It would seem obvious, too, that learning how to read involves real learning–receiving and internalizing step-by-step instructions on how to decode the symbols on the page, fit them to spoken sounds, and then link those sounds to meaning. Hence vocabulary lists and the old-fashioned technique of having novice readers “sound out” words by reading aloud in class in order to associate sounds and letters. Children also need to learn how to make all those connections quickly and almost unconsciously, or reading will always be difficult and unpleasant for them, which is why fluency and comprehension are key measures of reading skill. Learning how to read would seem analogous to learning how to play the piano, in which practicing scales, mastering fingering technique, decoding the notes, and developing a feeling for the rhythm and beauty of the music are simultaneous but separate components of the process.

All this common-sense intuition–much of which underlay the famous phonics-intensive McGuffey Readers of the 19th century–is in fact supported by decades of 20th-century scientific research into how people actually learn how to read, starting with the work of Jeanne Stern licht Chall, a psychologist with a special interest in fostering the literacy skills of poor children who founded the Harvard Reading Laboratory at Harvard’s graduate school of education in 1966. Starting in the 1970s, a flood of reading studies–an estimated 10,000 in all–applied quantitative analysis and experimental, control-group-based research to identify the instruction strategies that teach reading most efficiently. The researchers included not only specialists in education and early childhood development but also experts in such fields as linguistics, psychology, neurology, genetics, anthropology, and sociology.

The resolutely apolitical NICHD, part of the National Institutes of Health, has been funding studies of reading development since 1964, and has sponsored longitudinal studies of 44,000 children in more than 1,000 schools since the early 1980s, tracking some of those children and their reading progress for more than 20 years as they grew to adulthood. It was all that research which led the NICHD to identify the five components that appear in Reading First’s enabling legislation (phonemic awareness, phonics, fluency, vocabulary, and comprehension). The value of these studies, their proponents argue, is that like all scientific studies, they are based on rigorous methodologies–assessments, for example, not only of how well children can read using various instruction strategies, but even how they move their eyes as they scan a printed page. And, like all valid scientific findings, the results can be replicated.

“We know how reading is acquired,” says Louisa Cook Moats, a protégée of Jeanne Chall’s at Harvard and director of the NICHD’s Early Reading Interventions project from 1997 to 2001. “It’s learning to process very specific kinds of linguistic information and build networks that coordinate phonological processing to the patterns of printed symbols that the eye sees, and it’s also connected to meaning and the building of vocabulary. When I explain how the process works to teachers, I compare it to an unraveled rope with the strands sticking out. The strands are all those beginning skills to be woven together in the rope.” Reading looks automatic and natural, Moats explains, but only because skilled readers are practiced enough to decode the symbols at lightning speed.

There are many causes for the resistance of the education establishment not only to the conclusions that Moats and others have drawn about reading instruction but to the research that underlies those conclusions. One bedrock philosophical principle, however, unites all those who oppose the step-by-step teaching of literacy skills: the notion that learning how to read is not at all like learning how to play the piano. Instead, the proponents of “whole language” instruction contend, it is a natural process akin to learning how to speak–something that children don’t have to be taught formally but pick up automatically if exposed to a sufficiently print-rich environment. Stephen D. Krashen, a professor emeritus of education at the University of Southern California and self-described “staunch defender” of whole-language strategies, explained in an email: “[A]ny child exposed to comprehensible print will learn to read, barring severe neurological or emotional problems.” Or, as Krashen amplified in a telephone interview: “Kids learn to read by reading.”

Hence the antipathy of the whole-language proponents to having children read a story out of a reader such as Houghton Mifflin’s; that doesn’t count as “real reading,” to borrow a phrase from Krashen’s email. Indeed, textbooks or any other kind of formal instructional material are eschewed. In elementary-school classrooms across the country, reading instruction typically consists of what is called “shared reading.” The teacher reads a story aloud to the class, often from a “Big Book,” an oversized, large-type edition of an illustrated children’s book of the teacher’s choosing that is propped up on a table or on the floor in front of the class. The teacher might read the story out loud several times, pointing out words that may be difficult, and then have the class read the story aloud in unison while the teacher turns the pages. There is almost no individual reading aloud, and the sounding out of words phonetically is actively discouraged as tending to turn youngsters into rote parsers of syllables who fail to understand what they are reading.

As for phonics per se, both Krashen and Yvonne Siu-Runyan insist that they indeed incorporate phonics instruction into their reading strategies, but only in elementary fashion and on an as-needed basis–”basic phonics,” as Krashen puts it. Whole-language instruction also typically includes periods of independent silent reading–”Drop Everything and Read” is the name for these impromptu sessions–in which the children pick out and peruse material of their choice from a classroom library of “leveled books”–that is, books that the teacher deems appropriate for their reading level. During these sessions the teacher typically “models” the process by dropping everything and reading silently from a children’s book, too, on the principle that seeing other people read encourages reading. As for vocabulary, whole-language classrooms typically incorporate a “word wall”–an ever-changing collection of large-letter words written on posters that the children chant together cheerleader-style and then write out.

The instructional principles behind whole language–light on formal content and heavy on assumptions that children will learn to read by feeling enthusiastic about reading–are far from new. Indeed, they date back to the end of the 19th century, to the educational theories of John Dewey (1859-1952), the pragmatist philosopher and educational theorist who held that children learn best not by directly absorbing instruction from their teachers in specific subjects such as mathematics or history, but by interacting with the real world. School, in Dewey’s thinking, should offer a simulacrum of real-world experience in which learning takes place obliquely as the child explores his or her surroundings under the guidance of a teacher. Dewey was in turn influenced by the romantic philosophy of Jean-Jacques Rousseau, who believed that children were naturally perfect and that education ought to consist of allowing them maximum freedom to develop their innate talents.

In 1904 Dewey joined the faculty of Columbia Teachers College, regarded then as now as America’s premier education school (U.S. News currently gives Columbia Teachers its No. 1 rating). From there Dewey’s “progressive” theories of pedagogy profoundly influenced several generations of American teachers and school boards, right up until the Sputnik launch of 1957, when it suddenly looked as though the Soviet Union, whose Communist leaders had kept in place a decidedly non-progressive education system dating from czarist days, had the United States over a barrel in science and technology. The Dick and Jane readers widely used in American elementary schools from the 1930s through the 1950s were offshoots of a branch of Dewey-ism that held that phonics instruction was backward and proposed that the way to make children literate was to expose them to simple words repeated interminably. (”See Dick. See Dick run. See Dick run fast.”) This so-called “look-say” pedagogy (a forerunner to whole language in its emphasis on context and meaning rather than sounds and letters) met its end after Rudolf Flesch published his bestselling Why Johnny Can’t Read in 1955, two years before Sputnik. By the early 1960s it looked as though progressive education had run its course in all but the most outré private schools. Jeanne Chall’s 1967 book Learning to Read: The Great Debate, proposed a return to thorough grounding in phonics, but in up-to-date combination with interesting children’s literature.

Then came a revolution in pedagogy that swept through the K-12 grades in the 1970s and 1980s as thoroughly as its college-level sister, postmodernism, swept through the academy. The revolution was called “constructivism.” Like postmodernism, it had its grounding in arcane Francophone theory: the ideas of the Swiss cognitive psychologist Jean Piaget. Piaget proposed that children progress through distinct developmental stages during which they acquire knowledge not simply by learning it from the outside but by “constructing” it from within, building upon and reflecting upon what they already know in order to rise to new levels of knowing. In Piaget’s theoretical dialectic, the subjective process of learning was more important than any particular content learned. Indeed, Piaget argued, it was crucial that the developmental process taking place within each individual child’s mind not be interfered with, but rather nurtured and encouraged by the child’s teachers. As the ubiquitous mantra of Piaget-influenced educational theory later put it, the teacher should be “a guide on the side, not a sage on the stage.” The essential constructivist principle is that teachers should teach nothing directly, but rather function as coaches while their students basically teach themselves.

This was Dewey’s progressivism with a new, fashionably Continental face. “The idea is that education is growth, education is development, and that children grow all by themselves,” said Diane Ravitch, an education policy analyst and author of Left Back: A Century of Battles over School Reform, a mordant critique of constructivism. “The idea is that children figure everything out for themselves,” Ravitch added. “There’s no authority.”

Piaget acquired an army of American apostles at education schools and elsewhere. Chief among them were Frank Smith, an Australian journalist-turned university instructor, and Kenneth Goodman, an education professor at the University of Arizona. Smith, whose 1971 book Understanding Reading derided the teaching of phonics, and Goodman are credited as the creators of whole-language theory. In a 1967 article in an education journal, Goodman had described the process of learning to read as a “psycholinguistic guessing game” in which children decipher words on a page, not by decoding them phonetically as Chall maintained, but by following “cues.” The cues, Goodman maintained, can be the individual letters and sounds in the word–or they can be the larger context of the story in which the word appears, the artist’s illustrations, or even (and perhaps especially) the child’s own previously acquired knowledge. Like Smith, Goodman argued that phonics instruction was useless at best, downright harmful at worst. “Matching letters with sounds is a flat-earth view of the world,” he declared in a 1986 book, What’s Whole in Whole Language. Dramatically turning centuries-old principles of reading instruction on their heads, Goodman maintained that “a story is easier to read than a page, a page easier to read than a paragraph, a paragraph easier than a sentence, a sentence easier than a word, and a word easier than a letter.”

Both Smith, who had never taught reading in an elementary-school classroom, and Goodman, who had, derided the use of textbooks, worksheets, and other formal instructional material. Smith’s 1986 book, Insult to Intelligence: The Bureaucratic Invasion of Our Classrooms, complained about children being forced by their elders to memorize mountains of useless data. (Memorization is generally considered in constructivist theory to be developmentally inappropriate for elementary school.) In whole-language theory, the teacher’s job is to identify the child’s errors–or “miscues,” as they are called–and nudge the child in the direction of the correct cues. “Drill and Kill” is their derisive term for pedagogy that emphasizes the systematic teaching of content.

Thus began the practice, now a bedrock of whole-language pedagogy, of teachers’ encouraging beginning readers to look at the first letter of any difficult word they encounter in a story and guess the rest, or if that strategy fails to produce results, simply to skip the word and return to it later. Although Goodman refused to be interviewed for this article, stating in a pair of dyspeptic emails that he would not respond to “negative” criticism of his theories, Yvonne Siu-Runyan provided an example of how a whole-language reading lesson works in practice. “A child encounters the word ‘butterflies’ in a story,” said Siu-Runyan. “The first time he reads it as ‘b-flies.’ Maybe the next time he reads it as ‘butt-flies’ and the next time as ‘betterflies.’ For me to assume he’s not going to get it would be a mistake, because finally he’ll say to himself, ‘Does this make sense?’ He’ll look at the pictures of butterflies [in the book] and say to himself, ‘Oh, this is a story about butterflies!’ And he’ll get it right after that. It’s a lot more complicated a process than handing a child a list of words.”

Whole language and other aspects of constructivist theory swept through the education schools, starting with the flagship Columbia Teachers College, where Dewey’s progressive influence had never waned, where courses on reading pedagogy to this day concentrate on erecting a “theoretical framework” for instruction rather than teaching teachers what actually works in classrooms, and where the school’s publishing affiliate, Teachers College Press, churns out dozens of constructivist treatises every year. Smith and Goodman crisscrossed the country on the ed-school lecture circuit, where they were welcomed with open arms and standing ovations by professors and students alike. Whole language clearly appealed because it allowed teachers to do essentially what they liked in their reading classes, and it relieved them of the arduous work of ensuring that their students had mastered specific literacy skills. Teachers and administrators rushed to create “child-centered” and “learner-centered” curricula in every field and at every grade level (”learner” being the fashionable ed-speak word these days for “student,” as it connotes the constructivist idea that children take charge of their own education).

Sandra Wilde, an education professor at Portland State University in Oregon, deemed that learning how to spell, like learning how to read, “should ultimately be as natural, unconscious, effortless, and pleasant as learning to speak,” so spellers went the way of readers in classrooms across the country. Teachers encouraged youngsters to make up their own “invented” or “independent” spelling, also under the influence of Wilde’s self-described “holistic” approach, which theorized that children could learn from their spelling “miscues.” Wilde drafted a “Speller’s Bill of Rights” that included “the right to be valued as a human being regardless of your spelling.” Whole-language advocates and other constructivists also abandoned conventional tests and letter grades, which they thought slighted youngsters’ individuality, in favor of what they called “authentic assessment.” That usually means having students assemble samples of their work in a “portfolio” (the oversized envelope that artists take to job interviews) that the teacher then evaluates verbally.

Systematic lessons in grammar, handwriting, and punctuation also went by the boards, thought to be developmentally inappropriate for young children. The teaching of writing completely changed focus. Teachers in the primary grades had traditionally taught their students first how to construct grammatical and properly punctuated sentences, then how to form paragraphs, and finally how to build paragraphs into simple essays and stories. All this was abandoned in favor of a kind of writers’ workshop approach that focused on students’ self-expression and personal reactions. “Journaling,” which allows youngsters to choose their own topics to write about, became a favored classroom writing activity, even for kindergartners and first-graders. Students were encouraged not to worry about grammatical and spelling errors, as these could be cleaned up in an “editing” process with the teacher. Imitating the graduate writing program at the University of Iowa and the copy-desk procedures at the New Yorker was supposed to turn 6-year-olds into sophisticated writers, critics, and thinkers.

Two education professors at Indiana University of Pennsylvania, Donald A. McAndrew and C. Mark Hurlbert, in an award-winning 1993 article in the journal of the National Council of Teachers of English, went so far as to urge students to indulge in “intentional errors” of syntax and usage as a way of rebelling against the “tyranny” of standard English usage. In 2003 the National Council took its own insurrectionist stand against standard English, voting to endorse a manifesto titled “Students’ Right to Their Own Language”–namely the right to write their homework in hip-hop-ese, Spanglish, Valley Girl talk, or whatever other nonstandard dialect they believe best expresses their “community” or “personal” identities. Many whole-language teachers do not bother to prepare lesson plans or syllabi, relying instead on querying their students on what they would like to learn on any particular day.

Like their opposite numbers in the reading science community, whole language advocates can point to plenty of published research, fattening the education journals and bolstering what the whole-language proponents insist is their superior approach to teaching literacy. That research, however, almost uniformly consists of anecdotal recollections by its authors of eureka! moments in their classrooms. The story that Siu-Runyan narrated about the child who finally deciphered the word “butterflies” is a perfect example. The education-school slang term for such “qualitative” (in contrast to quantitative) observations, analogous to the material that anthropologists record in their field notebooks, is “kidwatching.” Almost all kidwatching research consists of teachers’ first-person success stories–because whole-language advocates are human and they almost never report their classroom failures. “But they’re sure that those reports [in the education journals] are 100 percent scientific,” says Patrick Goff, a professor emeritus of education at San Diego State University in California and reading science advocate. “That’s because you can get a Ph.D. in education without ever having to read a single quantitative study. Even my own university would not teach its students about the empirical evidence concerning the teaching of reading.”

Fortunately, perhaps, for about 40 to 50 percent of children–the socioeconomic top 40 to 50 percent hailing from upper-middle-class-to-wealthy “print-rich” homes where the reading of books, magazines, and newspapers is an everyday occurrence–whole-language reading pedagogy does little if any harm. The most verbal of these youngsters, the gifted offspring of lawyers, college professors, and Hollywood screenwriters, either already know how to read by the time they get to kindergarten or pick up reading quickly no matter how they are taught. Others who are not so naturally verbal struggle with whole language’s guessing games and unsystematic instruction but eventually manage to read at grade level and to write and spell passably. Furthermore, many whole-language proponents, such as Siu-Runyan and Krashen, are clearly patient, gifted, imaginative teachers sensitive to their students as individuals (Siu-Runyan says she slips structure into her student-interest-driven lesson plans, and Krashen, who currently teaches in a suburb of Portland, Oregon, where whole language is officially verboten, runs his classes as a kind of Dead Poets Society, ignoring the ban while the administration looks the other way).

Indeed, even the staunchest supporters of the five-component scientific approach to literacy acknowledge that whole language’s emphasis on child-friendly classrooms and high-quality children’s literature are valuable contributions to pedagogy. Those desks arranged in clusters, not rows, the children sitting on the floor, and the plethora of stimulating books in Laverne Johnson’s classroom at Ginter Park represent some of the best of whole language’s legacy. Finally, many affluent parents with progressive political leanings actually prefer the unstructured, arts-and-crafts-oriented methodology of constructivism, which is why private progressive elementary schools such as the Dalton School in Manhattan and the Peninsula School near San Francisco continue to flourish (by the time those children enter high school, though, SAT cram courses and the rat race for Ivy League admissions are the order of the day; few of America’s top private prep schools operate on progressive pedagogical principles).

The children who suffer from the whole-language revolution are that bottom 40 percent of American children, the poor and near-poor who come from households where books are seldom seen and where unschooled parents have starved their offspring of the rich vocabulary and cultural exposure to which better-off children are accustomed as a matter of course. Children whose parents don’t speak English at home fare worst of all in whole language. This group of low-income, print-deprived children is the group that needs direct reading instruction most desperately, and as the results in Richmond indicate, benefits from it most dramatically.

Long before Reading First became law in 2002, there had been a backlash against whole language by parents and school superintendents unimpressed by their students’ low test scores despite being assured that their children were being taught according to the most up-to-date ideas. In 1987 the state of California mandated a whole-language approach to reading and writing. Within a few years California’s reading scores on the NAEP test plummeted to third-lowest in the United States and its overseas territories; only Louisiana and Guam ranked lower. The decline stretched across the socioeconomic board, among the offspring of the college-educated as well as the offspring of Hispanic immigrants.

Jill Stewart, a writer for the Los Angeles Weekly, visited a second-grade classroom at a highly regarded school on Los Angeles’s wealthy Westside. There she met a little girl who wrote “I go t gum calls” for “I go to gym class” in a journal that was entirely free of punctuation (which hadn’t been taught yet). In another classroom, a 7-year-old boy had gotten by with memorizing the “shared reading” story that the teacher had read over and over but could not actually read a single word of the story on his own. At one Los Angeles school parents held nacho sales to buy their classrooms forbidden spellers. In Charles Sykes’s book Dumbing Down Our Kids, a mother complained that her fourth-grade daughter had received a grade of check-plus (above average) and a teacher’s notation of “Wow!” for these sentences: “I’m goin to has majik skates. Im goin to go to disenelan. Im goin to bin my mom and dad and brusr and sisd. We r go to se mickey mouse.”