Proust and the Squid: The Story and Science of the Reading Brain

by Maryanne Wolf

protean

Computer scientists use the term “open architecture” to describe a system that is versatile enough to change—or rearrange—to accommodate the varying demands on it. Within the constraints of our genetic legacy, our brain presents a beautiful example of open architecture. Thanks to this design, we come into the world programmed with the capacity to change what is given to us by nature, so that we can go beyond it. We are, it would seem from the start, genetically poised for breakthroughs.

Thus the reading brain is part of highly successful two-way dynamics. Reading can be learned only because of the brain’s plastic design, and when reading takes place, that individual brain is forever changed, both physiologically and intellectually.

a person who learns to read in Chinese uses a very particular set of neuronal connections that differ in significant ways from the pathways used in reading English. When Chinese readers first try to read in English, their brains attempt to use Chinese-based neuronal pathways. The act of learning to read Chinese characters has literally shaped the Chinese reading brain. Similarly, much of h...

To take one example from the language domain, when you read the 233 words in Proust’s passage, your word meaning, or semantic, systems contributed every possible meaning of each word you read and incorporated the exact correct meaning for each word in its context. This is a far more complex and intriguing process than one might think.

The richness of this semantic dimension of reading depends on the riches we have already stored, a fact with important and sometimes devastating developmental implications for our children. Children with a rich repertoire of words and their associations will experience any text or any conversation in ways that are substantively different from children who do not have the same stored words and concepts.

Proust’s unusually sequenced grammatical information had to be connected to the meanings of individual words without losing track of the overall propositions and context of the passage.

As you linked all this linguistic and conceptual information, you generated your own inferences and hypotheses based on your own background knowledge and engagement.

Unlike its component parts such as vision and speech, which are genetically organized, reading has no direct genetic program passing it on to future generations. Thus the next four layers involved must learn how to form the necessary pathways anew every time reading is acquired by an individual brain. This is part of what makes reading—and any cultural invention—different from other processes, and why it does not come as naturally to our children as vision or spoken language, which are preprogrammed.

When confronted, therefore, with the task of inventing functions like literacy and numeracy, our brain had at its disposal three ingenious design principles: the capacity to make new connections among older structures; the capacity to form areas of exquisitely precise specialization for recognizing patterns in information; and the ability to learn to recruit and connect information from these areas automatically.

Dehaene suggests that the visual areas in our ancestors’ brains responsible for object recognition were used to decipher the first symbols and letters of written language by adapting their built-in system for recognition. Critically, the combination of several innate capacities—for adaptation, for specialization, and for making new connections—allowed our brain to make new pathways between visual areas and those areas serving the cognitive and linguistic processes that are essential to written language.

Becoming virtually automatic does not happen overnight and is not a characteristic of either a novice bird-watcher or a young novice reader. These circuits and pathways are created through hundreds or, in the case of some children with reading disabilities like dyslexia, thousands of exposures to letters and words. The neuronal pathways for recognizing letters, letter patterns, and words become automatic thanks to retinotopic organization, object recognition capacities, and to one other extremely important dimension of brain organization: our ability to represent highly learned patterns of information in our specialized regions.

For the expert reading brain, as information enters through the retina, all the physical properties of the letters are processed by an array of specialized neurons that feed their information automatically deeper and deeper into other visual processing areas. They are part and parcel of the virtual automaticity of the reading brain, in which all its representations and indeed all its individual processes—not just visual ones—become rapid-fire and effortless.

From the child’s first, halting attempts to decipher letters, the experience of reading is not so much an end in itself as it is our best vehicle to a transformed mind, and, literally and figuratively, to a changed brain.

At its root the alphabetic principle represents the profound insight that each word in spoken language consists of a finite group of individual sounds that can be represented by a finite group of individual letters.

If there are no genes specific only to reading, and if our brain has to connect older structures for vision and language to learn this new skill, every child in every generation has to do a lot of work.

Learning to read begins the first time an infant is held and read a story. How often this happens, or fails to happen, in the first five years of childhood turns out to be one of the best predictors of later reading.

A prominent study found that by kindergarten, a gap of 32 million words already separates some children in linguistically impoverished homes from their more stimulated peers. In other words, in some environments the average young middle-class child hears 32 million more spoken words than the young underprivileged child by age five.

Children who never have a story read to them, who never hear words that rhyme, who never imagine fighting with dragons or marrying a prince, have the odds overwhelmingly against them.

Today, we possess sufficient knowledge about the components of reading to be able not only to diagnose almost every child in kindergarten at risk of a learning difficulty, but also to teach most children to read. This same knowledge underscores what we do not wish to lose in the achievement of the reading brain, just as the digital epoch begins to make new and different demands on that brain.

My colleagues and I use a variety of tools, from naming letters to brain imaging, to understand why so many children with dyslexia, including my own firstborn son, have difficulty not only with reading but also with seemingly simple linguistic behaviors like discriminating individual sounds or phonemes within words, or quickly retrieving the name of a color.

For we are also in the exciting early stages of understanding the little-studied benefits that accompany the brain development of some persons with dyslexia. It is no longer reducible to coincidence that so many inventors, artists, architects, computer designers, radiologists, and financiers have a childhood history of dyslexia.

we must be vigilant not to lose the profound generativity of the reading brain, as we add new dimensions to our intellectual repertoire.

The chance discovery of little clay pieces, no larger than a quarter, marks the birth of modern efforts to learn about the history of writing. Called tokens, some of these pieces came enclosed in clay envelopes (see Figure 2-2) that bore markings representing their contents.

The tokens primarily recorded the number of goods bought or sold, such as sheep, goats, and bottles of wine. A lovely irony of our species’ cognitive growth is that the world of letters may have begun as an envelope for the world of numbers.

along with cave drawings like those in France and Spain, tokens reflected the emergence of a new human ability: the use of a form of symbolic representation, in which objects could be symbolized by marks for the eye.

The neuroscientists Michael Posner and Marcus Raichle, and Raichle’s research group at Washington University, conducted a pioneering series of brain imaging studies to observe what the brain does when we look at a continuum of symbol-like characters with and without meaning. Their range of tasks included meaningless symbols, meaningful symbols that make up real letters, meaningless words, and meaningful words. Although clearly designed for other purposes, these studies provide a remarkable glimpse of what happens when the brain encounters ever more abstract and demanding writing systems—millennia ago and today.

When we see these same circles and lines and interpret them as meaningful symbols, however, we need new pathways. As Raichle’s work showed, the presence of real-word status and meaning doubles or triples the brain’s neuronal activity. Becoming familiar with the basic pathways used in a token-reading brain is an excellent foundation for understanding what happens in more complex reading brains. Our ancestors could read tokens because their brains were able to connect their basic visual regions to adjacent regions dedicated to more sophisticated visual and conceptual processing. These adjacent regions are found in other occipital and nearby temporal and parietal areas of the brain. The temporal lobes are involved in an impressive range of auditory and language-based processes, which contribute to our ability to comprehend the meanings of words. Similarly, the parietal lobes participate in a wide range of language-related processes, as well as spatial and computational functions.

Symbolization, therefore, even for the tiny token, exploits and expands two of the most important features of the human brain—our capacity for specialization and our capacity for making new connections among association areas.

The nineteenth-century French neurologist Joseph-Jules Déjerine observed that an injury to the angular gyrus region produced a loss in reading and writing. And today, the neuroscientists John Gabrieli and Russ Poldrack and their groups at MIT and UCLA find, through their imaging research, that pathways to and from the angular gyrus region become intensely activated during reading development.

The word “cuneiform” derives from the Latin word cuneus, “nail,” which refers to the script’s wedge-like appearance. Using a pointed reed stylus on soft clay, our ancestors created a script that looks, to the untutored eye, a lot like bird tracks (Figure 2-4).

Symbols rapidly became less pictographic and more logographic and abstract. A logographic writing system directly conveys the concepts in the oral language, rather than the sounds in the words. Over time many of the Sumerian characters also began to represent some of the syllables in oral Sumerian. This double function in a writing system is classified by linguists as a logosyllabary, and it makes a great many more demands on the brain.

John DeFrancis, a scholar of ancient languages and Chinese, classifies both Chinese and Sumerian as logosyllabic writing systems, with many similar elements, though of course also some dissimilar ones.

logographic systems appear to activate very distinctive parts of the frontal and temporal areas, particularly regions involved in motoric memory skills. The cognitive neuroscientists Li-Hai Tan and Charles Perfetti and their research group at the University of Pittsburgh make the important point that these motoric memory areas are far more activated in reading Chinese than in reading other languages, because that is how Chinese symbols are learned by young readers—by writing, over and over. This is also how Sumerian characters were learned—on little clay practice tablets, over and over again: “it verily was so.”

felicitous

Only around 600 BCE did Sumerian writing disappear. And even then, its impact remained, inside some of the characters and within the methods of Akkadian, the lingua franca from the third to the first millennium BCE.

What distinguished our ancestors in ancient Greece from us was the great value the Greeks placed on an oral culture and memory. Just as Socrates probed his students’ understanding in dialogue after dialogue, educated Greeks honed their rhetorical and elocutionary skills, and prized above almost everything else the ability to wield spoken words with knowledge and power. The astounding memory capacities of our Greek ancestors are one result. They

remind us of the significant effects of culture on the development of presumably innate cognitive processes, such as memory.

Specifically, the alphabet reader learns to rely more on the posterior of the left hemisphere, specialized regions with less bihemisphere activation in these visual regions. By contrast, the Chinese (and Sumerians) achieve efficiency by recruiting many areas for specialized, automatic processing across both hemispheres.

alexia

In a pathbreaking meta-analysis of twenty-five imaging studies of different languages, cognitive scientists from the University of Pittsburgh found three great common regions used differentially across writing systems. In the first, the occipital-temporal area (which includes the hypothesized locus of “neuronal recycling” for literacy), we become proficient visual specialists in whatever script we read. In the second, the frontal region around Broca’s area, we become specialists in two different ways—for phonemes in words and for their meanings. In the third, the multifunction region spanning the upper temporal lobes and the lower, adjacent parietal lobes, we recruit additional areas that help to process multiple elements of sounds and meanings, which are particularly important for alphabetic and syllabary systems.

one, reading in any language rearranges the length and breadth of the brain; and two, there are multiple pathways to fluent comprehension, with a continuum of efficiency taking varied forms among the varied writing systems.

By taking a meta-view of this entire history, we can see that what promotes the development of intellectual thought in human history is not the first alphabet or even the best iteration of an alphabet but writing itself.

First, Socrates posited that oral and written words play very different roles in an individual’s intellectual life; second, he regarded the new—and much less stringent—requirements that written language placed both on memory and on the internalization of knowledge as catastrophic; and third, he passionately advocated the unique role that oral language plays in the development of morality and virtue in a society. In each instance Socrates judged written words inferior to spoken words, for reasons that remain powerfully cautionary to this day.

From this larger interconnected view of language, memory, and knowledge, Socrates concluded that written language was not a “recipe” for memory, but a potential agent of its destruction. Preserving the individual’s memory and its role in the examination and embodiment of knowledge was more important than the indisputable advantages of writing in preserving cultural memory.

Gradually, each child in most literate cultures begins to acquire a repertoire of frequently seen letters and words before ever learning to write these letters. This phase of reading is like a “logographic” stage in the child’s development: what the child understands, not unlike our token-reading ancestors, is the relationship between concept and written symbol.

The behavioral neurologist Norman Geschwind suggested that for most children myelination of the angular gyrus region was not sufficiently developed till school age—that is, between five and seven years. Geschwind also hypothesized that myelination in these critical cortical regions develops more slowly in some boys; this might be one reason why more boys are slower to read fluently than girls.

They found across three different languages that European children who were asked to begin to learn to read at age five did less well than those who began to learn at age seven. What we conclude from this research is that the many efforts to teach a child to read before four or five years of age are biologically precipitate and potentially counterproductive for many children.

Writing and listening to poetry, for example, sharpen a child’s developing ability to hear (and ultimately to segment) the smallest sounds in words, the phonemes. Such first attempts to write reflect a sequence in a child’s growing knowledge about the connection between oral and written language. First, letters are written (or drawn) in imitation. To be sure, there is often more scribbled “art” than concept here. Next, letters begin to show off children’s evolving concept of print, particularly the letters in their own names. Gradually, other letters capture how children think words are spelled, with many a letter name used, let us say, ingeniously.

One of the singular accomplishments of the creators of the Greek alphabet was this dimension of awareness of speech sounds. It is one of the most powerful contributions of the alphabet, and also is one of the two best predictors of later reading achievement, the other being rapid naming.