The Reading Wars, How Whole Language Lost a Generation and Phonics Won Back

Imagine a hospital that told surgeons to try the new method first and read the research afterward. Now imagine it took forty years for the research to win. That is the story of American reading instruction, and the patients were children.

The argument of this post is simple, and I will state it before I defend it. For roughly half a century, many American classrooms taught beginning readers to guess at words from pictures and context, the evidence that explicit, systematic phonics works better was available the whole time, and the correction, where it has come, came from legislatures and parents, not from the education establishment. Let me define the terms, number the premises, give the other side its best case, and then walk through the record.

Defining the terms

Phonics is instruction that teaches children the relationship between the sounds of spoken language and the letters that represent them. Systematic means the relationships are taught in a planned order, not whenever a word happens to come up. Explicit means the teacher says it outright: this letter, this sound, now blend them.

Whole language is a philosophy, not a method. It holds that children learn to read the way they learn to talk, by immersion in meaningful text, and that reading is a natural process driven by meaning. Phonics, in this view, is a tool of last resort and best taught in passing.

Three-cueing is the practical tool that grew out of that philosophy. When a child hits a word he cannot read, the teacher asks him to consider three cues: meaning (does it make sense?), structure (does it sound like a sentence?) and visual information (does it look like the word?). In practice the third cue often came last, so a child looking at a picture of a horse might "read" the word pony off the page and be praised.

Balanced literacy is the later and softer name. It promises a balance of approaches, and in many classrooms it kept the guessing habits while adding a little phonics on the side.

The first warning, 1955

In 1955 an Austrian-born writing consultant named Rudolf Flesch published Why Johnny Can't Read. It was a polemic, and it was not subtle. Flesch argued that American schools had abandoned the teaching of letter-sound correspondences in favor of the "look-say" method, in which children memorized whole words by sight, and that this was producing readers who could not decode anything they had not been shown.

Flesch's book sold well and made the teaching establishment furious. Notice the pattern, because it repeats. The critique came from outside the profession, it was written for parents, and it was dismissed as unscholarly. Professionals in the field answered that Flesch was not a reading teacher, which was true, and that his claims were oversimplified, which was sometimes true. What nobody did was run a decisive comparison and settle it. The argument went back into the faculty lounges and stayed there.

The guessing game, 1967

Twelve years later a scholar gave the other side its most influential sentence. In 1967 Kenneth Goodman published a paper titled "Reading: A Psycholinguistic Guessing Game." Goodman's claim was that skilled reading is not a matter of seeing every letter and sounding it out. Skilled readers, he argued, sample the text, predict what comes next from context and grammar, and confirm or correct. Errors a child makes, which Goodman called miscues, were not failures to decode but windows into the child's thinking.

Notice what is true here. Skilled adult readers do use context, and good readers predict. That is why the idea persuaded so many serious people. What does not follow is the conclusion that beginners should be taught to read that way. The later research on eye movements, which is easy to find and worth reading, found that fluent readers fixate on nearly every word and process the letters within it. They do not skip around guessing. Context helps them with meaning. It is the decoding that does the heavy lifting, and decoding is what the guessing game left out.

Here is the difference in one sentence. Good readers use context to confirm what they decoded. Poor readers use context because they cannot decode. Teaching a child to behave like a poor reader and expecting a good one to emerge is a strange plan.

The steelman

The other side deserves its best case, in words its defenders would accept.

Whole-language teachers cared about something real. They did not want children to hate reading. They had seen classrooms where reading meant dull drills, worksheets, and basal readers with sentences nobody would say aloud. They wanted children surrounded by real books, reading for pleasure, writing their own stories, and discovering that print carries meaning and joy. They were right that motivation matters, that vocabulary and background knowledge matter, and that a child who can sound out words but never reads a book has not really learned to read. They also argued, with some justice, that phonics worksheets in the hands of a poor teacher can be as deadening as anything.

That is a good case, and I would add one thing: most whole-language teachers were sincere, hardworking, and loved their students. The failure is not a failure of character.

Now the answer, in numbered points.

1. The goals were right and the method was wrong. Nobody opposes joy in reading. The question is whether joy comes from decoding fluently or from being told to guess. Children who can decode read more, because reading is no longer painful. Motivation follows competence.
2. Phonics and real books are not rivals. A well-run phonics program is a few minutes of explicit teaching followed by decodable text and then rich read-alouds. The false choice was manufactured by the philosophy, not by the research.
3. Learning to speak is not like learning to read. Humans are wired for spoken language and acquire it with no formal teaching. Written language is a cultural invention a few thousand years old. Alphabetic writing has to be taught, because nothing in a child's brain evolved to find letters on a page. The analogy at the heart of whole language fails on the biology.
4. The harm falls unevenly. Children from homes full of books and talk often learn to read in spite of weak instruction. Children without that cushion rely on the school entirely. Guessing at words is a luxury only the lucky can afford.

California, 1987 and after

The movement did not stay in the faculty lounge. In 1987 California adopted a new English-language arts framework that put literature, immersion and whole-language ideas at the center of instruction and pushed phonics to the margins. California was the largest textbook market in the country, and publishers followed it. The reasoning in Sacramento was well-meant: the previous approach had bored children, and a literature-rich program would inspire them.

Then came the measurements. In 1992 and 1994 the National Assessment of Educational Progress tested fourth-grade reading state by state, and California's fourth graders landed near the bottom of the states that participated. I will not trust my memory for the exact ranking, and you do not need it. Look up the 1994 results for yourself, because the pattern was unmistakable and it embarrassed Sacramento.

The reaction is one of the more instructive episodes in modern education policy. By the mid-1990s the legislature and a state task force were calling for explicit instruction in phonics and decoding, and a new set of state requirements restored them to the early grades. California reversed course within about a decade. The state that exported the whole-language approach to the nation's textbook publishers imported the correction from a legislature that had seen the scores.

Mark that sequence for what it teaches. A philosophy was adopted statewide on the strength of its appeal. It was measured against real children. It failed the measurement. And a political body corrected it, because a political body can be voted out, which a faculty consensus cannot.

The National Reading Panel, 2000

By the late 1990s Congress had grown tired of the argument and asked for an answer. In 1997 it directed the National Institute of Child Health and Human Development, working with the Department of Education, to form a National Reading Panel to assess the research on how children learn to read. The panel reported in 2000.

It looked at the research literature through the lens of experimental and quasi-experimental studies, the kind that can show cause and effect, and it examined five areas: phonemic awareness, phonics, fluency, vocabulary and comprehension. On phonics, the panel found that systematic phonics instruction was more effective than non-systematic or no phonics instruction for kindergarten and first-grade children, and for children at risk of reading difficulties. It found benefit across the early grades.

You can fairly criticize the panel. Its charge was narrow, it did not study everything, and some of its critics argued that its summary of its own findings overstated the case for phonics. Those are fair points about emphasis. They are not a reason to ignore the central result, which several later reviews have supported: teach the code explicitly and in order, and more children learn to read.

Reading First, a lesson in federal limits

The National Reading Panel's findings fed directly into the federal No Child Left Behind Act of 2002 and its Reading First program, which sent money to states and districts to adopt reading programs grounded in "scientifically based reading research." On paper it was a victory for evidence.

In practice Reading First is a case study in the limits of Washington. The program was criticized in congressional hearings and in the Education Department's inspector general reports over how some contracts and advisors were selected. Then the federal evaluation, released in 2008, found that Reading First increased the time spent on the core elements of reading instruction in the early grades, by a few minutes a day, but did not produce a statistically significant improvement in students' reading comprehension. Congress zeroed out the funding soon after.

I want to be careful and fair here, because opponents of phonics have leaned on that evaluation for years. It does not show that decoding is unimportant. Decoding is a necessary condition for comprehension, not a sufficient one. A child who decodes fluently still needs vocabulary and knowledge of the world to understand what she reads, and a program that taught the code while starving the content would not move comprehension scores. That is a legitimate criticism of how the money was spent. It is also a reminder that national programs fail for ordinary reasons: contracts, compliance, paperwork and the difficulty of changing what thousands of teachers do on Monday morning.

The quiet persistence of the guessing habit

By the 2010s the research case was settled among reading scientists, but the practice was not. Popular commercial curricula built on balanced literacy and three-cueing stayed in wide use. Many teachers had been trained in education schools that never taught them the science of how the brain learns to read, and had no reason to doubt what they were taught. Parents who noticed a struggling child were often told to be patient, that children develop at different rates, and that the reading would "click."

The corrections began with parents. Families of children with dyslexia were among the first to organize, because their children could not survive on guessing. They learned the research on their own, hired tutors, and then turned to legislatures. Journalists then began asking why, if the science was known, so few teacher-preparation programs taught it. The answer was uncomfortable: because a philosophy is easier to keep than to confess you were wrong about.

Here I would like to be fair about what we do and do not know. The claim that whole language lost a generation is an inference about causes from a large pattern. We cannot rerun history and teach the same children by another method. What we can say is that the national reading results for fourth and eighth graders have been flat or worse for decades, that the gap between the best and worst readers has widened, and that the states that moved decisively toward the evidence have produced the best-documented improvements. Put together, those facts make the inference strong, though not mathematically certain.

Mississippi, 2013

The most striking case came from a state nobody had expected to lead. In 2013 Mississippi passed the Literacy-Based Promotion Act. It did several things at once: it required screening of young children for reading difficulty, it put reading coaches into low-performing schools, it required retraining of teachers in the science of reading, and it set a third-grade reading gate that ties promotion to a reading standard, with good-cause exemptions and extra support for children who fall short.

Look at what the bill did not do. It did not spend a fortune on a new building program, it did not wait for a federal grant, and it did not accept the explanation that poverty settles the question. Mississippi was and remains one of the poorest states in the country.

The results were closely watched. Mississippi's fourth graders had scored near the bottom nationally in reading in 2013, and by the 2019 NAEP they had risen to a level close to the national average, a gain that few other states matched. Remember that I told you to be careful about causation. Mississippi did several things at once, and economists and education scholars debate how much each one contributed. The retention policy in particular is argued over. But the part of the package aimed at teaching the code is the part with the best evidence behind it, and the outcome is hard to explain without it.

Other states have followed with laws requiring screening, structured literacy training and instruction aligned to the research. A number of them also banned the use of three-cueing as a program of instruction for beginning readers. I will not give you a count, because the list changes as legislatures meet. Check your own state legislature's website and see where it stands.

What this does for your family

If you are a parent, here is what the history means in practice. It is not medical, legal or financial advice, just a short list of questions you are entitled to ask a school.

• How does the school teach children to read words they have not seen? A good answer talks about sounds, letters and blending. A worrying answer talks about looking at the picture, guessing what makes sense, and "reading the first letter and skipping ahead."
• Is the program systematic, with a planned sequence of letter-sound relationships, or does it teach phonics only when a word seems to need it?
• Are the early books decodable, meaning they use the letter-sound patterns the child has been taught, or are they predictable texts a child can memorize?
• Does the school screen every child for reading difficulty in the early grades, and what does it do when a child falls behind?

We have Is Your Child Being Taught to Read With Phonics? A Parent's Checklist that goes through these questions in detail. For the data side of the story, see What the Nation's Report Card Measures and What Proficient Really Means, because the reading scores that started so many of these conversations come from there. If you want the older tradition of teaching reading and virtue together, read about The McGuffey Readers Taught Reading and Virtue to Millions of Americans, which taught millions of children from a sequence of lessons that started with letters and sounds. And the question of who decides is answered in Who Actually Runs Your School District, Board, Superintendent or State.

The Founders' stake in a literate people

There is a constitutional dimension here, and it is not a stretch. The founding generation believed self-government required an educated people. The Northwest Ordinance of 1787 declared that "religion, morality, and knowledge" being necessary to good government and the happiness of mankind, schools and the means of education should forever be encouraged. Madison, writing late in life, said a popular government without popular information is a prologue to a farce or a tragedy. You can read more in the post on Religion, Morality and Knowledge, the Northwest Ordinance's School Clause.

A republic of citizens who cannot read the Constitution, the ballot, the contract or the statute is not self-governing. It is governed by whoever reads on its behalf. That is the deepest reason the reading wars matter. A child who cannot read is not merely behind in school. He is dependent, for the rest of his life, on the people who can.

The case in brief

1. Whole language and three-cueing taught beginners to guess at words, a habit that skilled readers do not use, and the evidence for explicit, systematic phonics was available from the 1950s onward.
2. California's 1987 framework spread the guessing approach nationwide through the textbook market, and California's own scores forced a reversal in the 1990s.
3. The National Reading Panel of 2000 found systematic phonics more effective in the early grades, and later federal programs showed how hard it is to change classrooms from Washington.
4. Mississippi's 2013 law, which paired screening, teacher retraining and a third-grade gate, shows a poor state can move scores when it teaches the code.
5. Corrections came from parents and legislatures, which is where self-government says they should come from.

What you can do: ask your child's school in writing how it teaches children to decode a new word, and read the answer before the next school board meeting.