Home Articles Reader Opinion Editorial Book Reviews Discussion Writers Guide About TCRecord
transparent 13

Understanding Reading: A Psycholinguistic Analysis of Reading and Learning to Read, Sixth Edition

reviewed by Richard Allington - December 01, 2013

coverTitle: Understanding Reading: A Psycholinguistic Analysis of Reading and Learning to Read, Sixth Edition
Author(s): Frank Smith
Publisher: Routledge, New York
ISBN: 0415808294, Pages: 390, Year: 2011
Search for book at Amazon.com

Forty years ago, I received a copy of the first edition of Smith’s classic text, Understanding Reading. It was gift from a mentor, as I had just graduated with my Ph.D. degree. My mentor said something to the effect, “You are smart enough to read this and pay attention to the valuable points while also disregarding the stuff that simply isn’t true.” So, forty years ago I read Smith for the first time just as I was beginning my career as an academic. I have now just read the 6th edition of Understanding Reading as I near the end of my academic career. I considered bringing down the original edition from high upon one of my bookshelves to do a page-by-page, chapter-by-chapter comparison of the two editions.  In the end I left the original edition on the shelf, read the latest edition, and now offer a review.

First and foremost, I must begin by saying that Smith writes as clearly as anyone in the field, a lasting impression from the first edition. He writes largely without citations to the research in the main text but does provide research citations in the chapter “Notes” at the end of the book.  Notes is not a good title for this end matter if only because that section is nearly as many pages in length as the main part of the text.

I should also mention that Smith writes in the Preface that this edition is 25,000 words shorter than the previous edition and that over 200 new research citations have been added (and some 500 research citations eliminated). Given that the bibliography only lists 40 papers published since the year 2000, most of the new citations Smith mentions must be of older papers not cited in earlier editions.


If pressed to give Smith a label, I would offer anti-establishmentarian. Smith argues against almost every current educational practice, at least the practices associated with teaching children to read. For instance, Smith argues that teachers may make the task of reading incomprehensible by providing instruction focused on developing proficiencies at decomposing words into sounds or decomposing words into letters or attempting to relate letters to sounds. These activities, of course, are just the sorts of activities that the National Reading Panel indicated research supported.

Smith critiques that NRP report for suggesting that there are two types of research. The first is “scientific” research, randomized true experiments that support teaching the sorts of things noted above. The second type of research according to Smith was dubbed “unscientific” and by that the NRP meant all other research. He also notes that the NRP accomplished the political work required while also supporting both entrepreneurism and teacher bashing.  While Smith is hardly the first academic to criticize the NRP and its report (c.f., Allington, 2002), he does remain one of the very few who reject the report in its totality.

Smith argues for “autonomous” teachers who make daily decisions about how best to support children in their attempts to become literate. He is a big fan of both intentionality and interests as central to learning to read. However, beyond that Smith offers few clues as to how he might expect classroom opportunities to learn to read to be developed and sustained.

If there is an obvious limitation of Smith’s argument it is that one walks away after reading this book thinking, he’s right, but wondering what his arguments mean for teachers, schools, and schooling. If we toss out all of the reading programs and then take the doors off the school library is it likely that children will find the books that interest them and begin to read them, pausing only when they need a bit of help from a peer or an adult? The reason Smith seems so unhelpful in this case is his anti-instruction stance. He argues that reading cannot be taught and that most attempts schools provide to help children learn to read are actually more likely to hinder children in the acquisition of reading proficiency. An anti-establishmentarian.


Smith offers a top-down model of reading and reading development. He argues that good readers do not look at every word much less every letter in every word. Such close inspection is not needed according to Smith because of the redundancy in language that offers overwhelming cues to what the text means. Understanding what a text says, not reading a text accurately, is Smith’s measure of reading development.

What Smith rejects is the compelling evidence others have provided about the role that phonological sensitivity plays in early reading acquisition. He doesn’t just reject that evidence but also argues that attempting to foster the development of phonological sensitivity underlies the failure of many, if not most, children who struggle with learning to read.

Likewise, attempting to develop even basic letter-sound relationships is tagged as another primary reason why some children struggle with learning to read. One can be sympathetic with Smith on this issue because there is so much misinformation around when the issue of developing decoding proficiencies comes up. There are also multiple entrepreneurs selling packages that claim to be the solution for developing decoding proficiencies. The fact that no research evidence exists supporting such claims supports Smith’s various assertions about decoding instruction. At the same time there is evidence that developing phonological sensitivity and decoding proficiencies does, in fact, advance reading development but none of this evidence is even introduced by Smith much less critiqued.

In other words, Smith largely cherry picks the evidence he cites such that only evidence that supports his argument is provided. However, that doesn’t mean that no evidence contrary to his arguments is available. This seems part of what has made Smith such an easy target for folks on the other side of his arguments, the folks who argue that there are sub-skills that must be taught including phonological sensitivity and letter-sound relationships.


While in my view Smith overstates much, I do believe in his emphasis that much of what is learned when learning to read is learned by reading.  While Smith doesn’t review studies demonstrating the power of invented writing in fostering phonological sensitivity and letter-sound relationships that evidence is available if generally ignored (Clarke, 1988). He doesn’t review studies of self-teaching, of how early readers learn much that is useful about decoding and word recognition simply from engaging in reading and writing. Smith doesn’t review the research, both correlational and experimental research, on the powerful role volume of reading plays in the development of proficient readers even though extensive independent reading would seem central to his top-down model.

It was this research that led the committee of the National Research Council of the National Academy of Education  (Snow, Burns, & Griffin, 1998) to argue for considering both the top-down view of reading acquisition that Smith advocates as well as the bottom-up view of reading acquisition that Smith argues so vociferously against.  As it turned out proponents of the bottom-up view created the NRP to selectively cite the evidence for their point of view (Cunningham, 2001; Pressley, Duke, & Boling, 2004). That point of view then became established in the framework for the Reading First program of the No Child Left Behind legislation.

What was left out of the NRP report was any suggestion that volume of reading or even independent reading had any research-based role to play in developing proficient readers. Perhaps intended as a shot to the heart of the ideas that Smith advocates, the widely distributed ‘plain language’ version of the NRP report (Armbruster, Lehr, & Osborn, 2001) argued against having children spending school time reading independently: "Rather than allocating instructional time for independent reading in the classroom, encourage your students to read more outside of school" (p. 29).

Thus, the intended message was that instructional time was too important to waste by having children just engaged in independent reading. This message went further than the NRP report had but it was also far more widely read than the NRP report. Thus, the idea—extensive independent reading—central to Smith’s top-down view of reading development, was crushed.

Given both the wide evidence, although little of that evidence is experimental, indicating the powerful role that engaged independent reading plays in reading development the crushing of independent reading as an important, if not critical, factor in reading development was accomplished in much the same way Smith “crushes” the evidence on phonological sensitivity and developing decoding proficiencies—by ignoring or only selectively reviewing the research available. The difference, however, was that the crushing of the evidence on independent reading came at the behest of the federal government and, thus, the crushing was widespread and potent.

It may be that the primary effect of largely eliminating independent reading from the school day while emphasizing the importance of decoding instruction was that the federal evaluation of the Reading First program found no positive effects on reading achievement for schools participating in the Reading First initiative (Gamse, Jacob, Horst, Boulay & Unlu, 2009). I should note that children in Reading First schools could read nonsense words faster and more accurately than children in other schools but that small gain did not improve their performances on standardized measures of reading achievement. Smith might have used this large-scale evaluation of the Reading First program to support his arguments on the futility of focusing on sounds and letter-sound relationships. But he didn’t.


Smith presents a very readable argument for his top-down model of reading and reading development. Unfortunately, while immensely readable it is also largely wrong about how to best develop readers. Smith’s emphasis on the power of self-teaching and the power of wide and independent reading are correct. It is his insistence that letters and sounds and even individual words are unimportant that is wrong.

As any number of other authors have argued, decoding lessons only make sense to children if they want to learn to read. Then only fairly small allocations of instructional time (6 to 10 minutes a day) to decoding instruction are necessary and only necessary in the beginning stages (kindergarten and first grade) for developing readers. As important as these aspects of beginning reader might be, the importance of wide and successful reading experiences cannot be underestimated, although in current school practice they usually are.

I’ll close with much the same advice I was given 40 years ago when given a copy of Smith’s book. Read it for the interesting and entertaining writing about the reading process. Read it for what he has to say about the power of making sense while reading and of independent reading activity, but largely ignore the points he makes about the pointlessness of phonological sensitivity and decoding lessons.


Allington, R. L. (2002). Big brother and the national reading curriculum: How ideology trumped evidence. Portsmouth, NH: Heinemann.

Armbruster, B., Lehr, F., & Osborn, J. (2001). Put Reading First. Washington, DC: National Institute for Literacy.

Clarke, L. K. (1988). Invented versus traditional spelling in first graders' writing: Effects on learning to spell and read. Research in the Teaching of English, 22, 281-309.

Cunningham, J. W. (2001). The National Reading Panel report. Reading Research Quarterly, 30(3), 326-335.

Gamse, B. C., Jacob, R. T., Horst, M., Boulay, B., & Unlu, F. (2009). Reading First Impact Study: Final Report (NCEE 2009-4038) (No. (NCEE 2009-4038)). Washington, DC: National Center for Education Evaluation and Regional Assistance, Institute of  Education Sciences, U.S. Department of Education.

Pressley, M., Duke, N. K., & Boling, E. C. (2004). The educational science and scientifically based instruction we need: Lessons from reading research and policymaking. Harvard Educational Review, 74(1), 30-61.

Snow, C. E., Burns, M. S., & Griffin, P. (1998). Preventing reading difficulties in young children: A report of the National Research Council. Washington, DC: National Academy Press.

Cite This Article as: Teachers College Record, Date Published: December 01, 2013
https://www.tcrecord.org ID Number: 17336, Date Accessed: 5/20/2022 1:04:06 AM

Purchase Reprint Rights for this article or review
Article Tools
Related Articles

Related Discussion
Post a Comment | Read All

About the Author
  • Richard Allington
    University of Tennessee
    E-mail Author
    RICHARD ALLINGTON is a Professor of Literacy Studies at the University of Tennessee. Recent publications include; Handbook of Reading Disability Research, Routledge (co-edited with Anne McGill-Franzen), Summer Reading: Closing the rich/poor reading achievement gap, Teachers College Press (co-edited with Anne McGill-Franzen) and What Really Matters for Struggling Readers (Pearson). He has served as president of the Literacy Research Association and the International Reading Association. Current work includes developing a research summary of the role of text complexity in learning to read.
Member Center
In Print
This Month's Issue