The process of automatization is tangible in the example of arithmetic. When children first learn to add 3 + 4, they count on their fingers, making the parietal cortex work hard. But at some point in the learning ‘three plus four is seven’ almost becomes a poem. Their brains are no longer moving imaginary objects or real fingers one by one, but going to a memorized chart. The addition has been outsourced. Then a new phase starts. Those same children begin to solve 4 × 3 in the same slow, laboured way–using the parietal and frontal cortex–‘4 + 4 is 8. And 8 + 4 is 12.’ Then they develop another way to out-source, automatizing the multiplication in a memory table in order to move on to more complex calculations.
An almost analogous process explains the virtuosos we looked at earlier. When chess masters solve complicated chess problems, what activates most differently is their visual cortex. We can summarize by saying that they don’t think more, but rather they see better.* The same thing happens with great mathematicians who, when solving complicated theorems, activate their visual cortex. In other words, the virtuoso managed to recycle a cortex ancestrally devoted to identifying faces, eyes, movement, dots and colours in order to utilize it in a much more abstract realm.
Automatizing reading
The principle we can infer from the triangle experts explains what is perhaps the most decisive transformation in education: turning visual scribbles (letters) into the voices of words. Since reading is the universal window on to knowledge and culture, this gives it a special pertinence over the rest of the human skills.
Why do we begin to read at five years old and not at four or six? Is that better? Is it best to learn to read by breaking down each word into the letters that make it up or the other way around, by reading the whole word and associating it with a meaning? Given reading’s pertinence, these aren’t decisions that should be made from a stance of it looks to me like, but rather they should be built on a body of evidence that brings together the experience of years and years of practice and a knowledge of the cerebral mechanisms that support the development of reading.
As in other realms of learning, the expert reader also outsources. Those who read poorly not only read more slowly; what most holds them back is that their system of effort and concentration is focused on the reading and not on thinking about what the words mean. That is why dyslexia is often diagnosed by a deficit in reading comprehension. But it doesn’t have anything to do with intelligence, rather simply with the fact that the effort is being put somewhere else. In order to be able to empathize with this, try to remember these words while you read the next paragraph: tree, bicycle, mug, fan, peach, hat.
Sometimes, when reading in a foreign language we’ve just started to learn, we come to realize that we’ve understood very little, because all of our attention was focused on translating. The same idea applies to all learning processes. When someone starts to study percussion, their focus is on the new rhythm they are learning. At some point that rhythm becomes internalized and automatic, and only then can they concentrate on the melody that floats above it, on the harmony that accompanies it or on other rhythms that are in a dialogue with that one simultaneously.
Do you remember the words now? And if you do, what was the last paragraph about? Successfully completing both tasks is difficult because each one occupies a limited system in the frontal and parietal cortex. Your attention chooses between juggling those six words so they don’t vanish in your memory or following a text. Rarely can it focus on both.
The ecology of alphabets
Almost all children learn language very well. As a mature adult, I arrived in France without being able to speak more than the simplest words of French and it seemed strange to me that a small child I met who knew nothing of Kant’s philosophy or calculus or the Beatles could speak French perfectly. It must surely have seemed strange to the same kid that a grown-up was incapable of something so simple as correctly pronouncing a word. This everyday example demonstrates how the human brain can display a mental virtuosity that has very little to do with other aspects we associate with culture and intelligence.
One of Chomsky’s ideas is that we learn spoken language so effectively because it is built on a faculty for which the brain is prepared. As we have already seen, the brain is not a tabula rasa. It already has some built-in functions, and the problems that depend on them are more easily resolved.
In the same way that Chomsky argued that there are elements common to all spoken languages, there is also a common thread to all alphabets. Of course, the thousands of alphabets that exist, many already in disuse, are very different. But if we look at them all together, we immediately notice some regularities. The most striking is that their construction is based on just a few hand strokes. Hubel and Wiesel won their Nobel Prize for discovering precisely that each neuron of the primary visual cortex detects strokes in the small window it is sensitive to. The strokes are the basis of the entire visual system, the bricks of its form. And all alphabets are built with these bricks.
There are horizontal and vertical lines, angles, arches, slashes. And when you count the most frequent strokes in all alphabets there is extraordinary regularity: those strokes that are most common in nature are also the most common in alphabets. This isn’t the product of a deliberate, rational design; alphabets just evolved to use a material that is quite similar to the visual material we are used to dealing with. Alphabets usurp elements for which our visual system is already fine-tuned. It’s like starting with an advantage, since reading is close enough to what the visual system has already learned. If we try to teach with alphabets that have no relationship to what our visual system naturally recognizes, the experience of reading would be much, much more tedious. And, the other way around: when we see cases of reading difficulties, we can ease that process by making the material to be learned into something more digestible, more natural, more easily consumed–something for which the brain is prepared.
The morphology of the word
New readers pronounce a letter as if in slow motion. After many repetitions, this process becomes automatic; the ventral part of the visual system creates a new circuit able to recognize letters. This detector is built by recombining previously existing circuits that identify strokes. And in turn these become new bricks in the visual system that, like Lego pieces, are recombined to recognize syllables (of two or three successive letters). The cycle continues, with the syllables like new atoms of reading. At this point, a child reads the word ‘father’ in two cycles, one for each syllable. Later, when reading is firmly established, the word is read in just one sweep, whole, as if it were a single object. In other words, reading goes from being a serial process to a parallel one. At the end of the process of reading, readers form a brain function capable of extracting most words–except for extremely long and compound words–as a whole.
How do we know that adults read word by word? The first proof is in a reader’s eyes, which move and stop on each word. Each one of these stops lasts about 300 milliseconds and then jumps abruptly and rapidly to the next word. In writing systems like English, that move from left to right, we focus very close to the first third of the word, since we mentally sweep from there to the right, towards the future of reading. This very precise process is, of course, implicit, automatic and unconscious.
The second proof is in measuring the time it takes to read a word. If we read it letter by letter, the time would be proportional to the length of the words. However, the time a reader takes to read a word made up of two, four or five letters is exactly the same. This is the great virtue of parallelization; it doesn’t matter if there are one, ten, a hundred or a thousand nodes to which we must apply the operation. In reading, this parallelization has a limit in very long and compound words, such as sternocleidomastoid.* But within a range of two to seven letters reading time is almost identical. On the contrary, for someone who is just learning to read or for dyslexics, the reading time increases proportionally to the number of letters in a word.
; We saw that the talent students have when starting their studies is not a good predictor of how good they will be after many years of learning. We now understand why.
In France, based on the finding that expert readers read word by word, one group concluded–erroneously–that the best way to teach children how to read was through holistic reading which, instead of starting by identifying the sounds of each letter, starts by reading entire words as a whole. This method quickly spread in popularity, probably because it has a good name. Who doesn’t want their child to learn with the holistic method? But it was an unprecedented pedagogical disaster that led to many children having reading difficulties. And, with the argument sketched here, you can understand why the holistic method didn’t work. Reading in a parallelized way is the final phase that can only be reached by first constructing the intermediate functions.
The two brains of reading
Throughout this book we have focused on two different brain systems: the frontoparietal, which is versatile but slow and demands effort; and the ventral system, which is devoted to some specific functions that are carried out automatically and very quickly.
These two systems coexist, and their relevance varies over the course of the learning process. As seasoned readers, we primarily utilize the ventral system, although the parietal system is working residually, as becomes evident when we read complex handwriting or when the letters are not configured in their natural form, either vertically, from right to left, or separated by big spaces. In these cases, the circuits of the ventral cortex–which are not very flexible–cease to function. And then we read similarly to how a dyslexic person does.* In fact, it is harder to read a CAPTCHA** because it has irregularities that make the ventral system unable to recognize it. That is a way to dial back into the latent serial system of reading and find ourselves in the same situation we were long ago when we learned to read.
The temperature of the brain
When we learn, the brain changes. For example, the synapses–from the Greek for ‘join together’–that connect different neurons can multiply their connections or vary the efficiency of an already established connection. All of this changes the neuronal networks. But the brain has other sources of plasticity; for example, the morphological properties or genetic expression of its neurons can change. And in some very specific cases, the number of brain cells can increase, although this is very rare. In general the adult brain learns without augmenting its neuronal mass.
Today the term ‘plasticity’ is used to refer to the brain’s ability to transform. It is a popular metaphor but the term leads to an incorrect assumption: that the brain is moulded and stretched, gets crumpled up and smoothed out, like a muscle, although none of that actually happens.
What makes the brain more or less predisposed to change? With materials, the critical parameter that dictates their predisposition to change is temperature. Iron is rigid and not malleable, but when heated it can change shape and later be reconfigured into another form that it retains when cooled. What is the equivalent in the brain to temperature? First of all, as Hubel and Wiesel proved, there is the stage of development. A baby’s brain does not have the same degree of malleability as an adult’s. Yet, as we have seen, this is not an immutable variable. Is motivation the fundamental difference between a child and an adult?
Motivation promotes change for a simple reason that we have already discussed: a motivated person works harder. Marble is not exactly plastic but, if we go at it for hours with a chisel, it will eventually change shape. The notion of plasticity is relative to the effort we are willing to put in to make a change. But this doesn’t yet bring us to the notion of temperature, of predisposition to change. What happens in the brain when we are motivated that predisposes it to change? Can we emulate this cerebral state in order to promote learning? The answer is in understanding which chemical soup of neurotransmitters promotes synaptic transformation and, therefore, cerebral change.
Before getting into the microscopic detail of brain chemistry, we should have a look at a more canonical way of learning: memory. Almost all of us remember the events of 9/11, the planes crashing at the North and South Towers of the World Trade Center. What’s most surprising is that, fifteen years later, we not only remember those images of the burning towers but we also can recall with striking clarity where we were and who we were with when it happened. That deeply emotional moment makes everything around it, both the relevant–the attack–and the irrelevant, stick in the memory. This is why those who have been through a traumatic experience often have a very difficult time erasing that memory, which can be activated by fragments of the episode, the place where it happened, a similar smell, a person who was there, or any other detail. Memories are formed as episodes; in the moments our neuronal register is most sensitive, we vividly remember not only what activated that sensitivity, but also everything that formed around that episode.
This is an example of a more general principle. When we are emotionally aroused or when we receive a reward (monetary, sexual, emotional, chocolate) the brain is more predisposed to change. To understand how this happens, we have to switch tools and enter into the microscopic world. And that voyage will take us to California, to the laboratory of a neurobiologist, Michael Merzenich.
In his experiment, monkeys had to identify the higher pitched of two tones, as when we tune an instrument. As the two tones became more similar, they began to perceive them as identical even though they weren’t. This allowed him to investigate the limits of the resolution of the auditory system. Like any other virtue, this can also be trained.
The auditory cortex, just like the visual one, is organized into a pattern, a cluster of neurons grouped into columns. Each column is specialized to detect a particular frequency. So, in parallel, the auditory cortex analyses the frequency structure (the notes) of a sound.
In the map of the auditory cortex, each frequency has a dedicated region. Merzenich already knew that if a monkey is actively trained to recognize tones of a particular frequency, something quite extraordinary happens: the column representing that frequency expands, like a country that grows by invading its neighbouring territories. The question that concerns us here is the following: what allows that change to happen? Merzenich observed that the mere repetition of a tone wasn’t enough to transform the cortex. Yet if that tone occurs at the same time as a pulse of activity in the ventral tegmental area, a region deep in the brain that produces dopamine, then the cortex reorganizes itself. It all comes together. In order for a cortical circuit to reorganize, there needs to be a stimulus occurring in that window of time which releases dopamine (or other similar neurotransmitters). In order for us to learn, we need motivation and effort. It isn’t magic or dogma. We now know that this produces dopamine, which lessens the brain’s resistance to change.
We can think of dopamine as the water that makes clay more pliable, and the sensory stimulus as the tool that marks a groove in that damp clay. Neither can transform the material on its own. Working dry clay is a waste of time. Wetting it if you aren’t going to sculpt it is as well. This is the basis of the learning programme that we began discussing with Galton’s idea: the brain learns when it is exposed to stimuli that transform it. It is a slow and repetitive task to establish those grooves of new circuits that automatize a process. And the transformation requires, in addition to effort and training, a cerebral cortex that is in a state sensitive to change.
To sum up, we looked at Galton’s error in order to understand how learning is forged: the ceiling is not as genetically established as is commonly assumed and the path is also social and cultural. We also saw that virtuosos carry out their expert tasks in a qualitatively different way, not only by improving the original procedure. And that to persevere in learning we have to work with motivation and effort, outside our comfort zone and the OK threshold. What we recognize as a performance ceiling is usually not. It is an equilibrium point.
In short, it is never too late to learn. If something c
hanges in adulthood it is that our motivation gets stuck in what we’ve already learned and not swept up in the whirlwind of discovering and learning. Recovering that enthusiasm, that patience, that motivation and that conviction seems to be the natural starting point for those who truly want to learn.
CHAPTER SIX
Educated brains
How can we use what we have learned about the brain and human thought to improve education?
Every day, more than two billion children around the world go to school, in what is perhaps the largest collective experiment in the history of humanity. There they learn to read, forge their closest friendships, and build themselves as social beings. And at school, in a highly intense learning process, the brain is developed and transformed. However, neuroscience has crassly ignored this close link, remaining distanced from classrooms for years. Perhaps now is the prime time to establish a bridge between neuroscience and education.
The philosopher and educator John Bruer warned that this bridge connects distant worlds; what neuroscience considers relevant is not necessarily or usually pertinent to education. For example, understanding that a region of the parietal cortex is key to the numerical process can be important for neuroscience but doesn’t help a teacher to teach maths better.
In this effort, we must remain, more than ever, sceptical about the use of vague and imprecise scientific terms. I was once in a conference in which a supposed expert in neuroscience argued–as so many do today–that we should use the right hemisphere more. I raised my hand (the left one to be compliant …) and mentioned that even if I agreed that using the right hemisphere was useful, I simply did not know how to do it. Should I turn my head to the right to increase blood flow to the right hemisphere? His ‘neuro-expert’ response was that I should focus on drawing, colouring books and creative arts and forget about language. And then my question was, why didn’t he just say that straight out? I knew the answer. He was employing an artful yet useless metaphor. Referring to the brain and the hemispheres only served to appropriate the prestige of a scientific field for marketing purposes.
The Secret Life of the Mind Page 21