by Art Levine
IN education no question has produced so much bitter debate for so long as this one: What is the best way to teach children to read? It is a critical issue, because there is clearly a need for drastic improvement in the way our schools do this essential job. As many as 20 percent of Americans above the age of sixteen are classified as functionally illiterate—unable to use print to perform essential tasks—and the ranks are growing every year. Even as the literacy crisis deepens, partisans of different reading approaches square off against each other. New teaching fads come and go.
The latest of these fads is known as the whole-language approach to literacy. It exposes children to interesting reading and writing at the expense of systematically teaching specific reading and writing skills. Whole-language teachers, for instance, encourage young students to recite along with them as the teachers read aloud from entertaining big-print books. One of their central beliefs is that language should be learned from “whole to part,” with word-recognition skills being picked up by the child in the context of actual reading, writing, and “immersion” in a printrich classroom. It is a philosophy that has won the backing of influential teachers’ organizations, state and local education agencies, and tens of thousands of enthusiastic teachers. In Northfield, Illinois, for example, a reading specialist, Kenneth Smith, says that ever since the town’s schools adopted the whole-language approach, a few years ago, “our kids are more motivated to read and write, and there’s been an upturn in comprehension.”
Unfortunately, there’s little hard evidence to back up such claims across the board—and good reason to be concerned about the whole-language method when it’s used to the exclusion of other approaches. It can deprive children, particularly low-income and other disadvantaged students, of the intensive instruction in phonics—the study of sound-letter relationships—that they need to master reading. “The widespread rush into whole language could have real dangers for certain kinds of kids, and it’s scary to me,” says Robert Slavin, a co-director of the Center for Research on the Education of Students Placed at Risk, at Johns Hopkins University.
Evidence is mounting that the underlying premises of the wholelanguage approach may be false.
Evidence is also mounting that the approach may not be all that effective, and that its underlying premises may simply be false. Some of the many school districts that have tried the whole-language method have been attacked by educators and parents for a subsequent drop in test scores. In Texas a significant minority of schools have elected to spend their own money on phonics-based reading programs rather than use whole-language courses funded by the state; in Canada several citizens’ groups have been formed to protest the widespread adoption of whole-language methods; and in England in late 1992 the British Education Secretary, alarmed at declining achievement scores, called for a return to instruction in basic skills. His action followed an advisory panel’s report that derided “foolish methods that often lead to poor results,” including Britain’s version of whole-language teaching. Yet the whole-language philosophy continues to gain adherents among educators.
To be sure, even some critics of the method concede that traditional phonics instruction could benefit from several of the whole-language method’s innovations, and they are seeking to combine the best of both approaches. Common sense would seem to suggest that most teachers should, in fact, blend systematic skills instruction with the use of appealing literature—and many, in fact, do. But in education, a field too often dominated by zealots and crazes, balanced approaches are sometimes harder to find than those shaped by the swinging pendulum of pedagogical fashion.
THE whole-language method is the latest incarnation of the “meaningfirst” approach that has warred with the “phonics-first” approach to reading for more than a century. The traditional practice was to teach children the alphabetic code—the translation of abstract letters into sounds and words—before turning to actual reading. This approach was challenged in the mid-nineteenth century by Horace Mann, the secretary of the Massachusetts Board of Education. With a vehemence not unlike that of today’s wholelanguage proponents, Mann denounced the letters of the alphabet as “bloodless, ghostly apparitions” that were responsible for “steeping [children’s] faculties in lethargy.” He argued that children should be taught whole, meaningful words first, and promoted the “look-say” method that by the 1920s had come to dominate American education.
In the 1950s, however, the meaningfirst approach was itself denounced, in bitter, still-resonant political terms, in Rudolf Flesch’s best-selling book Why Johnny Can’t Read. Flesch’s broadside led to a wave of authoritative new studies. These concluded that reading programs that include systematic, intensive phonics instruction work better than those that do not. By the early 1970s most schools had returned to an essentially phonics-based program. But as the pendulum swung once again, these programs found themselves being criticized by some teachers and academics for killing off children’s interest in reading.
Most of the critics initially directed their fire, with some justification, at conventional teaching tools. They denounced worksheets used to drill students with tedious rows of words, the “basal readers” with their dull Dick-and-Jane-style stories, and reading groups organized by ability which stigmatize children and often promote failure.
Whole language supposedly offers an easy way to literacy. Ken Goodman, a professor of education at the University of Arizona, is one of the whole-language movement’s leading academics. He and his colleagues have argued over the years that learning written language can be as natural as acquiring spoken language, and that children can learn to read primarily by figuring out the meaning of words in context. “Good readers don’t read word by word,” Goodman says. “They construct meaning from the [entire] text.” Indeed, “accuracy is not an essential goal of reading.”
Whole-language advocates believe that children should learn at their own pace and shouldn’t be pushed too hard too soon. “We don’t believe that phonics should be taught through ‘skill and drill,’” says Bess Altwerger, an associate professor of education at Towson State University, in Baltimore. Phonics and related skills are supposed to be learned by the child in the context of actual reading and writing. The teacher should intrude only minimally into this process of discovery, at most waiting for the occasional “teachable moment” to pass along reading tips. As Goodman has said, “One cannot reconcile direct instruction with natural learning.”
Adherents also contend that whole language is the best approach for minorities and disadvantaged students. “We don’t lose struggling readers, because they don’t get a defeatist attitude,” says Mary Katsafanas, a talented and dedicated first-grade teacher in suburban Carroll County, Maryland. A movement publication, The Whole Language Teachers Newsletter, underscores that attitude: it recommends teaching children confronted by an unfamiliar word to “skip it, use prior information. . . . or put in another word that makes sense.” It warns, “Don’t sound-it-out.” Although Katsafanas does sometimes gently suggest that students use letter sounds to figure out words, she sees the mistakes her students make as part of a journey toward literacy rather than as errors to be criticized. During a visit to her class I heard a child read aloud from a rudimentary picture book; the girl wasn’t corrected when she read the word “hug” instead of “cuddle.” “She has the concept of the story,” Katsafanas said later. This child lagged behind her classmates in both reading and writing.
The need for any explicit, systematic phonics instruction is a “myth,” says Marie Carbo, the president of the National Reading Styles Institute, in Syosset, New York. Karen Smith, the associate executive director of the National Council of Teachers of English (NCTE), says, “Phonics is just one part of the reading process. The child learns surrounded by language. It’s just like oral language: through interaction with other language users, all of a sudden things start making sense.” Besides the NCTE, the International Reading Association and other teachers’ groups have endorsed the tenets of the whole-language movement. An independent organization, The Whole Language Umbrella, claims about 25,000 members and 450 local chapters (or teachers’ support groups) in North America. The whole-language method is now used by perhaps a fifth of all teachers of reading, with an even larger proportion adapting elements of it, such as the greater use of authentic children’s literature and the decreased use of intensive phonics.
The whole-language approach is also favored by most teachers’ colleges and university-level education programs. Its influence is so pervasive that in 1987 a survey of forty-three texts used to train teachers of reading found that none advocated systematic phonics instruction— and only nine even mentioned that there was a debate on the issue. Some form of the whole-language ideology has been adopted by more than a dozen state education agencies.
DESPITE the spread of whole-language instruction, research and experience not only fail to demonstrate its superiority but also make a persuasive case for the importance of phonics.
Since 1981 only a few dozen studies in reputable education journals have even attempted to compare the reading scores achieved by the whole-language method with those of other methods of teaching reading, and the conclusions are contradictory, at best. A 1989 overview paper in the Review of Educational Research found that the scanty whole-language research and the more extensive studies on its predecessor, the “language experience” approach, when taken together, show that they produced somewhat worse results in reading comprehension and worse results with disadvantaged students than traditional methods did. A similar 1991 University of Maryland survey, published this fall in Educational Psychologist, looked at whole-language research and found that most of the studies examined demonstrated no difference in outcomes—but also were too poorly designed to have much validity in any event. (There are exceptions: Penny Freppon, at the University of Cincinnati, for example, has done credible controlled studies that seem to support the idea that whole language can affect some reading measurements.)
The federal government, philanthropic foundations, and universities have sponsored other major studies on reading. Though these were not intended to compare whole language and phonics, the findings have been generally supportive of the intensive phonics programs derided by whole-language proponents. Low-income and slow students appear to benefit especially from explicit phonics instruction. These findings have been summarized in such reports as The Great Debate, by the Harvard education professor Jeanne Chall, the Commission on Reading’s Becoming a Nation of Readers, and a 1990 report sponsored by the U.S. Department of Education, Beginning to Read, by the psychologist Marilyn Adams. This last report concludes that the vast majority of [program comparison] studies indicated that approaches including intensive, explicit phonics resulted in comprehension skills that are at least comparable to, and word recognition and spelling skills that are significantly better than those that do not.
Phonics alone, of course, is not sufficient to promote strong reading skills. As Adams emphasized in Beginning to Read, children need both direct skills instruction and exposure to interesting reading.
Adams’s report sought and found common ground between the wholelanguage and phonics camps —but it didn’t heal the rift between the extremists. Those partial to phonics are not, for the most part, the ones who have proved intransigent. At a meeting of the International Reading Association four years ago. for instance, Ken Goodman attacked Adams as a “vampire” who threatened the literacy of America’s youth. Constance Weaver, of Western Michigan University, and other whole-language proponents question the reliability of pro-phonics research. Indeed, many whole-language advocates don’t accept conventional standards of proof, and criticize virtually any reading test for being an artificial assessment of a narrow set of skills. Professor Carole Edelsky. of Arizona State University, in a widely discussed paper in Educational Researcher, rejected calls for rigorous quantitative or test-score-based research on the whole-language approach as an “outrageous” effort to impose an old “paradigm” on the movement. “Testscore evidence doesn’t tell you what students actually do when they read,” Edelsky argues. She and others favor a qualitative approach called “kid-watching,” which uses such devices as freeform essays and portfolios that measure student progress over time—evaluation tools that are gaining mainstream acceptance. But many of the movement leaders remain hostile to controlled studies or other concrete measures of proof.
There have been troubling reports from the field about whole language. Reading scores of first-graders in San Diego dropped by about half when whole language was introduced there in 1990. One school-board member complained to the San Diego Union, “From parents I get lines like, ‘You’re experimenting with my kid.’ I hope we didn’t make a mistake—but I feel in my gut that we have.”
In Houston in the fall of 1991, eight inner-city elementary schools asked the school district to allow them to return to phonics-based instruction after a few years of trying a whole-language-style program. “Last year’s reading scores were the lowest in twenty years,” LaSalle Donnell, the principal of Houston’s Douglass Elementary School, said at the time. “If the kids can’t decipher words, they can’t enjoy a whole-language program.” During the 1992-1993 school year, after an intensive phonics program was adopted, the school’s scores on a state reading test rose an astonishing 48 points. At that time 98 percent of third-graders were reading at or above grade level, and the school’s reading scores are still among the best in the city.
The most disturbing reports have come from urban school districts with significant numbers of low-income and minority students (not particularly surprising, given the studies showing that disadvantaged students do worse in programs that make little use of phonics), in large measure because, unlike most middle-class children, they aren’t getting much lettersound instruction at home. Yet there are some signs that even middle-class students, who often have received informal tutoring in phonics from their parents, may suffer in certain whole-language classrooms.
For this reason the tale of the Cape Elizabeth, Maine, schools merits attention. For about five years a whole-language program—without any systematic phonics instruction at all before third grade—was offered in this small middleclass suburb. Then some parents began complaining that their younger children weren’t reading as well as their older ones had. A new superintendent, Constance Goldman, responded to those concerns by conducting district-wide standardized testing for second-grade students in 1991. The tests, although given several months earlier than usual, offered a worrisome, if rough, snapshot of students in trouble: 42 percent of the second-graders read below grade level, a far higher proportion of low-scoring students than in comparable communities. Goldman, disturbed by the results, moved to create a more balanced curriculum, keeping elements of the whole-language philosophy but adding more direct skills instruction. Reading scores improved. “At some point you have to set standards,” she says now.
IF questions can be raised about an exclusive reliance on the whole-language method in practice, it should not be surprising that they can also be raised about it in theory.
Since the 1960s Ken Goodman and Frank Smith, for many years a psychology professor at the University of Victoria, in British Columbia, have argued that skillful reading involves getting meaning largely from the context of the entire passage, rather than reading word by word. By sampling words and parts of words, readers engage in a “psycholinguistic guessing game” that enables them to predict the words they’ll encounter. Sounding out words plays only a small role, Goodman and Smith say, because it’s too cumbersome for the human mind to process every letter and every word.
Given this theory, the playing down of phonics and the emphasis on guessing words makes a kind of sense. But in the two decades since the whole-language philosophy first became influential in academic circles, a considerable amount of research has been published showing Goodman’s and Smith’s psycholinguistic theories to be wrong. Here is a brief survey.
• Contrary to the whole-language belief that readers “sample” text, skillful readers actually read virtually every word and every letter of the texts they encounter. That’s the basic conclusion of computermonitored eye-movement research by Keith Rayner, of the University of Massachusetts, and other scholars. Skilled readers are able to breeze through texts effortlessly because their knowledge of phonics has made word recognition automatic—thus freeing their minds to focus on meaning.
• The critical reading ability is word identification, rather than the use of context to figure out words. Charles Perfetti, of the University of Pittsburgh, for instance, has shown that readers who perform poorly when reading words out of context also do poorly at reading comprehension. “The skilled reader has such automatic facility at identifying words that he doesn’t need context,” says Frank Vellutino, the director of the Child Research and Study Center at the University of Albany. Philip Gough, of the University of Texas, has found that even skilled readers correctly guess no more than one in four words missing from a text. Whole language’s emphasis on context was undermined in a different way in December of 1991, when the Journal of Educational Psychology published a study challenging an oft-cited 1965 experiment by Ken Goodman. The original experiment had found that children made 60 to 80 percent more errors when reading words out of context than when reading them in context. But Tom Nicholson, a senior lecturer in education at the University of Waikato, in New Zealand, repeated Goodman’s work and determined that Goodman had failed to distinguish properly between good and poor readers; it was only the younger and poorer readers who made significantly more errors reading words out of context.
• Understanding letter-sound correspondences is indispensable to skilled reading. Frank Smith has argued. “To the fluent reader, the alphabetic principle is completely irrelevant.” In fact Jeanne Chall, of Harvard, and other researchers have demonstrated that two powerful predictors of future reading success are a knowledge of the alphabet and an awareness of the speech sounds that make up words—the very skills that are often lacking in disadvantaged students when they enter first grade.
Research of this kind, though, hasn’t prompted whole-language theorists or their followers to modify their beliefs. Goodman argues that researchers in the phonics camp use short written passages and artificial settings that do not reflect real-world conditions. “They can play games in the laboratory,” he says, “but teachers know that that’s not the way people read.” Frank Smith dismisses most of the computer-based research on thinking and reading as “irrelevant and misleading.”
WHATEVER its limitations, the whole-language approach does have something to offer. Its stress on reading enjoyable children’s literature rather than dull primers is surely worthwhile (though little more than common sense). Whole language also offers a supportive and tolerant atmosphere in which to learn to read. The emphasis on early writing wins broad support among the various reading factions, because it can help improve reading and thinking skills.
For its part, traditional phonics instruction also has limitations. Although the use of genuine literature and writing assignments in schools has grown, a majority of classrooms still devote too much of their reading-instruction time to simplistic workbooks and mind-numbing drills. Learning phonics this way is dull and time-consuming. Equally dubious is the widely advertised “Hooked on Phonics” program, which promises miracles for beginning readers, young and old alike. In 1991 a panel of reading experts questioned the effectiveness of this method, since it concentrates on teaching letter sounds without offering any meaningful reading.
Fortunately, a few promising programs have combined quick, accessible phonics instruction with whole-language-style activities. Perhaps the clearest demonstration that children benefit from such a combination is a program for troubled readers called Reading Recovery, designed by the New Zealand educator Marie Clay. For thirty minutes a day over an average of sixteen weeks, a specially trained teacher works with children individually to build their confidence by exposing them to short, charming, gradually more difficult books; mini-lessons in writing; and magnetic letters that the children play with to form words. The program is widely used in Ohio, Illinois, California, Texas, and New York, among other places. Early research showed that roughly 85 percent of Reading Recovery graduates were reading at levels comparable to their peers’ up to three years later—a rate of success that no other remedial program comes close to matching.
Reading Recovery costs a few thousand dollars per child, but its eclectic principles could be applied less expensively across entire classrooms. Unfortunately, whole-language extremists aren’t eager to let that happen. The program itself, admittedly, has been embraced by both phonics and whole-language advocates, with the latter being particularly likely to cite Reading Recovery’s success as another example of their achievements. But Marie Clay says, “No, this isn’t part of whole language.” The truth that whole-language purists don’t want to admit is that Reading Recovery’s success is due to its shrewd combination of teaching strategies from different pedagogical camps, including the use of phonics. So they deride those who are trying to forge just such a blend of methods. “It’s asinine,” Ken Goodman says. “We’re not going to solve anything by trying a little of this and a little of that.”
But it may be exactly that diverse approach that stands the best chance of creating classroom after classroom in which everyone can read.