The science and politics of reading instruction

« previous post | next post »

Just out: Mark Seidenberg, "Politics (of Reading) Makes Strange Bedfellows", Perspectives on Language and Literacy, Summer 2012. The article's opening explains the background:

In 2011, Wisconsin Governor Scott Walker created the Read to Lead Task Force to develop strategies for improving literacy. Like many states, Wisconsin has a literacy problem: 62% of the eighth grade students scoring at the Basic or Below Basic levels on the 2011 NAEP; large discrepancies between scores on the NAEP and on the state’s homegrown reading assessment; and a failing public school system in the state’s largest city, Milwaukee. The task force was diverse, including Democratic and Republican state legislators, the head of the Department of Public Instruction, classroom teachers, representatives of several advocacy groups, and the governor himself. I was invited to speak at the last of their six meetings. I had serious misgivings about participating. Under the governor’s controversial leadership, collective bargaining rights for teachers and other public service employees were eliminated and massive cuts to public education enacted. As a scientist who has studied reading for many years and followed educational issues closely I decided to use my 10 minutes to speak frankly. What follows is a lightly edited transcript of my remarks.

From the beginning of those remarks:

Educators are not the only people who think about how people learn to read. There is a vast body of scientific research on reading conducted by scientists, like me, in disciplines including psychology, neuroscience, and cognitive science. We investigate how children learn and develop. We study thinking, reasoning, language, memory, seeing, hearing, attention—all of the human capacities that underlie reading and other kinds of intelligent behavior. We study people in our labs, in schools, in prisons, and on the Internet.

You might not have heard much about this research because scientists usually aren’t included in the conversation. There is no scientist on this task force, for example. Here is what you are missing:

Psychologists have been studying reading since the 19th century. We focus on how children learn to read, how skilled reading works, and the brain structures that support reading. We have identified what makes learning to read hard for many children and impairments that underlie reading and learning disabilities.

This research is conducted by scientists in Wisconsin and every other state. It is studied in dozens of countries around the world, with different languages and writing systems. This research has yielded a remarkably consistent body of findings. They do not depend on the findings of any one lab or funding agency.

Much of this research is about how beginning readers acquire basic skills and about how the ability to comprehend different kinds of texts, for different purposes, develops in later grades.

Very little of this research has had any impact on how reading is taught. There is an enormous disconnect between science and educational practice. We occupy two different worlds. I believe this is an enormous waste.

Many people on the education side dismiss this research as completely irrelevant to their mission. Teachers aren’t exposed to this research as part of their training. From the schools of education, the attitude is move along, nothing to see here. What is the result? The way reading is taught makes learning to read much harder than it should be for many children. And it makes it harder to become proficient. That is truly a way to leave children behind.

Towards the end:

I know that people like me are saying these things to people like you in many meetings at many places around the country.

At this point, the educators start to boil. They feel disrespected and patronized. Teachers know more about teaching children than laboratory scientists. Good teachers discover what works through experience on the front lines. Raising questions about how reading is taught can be seen as disparaging teachers themselves. If there is a problem—and some educators insist there isn’t one—many would argue it could be solved by getting out of teachers’ way. First teachers had to cope with No Child Left Behind, which put the federal government in their classrooms. On top of that they get directives from the state department of public instruction, the local school district, and their principal. Now the scientists are telling them what to do. Perhaps everything would be fine if we would only let teachers teach.

As a scientist, I am not challenging anyone’s integrity, commitment, motivation, effort, sincerity, or intelligence. But scientists are challenging educators to examine their beliefs and open their eyes. The single most influential educational theorist in America is a man named Lev Vygotsky, who lived in the Soviet Union, wrote in Russian, and died in 1934. He never saw an American classroom, or a television, computer, calculator, video game, or smartphone. It is as though we haven’t learned anything relevant to educational theory since the 1930s. Imagine if physicians still relied on the medical authorities of that era.

If teachers really could figure out how reading works just by observation and experience, we would not be having this discussion. But what we can learn about reading from observation is limited. Most of what we do when we read is subconscious. You cannot inspect what happens in your brain while you’re reading. All you know is the result: whether you understood the text or not and whether you got the information you were seeking. People’s intuitions about reading are limited and often misleading. They are also biased: what people observe is affected by what they believe. That is why we conduct research—to understand components of reading that would otherwise be hidden from view and to do it in an objective way.

We aren’t trying to tell people how to teach. We are struggling to get the science into the conversation—about teacher training, curriculum planning, and policy making.

At the end of his presentation, Mark made two specific recommendations for how to improve the situation. To his surprise,

The Read to Lead Task Force eventually endorsed both recommendations. Given the toxic political climate in our state, and a panel of stakeholders with very different views and constituencies, this outcome amazed me. For the first time in many years I felt a little hope that the political process could sometimes manage to yield consensus on an important issue. A bill was written and, with much additional politicking and tweaking, passed and signed into law on April 2, 2012. Whether its provisions will be implemented in the intended ways remains to be seen.

Read the whole thing — it's not long.


  1. Ryan Miller said,

    August 17, 2012 @ 8:37 am

    Can you point to a good review article on how reading should be taught?

    [(myl) One review: Keith Rayner, Barbara Foorman, Charles Perfetti, David Pesetsky and Mark Seidenberg, "How Should Reading Be Taught?", Scientific American 2002.]

  2. Mark Seidenberg said,

    August 17, 2012 @ 8:51 am

    I should say that the two recommendations (change teacher certification, screen for children at risk) were already under discussion by the time I spoke. I had lobbied to get them on the table and was there to publicly endorse them, but the committee already discussed them extensively.

  3. Jonathan Mayhew said,

    August 17, 2012 @ 1:37 pm

    I was taught to read in the 1st grade in 1966 through an experimental program that used books printed in a phonetically consistent alphabet. I remember reading Milne's "When We Were Very Young" this way. There was a special symbol for the "wh" or "when," for example.

    Then, in 2nd grade, already knowing how to read, more or less, I had to learn to read English words as conventionally spelled.

  4. Barbara Phillips Long said,

    August 17, 2012 @ 2:11 pm

    @ Jonathan —

    Did you get instruction with Distar, possibly with the teacher using a clicker?

    One of the schools in my district used Distar when my daughter was entering kindergarten (about 25 years ago), and the only books in the classroom were in their custom phonetic alphabet.

    No "Dick and Jane," and no "Dick and Jane and Vampires," either.

  5. Harold said,

    August 17, 2012 @ 2:14 pm

    Is it really true that dyslexia can be treated if caught early? I knew two children (of family friends), both girls, who had dyslexia quite badly, and in one case it was caught early – she had to go to summer school and repeat kindergarten, and in the other it wasn't addressed until age seven. In neither case did it seem to make much difference how the school and family reacted. Both learned to read eventually (by sixth grade), though not very well, and both ended up at long last going to college and doing fine. The one who was given help at seven is now in veterinary school.

    The father of the other child — the one who was "caught" early, was very supportive, since he himself had not learned to read until age 12 or so. He said the breakthrough was when he took Spanish in seventh grade. He majored in German in College and went on to become a computer programer. Of course, when he was growing up no one knew anything about dyslexia. His mother tried to help him as well as she could, but It was basically sink or swim.

  6. Patricia said,

    August 17, 2012 @ 2:54 pm

    In California, scientists cannot qualify to sit on the state board that makes recommendations for implementing the new California State Standards. The application asks for your years of experience teaching (k-12) and your experiences with lessons that do implement the standards and your opinion about how they would operate in a situation with differentiated instruction, and similar questions. There was simply no room on the application for someone who was a scientist whose research was relevant to the standards– no way of asserting any kind of qualification to the board in that case.

    Along related lines, the candidates to take the RICA (Reading Instruction Competency Assessment) that I have casually spoken with don't seem to have any deep insight into the linguistic aspects of reading instruction. For instance, they didn't understand syllable structure, what a morpheme is, how word structure might play a role in the syllabification of a word. They couldn't articulate what phonemic awareness is and how exactly it might play a role in reading although they understood it was something important. They didn't know the difference between phonics and phonetics. They had no clue why working memory might be impacted by fluency of reading. They felt like they had to memorize a great deal of things that they felt were disconnected since they didn't understand the scientific bases for those practices. I should add, there is a linguistic component in the state standards for English subject matter programs– and students *still* feel like this. I don't know how widespread this feeling is among single subject candidates in California, but I found it quite alarming.

  7. Dan Hemmens said,

    August 17, 2012 @ 2:56 pm

    There was a special symbol for the "wh" or "when," for example.

    I'm confused by this. Surely in a phonetically consistent alphabet you wouldn't need a special symbol for "wh" because (in most dialects I'm familiar with) there is no difference in pronunciation between "wh" and "w" anyway. I'm pretty sure that (in my dialect at least) "wine" and "whine" (for example) or "white" and "wight" are perfect homophones.

  8. Patricia said,

    August 17, 2012 @ 3:19 pm

    Repeating a grade or taking summer school is not a dyslexia intervention. There are systematic intervention programs that target phonological awareness and other aspects of reading.

    it does seem that appropriate interventions can affect phonological awareness and phonological awareness (sorry– don't know how this was measured in the relevant study(s)) is correlated with reading skill in the early years. By grade 4, phonological awareness is no longer correlated with reading skill, instead IQ is. Phonological awareness can be taught or at least influenced through instruction.

    Along other lines, improving SES would make a huge difference in reading and other scholastic achievements: SES can be influenced through public policies.

  9. Cameron said,

    August 17, 2012 @ 4:06 pm

    I wonder if the "special symbol for the 'wh'" was the historic 'ƕ' .

    The wine-whine merger is not yet universal. Some of us still resist.

  10. Lillian said,

    August 17, 2012 @ 5:01 pm

    In my teacher training there certainly was an inclusion of this kind of research, but there needs to be more of it. I'm fortunate to be in a field that considers formal study of pedagogy to be essential, but a lot of teachers, paradoxically, aren't.

    There are a few significant problems that need to be seriously grappled with.
    – K-12 teachers often don't get enough study of pedagogy in general or specific study related to the discipline they teach.
    – College and university teachers often have essentially zero pedagogical training, which verges on criminal as far as I'm concerned.
    – Many educators/educational researchers working in the ed paradigm are a little too into action research, case studies, and quantitative research that can't be replicated. These can be valuable, but it can't be all we do. It's of very limited use to read journal after journal consisting of not much more than "here's what I did in my classroom! yay!"
    – Many scientists are a little too into "white lab" research that can't easily translate to a real-world classroom. Again, this stuff can be valuable, but it can't be all they do (or they can't expect it to be accepted easily). It's a bit like testing medications in computer simulations or mice only and skipping the human subjects stage before putting it on the market. And before you suggest that scientists researching ed-relevant topics wouldn't do that, they would. Some insist their ideas can't be tried out in classrooms, only in formal lab settings. (I have encountered this.) It's a bit of a problem. Clearly, we need to learn to work together more.

  11. Blair said,

    August 17, 2012 @ 5:39 pm

    Re: Harold

    Treatment of Dyslexia is complex. The first thing to understand is that the current diagnosis of dyslexia is based on a behavioral outcome, poor reading skills. However, there are many different possible underlying causes of poor reading skills. There are many steps involved in reading – visually recognizing the letters or characters, appropriately scanning across the line in order, assigning a phoneme to a letter or combination of letters, assembling these in order into a meaningful word, combining words into meaningful sentences. This is a complex task that requires the use of several brain areas in proper coordination, and deficits in any step will affect the final 'reading ability'.

    Unsurprisingly, research indicates that among people diagnosed with dyslexia, there may be several different groups of people with different underlying problems causing difficulties in the acquisition of reading skills. (For example, see

    However, there have been some interventions tailored to the particular underlying problem. If a child is diagnosed with dyslexia, and is tested on a variety of language and auditory perception tasks, and it is discovered they have difficulty with the temporal ordering of sounds and distinguishing sounds that have very brief time between them (auditory processing deficits are common), there are training programs that can improve their skills in these areas. There is some evidence these training programs may be equally or more effective than 'traditional' interventions (eg or

    So to the question "Can dyslexia be treated if caught early?" I would reply:

    1) Reading skills, like skills in anything, can be improved at any age. Older is harder, younger is easier.
    2) There do currently exist specifically designed training strategies for improving the language and reading skills of children that fall at the bottom of the normal curve.
    3) These and more general (but also intense) training programs in reading have been shown to bring children's reading skills closer to the average for their age group.

  12. Eric P Smith said,

    August 17, 2012 @ 6:12 pm

    @Barbara Phillips Long, Dan Hemmens, Cameron

    Only Jonathan Mayhew himself could confirm this, but I think he is referring to Initial Teaching Alphabet, or ITA. (DISTAR reached the schools a little later.) ITA was a compromise between a phonetic alphabet and standard English spelling (Traditional Orthography or TO), designed so that the first steps in reading were phonetic and yet the transition to TO was smooth. Its symbol for the 'wh' combination resembled a 'w' and an 'h' stuck together.

    It was a disaster.

  13. Patricia said,

    August 17, 2012 @ 6:42 pm

    @Lilian, forgive me if this seems like ranting. I am trying to respond to your interesting post in an efficient and abbreviated way.

    I find it bordering on criminal that people have to get lawyers to get their children's special educational needs met because they are being educated and educationally administered by people who don't understand their disabilities despite the myriad of research and laws that would point them in useful directions. I have seen children with hyperlexia given IEP's that focused on *phonics*. That is absolutely insane. I have seen expensive research based reading programs (for children with significant reading delays) be implemented haphazardly (e.g., starting in the middle, not doing the proscribed assessments before deciding where to start in the program, etc.) and therefore certainly ineffectively because it was convenient to do it that way or perhaps out of ignorance–again necessitating the hiring of lawyers to get children's educational needs met. I find it bordering on inexcusable ignorance when teachers tell children with dyslexia that they are lazy or didn't study and that is why their spelling tests are so bad. I have witnessed this so I am not exaggerating here. I find it bordering on professional incompetence for kindergarten teachers who are supposedly literacy specialists to use teacher-made big books (a good thing and research based as effective) that have bizarre, highly unfamiliar fonts set on busy, distracting backgrounds for kids who are just learning to read because the educators really don't understand what the research is saying.

    I am also find it deplorable to overhear Liberal Studies students say they are going into special ed because that is where the money and jobs are– they seem to have no sense of being called/really caring for the students they will be educating. And then they sulk during their entire intro to linguistics/language structure because their minds are challenged by something that they believe isn't directly applicable to what they will do in a classroom.

    Also, it seems to me that the point is not that all scientists should be discovering interventions but they are not. The point is the results of scientific research into all aspects of the reading process is pretty much ignored in the education given to educators.

    These are some of the problems as I see them.

  14. Daniel Barkalow said,

    August 17, 2012 @ 11:28 pm

    Beyond causes of dyslexia which are linguistic, there's straightforward vision problems as well as an interesting area of subtle vision problems. Jerry Lettvin once told my class that he'd had tremendous success with an intervention consisting of giving people with dyslexia a piece of red plastic with a hole cut in it which would mask out all of a page of text except for the word or letter the person aimed it at. (He didn't do a study of this technique because he'd already retired at the time.) It's entirely possible for people to be unable to see text usefully, while they can see to navigate, recognize faces, identify objects, etc. perfectly well, and that condition essentially means they're trying to learn to read full pages of text at once, without being able to learn to read letters sequentially first.

  15. Ben Hemmens said,

    August 18, 2012 @ 3:19 am

    I like the emphasis (e.g. in the Scientific American article linked to by myl in response to the first comment) that there is actually an element of mentally sounding words in all of our silent reading (and therefore phonics is important).

    Once I got a book written by a relative whom I knew as a dinner guest of my parents, so that I was used to his voice telling stories. I couldn't read the book, although the texts were fairly matter-of-fact, without vividly hearing his voice speaking the words. That made me wonder about the extent to which that particular rhythm and intonation was coded into the text – would other readers generate a similar result? – or was being added by me.

    In the meantime, I have also written (or translated) a lot of material, both for silent reading and some for voiceovers, and I'm curious about the contrasts and similarities between the two. I certainly write voiceovers differently; but I think the differences may have to do with writing to be understood by listening rather than just to be speakable (e.g. building in repetition of important/new words and phrases). I think "speakability" remains a good style ideal for any kind of straightforward prose, even if it is not meant to be read aloud.

  16. Boris J. said,

    August 18, 2012 @ 3:42 am

    @ Dan
    @ Cameron

    Some of us still resist indeed! :) Besides, here's what the New Oxford American Dictionary has to say on the issue :

    + whine |(h)wīn|
    + wine |wīn|

  17. Ben Hemmens said,

    August 18, 2012 @ 3:44 am

    @ Dan Hemmens:

    Does calling the letter h haitch rather than aitch correlate with making a difference between wine and whine? I (Irish) say haitch and whine, though I suspect my Hemmens great-grandfather, who was from Wiltshire, probably didn't.

    My wife, who teaches in an English department, only recently found out that some of us call the letter haitch: and it was our 2-year-old daughter who told her ;-) – having learned it from me, probably in one of our sessions of surfing kiddies' video clips on YouTube.

    Talking of which, I love the way they've made a "zed" version of "We are the alphabet": – but they've left the elevator as one of the examples for E. My daughter makes me call out all the examples as the thing goes along, and that trips me up every other time.

  18. Philip said,

    August 18, 2012 @ 9:28 am

    And if you think that reading instruction is a mess, how about "grammar" lessons from English teachers whose background is entirely in the history of British and American literature?

  19. Philip said,

    August 18, 2012 @ 9:53 am

    I read the Scientific American article which stresses the importantance of teaching phonics. But phonics presupposes that reading is a process in which graphemes (the squiggles on the printed page) become sound which becomes meaning. If that were essential to reading, then how would deaf children ever learn to read?

    [(myl) For children who are profoundly deaf from birth, learning to read is at best a big challenge. Thus Susan Goldin-Meadow and Rachel I. Mayberry, "How Do Profoundly Deaf Children Learn to Read?", Learning Disabilities Research & Practice, 2001:

    Only 15 percent of white deaf students who graduate from high school, and only 5 percent of AfricanAmerican and 6 percent of Hispanic deaf high school graduates, read above the sixth-grade level (Allen, 1994). Indeed, the median reading level of deaf high school graduates is fourth grade (Allen, 1986; Trybus & Karchmer, 1977). This level barely approaches newspaper literacy, and does not actually require the reader to have cracked the print code.

    And those statistics include many students who were not profoundly deaf from birth. It's also worth noting that deaf children who learn to read English (or other alphabetic languages) will know finger-spelling, in which each letter is spelled out with a finger shape; and will use finger-spelling fluently in their own communication. This is in effect "phonics for the hand", not "whole word" learning. Also, see Margaret Harris and Constanza Moreno, "Speech Reading and Learning to Read: A Comparison of 8-Year-Old Profoundly Deaf Children With Good and Poor Reading Ability", Journal of Deaf Studies and Deaf Education, 2006:

    Nine children with severe-profound prelingual hearing loss and single-word reading scores not more than 10 months behind chronological age (Good Readers) were matched with 9 children whose reading lag was at least 15 months (Poor Readers). Good Readers had significantly higher spelling and reading comprehension scores. They produced significantly more phonetic errors (indicating the use of phonological coding) and more often correctly represented the number of syllables in spelling than Poor Readers. […] [A]ll 9 children were good speech readers, suggesting that a phonological code derived through speech reading may underpin reading success for deaf children.


    And sometimes you have to know the meaning of a word before you know how it's pronounced: He read the minute print in the minutes.

    [(myl) Sometimes you do — this is one of the things that makes English a hard language to learn to read, compared to languages in which the orthography is more phonologically transparent. But as an argument that spelling-sound correspondences are irrelevant, this is preposterous.]

    Fxnxlly, hxw dxex phxnxcxs xxplxxn thx fxct thxx expxrxxncxd rxxdxrs cxn dxcxdx thxx pxssxgx rxlxtxvxlx exsxly?

    [(myl) In general, arguments about how skilled readers perform are not really relevant to discussions of how to teach beginning readers. But there's good evidence that in most cases, even skilled readers making guesses from incomplete information are relying in part on letter-to-sound associations. This evidence is discussed in the cited Scientific American article, or see e.g. Mark Seidenberg, "Reading in Different Writing Systems: One Architecture, Multiple Solutions", 2007.

    Overall, thank you for bringing up in a small space so many of the uninformed prejudices that have done so much damage in fostering ineffective methods for teaching reading. ]

  20. Jonathan Mayhew said,

    August 18, 2012 @ 9:59 am

    Yes, the ITA! Thanks. I've been looking for it but I had no idea what it was called, since I haven't thought about it much after the age of 6. It does indeed have a separate symbol for "wh" for those who don't have that merger. I learned to read very easily with this system and to this day, I am a voracious reader.

  21. Mary Bull said,

    August 18, 2012 @ 10:13 am

    @Blair in re Harold
    @Daniel Barkalow

    I'd like to add spatial orientation and perceptual-motor disabilities to the list of possible underlying causes of dyslexia.

    And I hope discussions like this at LL will somehow help with getting both better and earlier screening and more effective interventions into today's classrooms Things don't seem to have gotten much better since the days almost 50 years ago when my school district got a grant under the NDEA'S Title I provisions and pulled me out of my fourth-grade classroom for a new position as a special-reading teacher.

    I had rather mixed success with that job, and, after three years, I requested and received a transfer to a first grade, because I had concluded that earlier intervention would be better. I still believe that it is.

  22. Yet another John said,

    August 18, 2012 @ 10:45 am

    @Phillip: But the 2002 Scientific American article does cite evidence that (normal, non-deaf) readers do mentally sound out words as the read, via the fascinating homophone study. Does this evidence not convince you?

    Your question about how congenitally deaf people would learn to read English is really interesting. I'm not an expert at all in this field, but a quick Google search turned up an article by Vicki L. Hanson, "Phonology and Reading: Evidence from Profoundly Deaf Readers" (Haskins Laboratories Status Report on Speech Research 1989, SR-99/l00,172-179). Executive summary: there is evidence from various different studies that successful deaf-from-birth readers *do* use phonetic information to read English! Presumably they learn about the phonetic structure of words as they are learning to lip-read and/or when they learn to sound out words (as any well socialized deaf child would do these days).

  23. Aaron Binns said,

    August 18, 2012 @ 10:54 am


    Maybe the purpose of the "wh" symbol isn't to distinguish "whine" from "wine", but the "wh" in "when" and "who".

  24. Philip said,

    August 18, 2012 @ 11:08 am

    myl: I understand that the phonics vs. whole language pendulum swings back and forth every few decades, but certainly phonics doesn't explain EVERYTHING about how children learn to read, does it?

    [(myl) With respect, you could use a bit of remediation in reading comprehension yourself. None of the quoted or cited articles claims that "phonics explains EVERYTHING about how children learn to read" — rather, the view is that teaching reading without an appropriate role for phonics instruction is a recipe for (statistical) failure.]

    One of the things that caught my attention in the Scientific American article was the mention of "120 or so patterns" taught to first graders using McGraw Hill's Open Court Reading program Sorry, but that seems like overkill to me–especially since there will still be hundreds and hundreds of exceptions to even 120 rules. How does someone know when to apply a rule ("silent e" makes the preceeding vowel say it's name: cod—>code") and when not to apply it in a word like "have"?

    [(myl) On the question of which patterns, and how many patterns, to use in teaching beginning readers, you should look at the extensive literature on the subject. And the idea, in general, is not to teach children to apply "rules" consciously — though there may be some of that — but to assimilate the correspondences between letter patterns and sound patterns so that they can use them unconsciously and effortlessly. That most of the correspondences are partial regularities rather than exceptionless ones is a fact of life; but so is the central role that such correspondences play in in reading at all skill levels, and the poor outcomes that generally result from trying to teach reading without helping readers to learn them.]

    My real prejudice is that extralinguistic factors have a lot to do with children's reading abilities. I'd guess that there's a strong correlation between kids' reading levels and their parents' socioeconomic status. But if this is another of my "uninformed" prejudices, I'm sure you'll let me know.

    [(myl) Such a correlation certainly exists, though there are also some interesting interactions. But poor teaching methods yield poorer outcomes for both rich kids and poor kids; and good teaching methods yield better ones.

    Do you really think that (for example) because poor kids in general have worse reading scores, we should give up trying to teach them to read? Or do you really think that because rich kids in general have better reading scores, it doesn't matter how (or even whether) we teach them to read? If so, then your prejudices are dangerous as well as uninformed.]

  25. J.W. Brewer said,

    August 18, 2012 @ 12:05 pm

    There is an inevitable tension (in terms of focus, resource allocation etc.) between better/earlier screening to identify and assist the minority of children who have or are predisposed to dyslexia or other neurological-type obstacles to full literacy and doing a better job (by taking into account decades of research that has allegedly been ignored by the Ed School establishment) in teaching average/median kids who are not faced with those obstacles to read fluently.

    I do think the SciAm article touched on an important issue toward the end. Phonics-type approaches as traditionally structured may be less "fun" and more boring not so much for the kids but for the teachers themselves. That I think helps explain (with only a modest amount of cynicism about human nature and the institutional dynamics of teachers' unions and Ed Schools) a lot of the status quo enthusiasm for the sort of whole-language approach that the science apparently keeps showing is inadequate. Thus, the suggestion that newer generations of teachers be trained in how to implement more phonics-based approaches that *they* will find fun and engaging to carry out (with the side effect of being more beneficial to the students . . .) is very very important.

  26. bianca steele said,

    August 18, 2012 @ 12:27 pm

    As the mother of a preschooler (and an early reader myself, though often only a B+ student in school English, and with a sibling diagnosed with dyslexia at one point), this is something I think about a lot. I don't want to push her excessively, but on the other hand I've offered her opportunities to learn when it seemed like she was ready. Some people might think this is already "pushing," but it doesn't seem so to me. The fact is I have no experience with in-school reading instruction (since I could read already when I started school), I have no idea how actually teaching a class of 30 six and seven year olds from scratch, in a school day shared with math and other academic subjects, and the breaks kids that age still need, would be done.

  27. Philip said,

    August 18, 2012 @ 12:51 pm

    myl asked "Do you really think that because poor kids . . . have worse reading scores, we should give up trying to teach them to read"?

    Of course not, but my absolutely uninformed prejudice is that remediation for kids who are poor readers consists of more and more phonics and less and less real reading.

    [(myl) Evidence? Certainly this is not true of the formal, branded remediation programs that I'm aware of.]

    I don't think you and I disagree that much. Of course there are grapheme/phoneme correspondences in English that readers recognize. How much this is necessary to teach kids to read is a matter of debate, but certainly some of it is necessary, even though you yourself state that "the idea, in general, is not to teach children to apply 'rules' consciously — though there may be some of that — but to assimilate the correspondences between letter patterns and sound patterns so that they can use them unconsciously and effortlessly." That sounds pretty whole-languagey to me: Kids aren't taught to consciously apply one of 120 phonics rules or how to know when to apply one or the other of several conflicting rules, but they "assimilate"–through real reading–the correspondences and apply them "unconsciously."

    [(myl) But there's excellent evidence that "whole language" instruction doesn't reliably result in assimilation of such knowledge.]

    The last paragraph of the Scientific American article stressed the need for teachers to "strike a balance" between whole-language teaching methods and phonics. My sense is that we disagree about where that balance lies.

    My own kids–and I suspect this is true of the children of many or even most LL readers–learned to read by themselves because their parents spent hours reading to them from infancy on. By the time they got to kindergarten or first grade, they'd figured out how reading works by themselves, just like they learned their native language.

    [(myl) Some children more or less teach themselves to read, especially if they have the kind of early experience that you cite. Some children have serious problems learning to read, no matter what kind of exposure and experience they get. The issue is the large majority of kids in the middle, who are likely to learn to read well with proper teaching methods, and are likely to lag or fail with bad teaching methods.]

    I'm also sure that if this hadn't happened to any of my kids, if one of them had had trouble with reading and was in a remedial reading group, I'd have spent some time working with him/her on some simple grapheme/phoneme correspondences, but I've have spent a LOT more time–ten times more–just reading together.

    I can't claim to know for sure, but I doubt that this kind of remediation is what happens in our standards-based, no-child-left- behind, inner-city schools.

    Finally (and I'm sorry for the long post), Stephen Krashen points out the strong correlation between kids' reading levels and something absolutely simple like the number of books available in their elementary school libraries.

    [(myl) You seem to be stuck on the idea that if development of skill X correlates with influence Y, then it can't possibly be relevant to consider the effect of influence Z. This is (if I've counted right) the third argument of that kind that you've made — but no matter how many times you make the argument, it's still a fallacious one.]

  28. Mother of child with dyslexia said,

    August 18, 2012 @ 2:43 pm

    Philip, As the mother of a child with dyslexia, I can assure you I read literally several thousand books aloud to her by the time she started kindergarten. Maybe you think autism is caused by emotionally cold mothers as well.

    My daughter was done a huge disservice in Californian schools by whole language instruction and educators who could not recognize dyslexia. I hold Steve Krashen personally responsible for the catastrophe in California that was Whole Language– implemented on a large scale without even a pilot program. I don't know if reading instruction was more fun for teachers, but it was certainly easier because the catchphrase was: you don't teach reading, kids just pick it up by being exposed to (illustrated) books.

    She learned to read when I a) taught her phonics systematically. There were letters she had trouble with– for example "y" which is named [wai]– which is nothing like it sounds, she sometimes confused capital B and lower case "m" which you can see are both phonetically and visually related. b) made a reading aide of 2 3×5 cards put together perpendicularly like an capital L and an additional 3×5 card to "sweep" across the page. As long as I could visually syllabify the words for her she had no trouble reading. c) worked on phonemic awareness with games like giving her words she couldn't spell and then had her put first leafs on a pretend tree for each sound and later in her education put blocks in a row for each sound she heard, giving vowel blocks a different color. Later I found a computer program that was much cheaper than "Fast ForWords" where she could improve her phonemic awareness by prolonged and frequent exposure with feedback for accuracy from the computer "game".
    Just reading together was NOT enough. Even though we read massive numbers of books together. Standards based reading instruction does not work in the way you imagine.

  29. Philip said,

    August 18, 2012 @ 2:44 pm

    As I said, I have no evidence of what goes on in every remedial reading program in every elementary school district, but thank you for the link ("formal, branded remediation programs").

  30. Peter Bowers said,

    August 18, 2012 @ 8:59 pm

    A powerful talk by Seidenberg, definitely a major figure in the reading research. I agree wholeheartedly with his targeting of teacher training, not teachers for his criticism of the persistence of seriously flawed instructional practice.

    I would like to add one key point to what is missing from that training, but which often doesn’t receive the attention it deserves. Very simply, if we are going to strive to develop the most effective literacy instruction, one necessary foundation of that instruction should be an accurate understanding of how the English writing system works.

    I taught elementary school for 10 years before I started graduate school and began to look at the science of reading myself. For the first 9 of those 10 years, I taught under the widespread assumption that English spelling is a highly irregular spelling system, leaving learners with many words that simply have to be memorized. I was a terrible speller myself all those years. In that last year of teaching, I encountered a resource that built on a very different understanding of English spelling from the scientific field of linguistics. For the first time I started to understand English spelling because I was studying the structures and conventions of how English spelling works to represent the meaning of words. I was learning about how the structure, history and pronunciation of words are represented by English spelling. These were principles that had been explained by linguists such as Noam Chomsky, Carol Chomsky, and Richard Venezky since the late 60’s and early 70s, but somehow key features of what they described failed to be emphasized in the science of reading or teacher training. This is starting to change, however.

    It would probably surprise many readers of this blog to know that words like does, business, or sign are totally conventional, understandable spellings once we understand that the meaning elements of words called morphemes (bases and affixes) often use the same spelling despite pronunciation shifts and that we therefore need to look at words related in structure and meaning before we can explain the spelling of a word to a child who asks, "Why is there a 'g' in 'sign'?"

    For those interested in a more detailed description of how English spelling works (including some of those words) and how it can be taught I offer this link to a public lecture I gave on this topic at the University of Alberta not too long ago:

    It may seem crass, but lots of free resources and links to videos of this instruction in action can be found at my website here:

    Mark Seidenberg was a co-author on a seminal article "How psychological science informs the teaching of reading." in 2001 which stated the following:

    [L]earning to read is learning how to use the conventional forms of printed language to obtain meaning from words…This view implies that the child learning how to read needs to learn how his or her writing system works” [emphasis added] (p. 34).

    I couldn't agree more! If children are going to learn how their writing system works, teachers need training that explains how it works.

  31. prasad said,

    August 18, 2012 @ 10:15 pm

    I'm not a whole language nut (very far from it). But I hadn't encountered Philip's question about deaf kids before, and didn't think "Overall, thank you for bringing up in a small space so many of the uninformed prejudices that have done so much damage in fostering ineffective methods for teaching reading" was deserved. Actually, it made me think – aren't there ancient languages we can read, for which no-one knows the sound-letter mappings?

  32. David Morris said,

    August 19, 2012 @ 6:46 am

    My field is ESL for adults, so I'm way out of my area here, but I'll just comment that I'm surprised there is so little said here about the role of parents (or other primary (and secondary) care-givers) in the pre-school years, or even the school years). Surely being read to, and participating in pre-reading discussion, forms a very large part of literacy acquisition.

  33. David Morris said,

    August 19, 2012 @ 6:57 am

    Several of the commenters have commented about the role of parents, but I am very surprised that there is essential nothing about it (the role) or them (the parents) in Mark Seidenberg's speech/article.

  34. Bill Walderman said,

    August 19, 2012 @ 9:54 am

    "aren't there ancient languages we can read, for which no-one knows the sound-letter mappings?"

    This is an interesting question, although veering somewhat off-topic. For Latin modern readers use sound-mappings that are fairly accurate in terms of the original spoken language. For ancient Greek there are also sound-mappings that are somewhat less accurate. (For both of these languages the sound-mappings may differ slightly depending on the native language of the modern reader.) Based on personal experience, those who read Latin and ancient Greek translate the written symbols into mental sounds as they read.

    From what I've seen of ancient Egyptian, there is also a system of modern sound-mappings used by modern readers, although the actual values of the ancient phonemes far from fully understood (and probably varied considerably over the lengthy time-span of Egyptian hieroglyphic writing).

    For other ancient languages (e.g., Linear B Greek, and to some extent, interpreting fragmentary ancient manuscripts or those written in obscure scripts), "reading" is a process of constant decipherment, not at all like reading a book or a newspaper. I'd be interested in how scholars approach various cuneiform languages and languages where no inferences can be made about phonetic values.

  35. Philip said,

    August 19, 2012 @ 10:49 am

    I'm almost afraid to comment (and thanks, prasad, for saying that you didn't think my "uninformed prejudices" were necessarily either), but–and this is a real question, so please don't jump all over me–how do children learn to read non-alphabetic languages like Chinese? And here's a question for a REAL specialist: What kind of remediation works for Chinese kids who are having problems learning to read?

    And none of my questions/comments about the role of phonics in remedial reading have anything to do with teaching dyslexic children.

  36. Bill Walderman said,

    August 19, 2012 @ 1:09 pm

    "how do children learn to read non-alphabetic languages like Chinese?"

    According to the 2002 Scientific American article, for about 50 years Chinese children have been first taught to read using a Romanized transcription.

  37. Ruben Polo-Sherk said,

    August 19, 2012 @ 8:50 pm

    It may be more natural for many people to sound out words when they read, but it's certainly not the only possible mechanism.

    In my own case (with learning Japanese), there are many words (written with Chinese characters) I understand completely but I have no idea how to pronounce, as I learned the words from reading.

    It's the same as math–you can do math without knowing how to read the entire equation aloud–and music (where the analogy is a a bit different): you can know the mapping between a written note and the sound it represents without knowing the name for the note.

  38. kerry hempenstall said,

    August 20, 2012 @ 12:58 am

    The issue of learning to read a pictographic language is quite different to that involved in an alphabetic language. See some quotes below

    "A child in Japan is expected to learn only 76 Kanji in Yr 1, and only 996 by end of Year 3". p.111
    Beck, I.L., & Juel, C. (1992). The role of decoding in learning to read. In S.J. Samuels and Alan E. Farstrup (Eds.). What research has to say about reading instruction. (2nd. ed. pp.101-123).. Newark, DE: IRA

    By contrast for English writing:
    "In the very first reading books in first grade, … the child is exposed to between 300 and 600 different words in running text … ."
    Juel, C. (1993). The spelling-sound code in reading. In S. Yussen & M. Smith (Eds.), Reading across the life span (pp. 95-109). New York: Springer-Verlag.

    "Chinese adults are said to have a working familiarity with only about 4000-5000". p.16.
    Adams, M. J. (1990). Beginning to read: Thinking & learning about print. Cambridge, MA: MIT Press.

    "Humans are constitutionally incapable of memorizing more than about 2,000 word-symbol pairs. This is an ultimate limit, taking years of study. The Japanese have 1800 word symbols (kanji) in their basic writing system, and it takes children about 12 years to learn them. Mastering an additional 2,000 kanji is the mark of a highly educated person and can take up to a decade more of study. A good college dictionary contains 250,000 words. In short, no one can ever be a whole-word (sight word) reader".
    McGuinness, D. (2002). A prototype for teaching the English alphabet code. Reading Reform Foundation, 49, Autumn.

  39. Mother of child with dyslexia said,

    August 20, 2012 @ 4:28 am

    Philip, just who do you think should get remedial reading if the kids with dyslexia don't get it? Aren't they "having trouble with reading"?

    One view in dyslexia research is that dyslexia is basically the lower end of a continuum of reading ability.

  40. Dom Massaro said,

    August 20, 2012 @ 8:37 am

    I have been researching reading and learning to read for over 4 decades, and have witnessed the recurring controversies and the inability of behavioral and brain sciences to solve the problem. My research and applications have led me to a new proposed solution described in:
    Massaro, D. W. (2012). Acquiring Literacy Naturally: Behavioral science and technology could empower preschool children to learn to read naturally without instruction, American Scientist, 100, 324-333. A PDF
    can be downloaded at

  41. Plane said,

    August 20, 2012 @ 10:20 am

    @Ruben, there's research suggesting kanji is (at least in part) phonologically mediated as well, see e.g. (1993) Wydell, TN., Patterson, KE. and Humphreys, GW., Phonologically mediated access to meaning for Kanji: is a rows still a rose in Japanese Kanji?, Journal of Experimental Psychology: Learning, Memory, and Cognition 19 (3) : 491- 514

    You can find plenty more research on the subject if you look. Whether you believe this has implications for your own learning, I will leave up to you, as I am not qualified to give an opinion.

  42. Andy Averill said,

    August 20, 2012 @ 10:37 am

    As far as reading languages you can't pronounce, you don't have to go back to dead languages. I read French fairly well, but my pronunciation is atrocious. I doubt very much that I'm mentally pronouncing the words when I read a book in French. (I learned to read English from Dick, Jane, and Spot if that makes any difference.)

  43. prasad said,

    August 20, 2012 @ 11:07 am

    if you pronounce French atrociously you still pronounce it, right? Presumably your french sounds vaguely english-y or something. I don't think for french-reading to occur via 'mental sound' representation you need for that sound representation to be something Academie Francaise would endorse! Similarly, there's no trouble understanding in this picture how people read Latin.

  44. Ruben Polo-Sherk said,

    August 20, 2012 @ 10:48 pm

    I don't doubt that the processing link for many people can be [image->sound->meaning], with the phonology as an aid to recall the meaning.

    I really don't think that in terms of processing, logographic scripts are significantly different from phonetic ones for advanced readers–when you read English, you don't read each letter: you read each word at once as a unit (perhaps even several words at once). That symbol (or collection of symbols) that you process has an associated sound and meaning, which may or may not require input from context to determine. In that sense, phonetic and logographic scripts are identical.

    What I was saying was that going though the sound to the meaning is unnatural for me, and I don't do it with Japanese. (I also don't do it in French or Spanish or English or whatever). And so my point was that the processing path [image->meaning] (without the sound being relevant) is also possible.

  45. Keith said,

    August 20, 2012 @ 11:52 pm

    I am quite sure I mentally pronounce words as I read: I am very conscious of it. I have problems, for instance, if an author gives a character a difficult name, as it causes me to (mentally) stumble every time I encounter it. (And I am a fast, proficient, self taught reader.)

  46. Sandra wilde said,

    August 25, 2012 @ 4:35 pm

    I'm a literacy education professor. The issues are complex, but the "scientific" perspective in the cited article represents somewhat of a fringe view in its overemphasis on phonics, which is just a small part of literacy learning. Not referring so much to Seidenberg's article but to how it's played out in practice, particularly with mandating scripted phonics programs under NCLB. This is hotly contested terrain in the field. Richard Allington's book Big Brother and the national reading curriculum explores how it became politicized.

    I'd be interested to see linguists' reaction to the Massachusetts test for teachers that was cited.

    One note on ITA: manynof the materials in it reflected British pronciations but were sold in America too.

  47. Joanne Yatvin said,

    August 25, 2012 @ 8:36 pm

    After reading the speech by Seidenberg and all the comments, I still haven't the slightest idea what they are recommending to teachers of reading.

  48. Chandra said,

    September 18, 2012 @ 4:36 pm

    Again, coming to this discussion far too late (damn those busy summer vacation days… oh wait) but as an instructor of low-literacy adults, many of whom suffer from dyslexia and various other learning challenges, I am so very happy to see this discussion taking place.

    I use phonological awareness and phonics-based activities in my classes – not exclusively, but when I feel that they are needed – and have seen incredible improvement in people who swore that nobody could help them learn how to read. Yet I have colleagues who consider this nothing short of heresy.

    I would certainly not want to leave the entirety of curriculum development up to a group of scientists who had never set foot in a classroom. But the idea that we, as professional educators, should use scientific evidence as a basis for developing our techniques seems so perfectly and logically obvious to me that I can scarcely believe it is controversial.

    There is a place for the whole language approach, and there is a place for phonics. The two are not mutually exclusive. Learners – being, you know, human – come in a variety of types, and anybody who holds fast to one single approach as the be-all and end-all is doing a great disservice to their students.

RSS feed for comments on this post