What Good Readers Know | American Enterprise Institute - AEI
- Reading comprehension and higher-order thinking are inseparable from background knowledge.
- Knowledge compounds, so students with more background knowledge and vocabulary have an easier time learning new things.
- Teachers should spend far more time teaching students about science, history, geography, literature, art, and music, and school systems should ensure that curriculum is cumulative, sequenced, and coherent.
There is an unlikely figure who might have saved American schools a great deal of time and trouble teaching children to read over the past half century: the legendary physicist Richard Feynman.
In a famous 1974 commencement address at the California Institute of Technology, Feynman described the “cargo cults” that arose on some Pacific Islands after World War II.1 During the war, islanders had watched airplanes arrive bearing valuable supplies. When the war ended and the planes stopped coming, the islanders tried to lure the planes back by re-creating the conditions that had brought them: They carved runways, set fires along their edges, and built wooden huts for a controller donning makeshift headphones, with pieces of bamboo sticking out like antennae.
“They’re doing everything right,” Feynman observed. “The form is perfect. It looks exactly the way it looked before. But it doesn’t work. No airplanes land.”2
Well, of course it doesn’t work, you may think. It is hard to imagine anyone could believe otherwise, however remote their lives or limited their exposure to advanced technology.
In his speech, Feynman sought to warn his audience about scientific integrity, not reading instruction. Yet we have long indulged in something remarkably similar in American classrooms, mistaking visible behaviors for the underlying conditions that contribute to reading comprehension.
Visit an elementary school where struggling fourth or fifth graders are learning to read, and you are likely to see posters reminding children to use the “reading strategies” they’ve been taught:
“Good readers ask questions.”
“Good readers create pictures in their minds.”
“Good readers make connections when they read.”
You will also see teachers doing “think-alouds”—reading out loud to students and stopping to narrate features of a story, make predictions, ask questions, and model making inferences. These earnest, well-intentioned teachers are demonstrating to eager students the habits of good readers: showing children what proficient readers do and encouraging them to imitate those behaviors when they read independently.
This is cargo-cult reading. Push the Feynman analogy a little further, and it starts to do real work. The islanders copied only what was visible—runways, signal fires, the man with headphones—but the system that actually brought the planes depended on an enormous infrastructure that was foreign to them: radio networks, supply chains, trained personnel, command structures, and technical knowledge. What they could see was merely the visible features of a much deeper, more complex system.
Reading works the same way. Watch a fluent reader glide through a page, and you see almost none of what makes comprehension possible. Decoding has become so automatic that the reader scarcely experiences it as decoding. Thousands of words are recognized and understood almost instantly. And beneath it all lies an immense store of background knowledge, silently supplying context, filling gaps, resolving ambiguities, and allowing the reader to make inferences without consciously deciding to make them.
None of this is visible in the way that asking a question, making a prediction, or stopping to summarize is visible. And therein lies the fatal instructional mistake that has held back reading achievement for decades: We observed what good readers do and tried to teach struggling readers to imitate those behaviors, while paying far less attention to what good readers know.
Said differently, reading comprehension is not a skill we teach. It is a condition we create.
My students and I learned this lesson the hard way. When I became a fifth-grade teacher in a low-performing South Bronx elementary school more than 20 years ago, I walked in not understanding what struggling readers looked like. More than four in five of my students had failed the previous year’s state reading test; with minimal training and no experience, I expected to encounter children who were hard-pressed to read the words on the page.
Instead, in five years at PS 277, the lowest-performing school in New York City’s lowest-performing community school district, I never had a student who could not decode. Some read more fluently than others, certainly.3 But they could read the words. What they struggled to do was understand them. “I read it, but I didn’t get it” was a common refrain I heard again and again.
The distinction matters because reading is actually two different things that we routinely conflate. The first is decoding: turning the squiggles on a page into sounds and words. Decoding is a skill and a transferable one. We can agree on the pronunciation of made-up words like rigfap and churbit because once you understand the relationship between letters and sounds, you can apply that knowledge automatically even to invented nonsense words you’ve never seen before or memorized.
But reading comprehension is different. It’s inseparable from vocabulary and background knowledge. We tend not to notice this because when we read about familiar subjects, the knowledge we bring to the text does its work invisibly. Consider this sentence:
Trout hit into a 6-4-3 double play to end the game.
To a knowledgeable baseball fan, those few words communicate a remarkable amount of information. The Los Angeles Angels and their superstar hitter Mike Trout—not a fish—lost the game. There was a runner on first base and one out, and it was probably the ninth inning when Trout hit a ground ball to the shortstop, who threw to the second baseman, who relayed the ball to first base in time to retire the batter.
This may be an arcane example, but it’s not some peculiar feature unique to sports writing. It is how language works. “Writers leave out a lot of information that they assume readers will know,” explains University of Virginia cognitive scientist Daniel Willingham. “If they put all the information in, their writing would be tedious. But if readers can’t supply the missing information, they will have a hard time making sense of the text.”4 Like an iceberg with its vast bulk beneath the waves and hidden from view, written texts and spoken language run on assumptions that writers and speakers make about what readers and listeners already know. This is the core concept advanced for decades by E. D. Hirsch Jr., whose landmark 1987 book Cultural Literacy: What Every American Needs to Know diagnosed the struggle of students in my South Bronx classroom as if he’d been there.5
Hirsch famously described cultural literacy as “the network of information that all competent readers possess.” It’s taken-for-granted background knowledge “that enables them to take up a newspaper and read it with an adequate level of comprehension” and that writers and speakers assume their audience already has.6 When students possess that assumed knowledge, comprehension can feel as effortless as fluent decoding. When they do not, even a seemingly simple passage can become impenetrable. A piece of text is something like a Jenga tower, with each block representing a vocabulary word or a piece of assumed knowledge. Pull out one or two, and the structure remains standing. Remove one too many, and it collapses entirely. Meaning is lost.
The best-known demonstration of this phenomenon is Donna Recht and Lauren Leslie’s 1988 “baseball study.”7 They divided middle school students according to their pre-assessed reading ability and their knowledge of baseball, then asked them to read a story about a baseball game and reconstruct a passage describing part of the game. The result has become familiar in discussions of knowledge-rich curriculum, but it remains striking: Students judged to be “poor readers” who knew a lot about baseball substantially outperformed ostensibly “good readers” who knew little about the game.
This is worth dwelling on. If reading comprehension is a transferable skill that can be taught, practiced, and mastered—the way we have taught and tested it for decades—then the students who were “good” readers should have had no trouble outperforming the “poor” readers. The subject matter should be irrelevant. Either you’re a good reader or you’re not. But it turns out that it’s not that simple.
For example, you may be a voracious and sophisticated reader of history or politics but confused when you pick up a technical manual, a medical journal, or (if you’re not technologically adroit) instructions for installing the operating system on your computer. You didn’t suddenly become a poor reader. You’re simply reading out of your depth in an unfamiliar knowledge domain.
This is what my fifth graders were telling me when they said, “I read it, but I didn’t get it.” They could turn print into spoken language. But they often lacked enough of the knowledge and vocabulary the writer assumed to extract meaning from the text.
It’s not quite right to say that background knowledge improves reading comprehension. That makes knowledge sound like an enhancement—a helpful supplement that makes reading somewhat easier. Background knowledge supplies information that written language routinely leaves unstated. It allows us to resolve ambiguity, draw inferences, recognize references and idioms, and understand what a writer or speaker means without being told explicitly. In that sense, knowledge is not merely helpful to reading; it is an essential precondition.
If knowledge were merely necessary to understand text passages on topics we already know something about, its importance to reading would still be considerable. But background knowledge does something even more consequential: It increases our capacity to learn about new things.
Consider vocabulary. An educated adult knows tens of thousands of words. Obviously, we cannot acquire such a large volume of words through vocabulary lists, memorization for the SAT, and dictionary definitions. Most of the words we know and use are acquired incidentally, through repeated encounters with unfamiliar words in speech and text. We figure out what a new word means because we understand enough of everything around it to make a reasonable inference.
My former colleague at the Core Knowledge Foundation, Alice Wiggins, devised a wonderfully efficient demonstration of how this works. Consider the obscure word excrescence. See what happens when you encounter it in several different contexts:
To calculate fuel efficiency, the aerospace engineers needed an accurate estimation of excrescence drag caused by the shape of the plane’s cabin.Excrescences on the valves of the heart have been known to cause a stroke.
The wart, a small excrescence on his skin, had made Jeremy self-conscious for years.
By the third sentence, if not before, you probably formed a serviceable understanding of excrescence: a projection or outgrowth, often an abnormal one. In Wiggins’s demonstration, she displayed a final sentence to confirm it:
At the far end of the meadow was what, at first glance, I thought a huge domed building, and then saw was an excrescence from the cliff itself.8
Critically, notice what enabled you to learn the new and unfamiliar word: You have a sense of what aerospace engineers do; you’re also familiar with an airplane’s shape and probably what drag means in this context. You know what heart valves and strokes are. You certainly know what skin and warts are and what it means to feel self-conscious. That background knowledge enabled you to understand the gist of each sentence and progressively narrow the possible meaning of the new word.
Now imagine a child for whom aerospace engineer, drag, heart valve, stroke, and even self-conscious are unfamiliar terms. The same sentences offer far fewer clues. The opportunity to learn a new word is present on the page for both readers—one of whom possesses the requisite background knowledge, the other who does not—but only one possesses enough prior knowledge to take full advantage of it.
Cognitive psychologist Keith Stanovich, in a landmark 1986 paper titled “Matthew Effects in Reading,” pointed to the process by which relatively small differences in literacy can grow larger over time.9 The phrase comes from a passage in the New Testament’s Gospel of Matthew: “For unto every one that hath shall be given, and he shall have abundance: but from him that hath not shall be taken away even that which he hath.”10 In plain English, the rich get richer; the poor get poorer.
Background knowledge and vocabulary help explain the Matthew effect. Children who know more understand more of what they read and hear. Because they understand more, they can better infer the meanings of unfamiliar words and absorb new information. They grow richer. That new vocabulary and knowledge then become background knowledge for the next book, lesson, conversation, and experience. Knowledge, in other words, compounds.
And so too does its absence. Children who come to school with less knowledge understand less of what they encounter. That makes it harder to acquire new vocabulary and information incidentally, in context, leaving them still less prepared for what comes next. They grow poorer and fall further behind. Two children can sit through the same lesson or read the same book and emerge having learned different amounts—not because one is more intelligent or attentive, but because one had more knowledge available to make sense of the new material.
In sum, background knowledge cannot be treated as an enrichment to be added after children become proficient readers. It is the principal means by which they become proficient readers—and increasingly capable learners—in the first place.
There is a broader lesson here. The role of background knowledge does not end when reading comprehension ends. It extends to the cognitive skills that schools prize most highly: critical thinking, problem-solving, and the ability to learn new things.
We’ve already seen this in action. Think back to excrescence. Challenged to figure out the meaning of an unfamiliar word, you formed a hypothesis about its meaning. You drew inferences from what you already knew about airplanes, heart valves, strokes, and warts. As you encountered the word in different contexts, you weighed new evidence against your initial hypothesis, discarded possibilities that no longer fit, and gradually arrived at a more precise understanding.
We would describe this process as critical thinking, but it didn’t occur in a vacuum. You reasoned successfully and learned an unfamiliar word because you already possessed a large store of relevant knowledge.
The same principle applies far beyond reading. We often speak as though critical thinking and problem-solving are general-purpose cognitive skills that can be taught and practiced independently of subject matter. But a student cannot think critically about the causes of the Civil War without knowing a great deal about slavery, federalism, westward expansion, the Constitution, and sectional politics.
This is also why we should be wary of the earnest and oft-repeated idea that students should be taught to “think like a scientist” or “think like a historian.” Scientists and historians do not possess distinctive, transferable modes of thought that novices can simply be taught to imitate and apply. Expert thinking is inseparable from expert knowledge.
Because they possess enormous stores of historical knowledge, experienced historians know which facts are significant, which sources deserve skepticism, what precedents may illuminate an event, and which interpretations are reasonable. Physicians can reason through a difficult diagnosis because of years of accumulated knowledge that allows them to recognize significant symptoms and decide which possibilities are plausible. Remove that knowledge, and the “skill” disappears with it.
Novices, by definition, do not yet possess that knowledge. To ask or expect them to mimic the intellectual habits of experts is the same cargo-cult mistake we make in reading instruction, confusing the visible behaviors of expertise with the accumulated knowledge that makes those behaviors productive. This is why the familiar conflict between teaching knowledge and teaching children “how to think” is misconceived. Knowledge does not crowd out higher-order thinking. It’s what makes higher-order thinking possible.
Elsewhere in the American Enterprise Institute’s “The Practice Project” series, Daisy Christodoulou explores why skills like critical thinking and problem-solving prove stubbornly resistant to transfer from one domain to another.11 The lesson for reading is particularly clear. What looks from the outside like a general skill is actually the visible expression of substantial knowledge operating beneath the surface—the same mistake Feynman’s cargo cultists made when they confused the appearance of an airfield with the infrastructure and context that had brought the planes.
Knowledge is not merely the result of thinking. Knowledge is what we think with.
How should all this guide classroom practice?
For individual teachers, the most immediate implication is to recognize the opportunity cost of treating reading comprehension as a transferable skill. Every additional hour spent practicing finding the main idea, making connections, and identifying an author’s purpose is an hour that is lost learning about the solar system, ancient civilizations, ecosystems, mythology, music, art, and literature—the knowledge that will make future texts more comprehensible.
This does not mean abandoning instruction in comprehension strategies. A modest amount of explicit instruction can be useful, particularly for inexperienced readers who need to learn that reading is supposed to make sense and that confusion should prompt them to stop and repair their understanding.12 But there is little reason to believe that endlessly practicing such strategies will transform children into broadly proficient readers. Once students understand the strategies, the greater need is to increase the store of knowledge and vocabulary on which comprehension depends.
That means treating knowledge building as part of literacy from the beginning, not something children graduate into once they have “learned to read.” Strong decoding instruction and knowledge building should proceed simultaneously. Science, history, geography, literature, art, and music should not compete for instructional time with literacy: They are the means by which literacy develops. Willingham said it best: Teaching content is teaching reading.13
At the same time, candor demands we acknowledge the limits of what even the most diligent individual teacher can accomplish. If knowledge compounds, then a genuinely knowledge-rich education must compound too. A coherent curriculum should build deliberately and cumulatively from year to year. No individual teacher can create that coherence alone.
Principals, not individual teachers, determine schedules, set instructional priorities, and decide whether science and social studies are protected or sacrificed for additional reading “skills and strategies” practice. District leaders select curricula and determine what children learn in each grade. School boards and state leaders establish standards, assessments, and incentives that can either encourage cumulative knowledge building or devalue it. Everyone who influences the curriculum should understand that literacy is built not through reading instruction alone but through the entirety of a child’s education.
The stakes cannot be overstated for disadvantaged children. Children from knowledge-rich homes have many opportunities to acquire vocabulary and background knowledge through books, conversations, travel, museums, extracurricular activities, and countless informal experiences. Their less fortunate peers are more likely to be school-dependent learners, relying much more heavily on the instructional decisions made on their behalf. Yet in my former school and others like it, we responded to their reading difficulties by doing precisely the opposite: cutting back on science, history, art, and other curricular content to make room for yet more reading skills and strategies instruction and practice.
Finally, none of this means good readers don’t ask questions, make inferences, monitor their comprehension, and make connections. They do. Feynman’s cargo cults weren’t mistaken that functioning airfields had runways and control towers. They were mistaken about what made the planes arrive. The same distinction matters in reading. The habits of proficient readers are worth cultivating, but they cannot substitute for the vast, largely invisible infrastructure of knowledge and vocabulary that makes proficient reading possible.
We have spent decades teaching children what good readers do. Teachers must also understand and impart what good readers know. But ensuring that every child systematically acquires that knowledge is larger than the work of any one teacher. It is the work of a school—and ultimately of a school system.
Reading comprehension is not principally a skill we teach. It is a condition we create. And we create it by teaching children about the world.
Robert Pondiscio is a senior fellow and affiliate of the James Q. Wilson Program in K–12 Education Studies at the American Enterprise Institute, where he focuses on K–12 education, curriculum, teaching, school choice, and charter schooling.