons alive for days at a time in petri dishes. The researchers stimulate the neurons in different ways and watch how they respond.

  In the traditional phrenological model, different talents and tendencies existed within separate compartments of the brain, and for a long time people assumed that much of the evidence from brain damage suggested that language existed within specific spaces. But as knowledge about the workings of the brain accumulated, the idea that only one particular part of it was devoted to language progressively weakened and finally was rejected. No neuroscientist has found any specific area or tissue that controls language and language only. There are no obvious neural add-ons in the human brain, and of all its cell types there isn’t one that only humans have.

  As recently as twenty years ago it was taught that language specifically resided in Broca’s and Wernicke’s areas on the left side of the brain. It’s hard to even imagine now how confidently that belief was held, because as we know today, language function is spread throughout the brain. According to Fred Dick, a senior lecturer in psychology at Birkbeck, University of London, all the laboratories that have tried to find a language area have been successful in that they have indeed found dozens, even hundreds, of them.3

  If you look for activation in any cortex, when language is spoken or comprehended, you will find it. Lieberman’s studies of Parkinson’s patients and Everest climbers, as well as Pinker’s work on the past tense i
n English, show that there is an overlap between the parts of the brain that are used for speech and the parts that are used for syntax. In addition, the brain areas that are active when learning language are different from the ones that are active when using language once it has been learned. Moreover, different areas are activated depending on the specific language activity, like the comprehension of words, categorizing a word (in a new task versus a learned task), translating between languages, or making decisions about grammar.4 Modern brain imaging has also revealed that the spread of language activation across the two hemispheres of the brain can differ substantially for each individual.5

  Clearly, there is no one-to-one correspondence between an area in the brain and all language ability. Although the brain does contain identifiable areas, complicated behaviors are underwritten by many different groups of neurons, and these are linked together to form circuits.6 The activation that takes place within a small, identifiable part of the brain is often a part of a much larger circuit of activation that is distributed throughout the brain. Walking, striking a piano key, speaking, and listening to speech arise from these large neural circuits.7

  Summing up our understanding in 2002, Lieberman wrote: “Although our knowledge is at best incomplete, it is clear that many other cortical areas [other than Broca’s and Wernicke’s] and subcortical structures form part of the neural circuits implicated in the lexicon, speech production and perception and syntax, and the acquisition of the motor and cognitive pattern generators that underlie speech production and syntax.” He lists the cerebellum, the prefrontal cortex, frontal regions of the cortex, posterior cortical regions, the anterior cingulate cortex, and regions of the brain traditionally associated with visual perception and motor control.”8

  The belief that language was located in the left hemisphere was based primarily on the fact that when people suffered damage to Broca’s area, the aphasia they experienced appeared to destroy a lot of grammatical knowledge. But the data are inconclusive, and as Elizabeth Bates9 and Fred Dick have pointed out, people with Broca’s aphasia are still able to make certain types of grammatical judgments.10 In fact, it seems they retain a great deal of knowledge of their language’s grammar, but have trouble accessing it. Moreover, the symptoms of Broca’s aphasia have also been reported in other groups who do not have damage in that part of the brain. Dick adds that the problems that Broca’s patients have can be language-specific (though much of the original testing for Broca’s was done in English, the findings were thought to be true regardless of which language and syntactic system the subject used). While this doesn’t mean that Broca’s area isn’t important for language, it does show that it isn’t the only language-involved area of the brain.

  Not only are language and other higher mental abilities distributed throughout the brain, but Broca’s area has been shown to serve other functions as well.11 As Bates and Dick note, “Activation in Broca’s area is observed when subjects plan covert nonspeech mouth movements, make rhythmic judgments, or perform complex sequences with the hands and fingers. In fact Broca’s area is active when the subject merely observes such movements by another human being or reacts to static objects (tools) that are associated with such movements.”12

  None of this evidence against the language-is-a-box-in-the-brain model means that language is just a function of a homogeneous general intelligence. Bates explained: “There is no such thing as vanilla cognition…There are variations in computational style and computational power from one region to another, from one layer to another within a single region, and from cell to cell.”13

  Also, once a human brain has matured, the distribution of language functions across that brain is not random. Particular areas take on important parts of the overall task of perceiving and understanding language. It is widely accepted that different sides of the brain dominate in the processing of prosody (right hemisphere) versus syntax (left hemisphere). In 2005 Lorraine Tyler and colleagues published an experiment that compared the perception of verbs that were regular (“jump,” “jumped”) and irregular (“think,” “thought”) in their past-tense form. They demonstrated how the sound, meaning, and structure of a word all appear to be processed in different areas of the brain.

  Brain imaging showed that in the experimental subjects regular past-tense forms are processed by a neural circuit that includes the left superior temporal gyrus, Wernicke’s area, and connections to the left inferior frontal cortex.14 Irregular verbs, however, take a different path through the brain. It appears as if the stem and affix of the regular past-tense verbs are computed as the words are heard, but the irregulars, which have no special syntactic marking, are treated simply as whole words, like nouns or uninflected verbs. Accordingly, people who suffer brain damage have been shown to have trouble with one type of past-tense verb or the other—but not necessarily both.15 Fine-grained brain imaging reveals that even if parts of the brain, like Broca’s area, perform many nonlanguage functions, they may still be very important for specifically linguistic processing.16 Such findings underline yet again the way that what we experience as a single thing—language, words, tense—arises from an amalgam of more and less general strategies.17

  Dismissing the principles of phrenology doesn’t rule out the possibility that human children are born with some specialization for language. Those with particular types of brain damage do experience delays in acquiring language. The fact that these children are slowed down suggests that the damaged areas may have been particularly fertile ground for language acquisition before the damage. However, the same children often naturally catch up to a normal level of language use, also suggesting that there are mechanisms that help the brain to recover, to reorganize on the fly. So even if there are parts of the brain that are best suited for language acquisition from birth, other areas can sometimes step in if they fail. The way that a brain can take different routes to the same basic behavior—in this instance, turning language loss into language gain—is called plasticity.

  Brad Schlaggar, a pediatric neurologist and a professor at Washington University in St. Louis, says that the best way to think of plasticity is as a support structure. When he gives a talk about plasticity, he always shows students slides of the St. Louis Arch. “As the structure goes up,” he explains, “the relationship between the scaffolding and the leading edge of the two sides of the arch changes as they rise up to meet in the middle. The relationship between the scaffold and the emerging mature structure is dynamic, as opposed to a scaffold that surrounds a building and then comes down again.” So if damage occurs to the brain of a seven-year-old child, it occurs in a completely different context than if the child were much older or younger. “The scaffolding idea means that even in adults, the organization of the brain for learning a novel task or a challenging task is different from the organization of implementing that task once you have acquired the skill.” The scaffolding for language seems to be particularly flexible. Fred Dick describes the development of language as a moving target. If damage is sustained in one area, language may move, morph, and settle into another.

  In his doctoral work Schlaggar transplanted the visual cortex of one fetal rat brain into another, placing it in the spot where the somatosensory cortex, which normally controls the body as it moves through space, typically develops. Schlaggar found that the transplanted visual cortex grew into a fully functioning somatosensory cortex. The inputs into the new region came from the body as it moved in space, and as a result that neural tissue became wired to process that kind of information.

  We tend to think of the brain as developing on a completely separate trajectory from that of the body. Traditionally researchers imagined that the brain had some kind of central developmental controller instructing different parts to assume responsibility for different abilities (the visual cortex develops particular types of neurons, while the auditory cortex develops differently specialized neurons, and so on). But recent research has cast grave doubts on the existence of any kind of central
controller. It looks as if the brain tissue that ends up becoming part of different specialized regions is not necessarily fated to end up that way, and that input to the brain coming through the filter of the body contributes to its architecture.

  Leah Krubitzer, a professor of psychology at the University of California, Davis, also demonstrated how the immature brain isn’t fated to be mapped into the specific regions that are typical of the adult brain. She removed a big chunk of the brain of newborn marsupials, and then let them grow up and develop normally. After they reached adulthood, she took another look at their brains. The cortices had organized themselves into exactly the same areas as a normal brain would, all in the same spots relative to each other, but they were all slightly smaller, so as to fit within the smaller brain. While there is a default optimum map, it appears that the map can be drawn over different kinds of neural terrain.

  For all the apparent complexity of the human language-brain relationship, it’s important not to lose sight of the fact that some hard-to-pin-down behaviors and preferences appear to be completely controlled by the way genes have built the brain. In 2001, in a strange complement to the experiment in which chickens with transplanted bits of quail brain ended up producing some species-specific quail calls, Evan Balaban and colleagues at the Neurosciences Institute in San Diego transplanted a piece of brain from a Japanese quail into the brain of a domestic chicken, and likewise placed a piece of chicken brain into the head of a Japanese quail. With their new chimeric brains, the birds continued to produce the calls of their own species, but instead of responding to the maternal calls of their own species, they showed interest in the calls of the other.18 There’s no reason to believe that processes like these aren’t also relevant to the human experience, even if they can’t fully explain the complexity of language.