FOR DECADES, as bitter battles over how best to teach reading infected America's classrooms, most schools ignored the growing body of studies showing the superiority of phonics. But reasons for getting serious about teaching phonics earlier and more systematically get harder and harder to dismiss.

That's particularly true thanks to two Yale professors, Sally and Bennett Shaywitz, who have been analyzing functional magnetic resonance images of readers' brains. In the 1990s, the Shaywitzes identified differences in the brains of children with reading difficulties, showing a physiological basis for the problems dyslexics exhibit in decoding the sounds of words.


Now, in research published this month in the journal Biological Psychiatry, the Shaywitzes identify physiological differences between dyslexics who have learned to compensate for their reading problems and nonimpaired but poor readers - differences underscoring the need for phonics instruction.

This new research shows compensating dyslexics form alternate neural pathways to decode words. They end up as accurate readers, though still often slower than nonimpaired, fluent readers.


By contrast - and this is the breakthrough - the Shaywitzes say nonimpaired, poor readers rely on memory centers when reading, unlike fluent readers and compensated dyslexics who both read better.

The Shaywitzes say it's as though the normal wiring of these poor readers had never been activated by instruction in decoding or phonics, so they ineffectively use rote memorization. These readers tend to come from disadvantaged schools; the Shaywitzes consider them "environmentally influenced" dyslexics, not physiologically dyslexic.

This applies to a lot of people. Overall, about 40 percent of all beginning readers have some problems, and the Shaywitzes believe two-thirds of them - perhaps higher in low-income areas - aren't genetically dyslexic. They could be fast, accurate readers with early, intensive phonics instruction.

The Shaywitzes' latest work is one more indictment of the still-popular whole-language approach to teaching reading that slights decoding instruction. It argues for earlier, more systematic phonics instruction - and in turn adds even more weight to the Bush administration's welcome move to raise Head Start's educational standards.

The new brain study also comes as states are being pushed to recast their schools' instruction firmly based on such scientific research in order to qualify for $6 billion in federal "Reading First" funds.

But even as phonics programs rack up success in cities such as Baltimore, a thin grasp of phonics - and stunning resistance - remains among too many educators. Tellingly, many states, including Maryland, have had to rework their "Reading First" proposals to include more substantive phonics.

Phonics proponents have long viewed the legions of poor readers turned out by phonics-averse American schools as instructional casualties, victims of "dysteaching." Now the contrast, between the mounting case for phonics and its far too limited use in schools, is so great that the issue has become not merely instructional, but ethical.