Small Wonders

The history of education in the twenty-first century may turn out to be like the history of medicine in the nineteenth century. Both medicine and education have great moral urgency. Passing on what we know to our children is, after all, one of the few ways we have of genuinely defying death; medicine just postpones it. Both medicine and education invoke knowledge to justify their authority. Doctors have always justified their practices by claiming that they understand how our bodies work. Educators have always justified theirs by claiming that they understand how our children’s minds work. But for most of history those claims were based on scarcely any systematic research. At best, they were pragmatic generalizations, the outcome of a long process of empirical tinkering.

During the last 150 years we have gradually begun to integrate real biological science into our medical practice. This has been one of the great scientific success stories. Surely even the most adamant postmodern critics of science believe that vaccinating babies is not just an exercise in patriarchal control. But our new biological knowledge has also told us that organisms and their illnesses are individual, variable, and complicated. And biology itself can’t determine what kind of medicine is worth having, and how much we’re willing to pay for it.

A similar story could unfold in education. In the last thirty years, we have begun to develop a science of children’s minds. This new research might be the equivalent of the scientific physiology that has transformed medicine. But it is unlikely to lead to some simple educational panacea. In fact, helping our children to be both smart and wise is likely to be just as difficult, as complicated and demanding, though just as valuable, as helping them to be healthy. And turning the scientific findings into practice depends on broader political and economic decisions.

Throughout history, our ideas about how children learn have informed our ideas about what education should be like. Calvin’s small sinners, Rousseau’s idealized innocents, Locke’s blank slates—all those particular images of children led to particular kinds of schools. Calvin, Rousseau, and Locke, and later Freud and B.F. Skinner, shared some basic assumptions. They assumed that children’s knowledge was a kind of oxymoron, that children were creatures of passion rather than reason, instinct rather than intellect. For better or for worse, teachers self-consciously shaped these natural creatures into rational, knowledgeable, civilized adults.

However, no one actually studied children systematically until the 1930s. That first empirical research, by Jean Piaget and Lev Vygotsky, reversed the traditional view. Piaget concluded that even very young children spontaneously and actively reasoned about the world. Vygotsky concluded that adults naturally and spontaneously did things that helped children to reason about the world. In the Sixties, Piaget and Vygotsky were rediscovered as part of the new discipline of “cognitive science.” From the findings of that science, it turns out that our everyday behavior, the way we naturally see, speak, think, and act, depends on abstract and complex kinds of knowledge, including knowledge of how the natural world works and of the intentions of other people. An important branch of cognitive science studies how children develop this knowledge.

As a result, we have learned more about what children know and how they learn in the last thirty years than we did in the preceding two thousand years. In fact, Piaget and Vygotsky have received the ultimate accolade of science; they inspired research that overturned their own theories. Piaget, for example, still thought children’s reasoning was egocentric, “pre-causal,” and dominated by appearances. The new research shows that four-year-old children already are able to take the perspective of others, to infer causes, and to go beyond superficial appearances. Children actually know and learn more than Piaget and Vygotsky thought.¹

Yet you could read much of the current debate about educational policy and have no idea that this research exists. The debate is largely dominated by a “conservative” view on the one hand and a “postmodern” view on the other, both of which are equally uninformed by cognitive science. One side argues for the narrow inculcation of specific skills and facts. The other side assumes that education can’t pass on an objective body of knowledge, but can only provide therapy and “empowerment.”

For some time now, there have been psychologists and teachers who have tried to integrate science and education. They have taken on the heroic but difficult task of the nineteenth-century pioneers of scientific medicine. Most of these educators have been in the “progressive” tradition associated with John Dewey. Dewey himself had no empirical evidence about children, but many of his successors were heavily influenced by Piaget and Vygotsky. Most recently, Howard Gardner has taken on the job of trying to change our schools in the light of our new knowledge about children. In a series of lucid and humane books, Gardner has introduced the ideas of developmental and cognitive psychology to a broader audience, and has suggested how those ideas might lead to educational reform.

In his newest book, The Disciplined Mind, Gardner argues for what he calls “education for understanding.” First, he presents a survey of recent scientific research, something he has also done at greater length in his earlier books. He argues that the new research can be used to help design a new kind of curriculum. He proposes that children should be introduced in depth to a few unequivocally important topics in science, art, and history, and that this introduction could provide the basis for further extending their knowledge.

Gardner also continues the argument he has made in earlier books that there are “multiple intelligences”—different ways of understanding the world that include such things as “social” and “musical” intelligence as well as the more standard “logical” and “verbal” intelligence. Given these “multiple intelligences,” children could approach important topics by many different routes. Finally, he suggests that, in practice, a number of alternative curriculums, including his own, but also including more “conservative” or “post-modern” curriculums, might be taught at different schools.

Gardner grounds his ideas in what we might call developmental cognitive science. The basic idea of cognitive science is that the human mind is a computational device, though one that is profoundly unlike any existing computers. We’re computers that are made of cells rather than silicon; we’re computers that can produce conscious thoughts and reflections (though Lord knows how); and we’re computers that were programmed by evolution. We can speculate about these differences between the way our brains compute and the way current man-made digital machines compute; we simply don’t know now which differences will turn out to be important in the long run. But the undeniable fact is that the ideas of cognitive science have led to real discoveries about how the mind and brain work.²

The brain has to solve what we might call the problem of knowledge. We know about a world of people and actions, objects and events, words and meanings. But all that reaches us directly from that world are a few photons hitting our retinas, and a few air molecules vibrating at our ear drums. How do we get here from there? The premise of cognitive science is that we do this largely by computing. We rearrange and manipulate the information at our senses in systematic ways and we come up with an accurate picture of the outside world—what cognitive scientists call a representation.

For example, our visual system is designed to solve a difficult epistemological and computational problem. We live in a world of three-dimensional moving objects. But the information that actually reaches our eyes is just a two-dimensional pattern of photons. Somehow we have to reconstruct information about the objects from this pattern of photons. Our brains take the partial, incomplete, and fragmented information at our senses and use it find out the truth about the world we encounter.

We actually know a great deal about the machinery that allows us to do this. Vision scientists can determine which algorithms the visual system must use. They can work out the instructions that transform the information at our retinas into a representation of the outside world. They have even begun to see how the cells in our brain are connected in a way that lets them compute such algorithms. There are many different kinds of computations and representations involved in vision and they are combined in complicated ways. In many respects, evolution seems to have hit upon near-optimal strategies for solving the extremely difficult problem of vision. No existing computer vision program is even in the same ballpark. Our eyes are finely adapted to give us an accurate view of the world, and so are our brains.

Developmental cognitive science applies this model to children’s learning. Children also have to figure out the structure of the world from the photons at their eyes and sound waves at their ears. They have to solve the problem of knowledge. It turns out that even very young babies already construct remarkably accurate, abstract, and coherent representations of the world. But children also change their computations and representations as they gather more information about the world. And the new computations and representations are even more accurate than the earlier ones. Children are computers that can reprogram themselves. The adults who take care of children also seem designed to provide them with just the right kinds of information at the right time. The child’s computational system is really a sort of network; we adults are computers that are designed to help the child computers to reprogram themselves.

For example, one particularly difficult problem of knowledge is the “other minds” problem. How do we know that other people have minds, when all we can see are moving bodies? Recently, we have begun to understand how children solve this problem. There seems to be an innate basis for this kind of knowledge. When an infant looks at her mother’s face and a toy rattle, she already represents them quite differently. Infants already seem to know that other people have feelings that are like their own. But children also learn a great deal about minds. Before they are three years old, they have learned that, in addition to having feelings, people have perceptions and desires. By the time they reach school, they have also learned that other people have thoughts and that those thoughts may differ from their own thoughts. In the school-age years, without any explicit instruction, children go on to learn that different people may interpret information in different ways and that they have different personalities. Other people also help children to learn. The way that parents, and even older siblings, talk about thoughts and feelings seems to influence children’s ideas about the mind. Children learn about “everyday physics” and “everyday biology” in much the same way that they learn about “everyday psychology.” During the time they are at school, for example, they spontaneously develop new ideas about how weight and size are related, and what makes things alive. Children also spontaneously learn about language, numbers, and music.

We know much more about the computations that allow us to see than we do about the computations that allow children to learn. But we do know that children must have extremely powerful learning machinery. In comparison, the visual system itself is a glorified pocket calculator, and the very best existing computers are not much better than abacuses.

Developmental cognitive science raises a puzzle for educators. If children are so smart by nature, why do so many of them seem so stupid in the classroom? The research shows that children spontaneously learn and that adults spontaneously help them learn. And yet, paradoxically, when we explicitly set out to teach children we often fail dismally. Why can’t we harness their capacities in the institutions we call schools?

Gardner thinks we could do this if we changed what we do in school. His own example is a curriculum based on the study of Darwin’ s theory of evolution, Mozart’s Marriage of Figaro, and the Holocaust. But he argues that other combinations of important topics could also be used. Teachers would introduce these topics by relating them to the ideas that children already have, and to questions children might ask. Children would then pursue a serious and detailed understanding of each topic. They would learn, at least in part, how science, art, or history are done. They would produce examples of scientific investigation, of artistic work, or of historical inquiry—what Gardner calls “performances.” In turn, those performances would be rigorously assessed, criticized, and analyzed by teachers, and by the children themselves, who would then produce revised and improved versions.

Gardner’s curriculum is designed to make what happens in schools more like what happens in natural learning. First, in the natural case, children know a lot to begin with. That initial knowledge, in fact, is what allows them to learn. It leads children to make testable predictions about the world, and it allows them to interpret the confusion around them, and decide which events deserve their further attention. For example, children can take their initial understanding of human feelings, perceptions, and desires and elaborate it into an understanding of thoughts. Children are such efficient learners precisely because their learning is constrained by what they already know.

Yet schools typically take very little advantage of the ideas children bring to the classroom. In fact, often teachers don’t know that children have those ideas. The result, as Gardner points out, is that school knowledge and everyday knowledge are compartmentalized but coexist, instead of the one transforming or replacing the other. In Gardner’s proposal teachers would initially focus on relating the great ideas to the ideas children already have. Developmental cognitive science demonstrates that children already know something about the subject matter of Darwin, Mozart, or Primo Levi. They have complex ideas about plants and people, music and minds, life and death, before they even get to school. Teachers would try to transform those initial ideas into the more sophisticated ideas of the disciplines.

Children are endowed with innate knowledge, but the new research shows that their ability to learn is even more remarkable. When we study children over time, we see a succession of related pictures of the world. Each new picture builds on the child’s earlier ideas, but also revises those ideas in the light of new experiences. Children reshape their representations of the world as they explore, experiment, and learn. They actively seek out evidence that is relevant to the problems they face, and they actively try to construct coherent explanations of that new evidence. To many psychologists, the development of knowledge in childhood looks a lot like the development of knowledge in science. Children seem to construct successive theories of the world that are the product of both their earlier theories and new evidence.

For example, very young children explain human action by referring to feelings, perceptions, and desires. This theory makes some inaccurate predictions. If I know that Mom wants some milk, I may predict that she will go to the refrigerator. But that prediction will be wrong if Mom also thinks that there is no milk in the refrigerator. Later, children begin to explain action by referring to thoughts as well as desires, apparently because that theory makes more accurate predictions and provides better explanations.

But in a typical classroom children rarely get to formulate a theory, make a prediction, or construct an explanation. Children hear something about science, but they almost never actually do science. Gardner’s curriculum is intended to feed into this natural process of theory formation and revision. The process of generating “performances” and revising them parallels the natural processes of theory generation and revision.

Moreover, adults themselves seem to be designed to behave in ways that will help children understand the world. One dramatic early example is “motherese,” the special, high-pitched, slowed-down language that adults use when they talk to babies. It turns out that “motherese” gives babies just the information they need to decode the sound systems of their particular language. Later, adults often seem to teach more sophisticated cultural skills, like hunting or cooking, in the same spontaneous and unconscious way, in a kind of informal apprenticeship. In our culture, in fact, good sports coaches seem to use these kinds of teaching methods (including very demanding, old-fashioned, politically incorrect coaches, who are not likely to know the work of Howard Gardner). They watch a child swing a bat, demonstrate the particular problems with the child’s stance, swing themselves, get the child to swing again. This may help to explain why so many children are so passionate about sports when they are so indifferent to school.

The interactions between children and adults that seem so important in everyday learning are very different from the usual classroom interactions. Parents (and coaches) seem to unconsciously adjust their behavior to the children’s knowledge. It’s much like what happens when we carry on a conversation and unconsciously make fine adjustments that take our listener into account. But imagine trying to carry on a finely adjusted conversation on many disparate topics with twenty very different people at once. You can see why teachers are often so much less effective. Gardner’s scheme would put the teacher in a position more parallel to parents or coaches.

In fact, imagine if we taught baseball the way we teach science. Until they were twelve, children would read about baseball technique and occasionally hear inspirational stories of the great baseball players. They would answer quizzes about baseball rules. Conservative coaches would argue that we ought to make children practice fundamental baseball skills, throwing the ball to second base twenty times in a row, followed by tagging first base seventy times. Others would reply that the economic history of the reserve clause proved that there was, in fact, no such thing as “objectively accurate” pitching. Undergraduates might be allowed, under strict supervision, to reproduce famous historic baseball plays. But only in graduate school would they, at last, actually get to play a game. If we taught baseball this way, we might expect about the same degree of success in the Little League World Series that we currently see in science performance.

Gardner also applies his ideas about “multiple intelligences” to his curriculum. He suggests that teachers should use many different approaches to teaching tuned to the capacities of different children. Children might draw charts of evolutionary development or stage operas as well as reading books and writing reports.

The idea of “multiple intelligences” is also grounded in the new developmental research. That research suggests that children use many different kinds of representations and computations to solve the wide array of problems that confront them. They understand people, objects, plants, words, songs, and numbers in quite different ways. There is no single, across-the-board set of learning mechanisms, and, contrary to Piaget, there don’t seem to be general stages of development. There are probably also differences in the degree to which different people use different types of representations.

Gardner translates such findings into the idea of “multiple intelligences.” The special knowledge we use to understand other minds, for example, is “social intelligence.” The phrase “multiple intelligences” is designed both to evoke our ordinary ideas about intelligence and to revise them. To most people it seems intuitively plausible that knowledge can be explained by an entity called “intelligence” similar to what some psychologists have called g. If “intelligence” means our ability to understand the world and adapt our behavior to it, then cognitive science is the study of intelligence, in the same way that biology is the study of life. But when we pursue that study, “intelligence” turns out to have very little explanatory power, just as “life” has little explanatory power in biology. We can explain how we understand the world by describing a wide variety of computations and representations that are used in different ways at different times in different combinations, just as we can describe how living things function by describing the interactions of many cells, organs, and systems. But “intelligence” or “life” may never enter into those explanations.

Here the example of scientific medicine is again pertinent. Historically many people have had a “vitalist” view of physiology. They believe that there is some force or energy that keeps people alive and well. Something like this, for example, is the chi of Chinese medicine. The popularity of “holistic” medicine suggests that vitalism is still plausible. (In Berkeley, where I live, you often hear about how “negative energy” makes people ill.) A “vitalist” physiology also suggests that the force is stronger or better balanced in some people than in others. There’s nothing intrinsically silly or incoherent or unscientific about vitalism. But it turns out to be wrong. There are many complex and varied mechanisms that allow our bodies to function, but there is no life force, no chi.

There do, of course, seem to be differences in how generally healthy different people are: some people never even catch cold and others get everything that’s going around. We could probably construct reliable measurement scales that would determine the relative general “healthiness” or “sickliness” of different people and those scales might predict something about their future healthiness. Those scales might even turn out to be related to interesting biological facts, or to be useful for particular practical purposes, such as constructing actuarial tables. And, in some sense, those measurements would be scientific. But, by themselves, those measurements would be, at best, tangential to understanding biological functioning. They wouldn’t explain illness and, in particular, they would provide no evidence for the view that there was some biological entity like “life force” or “healthiness.” Nor would they be very relevant to the project of curing disease. (It’s hard to imagine a case for offering better medical treatment to healthy rather than sickly children.)

“Vitalistic” notions of biology are clearly tenacious, and this also seems to be true of the notion of “intelligence.” Imagine an advocate of scientific medicine trying to explain biology to a vitalist. He might say that there are actually multiple “life forces”—oxygen circulates in the blood, electrical impulses circulate in the nerves, and so on—and, in a way, this would be right. Gardner has used “multiple intelligences” in much the same way.

But there is a risk that our hypothetical vitalist would think that there were simply a dozen little chi’s, and many people seem to interpret Gardner’s idea as if there are a dozen little g’s. For them, the multiple intelligences are just like “intelligence,” except that there are enough so everybody gets one. On this assumption, Gardner’s view may be inaccurately assimilated to the view that all ideas are equal; that education is like Alice’s caucus-race where everyone has won and all must have prizes, with “social intelligence” as the Miss Congeniality prize. This is clearly not Gardner’s view and Gardner himself warns against this sort of misinterpretation in his current book. Of course, trying to convey scientific ideas always raises this sort of problem. Still, I think he might have been better off if he had said that there is no such thing as “intelligence” rather than saying that there are many things that might be “intelligences.”

Gardner’s curriculum, and others like it, hold the same hope and promise as scientific medicine. And yet there is another aspect of the parallel that makes the picture more pessimistic. The sad truth is that, even now, most of the illness in the world is the result of a failure of political will and economic resources, not a failure of knowledge.

Gardner’s vision of education, like good health care, also depends on political will and economic resources. It assumes, like much “progressive” education, that the role of school is to allow children to learn as much as possible about the world. But, as Gardner notes, different groups of parents (or policymakers) may instead want schools to raise children’s SAT scores or their political consciousness, make them compliant or rebellious, allow them to memorize facts or to think better of themselves. Cognitive science has little relevance to those goals. His solution is to propose that many different types of schools with different agendas should be created.

The relations between science, children, and democracy raise intrinsically difficult questions. Individuals clearly have the right to control their own minds and bodies. It is less clear how much they have the right to control their children’s minds and bodies. Parents’ rights to choose schooling or medical treatment for their children somehow have to be balanced against the children’s own rights. Gardner also faces this dilemma. His admirably pluralist endorsement of many different curriculums may, pragmatically, be the best way to ensure that at least some schools have a progressive curriculum. Still, it doesn’t seem to entirely square with his belief that his curriculum is genuinely better for children.

Moreover, any curriculum like Gardner’s demands small classrooms, good physical facilities, and highly motivated and sensitive teachers. And, as Gardner says, children have to be well fed and safe before they can learn any curriculum. Education that takes advantage of cognitive science won’t be easy or cheap, any more than scientific medicine is. In fact, in many American public schools, the walls are crumbling and the toilets are overflowing, teachers are undereducated, underpaid, and overworked, and most of the children are poor and scared.

How we could implement Gardner’s curriculum in these conditions is depressingly unclear. But then scientific medicine must often have seemed hopeless to the reformers who set up hospitals in Hell’s Kitchen. Cognitive science could help make schools into places where children exercise their astonishing natural capacities to learn. But all the scientific knowledge in the world won’t help children if we are content to let those capacities fade away.