University Place

Becoming a Reader

– Thank you for joining us today. I am Heather Daniels, secretary of the academic staff.

I am pleased to welcome Mark Seidenberg as today’s speaker. Mark is a professor in the Department of Psychology. He received his undergraduate and PhD degrees from Columbia University. He’s lived both in cold and warm climates, starting with a stint at McGill University in Canada before moving both very far west and south to USC or the University of Southern California. In 2001, tired of the predictability of southern California weather, he moved to Madison–
(audience laughs)
and joined the UW-Madison faculty. Mark has published in all the cool psychology and neuroscience journals and in 2017, published a book titled “Language at the Speed of Sight: “How We Read, Why So Many Can’t, “and What Can Be Done About It”. And I have to say I read that book two summers ago and, as a mom of a 10 year old, found it very enlightening.

He describes his research in the following way: “I study language and reading with the goal of understanding “how these skills are acquired and used “and the brain circuits that support them. “The work involves a combination of behavioral studies, “neuroimaging, and computational modeling. “If at all possible, I’d like this research “to make it easier for more people “to become better readers and for children who struggle “with reading to obtain effective help.” I mean, if that’s not a great mission statement of the lab, I don’t know what it is. Please join me in welcoming Mark Seidenberg with his talk “Becoming a Reader” as our closing round table speaker for this academic year.
(audience applauds)
– Thank you.

It’s really a pleasure to be here. It’s also very close to my office.
(audience laughs)
And my talk today is about children becoming readers and in particular how we could help more children to succeed.

A little bit about myself. We have a lab called the Language and Cognitive Neuroscience Lab where we study a variety of topics.

We use several kinds of measures. We’re always looking at behavior, for example, children, adults, in different languages.

We look at the brain using methods like fMRI.

And we use these computational models to try to get a sense of the mechanisms that actually are going on in the brain to produce behavior. The goal is to get an understanding at all of these levels together about what underlies becoming a skilled reader, the kinds of obstacles the kids might encounter, whether reading works the same in different writing systems and languages and so on.

I also teach. I have a course called “The Science of Reading “and Its Educational Implications.” That’s taught in the psychology department.

So, people have been studying reading for a long time in many labs around the world.

Much is known about skilled reading, learning to read, the brain bases of reading, the causes of reading impairments and ways to remediate them, in different writing systems and in different languages.

Much remains to be learned, of course, but there’s consensus about the basic phenomenon principles.

At a certain point or maybe a certain age, you begin to wonder, if we know so much about reading and the science is so great, why are there so many poor readers?

So, America does have a literacy problem.

Overall, literacy levels are too low. Many people have only very basic reading skills. And this really raises questions about our ability to function. For example, to run a democracy or tell truth from fiction.

The evidence from this comes from a variety of sources that all converge on this conclusion. They’re the cross- national comparisons that you often read about in the media like the PISA, the TIMSS, the PIRLS. There’s several of these in which the U.S. always kind of comes out in the middle of the pack, but there are also U.S. assessments like the NAEP, so called Nation’s Report Card that paint the same picture.

We got data in Wisconsin, too. So, not too long ago, there were results posted on the Forward state assessment on the third grade. And their website is very hard to navigate, but essentially you can look at performance of third graders from a variety of backgrounds. You can group them in different ways. And no matter how you do group the data, there wasn’t a single group that in which over half the kids performed at what’s called a proficient or advanced level.

More than half are performing under that, which is basic and below basic.

These are the data but which we don’t have to look at too much. But, if 39% of the third graders are– This is setting the bar not that high and they just have to be above basic and if we’re only getting 39% of our kids overall to that level, and a quarter of them at the minimal level, then we’re doing something wrong.

Literacy is endangered for many reasons. So is education for all.

And what I argue is that we need effective educational practices more than ever especially for children who are at risk for other reasons, like economic ones.

I also want to argue that we could be doing better than we are.

In particular, we could make better use of this reading science that’s been accumulating over many decades but doesn’t get in the classroom door.

So, in fact there is a big split between education and science where the findings from the science of how children develop and how reading works don’t make their way into how teachers are trained and instructional practices and so on.

This disconnection makes it harder for many children to succeed. And it makes it harder for teachers to succeed, too.

They’re not being given all the tools they might have, make use of.

So I was moved to write a book about this,

which just discusses the nine chapters on the science of reading and three chapters on the implications for education.

I won’t go through this but there’s this one a quote from The New York Times guy I really like which said, “He has that rare knack of sounding reasonable “and righteous at the same time”.
(audience laughs)
So, the first thing to say is we could be doing better. Is that actually true?

Maybe we’re already doing as well as possible given other circumstances. So, Diane Ravitch who’s certainly the most prominent educational advocate,

has been for quite a long time, has argued as follows, “Look we know how to teach kids to read.”

We know that because children from higher income backgrounds, at better schools, learn to read.

And so, she concluded in one of her popular books that low literacy is due to poverty. It’s a poverty problem.

So, rather than being an educational issue, low literacy and poor school achievement, which follows from that, it’s thought to be an economic issue not an educational one, which is basically saying look, some people are too poor to benefit from education.

Well, if this right, my talk is over and we can just finish lunch.
(audience laughs)

My question is, is it a good argument? It’s obvious that the economic circumstances affect educational outcomes. Opportunity gaps are real.

However, low achievement in this country is not just limited to low SES groups, socioeconomic status I got to use the term. So it’s kids from lower income backgrounds. These are data from one source, the OECD PISA comparison. These are United States scores. These are for 15 year olds. And these are by income levels. And the main point here is the United States isn’t just scoring poorly in this low-income area compared to for example, countries like Canada and Finland. They’re scoring lower across the board, the whole distribution is shifted.

So, it’s not just a low SES problem.

User data from the NAEP. So, everybody hears about achievement gaps.

Lot of focus on the black- white achievement gap.

So these are data from the NAEP which is the U.S. administered test.

And the data have been broken down by eligibility for free or reduced-price lunch.

So, there are kids who are, the poorest kids who are eligible for free lunch, reduced price lunch, or not eligible. It’s a very rough index of relative economic circumstances. So, you see two effects here.

One is there’s definitely an effect of income.

The kids who were eligible for free lunch do score more poorly and similarly as expected for these other two groups. That’s the effect of the economic circumstance. However there’s also a group effect where the white kids are scoring higher than the black kids and this has been consistently true over the history of the NAEP.

Also, low socioeconomic status is associated with poorer schools and teachers, fewer libraries and other resources. It’s associated with many things related to education. So, it doesn’t really make sense to me to say there’s an effect of poverty that’s separate from the effective education.

We could say low achievement is an economic issue and an educational one.

So, eliminating poverty would obviously help a great deal. Someone should do it. And while we are working on ameliorating or reducing or even eliminating to the extent possible these differences in opportunity, there are millions of kids from various economic backgrounds including low SES backgrounds who are in school and who need to learn to read.

And indeed for kids from these lower income backgrounds what happens in school matters even more than for kids who come from more advantaged backgrounds.

I don’t think we should be using poverty as an excuse to give up on these kids.

So, granting that this is not meant to be at the expense of efforts to address economic inequalities of various sorts, we could do other, look to see about what other things we could do now. So, what we want to do is look for conditions that are malleable, things we could do something about that would have a significant impact on outcomes. Are there any?

Today, I’m going to just mention three.

One is certain educational practices and policies.

A second would be connecting science and education so that we can make use of the research instead of leaving it on the table.

And the third is boosting children’s language experience before they get to school. These are all things that could have a potentially significant impact on reading outcomes.

Educational policies and practices. Well, essentially if you look at many things that are taken as standard procedure in American schools they magnify existing differences between us in SES. So for example, there’s a lot of outsourcing where instruction that used to be done in the classroom is now expected to be done in the home. Your kid might need to have extra practice on something that’s going to have to happen in the home. Your kid might need phonics. That’s going to have to happen in the home. Your kid might need various kinds of things that are not going to be covered in the classroom ’cause there’s not enough time in the day, it’s argued, but unfortunately this assumes that there’s a person in the home who can do this and that they have access, they speak the language, and that they have access to the resources for example, that are available on the internet or that they have the funding to pay for a tutor.

So, to the extent that things that used to be taken care of in the classroom are now expected to be taken care of in home, and there are a lot of them, especially for beginning readers, this is something that’s going to multiply existing differences between, associated with income. Also we’re looking at curricula and the kinds of assessments that are done of kids’ progress. They’re very middle class. Indeed they make assumptions about children’s backgrounds that may not be valid. Kind of makes it harder for certain kids to make use of those materials than others.

It is certainly the case that schooling doesn’t act as the great equalizer, in fact, gaps between groups that exist on the first day of school actually increase over the first several years of schooling. They’re not getting smaller, they’re getting bigger.

Second thing is what teachers are taught about how to teach reading and how children learn. So, a lot of attention being paid to this now is there’s frustration of how teachers are prepared for their job. With respect to reading, this has been going on for a long time.

They aren’t taught very much about how to teach reading and they’re certainly not taught enough.

We scientists look at what prospective teachers are taught about how children develop and how they learn and how reading works and what literacy is and we think this is just utterly archaic.

People rely on some old standard people like Dewey, Montessori, Vygotsky, Piaget, and a few others. These are psychologist who made important contributions. They’re all at least 100 years old.

And indeed much has been learned since they made their contributions. On the education side, they’re sort of treated as the source of eternal verities.

And so, as a consequence they’re not, teachers who are preparing to teach young children aren’t exposed as a part of their preparation to the kind of research I wrote about in the book.

So, think about it for a second, math teachers, what’s the preparation for being good math teacher? One thing is you need to know math.

History teacher, you need to know history.

So, what do teachers who are going to teach children to read need to learn?

The way things are done now, it’s placed in the context of literacy.

People don’t really talk about learning to read anymore in educational circles. It’s very much focused on literacy, what children are doing in first grade, second grade, when they’re developing their skills. It’s placed in the context of, it’s sort of like junior literacy.

It’s not really a question about the structure of books or the nature of a different kinds of genres.

The stuff that’s relevant to actually getting the child off the ground and into reading, so that they can benefit from those reading experiences, is actually the science about how reading works and children learn. That information is pretty systematically withheld in this country. The same story is true in the United Kingdom, Australia and New Zealand. We have very, very similar histories and similar sorts of arguments.

Again, it’s a big country here. There are many programs. There are exceptions to the generalization the I just made, but not many.

Currently, what happens for the most part is teachers learn to teach by teaching. It’s taken as a kind of philosophical point. So, there’s a commitment to this idea of discovery learning which is basically: It’s better to just engage in activities and be an active learner and discover, than have an expert sort of tell you things.

Well, discovery learning may or may not work well for various kinds of topics that children need to learn. But what has happened here is people have applied it to the profession of teaching itself. Basically, it’s learning on the job.

And what I argue is learning on the job is a really bad way to find out how reading works. Look, most of what’s going on in reading is going on unconsciously. You are aware of the result. You understood a text, you didn’t understand a text, you connected it to something else, you have a question, that’s all the conscious stuff. It’s the tip of the iceberg. The way that you got to that understanding is all going on under the hood. It’s the stuff that we do, our research, to kind of bring out into the light.

So, trying to figure out how reading works and how children learn in a classroom of 25 kids where many other things are going on.

I didn’t apologize to the mic.
(audience laughs)

It’s a very bad problem for learning on the job. And indeed I think it makes it harder for teachers to succeed and it’s one of the contributing factors to the burn out in the field where something like 50% of people who go into the field leave it within five years.

Okay, so those are some practices on the education side that might benefit from change. What about this disconnect between science and practice? So, there is a two cultures problem, happens on university campuses.

This is something we could fix tomorrow. But in fact, the barriers to it are enormous.

The kind of reading science I’m talking about has been going for literally decades. There have been regular attempts to knock on the door of educators and say, “We need to talk “about what the implications of this are for “how you might structure curriculum, “how you might interact with children, “how we might promote reading success.” Those conversations are very, very difficult to have.

And as a consequence, this research has had very little impact on actual practice. In this country, there have been reviews of the scientific literature relevant to learning to read and dyslexia. One of them is the National Reading Panel report which gets the most attention but there were several others that said essentially the same thing. In the United Kingdom, they had the Rose Report. There are similar things in Australia and New Zealand. For a very long period of time scientists like myself have been knocking on the door and saying look these things that we are pretty sure are true about children and learning to read aren’t really well reflected in your assumptions about how they learn to read and maybe we could actually make use of this information. That hasn’t happened.

I think this has been harmful. It’s been unnecessary. It’s been wasteful.

So, what I want you now to do is be a skeptic and say oh yeah, who cares? It’s science. It’s just scientism. Science is the solution to everything. What I want to do is give you an example which suggests yes, there is some information here that could be made use of. It’s just one example.

So a lot of research has focused on a very general question which is what’s the relationship between reading and spoken language?

One way you can frame it is, here’s print, and what’s its relationship to speech, spoken language?

And similarly, we know something about speech, what I’m doing now, what’s its relationship to print?

So, here’s an experiment. It’s the first from a paper that was the first one that I ever published. It was a really simple experiment with college undergraduates. In my undergraduate class we do this in class, We can’t really do it here. But you can get a sense of what the experiment was like. So, a student, a participant in the experiment has some headphones on and they’re listening to pairs of words, spoken clearly. One after another and they only have– Their task is just to decide, do these two words rhyme? There’s nothing tricky, but it’s really do these two words rhyme? Sometimes they do, sometimes they don’t. Person presses a button when they do, presses another button when they don’t.

So, the just familiar things like pie, tie, those rhyme. Rye, tie, those rhyme. Bed, tie those don’t rhyme. There are many pairs of this sort, of course, and they’re all mixed up and so on and the important point is they don’t see them. They’re not reading the words and they’re not writing them. They’re just hearing them. What we’re asking is the finding of importance is people are very accurate at this but it takes them longer to just decide that these two pairs like this rhyme.

And the reason is because they don’t overlap as much in spelling.

Basically, if you’re highly illiterate like Wisconsin undergraduates for the most part.
(audience chatters)
No, sincerely.
(audience laughs)

There’s a sense in which rye and tie are not as good a rhyme or just don’t overlap in sound as much as pie and tie because your knowledge of spelling has penetrated your representation of speech. These effects don’t occur in people who are illiterate. They are due to– The spelling is sort of this thing that infiltrates your knowledge of spoken language.

So, that’s the case of spelling having an impact on speech.

There something called phonemic awareness test. So, this is a test that people use with beginning readers typically like first graders. And the task is very simple. It’s like this is the audience participation part. If you want. So, the task goes like this. Say split without the puh.-
[Audience]
Slit.

– I can’t hear you.

–
[Audience]
Slit.

– Okay, so, I’m going to give you a couple of others. Easy task, right? Slice that sound out and then just say what’s left over.

Say crab without the buh.

–
[Audience]
Cra.- Yeah, sometimes they’ll be words, sometimes they won’t.

Say tub without the tuh.

–
[Audience]
Ub.- Say lamb without the muh.

–
[Audience]
Lab.

– What did you say?

–
[Audience]
Lab.

– You’re good. Okay, so lamb without the muh is actually la.

What you did was you used your knowledge of spelling and took the M out and looked at the spelling that was left over and used that to pronounce the word. Try one more. Say sword without the suh.

–
[Audience]
Word.

– Sword without the suh is ord. This is actually taken, of course from a real experiment in which good readers, who are like middle-school aged,

made errors such as say sword without the suh and they’d say word.

So, what’s happening here is there’s this close integration of the print and sound

that enters into a task again, no one’s reading anything here, they’re just listening to things. You think of the word bat as consisting of three sounds, buh ah tuh. Right? Those are called phonemes. Bat actually, the spoken signal, does not consist of three discrete sounds. The idea that it consists of three sounds is an illusion that’s based on how the word is spelt. What this is saying is that when you become literate, spelling and sound become intertwined. and in fact the shape each other.

This is an experiment that was done– I did my experiment on this a long time ago. This is the rhyming study again. Some decades later, neuroimaging and related methods for studying the brain activity in intact humans came about. This was a graduate student of mine who is now an eminent neuroimaging person. Basically he did the rhyme study, pie tie, and he used a method that allowed for interference with either parts of the brain that are related to speech or parts of the brain that are related to spelling. Long story short, what he found was that mismatch affect where it takes you longer to decide that pie and rye rhyme,

tie and rye, is actually arising from a part of the speech system, SMG. It’s a part of the circuitry that’s used for comprehending and producing speech. So, this was a modern method that confirmed what we’ve looked at– We suspected earlier which is that spelling really is seeping into the language system and changing it.

Okay, so that was all about print influence of speech. What about the opposite? So, now I want to think about how spoken language affects reading. This is silent reading. And the way to start is just you should maybe read this to yourself.

So, it’s a sentence if you get the syllabic stress right. The syllabic stress is a property of speech.

And I can demonstrate this by assigning the syllabic stress incorrectly.

“Permit me to give you a parking permit.” Makes no sense.

“Permit me to give you a parking permit,” makes sense. But the sentence is only interpretable. It can only be comprehended if you assigned the correct stress that is a property of the sound pattern of the word to these words. There are other studies that look at these phenomena very systematically. I’m just illustrating the effects here very, very crudely. So, there’s something like somebody’s reading “I’m hungry. “Hand me that tasty looking apple and pair.” Now, they’re reading this– They’re reading this silently. And the phenomenon that occurs under a variety of circumstances is that people will make errors or they’ll look longer or they’ll blink basically when they get to a word that sounds like the item that would fit but isn’t actually spelled correctly. What this is saying is look if you’re reading visually, as some people argue, and just going straight from print to meaning, then the fact that this word sounds like pear, the fruit, should be irrelevant.

We do these experiments over and over and over again and it’s a very robust effect. People will have an issue with in silent reading with the words that sound like they should be correct but aren’t. In fact, we set up circumstances where we really try to– We set it up so that if we discourage people from making these mistakes– We make it so that they’ll perform better or they’ll earn more money or whatever, if they just turn off the sound in their heads. They can’t do it.

Okay.

This is a study that came out of Haskins Laboratories which is affiliated with Yale and New Haven.

This is a study led by Don Shankweiler who is really someone I really highly respect as a person, who’s researched for, done research on struggling readers for a very long time. They asked a really simple question. They take kids, these are middle school to high school kids, so it’s not just the beginning readers. And they’re kids who are reading at a variety of levels.

They get some behavioral measures to assess their reading and other things and then they put them in the fMRI scanner

and as you I’m sure all know, what you get is some information about areas of the brain that are activated when people are performing various kinds of tasks.

Well, the tasks that they had here were they simply compared what brain activity when people were listening to a story versus reading the story.

And they counterbalanced everything so that they didn’t see the same story in both speech and reading and also the source of niceties. So, what you want to ask is when you’re reading a story and trying to comprehend it versus hearing the story and trying to comprehend it, how much of the brain circuitry is shared and how much of it is different?

And the findings from the study is really cool. It is that the degree of overlap in brain activity for listening and reading

increases as a function of reading skill.

The more skilled you are, the more overlap there is. It’s not complete overlap because one of them is heard and another one of them is read. There’s some vision and hearing, but the core things are overlapping.

From the brain perspective, one of the hallmarks of being a skilled reader is you’ve integrated print with spoken language.

Finally, this was a study from the same group where they looked at overlap between speech and comprehending speech and comprehending print in four languages where the languages are very different typologically. That is they’re dissimilar. And they also have different kinds of writing systems. They looked at Spanish, English, Hebrew, and Chinese Mandarin.

And again, the finding is when you are a skilled reader, the degree of overlap between print and speech is very high. And it was true across these languages and writing systems. So, they said this was a universal sort of signature of what it means to be a skilled reader.

What’s all this research telling you?

Reading and speech are deeply intertwined. It shouldn’t be a surprise. Writing systems are codes for representing spoken language. That’s what they do. In fact, in my book I go through the history of writing and how was discovered ’cause I thought it was a really interesting topic. How did anyone ever figure this out? And why did it take so long? So, is this a writing system? No, it’s a cave painting and it communicates something but it doesn’t– We don’t know exactly what beyond there was a horse there.

Is this a writing system? It’s the sign symbolism they you get in the airport.

Again, it’s communicative, it’s a sign, a communicative sign. It’s not a writing system. It doesn’t represent the words fork and knife. What does it represent? An eating area, food.

That’s a writing system. The reason is writing systems are systems for representing speech and the challenge to our ancestors was first to recognize that would be a good thing and the second was it took thousands of years to work out systems that did this adequately. So speech just kind of baked into writing.

Children learn to speak before they learn to read. Spoken language skills are very strongly related to early progress in reading. Spoken language deficits which can come from experience or be constitutional in origin have enormous impact on kids’ progress.

So, this is research that covered probably four decades.

If people had been paying attention to the science, would there have been reading wars? Everyone knows about the reading wars. What were the reading wars about? The reading wars were boiled down to whether people should emphasize phonics. Phonics is methods that emphasize teaching the kid how spelling relates to spoken language.

Big wars about this, disagreements. And indeed they continue sub rosa to this day.

This research said that the question was really a poor one because what happens when you were reading is you are learning to integrate, dock your knowledge of print with what you already know about speech. And to say, well actually, that’s a bad thing to do, it’ll make you a poor reader, was the exact opposite of what this research had been indicating over a long period of time.

And indeed most educators, people who are on the front lines, they’ve been socialized into the view that if you emphasize the link between print and spoken language, that’s the path to becoming a poor reader. You’ll be moving your lips while you read or something. The issue is really more than just phonics. It’s about your knowledge of spoken language, your knowledge of print, the ways in which those two things relate and indeed the way in which they relate to the world.

Okay, in fact we don’t hear very much about the reading wars anymore. The reading wars are over and the science lost.

Essentially, we’ve had balanced literacy and then other things to follow which essentially said, do what’s best for you in your classroom. Do what’s best for your children, your learners, your students.

The curricula that– Most of the curricula that are used now are big fat books in which they do have to include some phonics, but they basically are a kitchen sink approach. Everything that any teacher could possibly have learned about what could be relevant to learning to read is now in this book. And the teacher manual really leaves the teacher with far too many things to do

in a school year. There isn’t enough time to actually do everything. That yields the opportunity to just kind of create a curriculum out of these materials. People go with what they’re familiar with. What they were taught is that phonics is going to lead to poor readers. You should do it very sparingly, avoid it in large doses ’cause it’s toxic. What the science says is kids benefit from instruction on this. You want to get them in and get out so that they can move on to all the other things you need to do when you’re reading.

Okay, so the summary of this part is teachers’ beliefs about reading aren’t connected to basic facts. I should say here this is not a blame the teacher talk.

It’s not about the teachers. It’s not about their intelligence, their motivation, their willingness or anything. It is about how they were taught.

And that was not adequate. What happens instead is that there’s this exchange of information in what’s called communities of practice.

This is a way for like-minded people to share information.

And this is also a way to propagate misinformation. As I say in the book, the debates about how to teach reading really were an early example of the polarization over many ideas that we see in society now, that are characterized by groups of people who talk to each other, the echo chamber kinds of things, where you get magnification of misinformation

about there being pigs blood in vaccines, I read yesterday.

In Wisconsin, we have that kind of thing going on. So, in Wisconsin there’s essentially a lobby called the Wisconsin State Reading Association. It’s a teacher-led organization. I joined. They send out materials.

They’re woefully out of date. They say things that really are bad. They promote practices that are not good. And yet it’s very difficult to change.

Here’s another place. This is the page on reading in the DPI.

It’s astonishing. It contains links to websites I describe in the book that just have basic, basic historical facts wrong and really sort of collate for people things they should look at and then things they should not. Leave out things they should not. From my perspective, this is just perpetuating misinformation.

So, I think on the education side there’s a lot of things to answer for even if all the children’s issues in the schools are not educational ones.

And I’ve spent a good part of my career trying to build bridges. Part of the reason I wrote the book was because I became frustrated with how hard that was and indeed it’s not just about Wisconsin. It’s something that’s happening, exists across the country.

Okay, that’s education. Let’s talk about kids.

So, what’s something else that’s malleable? Remember we’re looking for things that are malleable, that we could do something about while we’re working on other problems. Well, here’s something that’s malleable. Children’s spoken language.

So, the starting point for reading is speech. Kids learns to talk before they learn to read.

A child does not relearn the language when they learn to read. They learn how print relates to the knowledge of language they already have. That’s derived from speech.

Many reading– I’ve talked about reading in instruction, things we could be doing better in the schools. It is also true that many children’s problems originate before they even start school. So, there are differences in how far along children are on day one. This points to things that we should be doing before, when the children are younger. One of the things that has come out as being very important is differences in their exposure to spoken language.

So, I think many people will have heard of this study ’cause it finally became famous after the authors died. It’s the Hart and Risley study. It was a heroic study for its era. It was published early ’80s. And they had collected data about the language in the home in a small number of families. And basically what the important part of the study is that they documented that there were very large differences in the range of vocabulary of items, the uses of language and the complexity of the sentences in the homes that they had visited and obtained data from. Now, in their study they related it to SES. So, it’s the poor kids who had a weaker language experience and it was the higher SES kids who had more.

This research launched many other studies. Now, the picture has been spelled out a bit more.

It’s this variability in how much spoken language is used and what types isn’t just a fact about lower income families, it also varies in middle income and upper income families. And also if you just look at lower income families as Catherine Snow, who’s at Harvard did, there’s huge variability in how those parents interact with their children using language.

So, nonetheless there is variability.

And variability and exposure, knowledge of spoken language, is a huge factor related to progress in reading and from there, in school.

The Hart and Risley subjects whose spoken language was recorded and that was that early study that established the phenomenon, they actually followed up on a subset of these kids when they hit third grade. So, they were able to say, “Well, we saw these differences in spoken language “among these kids and their parents.” Did that translate into differences in progress in learning to read, and the answer was yes.

There are lots of other studies about speech and later reading.

One really impressive kind of study is–

It’s called a prospective longitudinal study of children at risk. What is means is you find families where there’s already a history of dyslexia, reading impairment, struggling readers. It’s a heritable condition so usually there’ll be more than one family member who’s affected.

So, you find a bunch of these families

and then there’s a child who’s born and follow the child. And the question is, “Will this child

“develop dyslexia?” Will they struggle with reading and will we be able to see it emerge over the first few years of life before they get to school?

So, what you’re doing in these studies, the first one was done by a prominent researcher named Hollis Scarborough. Now, there are several of them. You record things about their vocabulary and how many utterances they produce and the complexity of the utterances and other things about their other kinds of abilities.

And then you follow them over a period of years, periodically going back to see how they’re doing.

And these studies have yielded just really incredible sorts of results. So, one of them that was done by a guy named Dennis Molfese.

His team recorded neonates’ responses to sound within 36 hours of birth. The way you do it is you put one of those caps on the kid, on the infant, with all the wires coming out. It records very, very low-level changes in voltage that are happening on your skull now, all the time. And what they were looking for was, they’re evoked potentials, they’re EEGs.

They’re looking at these infants’ responses to speech and to auditory tones. They record the data within the first 36 months of age and then they follow the kids.

And long story short, they followed them until they were eight and they found differences in the pattern of responding to sounds within 36 hours of births, explaining variance in their reading when they were eight years old.

Indeed, there were small kinds of differences in the wave forms that corresponded to small differences in their performance on various reading tasks. It was a brilliant study. Here’s another example of it. This is from some British folks starting from third grade, sorry, age three to age eight.

They look at the kids’ oral language, basic stuff like vocabulary and sentence structure. Then they look at kind of pre-reading skills. They look at all these things, but oral language is related to the emergence of pre-reading skills like knowing the alphabet and knowing the sounds associated with letters. Then you get to word reading, basically decoding words fluently and accurately. And then finally, you get to reading comprehension. These arrows aren’t just there for, aren’t decorative. These are a path analysis of the effects of one thing on the other. And the important thing here is to follow this trajectory and moreover differences in oral language recorded when the kid is three are predictive of differences in reading comprehension much later.

Okay, all this has pointed to the importance of early spoken language and in the U.S. awareness of this issue is growing and the question is what can be done?

There are efforts underway in various places. The most visible one has been in Providence, Rhode Island, where they were some community-wide sort of effort to essentially do language enrichment with children who were at risk for not having the spoken language skills that reading and other things could build on. And those projects, they’re ongoing and they may be successful eventually. They have not yielded very good results. And you can be– There’s many reasons for this. One of them is the amount of time on task you need to make up for differences in gaps. Another one is what you’re communicating to a parent when you say you need to change the way you talk to your kid. It’s a hard message and it’s not one that is, that will promote buy-in. There are further problems because you don’t want just more of the same, you want a broader variety. You don’t get that just from

getting the parents to talk more.

So, there was an old message about reading to children is essential. The message about talking to your kids is essential is spreading and that’s good. We might have to do other things to accelerate this process.

So, here’s something positive.

It’s not that easy to get parents to change the way they interact with their kids and the message is not really great.

However, there are other ways to kind of pour this language sort of over the kid.

So reading to children is good for a lot of reasons. But what psychologists have been recognizing and studying recently is– This idea initiated with my colleague Maryellen MacDonald who is in the psychology department and this is a really brilliant insight. When you’re reading to your kid, it is a form of speech to the child. You’re talking to them with them.

It turns out that the language of children’s books, even for the youngest kids, is different from what we say.

So here’s a book.

It’s a book that kids, like, throw up on. I mean, it’s for the youngest kids.

What’s the first line in the book? “In the light of the moon, a little egg lay on a leaf.”

People don’t talk that way.
(audience laughs)
In fact, it’s a more complex literary kind of construction and when you are reading that line to your kid for the nth time, you are actually– Those are learning trials where the kid is getting exposed to a range of expressions that they would not pick-up from everyday speech. We can do research on this now because we can collect vast databases of what parents say their caregivers say to children and we can collect data about what’s happening in the books and we can see where they differ and where they overlap. So, the idea here is that, in fact, these books have words, expressions

that the kid will not just pick up from everyday speech and therefore might be a source of language enrichment.

It’s going to depend on there being a parent or caregiver in the home who knows the language, who can read the book, who can interact with the kid in a way that makes it a productive learning experience. But none the less it is potentially another reason to read to your kid.

Here’s a couple of other ones that everybody can recognize. “The night Max wore his wolf suit made mischief.” It’s like a 32-word sentence. And the sentence structure of course, this is not that way we talk, again. And you’re exposing, in the context of this wonderful book, you’re exposing the kid to elaborate language, something they could never say. I like this one: “In an old house in Paris that was covered in vines.” That was covered in vines is a relative clause. It’s a relative clause. That’s said to be a kind of academic language that many kids have problems with. And it’s typically not dealt with until later in children’s reading curriculum. But in fact, we’re starting to say relative clauses to children when we read them Madeline book. In fact, they’re quite, again, there’s research on this, relative clauses and reduced relative clauses which are really hard are common in children’s books.

So, we think this is a potential way there needs to be more research. We think, in principle, this should be a way to insert a certain amount of additional knowledge of spoken language into the kid’s head when they’re very young, when they’re very good at language learning, young kids are great language learning,

in a situation that is a pleasurable, etcetera.

And could potentially leave the kid a little less gobsmacked for the kind of language they’re going to need to use when they get to school.

So, this is a potential.

Okay, so the summary of that point is that children learn spoken language mostly from adult speech.

Reading to children is a unique type of speech to children.

And potentially it can expand their knowledge of spoken language in ways that will be helpful to them when they get to school and actually have to learn to read.

Again, the circumstances have to be right and how big these effects will be and what other kinds of things they have to be coupled with, is stuff that we still need to do more research on.

Okay, I have about 2 minutes left or maybe I’ve gone over, I don’t know. I have a long section here that I’m going to skip and just summarize briefly.

I hesitate to go through it ’cause it is a sensitive and controversial issue and so talking about things just for a minute or two runs the risk of leaving something really important out. But since achievement gaps are important and also especially in Wisconsin and especially in Madison where they’re particularly large.

With Julie Washington and Maryellen McDonald and others, we’ve been doing research on African-American children who are speakers of African-American English which is a dialect of English.

So, the long story short about this if I could indulge you for a couple minutes on it is this.

There’s a huge amount of misunderstanding about dialects.

African-American English is a major dialect of English. There are others. And what the linguistic research on this established in about 1970

was that it is very unremarkable as an example of dialect variation. That is dialect variation is something that happens in language, spoken languages around world.

There are dialects everywhere. And when you look at their linguistic properties, they share certain kinds of properties. Everybody who learns to speak English, learns a dialect of the language. African-American English, contrary to previous racist characterizations of the dialect is just a normal dialect. It happens to have certain features and the use by an identifiable community. Here’s the issue.

There are differences between this dialect and the one that’s used in school or at least the one that’s used in books. There are differences in vocabulary and morphology, verb agreement, in sentence quantification, how quantities are expressed. There are a number of differences. They’re both versions of English but the one with the kid who’s speaking in the home, in the community, which is perfectly fine as a dialect, it’s just another dialect, actually differs from one that they’re going to need to have control over to succeed in school.

‘Cause you don’t learn to read African-American English, it’s an oral dialect, doesn’t have a written form. And moreover, it won’t have one because the written form that we use in this country is one that’s based on a more generic sort of general English.

Okay, the bottom line to that story is and of course, we’re doing research on this is–

There’s nothing wrong with AAE linguistically. It’s just a low prestige dialect. It’s the way people discuss it, talk about it, because of the people who speak it, because of the income disparities and so on. ‘Cause it’s not used in universities and it’s not used in government and business. It’s not a linguistic issue, it’s an issue about who’s using it, in what contexts. So here’s the basic thing that we’ve argued.

You look at the achievement gaps for kids that are caused by, granted, many factors. One of them that is overlooked is this fact that the kid’s own speech differs from the dialect that they’re going to need for school. What this means is that a child who’s speaking a minority dialect in the home has more to learn to reach achievement milestones than a kid who is already speaking this general English dialect in the home. They have to learn a second dialect. They have to learn alternative ways to say things that they already know how to say in their own dialect. Because they’re going to need these alternative ways to be able to read, to be able to understand the teacher potentially.

So, what we say is the comparison is not actually fair.

There’s two people. One group’s path to actually making progress in reading, expected progress is a lot longer than the other. That’s not taken into account when you just look at the scores.

You can use this assignment, take home assignment, think about how this compares to the bilingual situation. So, a child who’s got a different language, speaking a different language in the home, and who’s coming to school and is supposed to learn to read English, they fall behind. Their vocabularies in English are smaller, et cetera. There’s an obvious attribution to make there. The kid is behind ’cause they know Spanish and they’re learning English. Some of the same things happen when you are speaking a non, what’s called a nonstandard, but it’s just a minority dialect in the home. You have more things to learn, you need to learn alternative things, and you need to be able to switch between them. All of which are distracting you from learning to read, learning arithmetic, being able to focus on the content, being for able to focus on what the teacher is saying, and understanding the content rather than doing the translation into your own dialect.

Okay, so a part of the achievement gap is built in.

That’s even before SES and other factors are considered. Obviously, SES and so on are going to magnify these effects.

They can.

Okay.

So,

here’s the conclusions. In the United States and some other countries, educational practices make it harder for many children to succeed. Clearly, some children are going to succeed regardless. They may be getting taught in the home. They may be brilliant. They may have other advantages. For most kids, instruction makes a difference in terms of how fast they progressed and also whether they end up enjoying the activity of reading. If you make learning to read hard,

as I think current practices do, then you may end up with kids who eventually succeed but hate it.

Practices rely on external support

which discriminates against lower income families. Teachers are ill prepared for the job despite their extraordinary commitment.

And then the effects are compounded if the child is a speaker of a different language or a nonstandard dialect.

So, reading science has a lot to say about all of these issues and others.

And that’s why it’s deeply, deeply disturbing that we can’t get this stuff in the door.

It’s been hard to get into the door at the Capitol as they are discussing a bill that is supposed to establish procedures for screening kids with dyslexia and some other minor things. Wisconsin is way behind other states in terms of these kinds of actions. I could tell you about what’s going on in certain other states that’s a bit more positive. Right now, the educational establishment is pretty entrenched and an obstacle to change. Course it could change if we wanted it to. Thank you.
(audience applauds)