Author: Mark Seidenberg

Decoding “The Simple View of Reading” III

This is the third (and thankfully the last) of my posts on the Simple View of Reading and its relevance to instruction.  So far I’ve pointed out that the SVR didn’t address what was in the components, how they are learned, or the role of instruction, and that attempts to extend the SVR run into other conceptual and practical problems. 

As noted previously, I will prepare and post here a tidy PDF version that combines the three posts and includes the proper footnotes and references. Soon.

What About the 150 Studies?

The Simple View articles didn’t address instructional issues, but what about the many studies based on it, said to number 150? That is an impressive number. In the “science of reading” (SoR) approach people cite this number as though it justifies their programs, products and practices.6 That is not science. One has to actually look at the research to see what is there. Otherwise it’s just applying a scientific veneer.

The reason so many studies are consistent with the Simple View is because it states something that is true–the part about the child already knowing spoken language and needing to learn how print connects to it. It doesn’t take 150 studies to establish this point; it doesn’t take even one. It just takes a clear description of the situation, because then we can ask: is this ever not true? Is it ever the case that learning to read does not depend on existing knowledge of language (usually spoken; sometimes signed)? Is it ever the case that children do not have to connect print to this knowledge? Is it possible to comprehend texts without any knowledge of print? Can texts be comprehended without any knowledge of spoken language? No. The SVR is wholly correct about these general points.7

Empirical studies based on the SVR usually focus on two further claims: Claim 1) that the two components, print and spoken language, are sufficient (all that is required) to account for reading comprehension at every level of reading skill, and Claim 2) that the two elements make independent contributions to reading comprehension. My personal assessment, based on knowledge of this research and much other relevant work, is that Claim 1 is basically correct but for a pretty completely uninteresting reason, and Claim 2 is false.

The claim that everything related to reading can be fit into the two components can be taken as correct if only because there are few restrictions on what can go into them. The SVR is therefore consistent with a wide range of findings, but at a high cost.  It’s not that the SVR is correctly predicting specific factors that affect reading. Rather, once a factor is discovered it can be plausibly assigned to one of the components because they are so loosely defined. The theory can therefore account for any pattern of results. That is why I find the “supported by 150 studies” claim unimportant. If it is true, it’s because of a flaw in theory. 

Note also that studies that are consistent with Claim 1 aren’t necessarily consistent with each other. That is, they don’t show the same things at the level below print and spoken language. That’s why there are so many studies of the SVR: people disagree about what’s in the fine print.  Moreover, what is in the components at different ages can also differ, if needed. For example, at the onset of learning to read, vocabulary knowledge should go in the spoken language component, because that is how vocabulary is initially learned and the child can’t read yet. For people who are literate, however, new words are acquired mainly through reading. Accounting for these findings requires two versions of the SVR. If new research were to show that they are wrong, well, the SVR would be compatible with that outcome as well. That is the problem.

Or consider topic knowledge. People’s knowledge of the world obviously differs. We all have our areas of expertise and ignorance. By the start of school, children’s familiarity with topics that will be needed for school varies, which affects their progress in reading and other areas. Where does topic knowledge fit in the SVR? People who have similar reading abilities can certainly differ in their knowledge of a given topic, suggesting that world knowledge is distinct from reading and language. If that is true, the Simple View will need an additional component. Yet it isn’t hard to think of how topic knowledge could plausibly be assigned to the print/word recognition and spoken language components. For example, learning about animal habitats or parts of the body involves learning new words and expressions. Children need to learn to read and understand habitat and skeleton as well as does and said. That is just adding to the word recognition component. Topic knowledge also involves understanding sentences such as  “the ocean is a habitat for many species.” That can be considered an extension of language comprehension.  Voilà: two components are sufficient.

Do these exercises strike you as interesting or important? Me neither. Rather than fussing with the SVR one might focus directly on how topic-related words and ideas are learned and stored in memory, which would inform how they can be taught, efficiently and effectively. 

What The Data Look Like

I know that this isn’t everyone’s cup of tea–no, it is very few people’s cup of tea–but looking at the results from studies inspired by the SVR will be helpful if you can tolerate it. Figure 5 shows some of the findings from an excellent study by Ho et al. (2017) that examined the SVR in the context of learning to read Chinese. (If the SVR is correct it should apply to languages and writing systems other than English and alphabets.) The figure isn’t meant to be intimidating. The authors did a good job of presenting their results and their figures are useful for illustrating what the findings of such studies are like. This material will not be on the licensure exam.

Children in grades 1-3 performed a battery of behavioral tasks related to decoding, spoken language comprehension, and reading comprehension. The authors were also interested in a task called RAN (rapid automatized naming) in which children are presented with familiar visual stimuli to name aloud as quickly as they can. The stimuli can be letters, numerals, pictured objects, or color patches. The number of stimuli named correctly in a short time (e.g., 20 sec) is recorded. RAN performance is often found to be related to reading comprehension, although it is unclear exactly exactly why. (The study that I did on this was called “See Dick RAN” and it did not settle the issue.) Ho et al.’s study examined whether reading comprehension was predicted by print knowledge and spoken language, as in the SVR, and whether the predictions were improved by adding RAN as a factor. They called the components “decoding” and “linguistic comprehension,” and I will follow their usage here. The study includes numerous data analyses; I’m using two for illustrative purposes.

Figure 5 shows statistical relations among the factors and their relation to reading comprehension. The behavioral measures were chosen with these factors in mind. The numbers are estimates of the strength of the relation between two factors. A bigger number indicates a stronger effect; the stars indicate levels of statistical significance. The results show that Decoding and Linguistic Comprehension were both related to Reading Comprehension, as in the SVR. RAN had an impact on Decoding and Linguistic Comprehension but not Reading Comprehension. That’s not consistent with the SVR, although the reason why RAN has these effects remains somewhat unclear. 

The other big finding is that the Decoding and Linguistic Comprehension factors were related to each other, not independent (that’s the arrow labeled .754**), which is also inconsistent with the SVR.

Figure 6 shows the results for a statistical model predicting Reading Comprehension from the other factors. Decoding and Linguistic Comprehension both contributed to Reading Comprehension, as in the SVR. However, Decoding also contributed to Linguistic Comprehension, again indicating that they are not independent. In this analysis, RAN was associated with Decoding but not Linguistic Comprehension.

Conclusions? Even if your eyes have glazed over from looking at these figures, you can tell that the findings are not in a form that directly relates to instruction. The results involve factors (Decoding, RAN, etc.) derived from statistical analyses of children’s performance on behavioral tasks. The tasks themselves are not suitable for teaching. For example, word reading ability might be accurately assessed for the purposes of the study using a list of 40 words, but teaching those 40 words is obviously insufficient. You can’t go from these factors and the quantitative relations between them to a curriculum with an appropriate mix of instructional activities. The data are interesting for other reasons but the connections to teaching reading are minimal.  

Do the results support the SVR? Yes and no. Decoding and Linguistic Comprehension were both related to Reading Comprehension, as in the SVR. However, adding another factor, RAN, improved the prediction of Reading Comprehension, via its effect on Decoding. Decoding and Linguistic Comprehension were related, not independent. 

How do these results relate to those from other studies? Some have yielded similar results, showing that prediction of Reading Comprehension is improved by adding another factor to the SVR, or that Decoding and Linguistic Comprehension are related rather than independent. However, the literature also includes studies that did not yield one or the other of these outcomes. Some studies found that Decoding and Linguistic Comprehension were related, but in a different way than above. It’s a messy set of findings.8

Where does all this leave the SVR? If a person’s goal is to show that Claim 1 is correct, well, go for it. The components are so loosely defined it can be done, although it might involve some uncomfortable stretching. For the Ho et al. results, for example, one could decide that the RAN task is actually another measure related to decoding. It’s not implausible but I’ll spare you the arguments. I don’t assign much importance to this exercise, but it may be of more interest to others.

It is far more important to recognize that Claim 2, that the print and spoken language components are independent, is strongly contradicted not only by studies such as Ho et al.’s but from much other research unconnected to the SVR. Since about 1980 research on human cognition, language, and learning has emphasized interactions between different types of knowledge. That refers to ways in which different types of knowledge influence each other, in learning and in performing tasks such as reading. “Interactive” is the opposite of independent. Interactivity is seen at every level of reading and language.9

Unfortunately the idea that the two components are independent has taken hold in the SoR. Web sites and PD materials usually describe the two components as “interdependent;” however, there is also a lot of emphasis on how they differ. Learning to read requires instruction whereas learning a first language does not, it is said. This view can mainly be traced to Gough and Hillinger’s (1980) description of learning a first language as “natural,” whereas learning to read is “unnatural,” because it involves an artifact of recent invention, writing. Reading therefore requires instruction, whereas learning a first language does not. This view is represented by illustrations such as this one:

Educators may well realize that the two components are “interdependent” in some sense but in practice this line of reasoning has contributed to highly questionable approaches in which the two are taught separately. It has also led to an overemphasis on explicit instruction, a critical issue that I will take up in my next post. It is true that reading requires some explicit instruction. It does not follow that everything that underlies reading skill has to be explicitly taught because otherwise it will not be learned.

To summarize, once we take a glimpse under the hood we discover that the way the SVR divides things only makes sense at a very broad level. Gough and colleagues left aside many details in order to make one very basic point: that learning to read involves learning how print relates to spoken language. Teaching children to read requires getting more specific about what needs to be learned, when, and how. The SVR isn’t a good starting point for doing this.  Many findings are consistent with the SVR but that fact is irrelevant. What matters is whether such studies bear on specific proposals about teaching reading, all of which venture beyond the limits of the SVR.

The SVR + The Reading Rope = ?

Finally, let’s consider the relationship between the SVR and Hollis Scarborough’s Reading Rope (RR; from Scarborough, 2001), which are often paired. Scarborough called the RR “an illustration of the many strands that are woven together in skilled reading.” It’s a brilliant drawing and like the SVR it makes an important point–two, actually. One is that word recognition and language comprehension consist of parts. The figure provides additional detail about the makeup of the two components  (although, surprisingly, Scarborough did not mention the SVR). For example, the word recognition component in the RR includes decoding, sight word recognition, and phonology. The second point is that the parts, which are different types of knowledge, develop over time and become tightly interwoven in skilled reading. 

Both points are useful. As I have been arguing, it is necessary to unpack “language comprehension” and “word recognition” in order to decide what to teach. The RR also incorporates the fact that reading skill develops. Whereas the SVR is a snapshot of the child’s knowledge at the onset of reading instruction, the RR conveys that the parts develop and work together. 

Like the SVR, the RR is a credible place to start. Also like the SVR it isn’t an adequate basis for deciding what to teach, when, or how. Let’s leave aside questions about the particular components that Scarborough specified. Let’s leave aside questions about what “woven together” means. The fundamental question for education is how such knowledge is acquired, in particular the roles of instruction and other literacy-related activities. Scarborough captures the important idea that reading develops, but not how it is learned. The RR is still a recipe without the amounts or steps. The same concern arises for other lists of the components of reading of which there are quite a few. It’s useful to try to specify the knowledge and skills that underlie reading. Then we can look at how they are learned.

Scarborough herself has said much more about these issues. For example, in the article that included the RR drawing, she wrote:

It is customary to consider separately the strands involved in recognizing individual printed words from those involved in comprehending the meaning of the string of words that have been identified, even though those two processes operate (and develop) interactively rather than independently. 

Twenty years later, this observation is still correct. Both parts of it: that print and language are treated as distinct in many corners of education (see above), and that they develop interactively, from a very young age. 

She also said:

It is now abundantly clear that reading acquisition is a process that begins very early in the preschool period, such that children arrive at school having acquired vastly differing degrees of knowledge and skill pertaining to literacy. 

Twenty years later, this observation is also correct, the issue having been extensively investigated in different languages, cultures, and economic strata. The current focus on how to teach reading in school has overshadowed the enormous impact of early experience with language and the world. Many children struggle with reading for reasons other than how they are taught.

Scarborough has made numerous other contributions that extend well beyond what she could illustrate in a drawing. That work is at least as important as the Reading Rope.

Conclusions

Classic ideas such as the SVR and the RR are fine places for the “science of reading” to start and poor places to stop. If you don’t know about this work it’s new to you. If you do know about it, you’ll respect the fact that the studies don’t address basic questions about instruction or learning, and thus are consistent with many different approaches, including poor ones. I encourage people to embrace this work for what it offers–some important general insights about reading–and move on.

Rather than components of reading such as print and language we need an account of what, when, and how. We need a developmental perspective that considers the relationships between different types of knowledge, how the information is learned, and how learning changes as knowledge grows. Such questions have been the focus of reading research for a long time. Bringing more of this research into the “science of reading” is a reasonable goal and essential if it is to succeed.

Footnotes (continued from earlier posts)

  1. See, for example, Figure 1 from my first post.
  2. The exceptions are things like learning to read in a second language without knowing how it is spoken, or reading aloud in a transparent (highly consistent) writing system such as Finnish without knowing the meanings of the words.  Neither is relevant to people learning English.
  3. A few representative studies that assessed whether the SVR requires another component or if components are independent

answer no: Adlof et al., 2006, Braze et al., 2016; answer yes: Francis et al., 2018, Oullette & Beers, 2010, Ho et al., 2017.

Yes and no: Protopapas et al., 2013; Tunmer & Chapman, 2012. 

It’s complicated: Wagner et al., 2015

Overview: Catts, 2018

  1. The landmark article in this area was Rumelhart, 1977. See also Lesgold & Perfetti (1971, reprinted 2017).

References

Adlof, S. M., Catts, H. W., & Little, T. D. (2006). Should the simple view of reading include a fluency component? Reading and Writing, 19, 933-958.

Braze, D., Katz, L., Magnuson, J. S., Mencl, W. E., Tabor, W., Van Dyke, J. A., … & Shankweiler, D. P. (2016). Vocabulary does not complicate the simple view of reading. Reading and Writing, 29, 435-451.

Catts, H. W. (2018). The simple view of reading: Advancements and false impressions. Remedial and Special Education, 39(5), 317-323.

Francis, D. J., Kulesz, P. A., & Benoit, J. S. (2018). Extending the simple view of reading to account for variation within readers and across texts: The complete view of reading (CVR i). Remedial and Special education, 39, 274-288.

Ho, C. S. H., Zheng, M., McBride, C., Hsu, L. S. J., Waye, M. M., & Kwok, J. C. Y. (2017). Examining an extended simple view of reading in Chinese: The role of naming efficiency for reading comprehension. Contemporary Educational Psychology, 51, 293-302.

Lesgold, A. M., & Perfetti, C. A. (Eds.). (1971, 2017). Interactive processes in reading (Vol. 6). Routledge.

Ouellette, G., & Beers, A. (2010). A not-so-simple view of reading: How oral vocabulary and visual-word recognition complicate the story. Reading and Writing, 23, 189-208.

Protopapas, A., Mouzaki, A., Sideridis, G. D., Kotsolakou, A., & Simos, P. G. (2013). The role of vocabulary in the context of the simple view of reading. Reading & Writing Quarterly, 29, 168–202.

Rumelhart, D. E. (1977). Toward an interactive model of reading. In Attention and performance VI (pp. 573-603). Routledge.

Seidenberg, M. S. (1992). Dyslexia in a computational model of word recognition in reading. In P. B. Gough, L. C. Ehri, & R. Treiman (Eds.), Reading Acquisition (pp. 243–273). Lawrence Erlbaum Associates, Inc.

Decoding “The Simple View of Reading” II

This is the second of three posts examining the Simple View of Reading, one of the pillars of the “science of reading” (SoR) approach to reading instruction. In the previous post I noted that the SVR makes an important point–that reading involves the child learning how print represents words in the spoken language they already know–and that people expect it to have further implications for instruction. Here I look at the contrast between what the authors of the SVR proposed and how it is being used in the SoR. In the third post I look at what later studies of the SVR showed, and consider what is gained by linking the SVR to other pillars of the SoR such as the Reading Rope (Scarborough) and the report of the National Reading Panel. I’ll end with some broader observations about the use and misuse of research in deciding educational issues.

What Is and Isn’t in the SVR

Here is how the SVR is typically presented for educators. This illustration is taken from an explainer from The Reading League, an NGO whose mission is furthering the SoR. (Click on figures to embiggen.)

Think of this picture as a snapshot of conditions that exist as children begin learning to read. They understand spoken language at a 5-6 year old level. They will be able to comprehend texts (at that level) if they learn how the written code represents words they know from talking and listening. The multiplication sign indicates that both components are necessary: reading comprehension is zero if either of the components is zero.

What are the instructional implications of this statement, beyond the important general point about the need for print instruction? To go any further we’ll need to consider a few more things: what, when, and how.

What means specifying what’s in the circles. “Word recognition” is a convenient bit of jargon but we need to say more about what it involves in order to decide what to teach and how to assess progress. The same holds for language and reading comprehension, although at the outset the main concern is reading words.

When means considering when specific types of knowledge and skills are acquired over time, a crucial issue for instruction. How much the child knows, what they need to learn, and what they are able to learn change over time. The SVR focuses on one developmental fact, that spoken language (ability to speak as well as comprehend) develops ahead of learning about print; therefore learning to read turns on gaining knowledge of print and its connection to spoken language. It doesn’t speak to how reading develops over the next several years (or language, either).

How refers to learning: how different skills and types of knowledge are learned. Learning is conspicuously missing from the “science of reading”–remember, the picture of SVR in Figure 1 is just a snapshot of the state of the child’s knowledge. It’s not a description of how they got there or how they progress. Humans have two main ways to learn: via explicit instruction and via implicit learning. Explicit instruction you know. Implicit learning is the way brains pick up on patterns in the environment, without conscious attention or awareness. We update this knowledge all the time, as we do what people do: read and talk, write a paper, drive a car, make dinner. Achieving the right balance between these two types of learning, and adjusting it as students progress, is crucial to helping them learn efficiently and effectively.

The SVR doesn’t address these what, when, and how questions. That is why it is a huge leap to go from the SVR to conclusions about instruction. 

We could stop this discussion of the SVR here. Yes, it makes an important basic point that changes how people think about learning to read once they hear about it. No, it didn’t address other issues that are closer to the classroom. Fortunately, there are other places to look for that.

In my view, stopping here is the appropriate step, but I have to address some lingering concerns:

• Many people are trying to use the SVR as a framework, filling in some of the missing information. How is that going?

• What about the 150 studies supporting the SVR? Do they address the what, when, and how questions?

• How much is gained by coupling the SVR with other studies in the canon, such as the Reading Rope or the 5 Pillars of Reading from the NRP report?

Word Recognition and Decoding

Many people have recognized the limits of the SVR and tried to extend it in various ways. Doing this makes the limitations of the SVR as a framework for instruction more apparent. Dividing things up into print knowledge, language comprehension, and reading comprehension runs into problems–even though the basic point about the need for instruction about print is valid.

Let me illustrate this by asking: “word recognition”–what is that? Yes, I know: it’s recognizing words. But to take this to the classroom we need to get more specific. It’s not a trick question, but as you’ll see, it isn’t actually that easy to answer.  

The print-related component is called “word recognition” but that isn’t right because beginning readers already know how to recognize words, from spoken language (researchers call this “auditory word recognition”). Reading builds on this existing skill. No problem, we change “word recognition” to visual word recognition (VWR). The language comprehension part on the right in Figure 1 is really “spoken language comprehension”, which includes auditory word recognition. (It also includes language production, just as the print part also includes writing, but these parts weren’t included in the simple view.)

With these clarifications we have Figure 2 (I’ve omitted the = reading comprehension part to save space):

Are we good? Well, you can’t teach “visual word recognition.” That is a label for what we want learners to be able to do. Teaching has to focus on the kinds of knowledge that are involved, mainly words and their properties, which can be taught. 

Trying to do this within this version of the SVR creates additional instructional puzzles. Visual and auditory word recognition share a lot of parts–for example, meaning. We don’t represent the meaning of a word once for reading and another time for speech. Word meanings develop from using spoken and written language, and from experiencing the world (perceiving and acting). This common store of information is accessed in different ways in performing different tasks (reading, spelling, hearing, talking, others). Shouldn’t science-based instruction be consistent with these basic facts?

The same issue arises for phonology: it’s used in both reading and comprehending speech (also speaking). We therefore need phonology in both circles, but people don’t have one phonological code that they use in reading and a different one for listening and talking. Same for vocabulary and morphology. Same for essentially everything, even print.

Again: Trying to make sense of the SVR at the level of detail that’s required for instruction creates paradoxes and contradictions.

Bear with me just a little longer. I know this is a slog. 

Decoding “Decoding” 

What if we put “word recognition” entirely in the language comprehension part, and limit the print component to “decoding,” meaning using knowledge of the mappings between orthography and phonology (“phonics”) to pronounce letter strings (words like must and novel/nonce words like nust)? The purpose of phonics instruction is to enable this process. Decoding starts with overt pronunciation (“sounding out”), but readers soon begin to do this without speaking, using phonology, a mental code that is based on the pronunciations and sounds of words. I’ll call this “orth→phon” for short.

Now we’ve got this:

This is the other main interpretation of the SVR, as seen in illustrations like this one (from here), but also found elsewhere:

Call this SVR.D. This graphic does not say the same thing as Figure 1 because “decoding” as described here is not the same as “visual word recognition.” The latter refers to any process by which people recognize written words.  “Decoding” refers here to something more specific, generating phonological codes from print, which only works for some words in English. Such codes can be used to pronounce letter strings aloud. They can also be used to recognize words if the phonological code matches one for a word that is known from spoken language. 

“Decoding” in this sense and “word recognition” work with different kinds of spelling patterns. Here’s a summary:

Which version of the SVR are you using? 

If you’re feeling confused about the term “decoding,” you’re not alone. I just described the traditional concept of “decoding”, which predates the SVR and has been the basis for phonics instruction in the US since the 19th century. In English, only words that obey phonics rules (sometimes called grapheme-phoneme correspondences) can be decoded. “Decodable texts” are mainly composed of words that use grapheme-phoneme correspondences/phonics rules that the reader is assumed to have been taught.  

Decoding in this sense does not work for words whose pronunciations violate the rules. These were traditionally termed “sight words,” but they are also called “heart”, “trick”, or “snap” words, with various proposals about which words should be included, how they are read, and how they should be taught. Those details don’t matter here. Under all of these proposals there are exceptions that have to be learned and recognized some other way. If there is a rule for pronouncing but, cut, and nut, it will not work for put.

Gough and his colleagues used the term “decoding” in a different way: as a synonym for “word recognition.” 

For the simple view, skilled decoding is simply efficient word recognition: the ability to rapidly derive a representation from printed input that allows access to the appropriate entry in the mental lexicon, and thus, the retrieval of semantic information at the word level.  Hoover and Gough (1990), p.130.

This is how jargon gets in the way of understanding. Gough and colleagues were referring to the processes by which words–all kinds of words–are recognized. They were agnostic about the details. Unfortunately, they chose to label this component “decoding,” which already had a different, narrower meaning. This created massive confusion that has now spread to the “science of reading.”

For example: If you look closely, Figure 4 is attributed to Gough and Tunmer (1986) and Hoover and Gough (1990). The print component is labeled “decoding”, as in these articles. However, Gough and colleagues did not use “decoding” to mean the “ability to use sound-symbol relationships to read [some] words“. In fact, they pointedly questioned the relevance of this kind of decoding to beginning or skilled reading. Thus the illustration does not accurately represent Gough et al.’s own proposal. 

Does this matter? Well, yes. The SVR is one of the main research studies on which the science of reading is based. These observations suggest that people do not agree on what it means. In Figure 4 it is being used to justify teaching children to decode in the traditional sense, which Gough and colleagues did not propose or endorse. That assumption is coming from somewhere else and requires its own justification. In other contexts, the component is taken as “word recognition” and used to sanction other methods.

Well, so what? Isn’t Figure 4 a reasonable place to start? Readers need to learn to decode (original sense) or something like it. We’ll also need to add procedures for dealing with the exception words. This can go in the print component of the SVR, sure. But where’s the payoff?  Despite all this effort to amend the SVR, we are back at Square One: how do we teach children to read words quickly and accurately given the properties of written English? So far we’ve only managed to describe the problem, not how to solve it.  

Getting a little tired of this? Me too. And yet the same issues arise if we examine other targets of instruction such as vocabulary and morphology. The SVR again does not directly address these topics and so carries zero direct implications for instruction. If we were to expand the SVR to incorporate these types of knowledge, where would they go? Like meaning, they seem to belong to both components. Does that mean we teach each of them twice? Note that morphology presents the same challenges as phonics: “rule-governed” patterns with lots of exceptions (see, for example, the past tenses of verbs).  Syllables, too.

These are the puzzles that arose in 1980s research on learning to read. They are arising again in approaches to instruction based on this work–for example, debates about which words to treat as “heart” words or “sight words” and in how to teach phonics. The articles show Gough and colleagues struggling to describe “word recognition” and “decoding” using the concepts that were available at the time. They knew that readers needed to learn how spelling represents phonology and that at least some instruction was required in order to break the code. But instruction about what? They were skeptical of phonics rules, which seem like clumsy approximations of the complex system readers learn. Moreover, fluent reading seems very unlike applying rules to a series of letters. Having to memorize exceptions to the rules such as said and done creates an onerous burden for the learner and misses the fact that they overlap with many “rule-governed” words.  Gough and colleagues’ use of “decoding” was confusing but perhaps understandable. They were using it to refer to word recognition procedures that they correctly intuited had to exist but couldn’t quite figure out.

A solution came a few years later when McClelland and I applied new ideas about how brains learn to reading. In this account, people learn a network of mappings between spelling, sound, and meaning. These mappings reflect patterns that exist across words. Unlike rules that apply every time a pattern is encountered, the mappings vary in frequency and consistency. This knowledge can be closely approximated by simple neural networks that learn based on similarity: learning about one pattern facilitates learning other, overlapping patterns. Such networks can easily learn patterns that are hard to describe in words, for example that –ook is pronounced one way in book, cook, and look but differently in spook, which overlaps with spoon and spool, which pattern with boon and cool. The same procedures are used in reading all words; there isn’t a separate one for words that violate the rules. The network uses what it has learned from words to process letter strings it has never seen before.

Phonics rules, on this view, are a convenient fiction. They put parts of this complex web of knowledge into words. They are not accurate descriptions of what is learned, but they are useful because they make explicit instruction about the system possible. Explicit instruction has two purposes. First, it draws attention to what the child can learn: orthographic patterns that map, with varying frequency and consistency, to phonological patterns (and thence to meaning, if the pattern corresponds to a word known from speech). Learning the rules themselves is less important than getting clued into the fact that such patterns exist. The second function of explicit instruction is to get the orth-phon system off the ground quickly. With some patterns established, learning based on the similarity across patterns can proceed, with less and less reliance on a teacher for feedback as knowledge accumulates. 

This framework, which has many other bits, took a long time to fully develop. It runs counter to very strong intuitions about reading, and so took a while to gain broad acceptance among researchers. In my view, the “science of reading” is recapitulating this history. At the moment, it is focused on the Gough-era concepts and running up against the same puzzles. The good news is that there is a lot of very useful stuff to come.

Interim Summary 

The point of this excursion was to examine what happens if one takes the SVR seriously as a framework for early reading instruction. Attempts to be more specific about the print component run into theoretical and practical complications. We know that the reader needs to learn to read and understand words quickly and accurately; that this involves linking written and spoken language; and that the properties of written English make this challenging. After stepping through possible extensions of the SVR, we arrive back at Square One: how to teach children a system with productive patterns but many words that violate the main patterns in varying degrees, including most of the most common ones. (I’ve also noted that there is a possible solution just over the horizon but for now we are focused on the SVR and other classic SoR studies.)

There are still a few loose ends. First, it’s said that 150 studies support the SVR. Do these studies address the what, when, and how issues? Second, do other studies that are pillars of the SoR, such as the Reading Rope, provide details that are missing in the SVR? How far can we get with what we’ve got? We need to look. 

Decoding “The Simple View of Reading”

This is one in a series of intermittent posts about issues that arise in trying to use research in cognitive science and neuroscience about reading, language, learning, development and related topics to improve literacy outcomes. This “post” is more like an article, with footnotes and references. I’m going to post it in three bite-sized pieces, plus supporting materials. Some light summer reading! It expands on issues I raised in my previous post, and in talks and articles over the past few years.

As a scientist who advocates using research to inform instruction, curricula, and teacher training, I think it is important to assess what is being done under the rubric “the science of reading.” The first hurdle, convincing a critical mass of people that reading education could be improved by utilizing basic research, seems to have been cleared. That’s important progress. The next challenge is to develop a research-based approach that is also effective, efficient, and equitable. Connecting research to practice is difficult, and the “science of reading” is a work in progress. If the first step was recognizing what is possible, the next step is doing it well enough to achieve the goal of increasing literacy, especially among people who are usually left behind. That is the focus of these posts.

The “science of reading” approach in education is being carried forward by influencers whose background in cognitive science is limited. I don’t think teachers either want or need the amount of detail I provide here. This level of analysis is highly relevant, however, for the people who are creating curricula, assessments, professional development courses, and policies. 

My previous post ended with some concerns about the focus on a small number of classic studies in the “science of reading” (SoR) movement. These include the Simple View of Reading, the National Reading Panel report, the Reading Rope, and a few others. Some of the classic studies are useful for introducing basic concepts and findings to teachers and others who want to learn more. When I spoke to teacher groups about connecting research and instruction ten years ago, people would often say they were surprised to find out that this research even existed. Now many more teachers have heard about it, and in many places they are being required to learn more about it via PD courses. 

Here is a conundrum for the SoR:  What makes these classic studies good for onboarding teachers–their narrow focus on a few accessible concepts–is what makes them wholly inadequate as the basis for a science-based approach to reading instruction. The problem with these studies isn’t just that they are out of date, although they certainly are: research progress didn’t stop after the NRP report in 2000–in fact, it accelerated. Some of the later work was motivated by the limitations of these classic studies and led to theories and findings that superseded them. The further concern is that these studies simply do not speak to many of the most important issues, such as the impact of children’s experience of language and the world prior to school entry on learning to read, and the roles of different types of learning. This canon of studies is nonetheless being taken as the core science in the SoR and used to justify a variety of practices.

That is a mistake. If these are the findings and concepts you’re working with, you’ll have to make a lot of additional assumptions to link them to the classroom. Such assumptions could be based on other research, but currently they are being decided in an ad hoc way by the people, organizations, and companies that have stepped in to meet the demand for instructional materials and PD activities. New products and approaches are being assessed in terms of their consistency with the “science of reading,” but the studies taken as representing the science are consistent with many approaches to instruction, including ones that would be seen as ill-advised if other research were considered.

In the previous post I conjectured that this is happening because leaders of the SoR movement are themselves newcomers to reading research. Changes in teacher education, curricula, and practices are badly needed. Who is carrying the ball forward? Not the major schools of education.1 Not the cognitive scientists who conduct research. Mainly, people with expertise in allied areas such as school psychology, teacher education, curriculum and instruction, or software development. That’s different from being a cognitive scientist who reads and contributes to the research literature, understands the methods and data analyses that are used, and can assess evidence and claims appropriately. Teachers aren’t that far behind the people teaching them about the SoR.

The current situation reflects a simple mismatch between supply and demand. The demand for courses and materials based on reading research is high. The supply of quality resources is low, in part because few cognitive scientists look at instructional issues.  

We can do better, certainly. Capitalizing on the present opportunity to improve literacy outcomes by incorporating basic research demands that we do so. There is only a problem if we stop asking questions about what we are doing. Recognizing that current practices and materials are only interim solutions is preferable to institutionalizing what happened to be available when the “science of reading” caught fire. I will flesh out these points with a detailed example here. My goal is to make room for highly relevant later research that has yet to enter the discussion. 

You may be wondering why I don’t just cut to the chase and tell people what they should do instead of merely analyzing what is going on. In later posts I will describe other strategies for developing research-based approaches to reading instruction, curricula, and teacher education. They won’t be definitive, of course, but they should be helpful. My own contributions aside, it’s important to recognize that translating research into practice is not easy and that developing new ways to teach reading based on cognitive research is a process. This process includes formulating and assessing approaches and materials, lather rinse repeat. That is how we successfully transitioned from Whole Language/Balanced Literacy to the present, and it is how we will progress further.

Journalists such as John McWhorter and Nicolas Kristoff of the New York Times have popularized the view that “we know how reading should be taught,” as though the only challenge is replacing one method with another. It would be more accurate to say we know what is important for learning to read (and what to avoid), and are currently figuring out how best to teach it. Looking critically at new approaches to classroom instruction, teacher education, and curricula is part of the process. 

Finally, I view my primary role as teaching people about research, including what has been found and how to develop and assess implications for instruction. I am also trying to enable educators, publishers, and government officials to ask the kinds of questions I’m raising here.

Nobody Doesn’t Like ‘The Simple View of Reading’

Let’s start with the Simple View of Reading, because that’s where the “science of reading” usually starts. I’m reminded here of the discussion of reading to children in my book (p. 113). I didn’t want to ruin anyone’s memories of being read to by grandma or of reading to their children. Reading with young children is great but it’s not sufficient to insure they’ll learn to read, which almost always requires some amount of instruction. The Simple View of Reading is also great but not sufficient to decide important issues about how to teach reading. 

I remember hearing about the SVR at a conference at the University of Texas-Austin organized by Phil Gough in 1986. (My talk was about a neural network model of dyslexia.)2 The SVR was developed in a series of articles published in academic books and journals, starting with Gough & Tunmer (1986), and followed by several others. The SVR was an idea, an observation about learning to read, not a conclusion based on empirical research. I thought the idea was important and discussed the Simple View in my book (p. 118). The SVR says that reading is the product (in the arithmetic sense) of word recognition and language comprehension. Stated in words, the SVR holds that:

1. Reading comprehension can commence if children learn how the printed code represents words they already know from speech. They can then access everything they already know about such words: what they mean; how they combine with other words in sentences; images, actions, and sounds associated with them, and more.

2. Reading comprehension depends on both components being sufficiently developed. You can’t read without knowledge of print; you can’t understand print without knowledge of language. 

These points seems obvious–once stated this way. Children do not re-learn language in order to read; rather, they learn how to access knowledge of language and the things we use language to talk about in a new way, from print. Knowledge of both print and language are necessary.3 

Beginning readers need to learn about print, but how? Gough and colleagues argued that it requires instruction. A child has to learn that there is such a thing as reading, that it involves a code that represents spoken language, and how the code works. Compared to learning a first, spoken language it is an “unnatural act,” another of Gough’s pithy expressions, but more about that one soon.4

The SVR makes a compelling case for instruction to support reading words. It is very useful in debates with people who continue to believe that children will “pick it up” in a “literacy rich environment,” or that print instruction is only for poor readers. Once the basic point has been absorbed, focus can turn to how children acquire these types of knowledge, how instruction can promote progress, what kinds of obstacles can arise, how they can be addressed, and so on. 

The SVR doesn’t have much to say about these issues, as the authors were careful to acknowledge.5 The SVR is like a cake recipe that specifies the main ingredients but not the quantities or the procedures for combining them. It’s useful to have the ingredients spelled out this way; you just can’t bake the cake. Actually, the SVR is more like a cake recipe that specifies “wet ingredients” and “dry ingredients”, because the terms “word recognition” and “language comprehension” can signify many things. What kinds of information and mental computations are involved in each? How are they learned? How do they fit together? Why do learners differ in how they progress? The SVR doesn’t provide much guidance because these questions weren’t the focus of the work.  

Teachers nonetheless want to know what SVR says about other instructional issues (I’m paraphrasing actual posts here): 

“Print knowledge and language are separate components in the SVR. Does that mean they should be taught separately?”  

“Decoding and language comprehension are different components in the SVR. Does that mean I should teach decoding using nonsense words, which don’t involve comprehension?”  

“The simple view says that children already know how to comprehend spoken language and need instruction about print. That means they don’t need additional instruction about spoken language, right?” 

Such questions are raised repeatedly on social media. Why?

People clearly don’t know that the SVR doesn’t address these issues, probably because they’ve learned about the work via secondary sources that emphasize its importance without mentioning its limitations. They may not be concerned because they’ve been led to believe that this concept, along with a few others, is the science relevant to instruction. Every curriculum or approach inspired by the SVR is therefore consistent with the “science of reading”. People also may have heard that the SVR has been confirmed in over 150 studies, which sounds like it covers a lot of ground. And the SVR does state a fundamental truth: people learn to read by learning how the written code represents spoken language they already know. All of these factors have magnified the expectation that the SVR should also have further implications for instruction. It doesn’t. Let’s look.

—

This is the first of three parts. Coming up: What’s in the components; are they independent; is the SVR sufficient;  the 150 studies supporting the SVR; connecting the SVR to the Reading Rope; where else to look.

I am posting this in piecemeal fashion, but will compile the parts into a PDF along with complete references, footnotes, and acknowledgements. In the meantime, I have to point out that I am by no means the first person to have raised many of the issues discussed in these posts. This article by Hugh Catts and this post by Timothy Shanahan are particularly relevant.

Footnotes

1. Paradoxically, the demand for in-service training in the “science of reading” may allow schools of ed to forgo including this material in pre-service teacher education programs.  

2. Seidenberg, M.S. (1991).  Dyslexia in a computational model of word recognition in reading. My wife Maryellen MacDonald (whom I had yet to meet) was an undergraduate research assistant in Gough’s lab at the time.

3. Ideas very similar to the “simple view” are discussed by Jeanne Chall In her classic book “Learning to Read: The Great Debate”, who attributes them to Leonard Bloomfield, one of pivotal figures in modern linguistics. Chall wrote that Bloomfield “questioned the initial emphasis on ‘meaning’” in prevailing views of beginning reading and “called instead for making learning of the ‘code’ or the ‘alphabetic habit’ the first step. Since the child comes to school with a considerable command of spoken language, he reasoned, reading instruction should begin by teaching him [sic] the printed equivalents for his oral vocabulary. And since English spelling is irregular, he added, this is best accomplished by teaching those words that are spelled regularly.” (Chall, p. 24)  Gough and Hillinger (1980) discussed this work.

4. Gough & Hillinger (1980), “Learning to read: An unnatural act.”  Gough was fond of a clever phrase (and probably still is though I’ve had only occasional email contact with him since he retired around 25 years ago).  He also referred to the printed code as a “cipher,” as in “breaking the cipher,” but that term is so archaic it didn’t catch on. 

5. As Hoover has noted, ”The Simple View is neutral regarding how reading should be taught.” (Reading League Journal, 2021). The discussions of how reading works and children learn in the original SVR articles are thoughtful and nuanced (I only fault them for hijacking the term “decoding”). That nuance is not preserved in illustrations of the theory that circulate widely among educators. 

Recent talks

Here are the slides  from a few recent talks about the challenges facing the “science of reading,” and related topics. I’ve also included links to the talks that are available, and links to websites for the events.

Montag Lecture, Atlanta, March 2022,  “Efficacy, Efficiency, and Equity: The Goals of Early Reading Instruction”.  Slides. Video.

Path Forward Summit, Barksdale Reading Institute, November 2021, “The Science of Reading and Literacy Outcomes: Who Needs to Know What?”. Slides.  Info.

Barksdale talk with reading educators,  Barksdale Reading Institute, Nov 10,2021, “What we know about how reading works that’s relevant to teacher education”.  Slides. 

International Dyslexia Association Los Angeles, March 5, 2022, “How to Connect Reading Research and the Real World”. Slides. 

Thompson Center Summit on Early Literacy.  Madison, WI, February 8, 2023, “Using reading research to improve literacy outcomes”.  Slides.  Video.

 

Best symposium ever?

Here’s a link to the recording of a symposium that I highly recommend. The speakers were Kymyona Burk, Emily Hanford, Donna Hejtmanek, and me. It was organized by a center here at UW-Madison because the state legislature seems to be getting serious about legislation related to reading. Heretofore there hasn’t been any forward movement here because of our political polarization. This event turned out to be super. Really stimulating talks by people who know how to communicate (even mine went pretty well!). It’s a great team–I only wish we could take our show on the road. It wouldn’t be as big as Taylor’s Eras tour, but it would also be easier to get tickets! Cheaper too.

The organizers left a lot of time for talk among participants and Q&A from the audience, which included legislators as well as educators, parents, and other interested parties.

Take a look: it’s worthwhile. I say this having attended a lot of very good events of this sort over the years.

My talk starts at 30:00, but it’s preceded by Kymyona and followed by Emily and Donna. The talks take into account the situation here in Wisconsin, but that situation is similar to ones in other states, and the talks cover issues of general interest.

About the science in “The Science of Reading”

I was recently in a group zoom meeting (a groom? a zoup?) with some educators who meet to expand their knowledge of reading research. A guest speaker gave a rambling talk about “science of reading” (SoR) issues. Then an experienced educator whose work includes teaching other teachers, asked: “if a student is a good reader, do we need to continue with phonemic awareness instruction?” 

My heart sank. Why would a person need to ask this? The goal of teaching children to read is reading, not phonemic awareness. We know that learning to read does not require being able to identify 44 phonemes or demonstrate proficiency on phoneme deletion and substitution tasks because until very recently no one who learned to read had to do these things. Instruction in subskills such as phonemic awareness is justified to the extent it advances the goal of reading, not for its own sake. 

If a student is reading–if they’ve “broken the code,” as Phil Gough1 put it years ago–instruction can focus on the many more things that need to be learned to become a skilled reader. Instructional time is limited and the clock is ticking down to 4th grade.Time spent jumping through PA hoops could instead be spent on activities that expand the knowledge that supports comprehending texts of increasing complexity and variety.

These conclusions seem obvious to me, given my understanding of the relevant research literature and the conditions under which children learn to read. They weren’t obvious to the questioner, or to the people who posted to “chat” thanking them for asking, and they won’t be obvious to many people reading this document.

I’m going to lay my cards on the table here: The treatment of PA in the “science of reading”–the idea that a certain level of PA is prerequisite for reading, and that PA training should continue until the student becomes highly proficient at PA tasks regardless of how well they are reading–is emblematic of problems that have arisen within the SoR approach. It is an overprescription that reflects a shallow understanding of reading development, yet has become a major tenet of the “science of reading”. The PA situation and other developments suggest to me that the SoR is at risk of turning into a new pedagogical dogma, consisting of a small set of tenets loosely tied to some classic but dated research, supplemented by additional assumptions that are ad hoc and ill-advised. 

Worse, I see the “science of reading” approach as replicating several conditions that fostered the misbegotten practices of the recent past for which the “science of reading” is meant to be a corrective: well-intentioned individuals with limited research backgrounds who offer strong but flawed recommendations (and products) to teachers and school districts. They are treated as authoritative sources by an audience that cannot tell if they are or not.

The situation is not as bad as the stories that Emily Hanford covered in her programs. There is now much greater awareness of what was wrong with previous approaches and of the potential to create more effective methods using what is known from decades of research about reading, language, learning, development, and other topics. The Simple View of Reading is definitely a better place to start than the 3-cueing method, but a little knowledge can also be a dangerous thing. Especially if you aren’t aware that it only is a little knowledge.

I think that hard questions need to be asked about what is being taught in the name of “the science of reading”, and why. They are the same questions that were asked about Calkins, Fountas & Pinnell, Goodman and others whose views are now, deservedly, in disrepute. Applying the same criteria in evaluating all work regardless of one’s personal beliefs is a basic principle in science. It is especially important to examine the tenets of the SoR now because they are being implemented via legislation and regulation. As in these other cases, recognition of the flaws in an approach is an important step toward change.

After spelling out the concerns I will offer some suggestions about where to go next in the ongoing effort to bring cognitive research into reading education.

Before starting I need to address a couple of potential concerns. I want it to be clear that I think that effective methods for teaching reading need to be grounded in cognitive science and neuroscience research about how reading works and children learn, the causes of reading difficulties, and other relevant topics. Such methods (and their scientific basis) could then be taught to prospective teachers and incorporated in curricula and other instructional materials. There is no viable alternative as far as I can tell; approaches such as figuring out what works from personal experience or relying on the personal vision of a “thought leader” are inadequate. So is exposing teachers to a variety of perspectives on reading and leaving it to them to fashion a personal teaching philosophy.

I know this discussion is likely to be unpopular among people who have embraced the SoR. The main response I’ve gotten to talks in which I’ve raised some of these issues is that the comments were rude, even mean-spirited, and not what the audience wanted to hear. That is itself symptomatic of a problem. My concerns are not about loyalty to a movement or affinity to an author or product. They are not about whose team a person is on.They are not about the people involved though they are about their work. Mainly they are about improving literacy outcomes using basic research in cognition, language, and development–and avoiding the mistakes of the past. 

Insofar as they are critical of certain aspects of the SoR, they are also likely to be welcomed by people who have opposed research-based reforms.That would be a mistake. The research base is strong and the logic of using it to improve instruction is impeccable. However, connecting research and practice is difficult, and what is being offered right now are interim solutions. There is only a problem if we fail to recognize this. The movement is on the right path, but the destination is still off in the distance. Criticism of a particular product or activity that has been offered in the name of the science of reading doesn’t mean I oppose linking research and practice. It means I have a problem with that particular product or activity and think others should take a good look for themselves.

I recently saw a TikTok post that said “the reading wars are over–the science of reading won!” Nothing has been “won” until we see that many more children are being taught more effectively with more of them becoming more highly skilled. I am writing these comments with those goals in mind.

Finally, about the expression “the science of reading”:  The term isn’t in wide use among researchers. There isn’t a field called “the science of reading” and people rarely identify as “reading scientists,” in my experience.2 In reading education, the term has been taken up by a movement (often abbreviated SoR) to reform instruction, teacher education and curricula. This movement/approach is not the same as the body of research about reading. For one thing, the former has as yet incorporated very little of the latter. For another, the “science of reading” has to deal with real-world issues, such as how to change deeply-held but mistaken beliefs about reading, and how to make use of research on reading and related topics given teachers’ and other educators’ limited backgrounds and many other responsibilities. These are issues about using research in education, which is different from basic research on how reading works, children learn, etc. I will use SoR to refer to the approach to reforming reading education, and expressions such as “cognitive science research related to reading” to refer to the research itself. 

I have a lot to say and will be rolling it out in a series of posts. The issues raised below will be considered further in later posts.

1. Creating demand. We have been very successful in convincing people about the need to reform reading instruction, curricula, and teacher education. The “we” here includes numerous people and organizations. My book had an impact; Emily Hanford’s outstanding documentaries, which are both riveting and infuriating, have reached a much wider audience. Decoding Dyslexia and other advocacy organizations led the pursuit of legislative remedies after years of educational stonewalling. There have been other participants and forces at work. Let’s just say that a variety of events caused demand for reforms to skyrocket. Moreover, people want things to change starting now. Having convinced many teachers (administrators, legislators) that the methods in wide use are flawed, they want to know what to do instead.

This is wholly understandable. If you’re told you have a serious medical condition that requires surgery, you want to have it treated as soon as possible. Knowing that there is a potential treatment but having to wait months until there is an opening in the surgeon’s schedule would be excruciating (I imagine). This also holds for teacher education: recognition that it is inadequate has created demand for ways to address the gaps in people’s knowledge (e.g., PD for in-service teachers).

The demand for change is such that reforms began to be mandated via “science of reading” laws passed in many states, and via similar regulations at the local educational agency (LEA) level. 

Thus, we have a situation in which there is high demand for the “science of reading,” coming from both bottom up (teachers, parents) and top down (government, LEAs). This is a big advance, opening a path to overcoming resistance from schools and departments of education. The question then is how this demand is being satisfied. 

2. A Big Barrier. As I have noted since I began writing about these issues, incorporating scientific findings and attitudes into education is a monumental challenge. The educational establishment is a very large, complex ecosystem that evolved over many decades without incorporating cognitive research. (Indeed, the system instead evolved defenses against scientific incursions.) We are now observing, in real time, what happens when basic research is released into this environment.

The biggest barrier to successfully joining research and practice is the lack of expertise in the relevant science–how the research is conducted and assessed, how findings are weighed in drawing conclusions, what is known and isn’t known, with what certainty–at every level of the educational establishment. That includes the college and university programs responsible for preparing future teachers for their jobs; the educators who design curricula and the companies that publish them; the professional and labor organizations that represent teachers; and, finally, the state and local government agencies that fund the teacher education programs in public colleges and universities, establish the qualifications for being permitted to teach, operate the schools, and purchase, at great expense, the curricula, consumables, and support services. It also includes the many commercial enterprises and NGOs now serving the market for “evidence-based” materials consistent with “the science of reading.”3

The opposite is also true: few researchers have much expertise related to education–specifically, the conditions in schools and classrooms, expectations for teachers and students, the numerous other demands on teachers’ time, and other “real-world” considerations. In fact, there aren’t many cognitive science researchers who are even paying attention, which is also a big problem.

In short, stoking the fires created demand for curricula and other materials that incorporate “evidence-based” practices consistent with the “science of reading,” and for teacher education (professional development courses, workshops, symposia, conferences) on the topic. The solutions that are being offered must accommodate end-users (mainly teachers) whose unfamiliarity with this type of research limits what they can absorb. Teachers are not alone in this regard. Government education agencies are run by civil servants and elected officials (such as school board members) for whom familiarity with cognitive science related to reading is not in the job description. These people establish educational policies (e.g, qualifications to teach) and decide which instructional materials and approaches can be used. 

3. Starting points. Given teachers’ limited knowledge in the area, the SoR has had to start with a few basic findings and resources. These include the Simple View of Reading, the 5 recommendations from the National Reading Panel report, the 4-part processor, the Reading Rope, Ehri’s orthographic mapping theory, and a few others.4 My point here has three parts: First, these are excellent resources for the purpose of getting teachers (and others) on board and conveying some basic ideas from the research literature. Second, they are not adequate as the basis for making instructional decisions (about what to teach, when, how much, and so on), because the studies didn’t address these issues. Such decisions should and could be informed by other research–but that requires looking beyond these canonical studies. Third, this work is nonetheless being used to justify all manner of materials and practices, including ones that other research suggests are ill-advised.

In short, the narrow focus on a few classic studies is valid for one purpose: introducing some basic concepts to teachers and others who are just starting out. It’s understandable why teachers don’t know much about cognitive science related to reading: it wasn’t part of their professional training. (I hope it is obvious that I’m not blaming teachers here.) But it’s a mistake to extend this same beneficence to the people, organizations, and commercial enterprises that offer materials about the “science of reading” approach to instruction. They are supposed to be experts who are convey this material to teachers. However, they too have emphasized the classic studies as the justification for their work, filling in what’s missing with additional assumptions that aren’t supported by these particular studies and are called into question by other research.

That’s not right. It leads to wonky prescriptions for teaching phonemes, among other problems.

This point is so important I feel compelled to restate it. Teachers aren’t cognitive scientists. A PD course on research related to reading doesn’t make one a cognitive scientist, either, although it may be useful and informative. Teachers need to have enough background to be able to use curricula and other instructional materials that draw on the broader research literature–materials that have the relevant science baked in. Creating such materials requires knowing the research literature and being able to distill valid precepts from it.

This isn’t happening, and I think I know why: the people who are designing the materials, running the workshops, and selling the products aren’t cognitive scientists either. Like many others, they are learning as they go along. They too may be more comfortable with science at the level of the classic studies.

This conjecture explains another characteristic of the SoR. Let’s say a company’s goal is to create curricula and instructional software that is “evidence-based” and “consistent with the science of reading,” as stipulated in many states and LEAs. The people at the company read the classic studies carefully and use the findings and recommendations to guide product development. They explain the SVR, the Reading Rope, the 5 pillars of instruction on their web site. Therefore, they claim, their product is “evidence-based” and “consistent with the science of reading”.5

The reasoning isn’t valid, of course, because these classic studies are consistent with numerous practices. So, for example, from their review of the literature the NRP concluded that phonemic awareness was relevant to learning to reading and a teachable skill. By the flawed reasoning I’m referring to, any program/product that incorporates PA is consistent with this recommendation of the NRP, and thus “evidence-based” and in accord with the “the science of reading” (a phrase that doesn’t appear in the report). That includes teaching phonemes in the dark, teaching phonemes in a particular sequence to “mastery” levels of performance on specific tasks, and other questionable approaches whose efficacy is unknown.

Relying on classic studies sets the bar far too low for incorporating research into practice. In future posts I’ll discuss some ways to raise it.

Also coming up:

• If the ed schools don’t provide leadership, who does?

• Learning to read is only a little unnatural.

• Just because you can teach it doesn’t mean you should.

• Reading difficulties aren’t always (or even usually?) about reading

• Learning does not follow a scope and sequence

• Tenets for Teachers revisited

And other topics. I hope these observations are helpful in addition to provocative.

Footnotes

1. Gough and Hillinger, “Learning to read: An unnatural act.” 1980.

2. The term was introduced by Snowling and Hulme (2005), editors of “The Science of Reading: A Handbook;”  the second edition appeared in 2022.  

3. The corporations that contract to run public schools should probably also be included, but I have not looked into the reading curricula they use. 

4. Such as Share’s self-teaching hypothesis and the baseball study of the role of background knowledge. There is a canon of studies here, and as with canons of literature there can be debates about exactly which studies to include.

5. An example: “The Science of Reading refers to the pedagogy and practices proven by extensive research to effectively teach children how to read. To easily understand the complex combination of skills that result in reading fluency, you need two main frameworks: The Simple View of Reading and Scarborough’s Rope.“

Studying “Units of Study”

Lucy Calkins and her team have published the much-anticipated revisions to her popular K-2 reading curriculum. An EdWeek article asks, are the changes to the materials sufficient? A better question is, sufficient for what? 

1. Are the revisions sufficient to get the curriculum approved for adoption in states with “science of reading” laws that require instruction in areas that the previous version neglected or discounted? 

Probably.

I have not seen the revised curriculum yet. I did receive about 50 pages of what looked to be corrected proofs of the K and grade 2 curricula a few months ago, around the time of Dana Goldstein’s New York Times article. I couldn’t tell much from them because it wasn’t clear if they were final versions or how they were selected. But looking at other curricula that have been approved, it seems likely that the new materials will be acceptable because of where the bar has been set.

The lists of approved curricula vary by state, as summarized by Sarah Schwartz in another of her outstanding EdWeek articles. Most states are including comprehensive curricula such as Wonders (McGraw Hill) and Journeys (Houghton Mifflin Harcourt). These curricula are the legacy of Balanced Literacy, which held that neither phonics nor whole language worked for every child, and that effective instruction should incorporate elements of both. Teachers were encouraged to use the methods that worked best for them. Publishers responded with comprehensive curricula that cater to every taste. They include at least some phonics and sight word learning and other basic skills instruction; they also include ways to help children learn to read with minimal instruction of this sort. There’s far too much in these curricula to actually teach all of it. The teacher has to create a scope and sequence out of these blocks of marble. The teacher’s manuals are cluttered with options. For teachers who are strapped for time–and who isn’t?–this is an unreasonable burden. 

The fact that these curricula include basic skills instruction is apparently sufficient to get them on many of the approval lists. That doesn’t ensure that the material is used, let alone used effectively, because these same curricula support other approaches. The same issue arises with the new Calkins curriculum. The authors were compelled to incorporate specific kinds of basic skills instruction in order to check the regulatory boxes. If you are looking for keywords like “phonics” and “orthography,” they will be there. It is essential to examine what they have created, obviously. However, it is equally important to look at what teachers are told about using this material, in the teacher’s manuals, in the workshops and webinars on how to use the new materials, in the on-line support, on social media.* 

The revised materials may well satisfy the letter of the law(s). The further question is whether buying them is ethical. The authors and publisher are offering a replacement of unknown quality for the defective product they sold for many years. Allowing them to profit from the mess they created seems very wrong to me.

2. Are the changes sufficient to maintain the allegiance of the many teachers who have been adherents of Calkins’ approach?

Probably. 

Lucy Calkins is a guru, the longtime leader of a large community of teachers and other educators. From her books, videos, and social media presence, it’s apparent that she creates strong bonds with teachers by treating them as partners and taking their concerns personally. She speaks their language rather than the language of a reading researcher or educational bureaucrat. She clearly cares about teaching and about teachers. She is dedicated to them and many are dedicated to her. 

Calkins also gained a following by filling the gap that ed schools created by failing to prepare teachers adequately for the job. Many teachers learn about reading via materials offered by figures such as Calkins and Fountas & Pinnell, authors of another popular but misguided reading program. Teachers are grateful for the seemingly authoritative guidance these people provide. They join a supportive “community of learners” who share beliefs and goals. Ceding professional training to the producers of commercial materials is a fundamental mistake in my view.

Calkins has faced the tricky challenge of modifying her approach without alienating her base. Satisfying the states’ new adoption criteria required incorporating additional basic skills instruction, including phonics (which she had previously treated as a method of last resort), and removing discredited elements such as the “3-cueing method.” These modifications entail significant departures from views she had advocated. Calkins has been preparing her adherents for this shift via position pieces and social media activities. 

In November 2019 she circulated a manifesto “No one gets to own the science of reading,” which coupled a vigorous defense of her approach with an ugly attack on the “hype about phonics” and “phonics-centric” advocates of “phonics at the expense of everything else.” Teachers responded with heartfelt expressions of gratitude on social media for defending their beliefs and for challenging the scientists. I responded here. 

Calkins’ missive didn’t make reading research or concerns about her approach disappear, and states such as Arkansas began passing “science of reading” legislation. Calkins then pivoted to acknowledging that phonics and basic print skills are important and that her curriculum needed “rebalancing.” “Postcards From a Journey” (summer, 2020) announced what seemed to be a major shift in direction, which elicited heartfelt expressions of disappointment, confusion, and denial on social media. The outpouring was such that Calkins posted a long statement to her Facebook group addressing their concerns. 

Whereas “Postcards” described modifications of the curriculum significant enough to merit an announcement, the Facebook post said “What stays the same in our work with K-1 readers? 98% of it.” Reports that the revisions would be more extensive were fake news: “While the journalists will try to persuade you otherwise (controversy gets more eyes on the page than consensus), this is actually a small shift in our thinking….”

Now the final product has arrived. Is it a serious “rebalancing” or just a minor nod in the direction of “the science of reading” to satisfy outsiders’ demands? I look at these previous events and conclude that the answer will depend not only on what is in the curriculum itself but also on what is said to teachers about it. That will be important to monitor. My principal aim in writing this post is to raise awareness of this background and draw attention to the issues that have arisen since the revision was announced. 

3. Are the changes sufficient to enable teachers to teach reading more effectively, using materials and practices that incorporate modern research on how reading works and children learn?

This is the important question, of course. The answer is that we do not know. Much depends on the quality of the materials and how they are used. Having this group develop a curriculum that includes phonics and phonemic awareness may seem a little like asking a vegetarian to write the Complete Book of Meat, but they are smart, experienced people who were highly motivated to make this revision a success.

My main concern is probably obvious by now, that the authors may have only done as much as needed to retain access to a massive marketplace in which they have a lucrative share. Doing more than that would risk alienating more of the supporters who are critical to the product’s success as well as a source of revenue (via fees for workshops, webinars, and merchandise). Have they done a good job with material they would not otherwise have included? They could have. Is this material bolted on to their previous approach or an integral component of an improved approach? We’ll have to see. 

I recommend looking especially carefully at the treatment of phonics. Documents circulated in the run-up to the publication of the new materials (“No one gets to own,” “Notes from a Journey,” the Facebook group post; handouts from an October 2021 TCRWP workshop; the 50 pages from a version of the new curriculum) described an approach to phonics and other basic skills (such as memorizing “snap words,” i.e., sight words) that was seemingly patterned on the 3-cueing method. Here’s an example of “possible prompts for coaching readers” from a handout from the October 2021 workshop:

A 3-cueing approach to phonics inherits the problems with using the approach to identify words that I described in my 2017 book. This is not an efficient way to teach, given the properties of print and language, and how children learn. As in the original use of 3-cueing, the phonics document describes a scenario in which the reader engages in a series of successive approximations to the correct response. When an error is made they are given feedback in the form of a clue (or “cue”). The process repeats if additional errors are made. The inefficiency arises from the fact that learning occurs throughout this sequence, not only when the correct answer is finally produced. An alternative is to provide sufficient instruction and practice to minimize such errors and allow the student to benefit from simple and direct feedback when they occur. 

I will end the speculation here. Let’s see what Calkins and her crew have come up with.

Final disclaimer: I think that all of the big commercial curricula are part of the literacy problem. None of them were developed using modern cognitive science research as a foundation. Patching these products to meet the new state regulations is certainly better than nothing, but not as good as designing better curricula from the ground up. The existing ones are hard to use, and teachers work around them (using other materials) as well as with them. If it were up to me, I’d declare a moratorium on purchasing any of them. Invest the money in more direct support for teachers and students, such as coaches, reading specialists, tutors in the classroom, and professional development. I am not in charge; this isn’t going to happen any time soon, and so we are forced to decide which are the best of a flawed bunch.

——————————

* My question about how the materials will be used may be less of an issue in states that are also regulating the content areas that must be covered, such as the five identified in the National Reading Panel report. Such legislation raises other issues I will take up another time.

Back in the saddle again

Hi there. We are back. Did we miss anything? Molly and I have been recharging ourselves and the website, which is now focused on blog posts and a few other resources (talks, articles, our Reading Meetings). We have a lot of issues on our minds and hope to discuss them here. Consider this a (re)start.

If you’re reading this you probably believe, as I do, that literacy outcomes can be improved by using instructional practices and materials that incorporate findings from research on reading, language, learning, development and related topics. Many people are pursuing this goal under the rubric of “the science of reading.” 

As it has developed over the past several years, the First Principle in this approach is that learning to read requires instruction about print and the relations between print and oral language (speech).This idea has strong research support. The growing acceptance of this point is itself a big advance over what came before. Putting this insight into educational practice requires going further: deciding what to teach, when, how to teach it, for how long, taking into account differences among children (e.g., in language background, temperament), and environments (e.g., access to resources in home and community). Many people are laboring hard on these crucial issues. 

In the meantime, teachers have to teach. They need answers. This reality has resulted in the adoption of practices based on a few key ideas derived from reading research. For example, reading is treated as a set of teachable components as in the National Reading Panel report. Print skills and language skills are treated as distinct as in the Simple View of Reading. There is a lot of focus on instruction about spelling (orthography), sound (phonology), and the mappings between them (phonics). Then there is the issue of how children can learn enough “high frequency” words to get reading off the ground. Finally, assumptions are made about learning. For example, that components should be taught in sequence with mastery of one (e.g., all letter names) the prerequisite for moving to the next (e.g., letter-sound mappings). There is heavy reliance on explicit instruction and on assessments of student progress. 

There are many variants on these basic ideas, and others I won’t get into here. Details vary. The general approach is appealing to teachers because it addresses their questions about what to teach, when, how, and for how long. 

I view these practices as interim solutions. They can be justified on the basis of filling an immediate need; they can’t be justified on the basis of fidelity to reading research. They raise questions for me about efficacy (are the practices effective), efficiency (can they be implemented in reasonable time), and equity (do they work for all children regardless of background). There seems to be a misplaced emphasis on mastering components of reading rather than reading itself. 

In my view, we need to continue using research to identify questionable practices and develop alternatives that have as strong an empirical and theoretical basis as possible. This doesn’t require waiting until someone has a bolt of insight about what to do. It requires incorporating more of what has been learned from reading research. 

The great neglected topic is learning: How do children learn? What is the optimal balance between explicit instruction and implicit learning? Which types of knowledge are best learned by which procedures? Given the limits on instruction time and a clock that is ticking, how can children’s experiences be structured to facilitate rapid progress? How does learning change as the child develops? I’m not posing these questions to be philosophical. We know a great deal about these topics and can use that knowledge to create approaches that are more effective, more efficient, and more equitable. 

To be continued.

Coming soon to a screen near you

Some of you know that I gave a talk in Atlanta last week that created some, um, friction. It presented under poor conditions and wasn’t recorded properly. I have re-recorded it. It should be available via the Atlanta Speech School soon. Here are the slides. Update: talk is here.

I took advantage of the opportunity to make some revisions based on useful feedback. I agree that my comments about LETRS didn’t do justice either to LETRS or to the questions I wanted to raise. Through her research studies, Louisa Moats single-handedly raised awareness of the fact that most teachers lacked knowledge of basic properties of spoken language and print relevant to reading instruction. We know this is because of gaps in pre-professional training which continue to this day. She also acted on her findings, developing LETRS with Carol A. Tolman. Many in-service teachers have benefited from LETRS training and will continue to do so.

In place of those remarks about LETRS, I offer the following issues and questions that LETRS raised for me and perhaps others.

  1. No one method will work equally well for all teachers in all contexts. Having other approaches is a good thing. Here is a historical analogy that may be apt:

Time was, when people bought new computer software it came with a thick manual that was a comprehensive guide to using it. Everything you needed to know was in the manual, if you could find it. Some time in the 1990s, people working in HCI (human-computer interaction) realized that people didn’t want to read a manual and then start using the program; they wanted to start using the program as quickly as possible and consult the manual when necessary. This led to the  “minimal manual” approach: a shorter guide that provided the essential information needed to start using the program. This approach achieved a balance between explicitly telling users how the program worked and giving them a way to look up things as they were needed. 

The “minimal manual” approach might be adopted for PD related to reading. I think that is the philosophy behind Reading Simplified and Top Ten Tools. I am only noting a positive feature of these programs, not providing a global assessment of their value. 

  1. Are we conflating what teachers need to know to be able to teach with what children need to know to be able to read? Teachers should know about things like phonemes, onsets and rimes, inflectional and derivational morphology, relative clauses, collocations, and other basic components of language. “Know” here means being able to explain the concept and provide illustrations. Does a child need to know these concepts? I shudder when I see words like “phoneme” and “orthography” being used in teaching 6 year olds. This issue does not only arise with LETRS. 

  2. From material on social media and from my own discussions with teachers and people who teach teachers, it seems clear that many “science of reading” proponents ascribe to the following view or a close variant of it: Children need to learn about the components of spoken words in a sequence that starts with bigger units like words and ends with phonemes. Acquiring this knowledge is the prerequisite for learning about print and phonics. Phonemes are the minimal components of spoken words and phonemic awareness can be developed via spoken word activities. 

I do not think these assumptions are correct. I have explained several of the reasons in the lectures about phonemes, phonemic awareness and reading posted elsewhere on this website and in several talks, including one for the International Dyslexia Association (slides), and as the keynote speaker at the inaugural conference of The Reading League. Other reasons will be apparent from the principles/guidelines that Margaret Goldberg and I are developing. 

I have mainly focused on the common but mistaken idea that phonemes are properties of spoken words (they are an abstraction that depends on exposure to an alphabet), and that children need to demonstrate knowledge of all of them (44, 38, or some other number depending on your theory), prior to moving on to reading. I think it’s folly to devote precious time to teaching children the “correct” pronunciation of each phoneme. Since phonemes aren’t isolable segments of speech, whatever pronunciation we assign to, e.g. b is a bit of fakery, a useful fiction, but not how we talk.

I especially take issue with the idea that “phonemic awareness” refers to knowledge of the structure of spoken words independent of print. Evidence that the phonemic abstraction depends on exposure to an alphabet has been accumulating since the 1980s. The title of Read et al.’s breakthrough (1986) article (click to download) is “The ability to manipulate speech sounds depends on knowing alphabetic writing”. In the same year, Derwing et al. (Phonology Yearbook, 3) stated that their studies and others “go so far as to suggest that … the very ability to segment speech [into phonemes] may be a by-product of learning an alphabetic orthography.” The NRP (2000) review of PA also emphasized the role of print, but more important, the issues were greatly clarified by behavioral, computational, and neural research conducted over the subsequent 20 years (cf. my talks). 

The “phonemes first” approach seems to have coalesced out of several sources. LETRS K-3 Unit 2 is one of them. It says:

Although learning phonics requires phonemic awareness, the term phonics pertains to learning to read printed alphabetic symbols. Phonemic awareness activities, on the other hand, do not involve print.  They are listening and speaking activities; they can be done in the dark or with a blindfold on. (However, it is recommended that students watch the mouth of the speaker they are listening to.). p. 93

In other places the text includes statements about phonemic awareness that place greater emphasis on its connection to print. For example, this table about ages when children typically acquire knowledge of various phonological elements is very informative. I included it in the second talk here. The ability to use phoneme-level knowledge to perform tasks such as phoneme identification, deletion, blending and others typically develops relatively late, after the onset of reading instruction (and even longer after preschool children start learning letters, letter sounds, and letter names). The treatment of phonemes and phonemic awareness in LETRS is somewhat inconsistent and open to different interpretations regarding instruction.

The idea that children need to learn their phonemes before they learn to read is seen in other work that is prominent in education.  In “Equipped for Reading Success,” David Kilpatrick states that 

“PHONEME AWARENESS has to do with sounds in spoken words. It has nothing directly to do with letters. It is an awareness of the sounds in spoken language. It is a mental/linguistic skill.” p.15

A footnote on this entry says

“Phonological awareness is about understanding and being aware of the sounds we make when we say words. The minute you introduce letters, you have left the realm of phonological awareness and entered the realm of phonics.”

These observations (and others in the same chapter) are contradicted by research on how orthography, phonology, and the mappings between the codes develop. I’ll get into this in more detail in a separate post. Here I will only note that the assertions about the nature of phonemic awareness and phonics, and the boundary between them aren’t part of mainstream theories of learning to read. 

Finally, I should mention the Heggerty multi-year phonemic awareness curriculum that builds levels of phonemic knowledge through the use of PA tasks of increasing difficulty. 

In light of the extensive research evidence on the reciprocal development of orthographic and phonological knowledge from a very young age, I think these assertions and practices need to be carefully reconsidered. I am particularly concerned about the treatment in LETRS because it such a widely used resource.

  1. Finally, LETRS contains a version of this figure, the Tolman Hourglass, which is probably popular because it seems to convey reading science in an easily comprehended way. If the figure is meant as a literal representation of how the various types of knowledge come together to support reading, it is grossly incorrect. Among other things, it conveys the idea that orthography and phonology are separate domains, and that each of them consists of a strict hierarchy of discrete structures that somehow come together to create 1:1 (?) mappings between letters and sounds. It strongly suggests an instructional strategy of stepping through the levels, from the simpler ones to the harder ones, which is inefficient and unnecessary. The figure isn’t a good representation of the structure of language, the structure of print, or how children learn about this stuff on the way to reading and spelling. If it is merely meant as a mnemonic for all those terms, that could be done without the graphics.

I hope this post clarifies these concerns and the reasons I have mentioned LETRS in particular. Of course, this post settles nothing, but it could lead to fruitful discussions and positive changes. I will be pursuing the issues further in future posts and other media.

Don’t Stop Believin’

Recently people on the Internet have asked whether I “believe in the science of reading”.  

Funnily, I’ve never been asked this before.

I’ve devoted my professional career to using scientific methods to investigate reading, language, learning, and other questions. I made a conscious decision to use this approach to address questions that interest me. I consider this a rational choice not a matter of belief.  So, I engage in scientific research on reading and have a strong commitment to it, yes.

Am I committed to trying to use what we have learned from this research to improve how teachers are prepared for their jobs and how children are taught? Absolutely. I think it’s a really good idea. Much needed.

Do I support the “science of reading movement,” the efforts of a loosely-defined cohort of individuals and organizations to link reading research and educational practices?  Yes and no.

Yes, I support efforts to make the linkage and participate in them myself. No, I do not approve of all activities that are being undertaken under the rubric of the “science of reading”. My investment in using research to improve literacy outcomes is such that I will do what I can to identify policies and projects that seem ill-advised or ill-informed, given what I know about reading and the educational context.  And do what I can to help devise better approaches based on what researchers have found, taking into account the conditions that govern learning in the real world. That is certainly the more difficult task.  

I hope that answers that question.

I’m also being asked whether I (still) think that children require instruction in order to learn to read (yes) and whether I think phonics is important (yes). However, I differ from many people about exactly what to teach and how. 

At the moment, I am concerned about over-reliance on teaching children rules for things (among other issues). Some instruction of this sort is clearly necessary (e.g., so that children learn that print is a code that represents spoken words at a certain level of abstraction). However, many people seem to think that this is the only way that children learn and that if they aren’t explicitly taught everything that’s relevant to reading, they won’t learn it. Because, you know, reading is an unnatural act.

I think that’s wrong for many reasons. I’ve written about this already (this longer article, this short one in The Reading League journal) and will have some new material soon. Explicit instruction is important but it also contains the seeds of its own destruction. It enables the learner to begin learning with less and less, and ultimately no, reliance on explicit instruction, so-called “implicit” learning.  The two work together and the relation between them changes over time.  Explicit instruction at the outset “scaffolds” the implicit learning that does most (though not all) of the work in learning to read.

Think of vocabulary. You know tens of thousands of words. You were explicitly taught a tiny fraction of them. Somehow you were able to learn the rest (the statistical structure of language is the key).  Explicit instruction (which starts with infants being taught the names of people and things) is necessary to create the structures that will allow learning via the implicit mechanism. Same holds for learning about print. 

OK: I should really save this for a separate post. The main point is that learning a rule is a form of rote learning. The rule is an arbitrary thing to be memorized. If you learn the rule, there is a big payoff, generalization: The rule can be applied to new, novel patterns. But, learning a lot of rules is hard. It’s slow. It requires conscious attention and effort. So does using them, which is the antithesis of rapid, automatic, fluent word recognition. There are a lot of them in reading, spelling, and language, too. Finally, one of the biggest developments in the cognitive science of learning is that for many types of knowledge, which include things like orthotactics, phonotactics, phonics, and vocabulary, generalization doesn’t require rules. It can be done using knowledge acquired via implicit, statistical learning. For example, you learn about the spelling patterns that occur and do not occur and use that knowledge in processing novel words. No rules, which is good thing because who knows all the rules for combining letters?

As I’ve said elsewhere, the goal isn’t balanced literacy, it’s balanced learning. That means coupling the appropriate dose of explicit instruction at a given point in development to sufficient opportunities to expand what has been learned via the implicit, statistical learning mechanism.  

I’ve probably said both too much and too little.  I’ll return to “balanced learning” soon. It’s on the stack.