PreprintPDF Available
Preprints and early-stage research may not have been peer reviewed yet.

Abstract

Can the science of reading contribute to improving educational practices, allowing more children to become skilled readers? Much has been learned about the behavioral and brain bases of reading, how children learn to read, and factors that contribute to low literacy. The potential to use research findings to improve literacy outcomes is substantial but remains largely unrealized. The lack of improvement in literacy levels, especially among children who face other challenges such as poverty, has led to new pressure to incorporate the “science of reading” in curricula, instructional practices, and teacher education. In the interest of promoting these efforts, we discuss three issues that could undermine them: the need for additional translational research linking reading science to classroom activities; the oversimplified way the science is sometimes represented in the educational context; the fact that theories of reading have become more complex and less intuitive as the field has progressed. Addressing these concerns may allow reading science to be used more effectively and achieve greater acceptance among educators.
1
Reading Research Quarterly, 0(0)
pp. 1–12 | doi:10.1002/rrq.341
© 2020 International Literacy Association.
ABSTRACT
Can the science of reading contribute to improving educational practices,
allowing more students to become skilled readers? Much has been learned
about the behavioral and brain bases of reading, how students learn to read,
and factors that contribute to low literacy. The potential to use research
findings to improve literacy outcomes is substantial but remains largely unre-
alized. The lack of improvement in literacy levels, especially among students
who face other challenges such as poverty, has led to new pressure to incor-
porate the science of reading in curricula, instructional practices, and teach-
er education. In the interest of promoting these efforts, the authors discuss
three issues that could undermine them: the need for additional translational
research linking reading science to classroom activities, the oversimplified
way that the science is sometimes represented in the educational context,
and the fact that theories of reading have become more complex and less
intuitive as the field has progressed. Addressing these concerns may allow
reading science to be used more effectively and achieve greater acceptance
among educators.
Reading is a remarkably complex activity involving most of our
mental and neural capacities. As such it has been the focus of a
massive amount of research by scientists from numerous disci-
plines who study human behavior and its brain bases. This interdisci-
plinary body of research constitutes what is sometimes called the
science of reading (for reviews, see Castles, Rastle, & Nation, 2018;
Rayner, Foorman, Perfetti, Pesetsky, & Seidenberg, 2001; Seidenberg,
2017; Snowling & Hulme, 2005). Many scientists who conduct this
research have long believed that it could be used to improve educational
practices and literacy outcomes (e.g., Adams, 1990; Stanovich & Stan-
ovich, 2003). That would be valuable, given persistently low literacy lev-
els in the United States and other countries, especially among groups for
whom factors such as poverty create many additional obstacles (Rear-
don, 2013; Snow, Burns, & Griffin, 1998). Previous efforts to connect
this research and educational practice have failed for a variety of rea-
sons (Seidenberg, 2017). The lack of improvement in literacy outcomes
over many years has led to new pressure to incorporate the science of
reading in curricula, instructional practices, and teacher education
(Gewertz, 2020; Hurford, 2020). The pursuit of legislative remedies for
low reading achievement in nearly every state (Davis Dyslexia Asso-
ciation International, 2020; National Conference of State Legislatures,
2019) is indicative of frustration over the lack of progress in addressing
well-founded concerns.
Mark S. Seidenberg
Matt Cooper Borkenhagen
University of Wisconsin–Madison, USA
Devin M. Kearns
University of Connecticut, Storrs, USA
Lost in Translation?
Challenges in Connecting Reading
Science and Educational Practice
2 | Reading Research Quarterly, 0(0)
These actions have revived long-standing disagree-
ments about the causes of low literacy and how to address
them. The arguments are distressingly familiar from the
“Reading Wars” (for varied accounts, see Gunther &
Lindstrom, 2003; J.S. Kim, 2008; Lemann, 1997). According
to Seidenberg (2017), disagreements about reading educa-
tion are a manifestation of a disconnection between the
cultures of science and education, dating from the creation
of U.S. schools of education in the early 20th century.
Research on cognitive, linguistic, social, and emotional
development that is highly relevant to education has
been only fitfully incorporated in programs for teachers,
curriculum developers, administrators, and policy experts.
Educators working with scientists of an earlier era devel-
oped approaches to reading instruction based on assump-
tions that were falsified by extensive research, but these
findings have had little impact on what teachers are taught,
and widely used instructional materials continue to incor-
porate them (for discussion of one example, see Seidenberg,
2019).
The fact that the same conflicts have persisted under
different names (skills vs. literacy, phonics vs. whole lan-
guage, and phonics vs. balanced literacy) while literacy
levels have been stagnant indicates that a different ap -
proach is needed. Concerns about reading instruction and
teacher preparedness have been amplified via social media,
advocacy groups, books (e.g., Goldstein, 2015; Seidenberg,
2017), and investigative journalism (Hanford, 2018), cre-
ating opportunity for change. Several states have initiated
reforms centered on increasing teachers’ familiarity with
the science of reading, mandating the use of instructional
practices that are consistent with it. Such efforts are gain-
ing momentum (Gewertz, 2020; Goldstein, 2020).
Renewed interest in using reading research to improve
practices is a welcome development. The potential benefits
are substantial but remain largely untapped. The research
base is extensive. Yet debates about connecting science and
practice have hardly changed (compare articles in this spe-
cial issue of Reading Research Quarterly with Stanovich &
Stanovich, 2003, and J.S. Kim, 2008). Education is an
enormous enterprise with numerous stakeholders whose in -
terests often conflict: government, academia, business, vot-
ers, tax payers, teachers, advocacy groups, families, students—
and reading researchers. Change is exceedingly difficult to
accomplish.
Many observers (e.g., Blaunstein & Lyon, 2006; Steiner
& Rozen, 2004) have criticized the educational establish-
ment, focusing on the schools of education that provide
professional training for teachers and administrators and
are the homes for experts in curriculum and instruction,
policy, and other areas. The schools are not all alike: they
contain numerous departments that represent different
fields, and individuals’ views certainly vary greatly. Reading
science is conducted by some researchers in schools of
education. Historically, however, they have deflected the
influence of such science in teacher education, the devel-
opment of curricula and practices, and educational phi-
losophy, rationalizing why it lacks relevance and placing
greater emphasis on a canon of accepted findings from
earlier eras (Seidenberg, 2017). Scientific literacy—familiarity
with core research findings; the ability to critically assess
the quality of a research study, the validity of the conclu-
sions, and their relation to other findings—is still not
strongly emphasized in professional training, leaving
practitioners susceptible to discredited or unsupported
claims (e.g., the persistence of neuromyths; Dekker, Lee,
Howard-Jones, & Jolles, 2012). Findings are cherry-picked
from the vast literature to support personal beliefs and sell
products.
Many educators have rejected the premise that their
policies and practices are a major factor in poor reading
achievement. Ravitch’s (2011) argument that poor educa-
tional outcomes are due to external factors, principally
poverty and government interference in her view, was
enormously influential. It successfully deflected attention
away from improving quality of education for the children
for whom it matters most, it ignored the ways that educa-
tional practices magnify the impact of income inequality,
and it wrongly implied that low literacy is limited to people
in poverty (Seidenberg, 2017). Still, relative to poverty and
government policy, using research to improve outcomes
seems almost inconsequential. Similarly, the invention of
“balanced literacy” successfully diffused the reading wars
at their peak in the early 2000s without addressing the
underlying issues. Declarations about the relevance of
phonics by organizations historically opposed to it (e.g.,
International Literacy Association, 2019) could have a sim-
ilar effect unless coupled with actions that change policies
and practices. The pedagogical status quo is also sustained
via a closed loop that includes educational authorities (aca-
demia), government (local, state, and federal officials who
control budgets and policies), and educational publishing
and technology corporations (producers of instructional
materials). Many such authorities work closely with state
departments of education and create products for the vast
education market.
We do not wish to minimize the importance of these
conditions, which create real obstacles that demand con-
tinued attention with the goal of achieving significant
reforms. However, acknowledging other conditions affect -
ing educational outcomes does not obviate the need to
examine educational quality, which also has a strong
impact, especially for students subject to other risk factors
(Aikens & Barbarin, 2008). If the science of reading can
improve students’ learning and literacy, we need to use it,
other factors notwithstanding.
Our goal in this article is to examine ways to make
better use of science to improve outcomes, at a time when
interest in the possibility is growing. We are concerned
about uses of reading research that could undermine
Lost in Translation? Challenges in Connecting Reading Science and Educational Practice | 3
well-intentioned attempts to bring it to bear on pedagogy.
The main products of this science are findings—systematic
data about phenomena—and, more important, theories
that are our best explanations for such findings. In read-
ing, we have numerous theories because it is a complex
behavior, the product of multiple skills and capacities;
because reading is not a uniform activity but rather varies
depending on purpose, skill, type of material, and context;
and because it can be viewed from multiple intersecting
perspectives (e.g., biological, behavioral, social, develop-
mental, cross-cultural).
A theory of how students gain reading skills should
(minimally) address what, how, when, and for whom.
The what component is a characterization of the types of
knowledge and mental operations (processes) relevant to
tasks such as reading aloud and comprehending stories.
The how part is a characterization of how the what is
learned. The goal is a mechanistic account of how a
learner gets from point A (e.g., the student cannot yet
read) to point B (the student achieves escape velocity:
basic skills that can develop further without much addi-
tional instruction about them). The when part refers to
the fact that reading, like other acquired forms of exper-
tise (e.g., gymnastics, mathematics), develops over an
extended period of time. The nature of the skill demands
that elements be introduced over time. So does the nature
of the child: Capacities to learn change with development;
what a child is able to learn also depends on the current
state of their knowledge, which changes as they progress.
For whom refers to individual differences among children
that also determine answers to the other questions. For
example, a child who is a native speaker of a different lan-
guage or dialect than the one used in school has different
needs than a child who already speaks it.
Every teacher acts on the basis of a tacit theory of
what, how, when, and for whom, based on what they have
been taught, learned from peers, and discovered from
experience. The curricula they employ also instantiate
assumptions in each of these areas. Incorporating reading
science is valuable because it adds a vast amount to what
is known about how reading works and how children
learn, beyond what can be established by other means.
We have three concerns about current efforts to use
this science to improve reading outcomes. First, there is a
need for additional translational research to establish
closer connections between theory and practice. We know
more about the science of reading than about the science
of teaching based on the science of reading. Second, we
are concerned about how reading science has been char-
acterized in educational contexts: It can be oversimplified
in ways that slow progress by seeming to sanction prac-
tices that are only loosely connected to it. Finally, the sci-
ence of reading is a moving target because it continues to
progress. Theories have grown increasingly complex and
nonintuitive, creating additional translational challenges.
We raise these concerns because the extensive body of
research about reading may be used more effectively, and
achieve greater acceptance, if they are addressed.
Lost in Translation?
Reading science does not come with educational prescrip-
tions attached. Science is one kind of thing (empirical find-
ings and explanatory theories), and educational practice is
another (activities that promote learning in real-world set-
tings). Connecting the two is the function of translational
research. Given what is known about howreading works
and students learn, what should be taught,when, and how?
Which approaches are effective? Forwhich students from
which backgrounds and socio economic circumstances?
Much has been learned from studies that used scientific
theories and methods to investigatecomponents of effec-
tive reading instruction (e.g., Vellutino, Tunmer, Jaccard, &
Chen, 2007), devise effective interventions (e.g., McGinty,
Breit-Smith, Fan, Justice, & Kaderavek, 2011; Morris et al.,
2012), and identify factors that predict reading outcomes
(e.g., Y.-S. Kim, Petscher, Schatschneider, & Foorman,
2010). Our concern is that although reading science is
highly relevant to learning in the classroom setting, it does
not yet speak to what to teach, when, how, and for whom at
a level that is useful for teachers.
To illustrate, consider research on the effectiveness of
instruction that focuses on increasing students’ knowl-
edge of lexical phonology. Beginning readers who are pro-
gressing more rapidly exhibit better knowledge of the
phonological properties of words, as measured by phono-
logical awareness tasks such as deciding whether two
words rhyme, indicating the number of syllables in words,
and deciding whether two words end with the same sound
or contain the same vowel (Castles et al., 2018). We know
why: Reading depends on speech. Students do not relearn
language when they learn to read; they learn to relate the
printed code to existing knowledge of spoken language.
Writing systems are codes for representing spoken lan-
guage (Seidenberg, 2017). The structure of spoken words
in English—the fact that they consist of sequences of pho-
nemes, syllables, and morphemes that are associated with
meaning—is reflected in their alphabetic representations.
Learning about the written code is easier forstudents who
know more about characteristics of spoken words that it
represents. Individual differences in knowledge of such
properties of spoken language at the start of formal
instruction have an enormous impact onstudents’prog-
ress (Hulme, Nash, Gooch, Lervåg, & Snowling, 2015).
The translational question, however, is what to teach.
For example, is it effective to focus instruction on building
phonological awareness? Interventions of this sort have
yielded very mixed results (Bus & van IJzendoorn, 1999).
Sometimes improvement on the specific tasks that were
4 | Reading Research Quarterly, 0(0)
the focus of instruction does not carry over to other tasks,
such as reading comprehension (Blachman, 1997).
The picture changes if we consider the impact of such
instruction in conjunction with other activities. Many stud-
ies have indicated that phonological awareness instruction
is more effective when linked to instruction about print and
meaning (e.g., Ball & Blachman, 1991; Bowyer-Crane et al.,
2008; Byrne & Fielding-Barnsley, 1989; Cunningham, 1990;
Gillon, 2000; Hatcher, Hulme, & Ellis, 1994; Schuele &
Boudreau, 2008; Stuebing, Barth, Cirino, Francis, & Fletcher,
2008). Theories of reading can easily explain these results.
The goal is gaining proficiency in reading (i.e., compre-
hending words and texts). Reading comprehension is facili-
tated by using print to access existing knowledge of spoken
language. The development of phonological representations
of words relevant to reading depends on one’s experience
with both spoken language and print. Thus, phonologically
focused instruction is more effective when linked to knowl-
edge of print in the service of reading for meaning.
That is good science, but what are the implications for
instruction? Keeping in mind that the research in question
concerns students, not preschoolers, the implications are
something like these: Avoid teaching phonological aware-
ness in isolation; emphasize connections among spelling,
sound, and meaning; and link these activities to actual
reading, the development of which is the instructional
goal. Guidelines of this sort are useful. They might influ-
ence how teachers construe and pursue their instructional
goals, but they do not speak to how to accomplish these
goals. A teacher is more likely to seek that information
from Pinterest and Teachers Pay Teachers (https://www.
teach erspa yteac hers.com/). A lot of reading research has
this character. The science is excellent; it is how we have
learned so much about how reading works. Practitioners
should know about it. Yet, there is a need to go the final
translational mile to impact practice.
In short, one reason the science has not gotten into
classrooms is because it has not provided sufficient guid-
ance about what to do there. It is not only that cognitive
science is not a part of teacher education. If it were clear
to teachers how such science could improve their effec-
tiveness and their students’ progress, teachers would
clamor for it. Some already do.
The imbalance between basic and translational research
has created other problems. Consider phonics, for example.
Phonics is not an important concept in theories of reading.
Behavioral and brain evidence show that for skilled readers,
orthography and phonology become deeply integrated
(Seidenberg, 2017). For struggling readers, orthography
and phonology are more weakly connected (Shankweiler
et al., 2008). The obvious implication is that among other
activities, early reading instruction should include ones that
facilitate acquiring knowledge of the correspondences
between print and sound—phonics.
Phonics is a translational issue. There has been re -
search relevant to developing effective phonics instruction,
demonstrating the advantage of direct instruction over
indirect methods, for example (e.g., Foorman, Francis,
Fletcher, Schatschneider, & Mehta, 1998; Stuebing et al.,
2008). However, the research literature does not provide
detailed guidance about which spelling–sound patterns to
teach; how many to teach; whether patterns should be
taught in isolation, such as all the pronunciations of the
vowel o or in disambiguating contexts (e.g., words such as
cot, cold, cost, doll, off); or other issues that have to be adju-
dicated for instruction to proceed. The market is filled
with phonics curricula that fill the translational gap but
vary greatly in assumptions about what to teach, when, and
how, and thus are unlikely to be equally effective. Programs
are motivated by science—students need to learn these
mappings, which requires instruction, ergo phonics—but
research has not validated specific solutions. Yet, that is
what educators ask us: Which program does reading
science say we should use?
In the absence of sufficient translational research,
almost every reading curriculum can claim an equally
loose connection to the “science of reading.” The risk, of
course, is that such programs will prove ineffective, not
because the basic science is wrong but because the transla-
tion was poor. It has happened before. The unmet chal-
lenges involved in teaching phonics effectively, in the
service of literacy while maintaining student motivation
and interest, led influential educators (e.g., Clay, 2001;
Goodman, 1989; Krashen, 2002; Smith, 1999) to conclude
that beginning readers would be better off without it,
which was a profound mistake. The educational challenges
have no bearing on the validity of the science; being hard
to teach does not change how reading works or what stu-
dents need to learn. Worse, the alternative approach they
developed—using contextual cues and strategies to guess
words while discouraging the use of phonology by mini-
mizing phonics instruction—makes it harder to learn to
recognize words quickly and accurately, a basic ingredient
of reading skill (Seidenberg, 2017).
It might be thought that the science-to-practice trans-
lation would be achieved via the educational publishing
industry that produces curricula and other materials for
teachers. Popular curricula (e.g., Reading Wonders: McGraw-
Hill Education, 2014; Journeys: Houghton Mifflin Harcourt,
2014; Units of Study in Phonics: Calkins, 2019; Fountas &
Pinnell Literacy Continuum: Fountas & Pinnell, 2016) were
produced by teams of experts in education and science.
Determining how science can be incorporated in such
materials is presumably one of the tasks of such teams.
Commercial curricula do not accomplish this because
they are compromised by the need to appeal to a broad mar-
ket and to local authorities caught up in debates about best
practices. The texts instantiate the balanced literacy idea that
Lost in Translation? Challenges in Connecting Reading Science and Educational Practice | 5
there is no single way to teach reading and that teachers
should be free to use elements from different approaches.
Rather than offering a best-practices curriculum based on
an effective translation of research to practice, teachers are
left to construct their own. Their choices are likely to depend
on their beliefs about reading and abilities to teach different
kinds of material, not research-based recommendations
about what to teach, how, when, and for whom.
The Science of Reading
in the Educational Context
The science of reading that is the focus of legislation, in-
service teacher training, and other educational reform is a
simplified version of reading research. In these contexts,
the science of reading concerns a relatively small number of
key ideas and findings: the alphabetic principle (Liberman,
Shankweiler, & Liberman, 1989), the simple view of read-
ing (Gough & Tunmer, 1986), the four-part processor
(Moats & Tolman, 2009), stages in reading development
(Ehri, 2005), the five components of reading from the
National Reading Panel (NRP) report (National Institute of
Child Health and Human Development [NICHD], 2000),
and a few others. Simplification is necessary to make
research more accessible to teachers and other parties. It is a
reasonable place to start, especially because the ideas are
important yet still not universally known or accepted.
However, reading science is an active, ongoing endeavor,
not a canon of findings. Overreliance on simplified
accounts of this research risks reifying it into precepts that
do not incorporate much of what the science has to offer.
To illustrate, we turn to the NRP report (NICHD,
2000). As everyone knows, the report was “an evidence-
based assessment of the scientific research literature on
reading and its implications for reading instruction.” The
panel identified five components for which instruction
had been shown to be effective: phonemic awareness, pho-
nics, fluency, vocabulary, and comprehension. The report
has been discussed ad nauseum. It was a valuable review
that served several functions, including simply drawing
attention to the existence of a body of science relevant to
development and instruction. The reviews of evidence
concerning phonemic awareness and phonics were signifi-
cant at the time given how negatively they were viewed in
the whole language approach. The report was also widely
critiqued (see Allington, 2002; Shanahan, 2004).
For our purposes here, the main point is that the
report was not a sufficient basis for designing an effective
reading curriculum, but that is how it is frequently
taken—today. The report mentioned but did not evaluate
methods of instruction in each area. The panel did not
evaluate existing curricula or describe the structure of an
ideal curriculum based on their findings. That was not their
charge. The report identified several targets for in struction
supported by empirical research, but left open numerous
questions about what to teach, when, how, and for whom,
as the authors acknowledged.
Why are these observations relevant some 20 years
after the report was published? Because in the context of
the science of reading and education, it is often taken as
having established the scientific basis for early instruc-
tion. The five components have been codified as the “five
pillars of instruction” that reading curricula should incor-
porate (Cassidy, Valadez, & Garrett, 2010; J.S. Kim, 2008;
McCardle & Chhabra, 2005). It does not detract from the
historical importance of the report to note that it is not
suitable for this purpose. Leaving aside the report’s lim-
ited scope, the five components are not the same kinds
of things. Phonemic awareness is a type of knowledge.
Phonics is a type of instruction about correspondences
between spelling and sound. Fluency is a characteristic
of skilled reading. Vocabulary is a primary component
of language, and comprehension—well, that is the goal.
Whereas phonemic awareness is a very specific type of
knowledge, vocabulary and comprehension are broad
categories subsuming numerous types of information and
mental operations, including ones that are not specific to
reading (e.g., knowledge of what happens in a restaurant,
making inferences about people’s beliefs and intentions).
There is a further problem if these components of
reading are treated independently, as in the NRP report
(they were investigated by subgroups that wrote separate
reports; NRP, 2000) and in curricula based on them. This
is where the five-pillars approach seriously departs from
reading research. The science addresses types of informa-
tion and processes involved in reading and how they
develop. What is missing from the list of components is a
developmental account of how they are learned, informa-
tion crucial for instruction. Researchers have studied
these issues extensively. In fact, the components are highly
interdependent. Phonemic awareness, the ability to treat
words as consisting of discrete sound segments, is the out-
come of a developmental process that begins with learn-
ing a spoken language and is finished by exposure to print
(Bertelson & De Gelder, 1989). This process is affected by
vocabulary size: Young learners begin discovering the
internal structure of words via the overlap among them,
which depends on the range of words they know (Metsala
& Walley, 1998). Phonics, as it refers to learning spelling–
sound correspondences, depends on not only phonemic
awareness but also vocabulary, which allows learners to
determine whether the way they pronounce a letter string
matches a known word (Share, 1995). Vocabulary size and
quality (Perfetti & Hart, 2002) affect comprehension, but
comprehension is also a source of vocabulary learning
(Landauer & Dumais, 1997). If these parts come together,
learners gain fluency in identifying and understanding
words and texts, and if they are fluent, learners can read
more and learn more from what they read, about orthography,
6 | Reading Research Quarterly, 0(0)
phonology, morphology, vocabulary, grammar, the con-
nections among these types of knowledge, the ways that
language is used to communicate, and the things we
communicate about. In short, the components interact
(Rumelhart, 1977). Skilled reading is possible because of
the dependencies among these types of knowledge. Learning
to read is possible because learning about one affects what
is learned about others. Instruction based on these aspects
of reading science would have a different character than
practices based on separate components.
Our concern is not about the aged NRP report but
about the way it is being used. It is not a good overview of
the science of reading (too much omitted or out of date)
but has been taken as such. It is not a sufficient basis for
developing a curriculum but it has been taken as such. In
extreme cases that we have observed, first-grade reading
instruction consists of blocks of time spent on each of the
components, in the order they were presented in the
report. The focus on the components leaves little time for
reading and talking about books.
We have used the NRP report as our example, but the
same questions can be raised about the use of other classic
research. It is not that the ideas (e.g., the simple view of
reading, the alphabetic principle) are wrong or unimportant;
to the contrary, they are essential and need to be widely
assimilated. Rather, they are incomplete, especially with
regard to learning; they do not address individual differ-
ences adequately; and they do not include important ideas
and findings that came later, many of which they stimu-
lated. Together with insufficient translational research,
overreliance on canonical studies leaves the door open to
varied practices that reading science has not sanctioned.
The Science of Statistical
Learning in Quasiregular Domains
We have discussed the need to bring reading science into
closer contact with how learning occurs in educational
settings. We observed that understanding what needs to
be learned and how it is learned is a prerequisite for iden-
tifying effective practices and that the science of reading
is an active endeavor, not a canon of findings. Our final
concern centers on the challenges involved in making use
of research that has become highly technical and theories
that may not agree with intuition.
As sciences advance, the phenomena studied become
increasingly remote from everyday experience. Instruments
such as telescopes and microscopes have led to the discov-
ery of phenomena (e.g., galaxies, molecules) that would not
otherwise be known to exist. Advances in methods for ana-
lyzing data have revealed patterns that would not otherwise
be detected. Despite reading’s status as something we per-
sonally experience, such developments have occurred in
reading science. Reading is mainly an internal event. The
explanations for how we read refer to unobservable mental
and neural operations. We study behaviors such as students
performance in matching words to pictures, reading sen-
tences aloud with correct pronunciations and intonation,
and answering questions about texts to draw inferences
about these underlying events. However, the evidence now
also includes data collected using specialized instruments
(e.g., eye tracking, electroencephalography, neuroimaging),
analyzed using advanced statistical and computational
models that reveal latent factors, “hidden” knowledge, and
levels of neural activity. These tools have taken our under-
standing of reading beyond the realm of intuition and
direct observation. The question is whether discoveries based
on such methods can inform instructional practices.
We can again illustrate using phonics. Skilled readers
use their knowledge of the correspondences between print
and sound to generate the phonological codes for words in
silent reading. What is this knowledge and how is it
acquired? Given the properties of written English, logic sug-
gests that two types of information must be involved: rules
to produce patterns such as savepavegave, which are also
used in sounding out unfamiliar words (or, in research stud-
ies, pseudowords such as mave), and exception or sight
words whose pronunciations violate the rules (e.g., have,
said, bear) and must be memorized. For generations, dating
back at least as far as the use of phonics methods in the early
19th century (Emans, 1968), this was the only account of
how we manage to read words aloud. It is the core idea
underlying the dual-route theory of reading (Coltheart,
Davelaar, Jonasson, & Besner, 1977; Coltheart, Rastle, Perry,
Langdon, & Ziegler, 2001). The instructional implications
of the theory are straightforward: Teach students the rules
(or enough to allow them to “break the code”; Gough &
Hillinger, 1980) and help them memorize the exceptions.
Although rules plus sight words remains the basis of
phonics instruction, the approach is inadequate in several
respects. What are the rules for pronouncing written
English? No one knows. There are many ad hoc lists
of rules varying in number and coverage, and there is
little evidence that readers employ specific rules, such as
those proposed by Coltheart et al. (2001) and Vousden,
Ellefson, Solity, and Chater (2011). Beyond simple cases
such as the pronunciation of vowels in consonant-vowel-
consonant syllables, it is not clear what the rules are or
even which words are rule governed. Is spook an excep-
tion because of book and look, or rule governed because
of spoon and spool (Seidenberg, 2017)? Worse, it is unclear
how students master the rules given that only a subset of
them can be explicitly taught. Given these uncertainties,
what should a teacher teach? The answer will depend on
which phonics curriculum is being used or which instruc-
tional materials are downloaded from the internet.
Relating this theory to instruction raises a deeper
question: What do students need to know to read aloud?
The word know is ambiguous, of course. It cannot be that
Lost in Translation? Challenges in Connecting Reading Science and Educational Practice | 7
students need to know the rules of English pronunciation
in the sense of being able to state them explicitly, because
no one can. Moreover, the conscious application of rules is
slow and effortful, the opposite of fluent. Perhaps readers
use rules without being consciously aware of them. But
how does a person learn a rule without awareness? There
are several algorithms for deriving rules from language
data (e.g., Albright & Hayes, 2003), but they are not real-
istic accounts of human rule induction. In phonics
instruction, a subset of the rules is explicitly taught. How
does explicit instruction turn into implicit knowledge,
and how does instruction in a subset of rules enable learn-
ing of ones that are not taught?1
Given all of these concerns: one might ask, What if it
is difficult to state the rules and how they are learned and
decide on the sight words because the system is not rule
governed? What if 200 years of phonics instruction has
been based on a false dichotomy? This issue was moot
until an alternative theory was developed by Seidenberg
and McClelland (1989). Their work incorporated ideas
about artificial neural networks, the type of computa-
tional learning system that in later, more advanced forms
underlies the powerful form of artificial intelligence
called deep learning. The framework, implemented in a
series of computational models, has been described in
detail elsewhere (e.g., Plaut, 2005; Seidenberg, 2005).
Here, we can only briefly consider the what and how
questions from before. The what question is about the
knowledge and processes that underlie reading aloud. A
neural network model learns to perform this task, taking
a spelling pattern as input and producing its pronun-
ciation as output (see Figure 1). The model represents
knowledge of the correspondences between spelling and
sound as a set of statistical dependencies (e.g., -ave is usu-
ally pronounced as in save but differently in have). The
network learns these dependencies based on experience
(i.e., presentations of words and their pronunciations)
using a statistical learning procedure based on how such
learning occurs in the brain. The models are not taught a
prespecified set of rules or mappings but rather discover
them through learning to perform the task.
In this approach, words fall on a continuum of spelling–
sound consistency, ranging from the most predictable,
rule-like patterns to ones such as colonel and diarrhea,
which are pretty terrible. There is no distinction between
rule-governed forms and exceptions because they share
structure: Putative exceptions such as have, said, and glove
overlap with other words such as had, send, and globe,
respectively. In a neural network, what is learned about a
word carries over to other, overlapping words. Knowledge
of have affects performance on rule-governed words such
as save and gave and generalization to novel forms such as
mave, evidence that they are learned, represented, and
processed within a common system. Knowledge that is
rule-like but also admits patterns that deviate in varying
degrees is termed quasiregular (Seidenberg & McClelland,
1989). Quasiregularity is characteristic of language at many
levels, including syllables, morphology, words, and grammar
(Bybee & McClelland, 2005; Seidenberg & Plaut, 2014).
Models developed within this framework account for
many empirical phenomena related to spelling–sound
correspondences, including facts about learning, the
development of fluency, and characteristics of perfor-
mance at different skill levels. Such models have also been
applied to related phenomena, such as the computation of
word meanings from print or speech, and guided research
on the brain systems underlying reading (Compton et al.,
2019; Graves et al., 2014; Ueno, Saito, Rogers, & Ralph,
2011). This theory is not as intuitive as rules and sight
words. It is hard to explain and hard for researchers to
analyze how such systems work. Exploring this approach
requires considerable background knowledge. The mod-
els are also incomplete in many respects. Yet, if they are a
more accurate account of fundamental characteristics of
reading, they should be relevant to instruction, bringing
us to the how question.
Humans engage in at least two types of learning:
explicit and implicit, also known as declarative and proce-
dural (Ellis, 2005). The explicit system is associated with
conscious awareness and intention, and procedures that
can be described using language, such as the rules for
chess. It is slow and effortful (cf. Kahneman’s System 2
thinking, a related notion; Kahneman, 2011). The implicit
system operates without conscious awareness, occurs
automatically rather than by intent, and involves unlabeled
statistical patterns (cf. System 1 thinking; Kahneman,
2011). The two systems work together, but their relative
prominence depends on what is to be learned at what
point in development. Consider, for example, the contrast
between how children learn the grammar of a first, spo-
ken language and how older people, such as college stu-
dents, learn the grammar of a second language. A first
language is learned via observation of and experience in
using language to communicate. Children are not explic-
itly taught the rules of grammar; they pick up the struc-
ture of the language via statistical learning. Children have
little awareness of the patterns themselves but learn to use
them appropriately. We eventually gain awareness of some
patterns by studying them, but that is to refine the lan-
guage we have already learned (e.g., to conform to aca-
demic expectations).
Learning a second language in school is different.
Rules of grammar are usually explicitly taught. Learning
depends heavily on already knowing a language (e.g., to
translate from one language to another) and requires
intention and considerable effort. With extended study,
successful second-language learners eventually begin
encoding language statistics through usage. The first- and
second-language examples illustrate that both types of
learning are involved in both cases, but the balance
8 | Reading Research Quarterly, 0(0)
between them shifts: implicit statistical learning more
prominent in L1 and explicit rule learning and instruction
in L2. The characteristics of first- and second-language
learning differ, as do the capacities of learners at different
ages (Seidenberg & Zevin, 2006).
The question then is, What kind of task is learning
to read words? The models developed by Seidenberg
and McClelland (1989), Plaut, McClelland, Seidenberg, and
Patterson (1996), Harm and Seidenberg (1999, 2004),
and in related work suggest that it mainly involves implicit
learning of the statistical structure of mappings between
form (orthography and phonology) and meaning. This
learning occurs in the background as children engage in
silent reading or reading aloud or participate in other
activities that provide opportunities to update this knowledge.
The networks that support reading and language are up -
dated every time we use them.
We also know, however, that some explicit instruction
is necessary. Unlike very young children learning to talk,
children do not begin to read until we teach them about
reading, modeling it for them. Many aspects of written
language are arbitrary, such as letter names and associated
sounds. Children also must learn that print represents
some aspects of spoken words (e.g., phonemic structure)
but not many others that affect comprehension (e.g., voice
quality, syllabic stress). Some studies have suggested that
explicit instruction about a relatively small number of
spelling–sound patterns can facilitate learning other, par-
tially overlapping patterns (e.g., Steacy, Elleman, Lovett, &
Compton, 2016). Explicit instruction can be seen as
FIGURE 1
The Triangle Framewo rk and Its Bases in Per ception and Action
Lost in Translation? Challenges in Connecting Reading Science and Educational Practice | 9
enabling statistical learning, and timely, targeted instruc-
tion can further accelerate it (for a related view, see
Compton, Miller, Elleman, & Steacy, 2014). We do not yet
know the optimal balance between the two systems for
different learners, but learning phonics is more like learn-
ing a first language than learning a second one.
This framework has not been fully specified; much
research remains to be conducted on the balance between
the two learning systems and how to translate those find-
ings into effective practices. The framework already pro-
vides a useful perspective on the long-running debate about
phonics, which arose from different assumptions about the
types of learning involved. At one extreme, children could
be assumed to learn the correspondences on their own in a
literacy-rich environment, an extreme implicit learning
stance. This is not correct because it ignores the role of
explicit instruction in jump-starting the statistical learning
component and accelerating it along the way to expertise.
At the other extreme, phonics programs often assume
that these correspondences need to be exhaustively taught,
like the rules of an exceedingly complex version of chess in
which the allowable movements of the pieces are probabi-
listic and contingent on the surrounding pieces. Readers do
not pronounce words by explicitly applying rules; doing so
would be a conscious, slow, effortful process, the opposite of
fluent. Teaching phonics by teaching rules and memorizing
exceptions leaves out the statistical patterns that permeate
the system and drive the implicit learning process.
We argue that neither extreme is correct. The goal is
not balanced literacy but balanced learning: providing
experiences that engage the implicit and explicit learning
mechanisms to facilitate acquiring the statistical structure
in quasiregular domains such as spelling–sound corre-
spondences. This balance is not well understood but
could be the focus of translational research. We have
argued that instruction needs to cover what is necessary
for children to learn, not merely what is familiar or easy.
The same can be said of using the science of reading to
inform instruction: We cannot merely focus on what is
familiar or easier to digest.
Conclusion
We began by noting that the potential for using reading
research (the science of reading) to improve literacy out-
comes is substantial but has been largely untapped, and
welcomed the renewed interest in making this connection.
We identified three challenges to connecting science and
practice (there are others; see Table 1). These challenges
are serious but can also be addressed. Doing so is likely to
increase the utility of the science in the classroom and its
TAB LE 1
Future Ste ps in Relating the Science of Reading to Educational Pr actice
Step Description
1. Pursue cross-
disciplinary
collaborations.
Summarizing findings and expecting others to pursue the implications has not been an effective strategy in
reading science. Many types of scientific research require teams of individuals with complementary types of
expertise. Translating reading science into verifiably effective educational practices does as well. Such teams
are more likely to succeed at employing basic insights about reading and learning in ways that can be utilized
by educators in the classroom.
2. Work toward a
new science of
teaching.
The goal of a research-based science of teaching is to identify effective instructional practices given a
specification of what needs to be learned (skills and types of knowledge), at different points in development,
for children who differ in ways that affect progress. Proposed practices are hypotheses about effective
instruction whose validity must be empirically assessed.
3. Avoid a narrow
focus on
phonics.
Discussions about connecting the science of reading to education are often limited to phonics. The considerable
research on this issue is only one part of a much larger body of research that has addressed the many other
elements of skilled reading and its development, including the many factors that affect students’ progress.
The science speaks to the importance of integrating print and sound early in development and to the role of
instruction. However, it does so in the context of other skills and knowledge, their dependence on each other,
and the development of reading over time.
4. Invest in early
learning.
Many students are at risk for reading difficulties on the first day of school (Loeb & Bassok, 2008), largely
because of individual differences in knowledge of spoken language and the world that it is used to
communicate about (Hoff, 2013; Muter, Hulme, Snowling, & Stevenson, 2004). Increased translational research
about what can be done in early learning contexts prior to the start of school will help fill in knowledge of
what can be done, when, and for which learners.
5. Develop a science
of reading that
applies to all
readers.
Most research on the science of reading has been conducted with individuals from a narrow range of
backgrounds. Conclusions based on this research cannot be assumed to generalize to understudied groups,
including racial/ethnic minorities and individuals from low-income backgrounds. Deeper understanding of the
impact of these individual difference factors is necessary to advance the science and its impact on education.
6. Examine
existing systems
of learning.
Curricula and instruction can be assessed with respect to whether they are consistent with basic facts about reading
and development derived from modern science. Existing systems, from formal curricula to informal practices,
should be examined and augmented in ways that move them closer to what we know about how learners learn.
10 | Reading Research Quarterly, 0(0)
acceptance as a source of insight about instruction, bene-
fiting teachers and students.
NOTE
This work was supported by the Vilas Trust, University of Wisconsin–
Madison (Seidenberg and Cooper Borkenhagen); the Deinlein Language
and Literacy Fund (Seidenberg); a grant (R305B150003) from the Institute
of Education Sciences (Cooper Borkenhagen); and a grant 5P20HD
091013 from the National Institute of Child Health and Human Devel-
opment (Kearns). The opinions expressed are those of the authors and
do not represent views of the Institute of Education Sciences, the U.S.
Department of Education, or the National Institute of Child Health and
Human Development.
1 The same ambiguity arises about awareness, as in phonological, pho-
nemic, or morphological awareness. The term awareness is unfortu-
nate. Teachers need to be aware of what phonemes and morphemes
are, in the sense of being able to describe and identify them accurately,
which can then inform their instruction. Readers do not need to know
these things, however; they merely have to use them, rapidly and
unconsciously. People managed to learn to read for thousands of years
before linguists developed the concept of phoneme.
REFERENCES
Adams, M.J. (1990). Beginning to read: Learning and thinking about
print. Cambridge, MA: MIT Press.
Aikens, N.L., & Barbarin, O. (2008). Socioeconomic differences in
reading trajectories: The contribution of family, neighborhood, and
school contexts. Journal of Educational Psychology, 100(2), 235–251.
https://doi.org/10.1037/0022-0663.100.2.235
Albright, A., & Hayes, B. (2003). Rules vs. analogy in English past
tenses: A computational/experimental study. Cognition, 90(2), 119–
161. https://doi.org/10.1016/S0010 -0277(03)00146 -X
Allington, R.L. (2002). Big Brother and the national reading curriculum:
How ideology trumped evidence. Portsmouth, NH: Heinemann.
Ball, E.W., & Blachman, B.A. (1991). Does phoneme awareness training
in kindergarten make a difference in early word recognition and
developmental spelling? Reading Research Quarterly, 26(1), 49–66.
https://doi.org/10.1598/RRQ.26.1.3
Bertelson, P., & De Gelder, B. (1989). Learning about reading from illit-
erates. In A.M. Galaburda (Ed.), From reading to neurons (pp. 1–23).
Cambridge, MA: MIT Press.
Blachman, B. (1997). Early intervention and phonological awareness: A
cautionary tale. In B. Blachman (Ed.), Foundations of reading acqui-
sition and dyslexia (pp. 409–430). Hillsdale, NJ: Erlbaum.
Blaunstein, P., & Lyon, R. (2006). Why kids can’t read: Challenging the
status quo in education. Lanham, MD: Rowman & Littlefield.
Bowyer-Crane, C., Snowling, M.J., Duff, F.J., Fieldsend, E., Carroll, J.M.,
Miles, J., … Hulme, C. (2008). Improving early language and literacy
skills: Differential effects of an oral language versus a phonology with
reading intervention. Journal of Child Psychology and Psychiatry,
49(4), 422–432. https://doi.org/10.1111/j.1469-7610.2007.01849.x
Bus, A.G., & van IJzendoorn, M.H. (1999). Phonological awareness and
early reading: A meta-analysis of experimental training studies.
Journal of Educational Psychology, 91(3), 403–414. https://doi.org/
10.1037/0022-0663.91.3.403
Bybee, J., & McClelland, J.L. (2005). Alternatives to the combinatorial
paradigm of linguistic theory based on domain general principles of
human cognition. Linguistic Review, 22(2–4), 381–410. https://doi.org/
10.1515/tlir.2005.22.2-4.381
Byrne, B., & Fielding-Barnsley, R. (1989). Phonemic awareness and let-
ter knowledge in the child’s acquisition of the alphabetic principle.
Journal of Educational Psychology, 81(3), 313–321. https://doi.org/
10.1037/0022-0663.81.3.313
Calkins, L. (with colleagues from the Teachers College Reading and
Writing Project). (2019). Units of Study in Phonics. Portsmouth, NH:
Heinemann.
Cassidy, J., Valadez, C.M., & Garrett, S.D. (2010). Literacy trends and
issues: A look at the five pillars and the cement that supports them. The
Reading Teacher, 63(8), 644–655. https://doi.org/10.1598/RT.63.8.3
Castles, A., Rastle, K., & Nation, K. (2018). Ending the reading wars:
Reading acquisition from novice to expert. Psychological Science in
the Public Interest, 19(1), 5–51. https://doi.org/10.1177/15291 00618
772271
Clay, M.M. (2001). Change over time in children’s literacy development.
Portsmouth, NH: Heinemann.
Coltheart, M., Davelaar, E., Jonasson, J., & Besner, D. (1977). Access to
the internal lexicon. In S. Dornic (Ed.), Attention and performance
VI (pp. 535–555). Hillsdale, NJ: Erlbaum.
Coltheart, M., Rastle, K., Perry, C., Langdon, R., & Ziegler, J. (2001).
DRC: A dual route cascaded model of visual word recognition and
reading aloud. Psychological Review, 108(1), 204–256. https://doi.org/
10.1037/0033-295X.108.1.204
Compton, D.L., Miller, A.C., Elleman, A.M., & Steacy, L.M. (2014).
Have we forsaken reading theory in the name of “quick fix” interven-
tions for children with reading disability? Scientific Studies of
Reading, 18(1), 55–73. https://doi.org/10.1080/10888 438.2013.
836200
Compton, D.L., Steacy, L.M., Petscher, Y., Rueckl, J.G., Landi, N., & Pugh,
K.R. (2019). Linking behavioral and computational approaches to bet-
ter understand variant vowel pronunciations in developing readers.
New Directions for Child and Adolescent Development, 2019(165), 55–71.
https://doi.org/10.1002/cad.20294
Cunningham, A. (1990). Explicit versus implicit instruction in phone-
mic awareness. Journal of Experimental Child Psychology, 50(3),
429–444. https://doi.org/10.1016/0022-0965(90)90079 -N
Davis Dyslexia Association International. (2020, May 22). State dyslexia
laws. Retrieved from https://www.dysle gia.com/state -dysle xia-laws/
Dekker, S., Lee, N.C., Howard-Jones, P., & Jolles, J. (2012). Neuromyths
in education: Prevalence and predictors of misconceptions among
teachers. Frontiers in Psychology, 3. https://doi.org/10.3389/fpsyg.
2012.00429
Ehri, L.C. (2005). Development of sight word reading: Phases and find-
ings. In M.J. Snowling & C. Hulme (Eds.), The science of reading: A
handbook (pp. 135–154). London, UK: Blackwell.
Ellis, N.C. (2005). At the interface: Dynamic interactions of explicit
and implicit language knowledge. Studies in Second Language
Acquisition, 27(2), 305–352. https://doi.org/10.1017/S0272 26310
505014X
Emans, R. (1968). History of phonics. Elementary English, 45(5), 602–608.
Foorman, B.R., Francis, D.J., Fletcher, J.M., Schatschneider, C., &
Mehta, P. (1998). The role of instruction in learning to read: Prevent-
ing reading failure in at-risk children. Journal of Educational Psy-
chology, 90(1), 37–55. https://doi.org/10.1037/0022-0663.90.1.37
Fountas, I.C., & Pinnell, G.S. (2016). The Fountas & Pinnell Literacy Con-
tinuum: A tool for assessment, planning, and teaching. Portsmouth, NH:
Heinemann.
Gewertz, C. (2020, February 20). States to schools: Teach reading the right
wa y. Education Week. Retrieved from https://www.edweek.org/ew/
artic les/2020/02/20/state s-to-schoo ls-teach -readi ng-the-right.html
Gillon, G.T. (2000). The efficacy of phonological awareness interven-
tion for children with spoken language impairment. Language,
Speech, and Hearing Services in Schools, 31(2), 126–141. https://doi.
org/10.1044/0161-1461.3102.126
Goldstein, D. (2015). The teacher wars: A history of America’s most
embattled profession. New York, NY: Anchor.
Goldstein, D. (2020, February 15). An old and contested solution to
boost reading scores: Phonics. The New York Times. Retrieved from
https://www.nytim es.com/2020/02/15/us/readi ng-phoni cs.html
Lost in Translation? Challenges in Connecting Reading Science and Educational Practice | 11
Goodman, K.S. (1989). Whole-language research: Foundations and
development. The Elementary School Journal, 90(2), 207–221. https://
doi.org/10.1086/461613
Gough, P.B., & Hillinger, M.L. (1980). Learning to read: An unnatural
act. Bulletin of the Orton Society, 30, 179–196. https://doi.org/10.
1007/BF026 53717
Gough, P.B., & Tunmer, W.E. (1986). Decoding, reading, and reading
disability. Remedial and Special Education, 7(1), 6–10. https://doi.org/
10.1177/07419 32586 00700104
Graves, W.W., Binder, J.R., Desai, R.H., Humphries, C., Stengel, B.C., &
Seidenberg, M.S. (2014). Anatomy is strategy: Skilled reading differ-
ences associated with structural connectivity differences in the read-
ing network. Brain and Language, 133, 1–13. https://doi.org/10.1016/j.
bandl.2014.03.005
Gunther, N., & Lindstrom, C. (Producers). (2003). A tale of two schools:
The documentary [DVD]. Available from http://www.pbs.org/weta/
twosc hools/ docum entar y/
Hanford, E. (2018, September 10). Hard words: Why aren’t kids being
taught to read? APM Reports. Retrieved from https://www.apmre
ports.org/episo de/2018/09/10/hard-words -why-ameri can-kids-arent -
being -taugh t-to-read
Harm, M.W., & Seidenberg, M.S. (1999). Phonology, reading acquisition,
and dyslexia: Insights from connectionist models. Psychological Review,
106(3), 491–528. https://doi.org/10.1037/0033-295X.106.3.491
Harm, M.W., & Seidenberg, M.S. (2004). Computing the meanings of
words in reading: Cooperative division of labor between visual and
phonological processes. Psychological Review, 111(3), 662–720. https://
doi.org/10.1037/0033-295X.111.3.662
Hatcher, P.J., Hulme, C., & Ellis, A.W. (1994). Ameliorating early read-
ing failure by integrating the teaching of reading and phonological
skills: The phonological linkage hypothesis. Child Development,
65(1), 41–57. https://doi.org/10.2307/1131364
Hoff, E. (2013). Interpreting the early language trajectories of children
from low-SES and language minority homes: Implications for closing
achievement gaps. Developmental Psychology, 49(1), 4–14. https://doi.
org/10.1037/a0027238
Houghton Mifflin Harcourt. (2014). Journeys: Common Core. Orl ando,
FL: Author.
Hulme, C., Nash, H.M., Gooch, D., Lervåg, A., & Snowling, M.J. (2015).
The foundations of literacy development in children at familial risk
of dyslexia. Psychological Science, 26(12), 1877–1886. https://doi.org/
10.1177/09567 97615 603702
Hurford, D.P. (2020). The science of reading (a response to the New York
Times). IDA Examiner, 9(1). Retrieved from https://dysle xiaida.org/
the-scien ce-of-readi ng-a-respo nse-to-the-new-york-times/
International Literacy Association. (2019). Meeting the challenges of
early literacy phonics instruction [Literacy leadership brief]. Newark,
DE: Author. Retrieved from https://www.liter acywo rldwi de.org/
docs/defau lt-sourc e/where -we-stand/ ila-meeti ng-ch all en ges -early -
liter acy-phoni cs-instr uction.pdf
Kahneman, D. (2011). Thinking, fast and slow. New York, NY: Macmillan.
Kearns, D.M. (2020). Does English have useful syllable division pat-
terns? Reading Research Quarterly, 55(Suppl. 1).
Kim, J.S. (2008). Research and the reading wars. In F.M. Hess (Ed.),
When research matters: How scholarship influences education policy
(pp. 89–111). Cambridge, MA: Harvard Education Press.
Kim, Y.-S., Petscher, Y., Schatschneider, C., & Foorman, B. (2010). Does
growth rate in oral reading fluency matter in predicting reading
comprehension achievement? Journal of Educational Psychology,
102(3), 652–667. https://doi.org/10.1037/a0019643
Krashen, S. (2002). Defending whole language: The limits of phonics
instruction and the efficacy of whole language instruction. Reading
Improvement, 39(1), 32–42.
Landauer, T.K., & Dumais, S.T. (1997). A solution to Plato’s problem:
The latent semantic analysis theory of acquisition, induction, and
representation of knowledge. Psychological Review, 104(2), 211–240.
https://doi.org/10.1037/0033-295X.104.2.211
Lemann, N. (1997, November). The reading wars. The Atlantic. Re trieved
from https://www.theat lantic.com/magaz ine/archi ve/1997/11/the-readi ng-
wars/37699 0/
Liberman, I.Y., Shankweiler, D., & Liberman, A.M. (1989). The alpha-
betic principle and learning to read. Ann Arbor: University of Michi-
gan Press. Retrieved from https://eric-ed-gov.ezpro xy.libra ry.wisc.
edu/?id=ED427291
Loeb, S., & Bassok, D. (2008). Early childhood and the achievement
gap. In H.F. Ladd & E.B. Fiske (Eds.), Handbook of research in educa-
tion finance and policy (pp. 517–534). New York, NY: Routledge.
McCardle, P., & Chhabra, V. (2005). Motivation and commitment in
the interest of science and our children. Phi Delta Kappan, 86(6),
448–451. https://doi.org/10.1177/00317 21705 08600608
McGinty, A.S., Breit-Smith, A., Fan, X., Justice, L.M., & Kaderavek, J.N.
(2011). Does intensity matter? Preschoolers’ print knowledge devel-
opment within a classroom-based intervention. Early Childhood
Research Quarterly, 26(3), 255–267. https://doi.org/10.1016/j.ecresq.
2011.02.002
McGraw-Hill Education. (2014). McGraw Hill Reading Wonders: CCSS
reading/language arts program. New York, NY: Author.
Metsala, J.L., & Walley, A.C. (1998). Spoken vocabulary growth and the
segmental restructuring of lexical representations: Precursors to
phonemic awareness and early reading ability. In J.L. Metsala & L.C.
Ehri (Eds.), Word recognition in beginning literacy (pp. 89–120).
Mahwah, NJ: Erlbaum.
Moats, L.C., & Tolman, C. (2009). Language essentials for teachers of
reading and spelling. Boston, MA: Sopris West.
Morris, R.D., Lovett, M.W., Wolf, M., Sevcik, R.A., Steinbach, K.A.,
Frijters, J.C., & Shapiro, M.B. (2012). Multiple-component remedia-
tion for developmental reading disabilities: IQ, socioeconomic status,
and race as factors in remedial outcome. Journal of Learning Dis-
abilities, 45(2), 99–127. https://doi.org/10.1177/00222 19409 355472
Muter, V., Hulme, C., Snowling, M.J., & Stevenson, J. (2004). Phonemes,
rimes, vocabulary, and grammatical skills as foundations of early
reading development: Evidence from a longitudinal study. Develop-
mental Psychology, 40(5), 665–681. https://doi.org/10.1037/0012-1649.
40.5.665
National Conference of State Legislatures. (2019). Third-grade reading
legislation. Retrieved from https://www.ncsl.org/resea rch/educa tion/
third -grade -readi ng-legis lation.aspx
National Institute of Child Health and Human Development. (2000).
Report of the National Reading Panel. Teaching children to read: An
evidence-based assessment of the scientific research literature on read-
ing and its implications for reading instruction (NIH Publication No.
00-4769). Washington, DC: U.S. Government Printing Office.
National Reading Panel. (2000). Teaching children to read: An evidence
based assessment of the scientific research literature on reading and
its implications for reading instruction: Reports of the subgroups
Rockville, MD: National Institute of Child Health and Human
Development.
Perfetti, C.A., & Hart, L. (2002). The lexical quality hypothesis. In
L. Verhoeven, C. Elbro, & P. Reitsma (Eds.), Precursors of functional
literacy (pp. 189–213). Amsterdam, Netherlands: John Benjamins.
Plaut, D.C. (2005). Connectionist approaches to reading. In M.J. Snowling
& C. Hulme (Eds.), The science of reading: A handbook (pp. 24–38).
London, UK: Blackwell.
Plaut, D.C., McClelland, J.L., Seidenberg, M.S., & Patterson, K. (1996).
Understanding normal and impaired word reading: Computational
principles in quasi-regular domains. Psychological Review, 103(1),
56–115. https://doi.org/10.1037/0033-295X.103.1.56
Ravitch, D. (2011). The death and life of the great American school sys-
tem: How testing and choice are undermining education. New York,
NY: Basic.
12 | Reading Research Quarterly, 0(0)
Rayner, K., Foorman, B.R., Perfetti, C.A., Pesetsky, D., & Seidenberg,
M.S. (2001). How psychological science informs the teaching of
reading. Psychological Science in the Public Interest, 2(2), 31–74.
https://doi.org/10.1111/1529-1006.00004
Reardon, S.F. (2013). The widening income achievement gap. Educa-
tional Leadership, 70(8), 10–16.
Rumelhart, D.E. (1977). Toward an interactive model of reading. In S.
Dornic (Ed.), Attention and performance VI (pp. 573–606). Hillsdale,
NJ: Erlbaum.
Schuele, C.M., & Boudreau, D. (2008). Phonological awareness interven-
tion: Beyond the basics. Language, Speech, and Hearing Services in
Schools, 39(1), 3–20. https://doi.org/10.1044/0161-1461(2008/002)
Seidenberg, M.S. (2005). Connectionist models of word reading. Cur-
rent Directions in Psychological Science, 14(5), 238–242. https://doi.
org/10.1111/j.0963-7214.2005.00372.x
Seidenberg, M. (2017). Language at the speed of sight: How we read, why
so many can’t, and what can be done about it. New York, NY: Basic.
Seidenberg, M. (2019, December 6). This is why we don’t have better read-
ers: Response to Lucy Calkins [Web log post]. Retrieved from https://
seide nberg readi ng.net/2019/12/06/lucy-calki ns-on-the-attac k/
Seidenberg, M.S., & McClelland, J.L. (1989). A distributed, developmen-
tal model of word recognition and naming. Psychological Review,
96(4), 523–568. https://doi.org/10.1037/0033-295X.96.4.523
Seidenberg, M.S., & Plaut, D.C. (2014). Quasiregularity and its discon-
tents: The legacy of the past tense debate. Cognitive Science, 38(6),
1190–1228. https://doi.org/10.1111/cogs.12147
Seidenberg, M.S., & Zevin, J.D. (2006). Connectionist models in devel-
opmental cognitive neuroscience: Critical periods and the paradox
of success. In Y. Munakata & M.H. Johnson (Eds.), Processes of
change in brain and cognitive development: Attention and perfor-
mance XXI (pp. 315–347). Oxford, UK: Oxford University Press.
Shanahan, T. (2004). Critiques of the National Reading Panel report:
Their implications for research, policy, and practice. In P. McCardle
& V. Chhabra (Eds.), The voice of evidence in reading research (pp.
235–265). Baltimore, MD: Paul H Brookes.
Shankweiler, D., Mencl, W.E., Braze, D., Tabor, W., Pugh, K.R., & Ful-
bright, R.K. (2008). Reading differences and brain: Cortical integra-
tion of speech and print in sentence processing varies with reader
skill. Developmental Neuropsychology, 33(6), 745–775. https://doi.
org/10.1080/87565 64080 2418688
Share, D.L. (1995). Phonological recoding and self-teaching: Sine qua
non of reading acquisition. Cognition, 55(2), 151–218. https://doi.
org/10.1016/0010-0277(94)00645 -2
Smith, F. (1999). Why systematic phonics and phonemic awareness instruc-
tion constitute an educational hazard. Language Arts, 77(2), 150–155.
Snow, C.E., Burns, M.S. & Griffin, P. (Eds.). (1998). Preventing reading dif-
ficulties in young children. Washington, DC: National Academy Press.
Snowling, M.J., & Hulme, C.E. (2005). The science of reading: A hand-
book. London, UK: Blackwell.
Stanovich, P.J., & Stanovich, K.E. (2003). Using research and reason in
education: How teachers can use scientifically based research to make
curricular and instructional decisions. Washington, DC: National
Institute for Literacy.
Steacy, L.M., Elleman, A.M., Lovett, M.W., & Compton, D.L. (2016).
Exploring differential effects across two decoding treatments on
item-level transfer in children with significant word reading difficul-
ties: A new approach for testing intervention elements. Scientific
Studies of Reading, 20(4), 283–295. https://doi.org/10.1080/10888
438.2016.1178267
Steiner, D.M., & Rozen, S.D. (2004). Preparing tomorrow’s teachers: An
analysis of syllabi from a sample of America’s schools of education.
In F.M. Hess, A.J. Rotherham, & K. Walsh (Eds.), A qualified teacher
in every classroom? Appraising old answers and new ideas (pp. 119–
148). Cambridge, MA: Harvard Education Press.
Stuebing, K.K., Barth, A.E., Cirino, P.T., Francis, D.J., & Fletcher, J.M.
(2008). A response to recent reanalyses of the National Reading
Panel report: Effects of systematic phonics instruction are practically
significant. Journal of Educational Psychology, 100(1), 123–134. https://
doi.org/10.1037/0022-0663.100.1.123
Ueno, T., Saito, S., Rogers, T.T., & Ralph, M.A.L. (2011). Lichtheim 2: Syn-
thesizing aphasia and the neural basis of language in a neurocomputa-
tional model of the dual dorsal-ventral language pathways. Neuron,
72(2), 385–396. https://doi.org/10.1016/j.neuron.2011.09.013
Vellutino, F.R., Tunmer, W.E., Jaccard, J.J., & Chen, R. (2007). Compo-
nents of reading ability: Multivariate evidence for a convergent skills
model of reading development. Scientific Studies of Reading, 11(1),
3–32. https://doi.org/10.1080/10888 43070 9336632
Vousden, J.I., Ellefson, M.R., Solity, J., & Chater, N. (2011). Simplifying
reading: Applying the simplicity principle to reading. Cognitive Sci-
ence, 35(1), 34–78. https://doi.org/10.1111/j.1551-6709.2010.01134.x
Submitted May 10, 2020
Final revision received July 1, 2020
Accepted July 2, 2020
MARK S. SEIDENBERG (corresponding author) is the Vilas
Professor and Donald O. Hebb Professor in the Department of
Psychology at the University of Wisconsin–Madison, USA;
email seidenberg@wisc.edu.
MATT COOPER BORKENHAGEN is a graduate student in
the Department of Psychology at the University of Wisconsin–
Madison, USA; email cooperborken@wisc.edu.
DEVIN M. KEARNS is an associate professor in the
Department of Educational Psychology at the University of
Connecticut, Storrs, USA; email devin.kearns@uconn.edu.
... Thus, across languages and scripts, the importance of morphological awareness for word recognition is increasingly recognized (e.g., Carlisle & Stone, 2005). Even for English, researchers have called for a reduced emphasis on phonics in favor of a more general link of form and meaning in word reading and writing (Seidenberg, Borkenhagen, & Kearns, 2020). Syntactic aspects of words are additionally highlighted as important for meaning-building in word reading and writing across languages (Jongejan, Verhoeven, & Siegel, 2007;Plaza & Cohen, 2003). ...
... For example, the pattern TH, as it appears in words such as thumb or both, usually is not pronounced as /t/ in English. In fact, the way in which word recognition takes place involves sensitivity to statistical regularities of the script (e.g., Seidenberg et al., 2020). Many of us informally begin to recognize these statistical regularities. ...
... Longitudinal comparisons of children's literacy development across scripts are, therefore, imperative for future work. Greater considerations of visual-orthographic complexity vis-a-vis word recognition should be particularly valued in research and teaching moving forward from a basic phonology-meaning-orthography model (Seidenberg et al., 2020;Seidenberg & McClelland, 1989). ...
Article
We review cognitive-linguistic approaches to conveying meaning, sound, and orthographic information across scripts in order to highlight the impact of variability in written and spoken language on learning to read and to write words. With examples of word recognition and word writing from different scripts, including Chinese, Arabic, Persian, and English among others, we highlight 1) characteristics and boundaries of a word and how these sometimes present challenges for reading and spelling, 2) phonological sensitivity, including phonological omissions in print, suprasegmental processing, and “distance” between spoken and written forms, vis-à-vis literacy acquisition at the word level, 3) the importance of specific types of divergent visual-orthographic knowledge for the mastery of different writing systems, and 4) expanding understanding of visual-motor skills and their role in spelling across scripts. All of these aspects of variability in different writing systems should be more broadly integrated as theoretical models and intervention methods of reading or writing are tested across different writing systems.
Article
Full-text available
There is intense public interest in questions surrounding how children learn to read and how they can best be taught. Research in psychological science has provided answers to many of these questions but, somewhat surprisingly, this research has been slow to make inroads into educational policy and practice. Instead, the field has been plagued by decades of “reading wars.” Even now, there remains a wide gap between the state of research knowledge about learning to read and the state of public understanding. The aim of this article is to fill this gap. We present a comprehensive tutorial review of the science of learning to read, spanning from children’s earliest alphabetic skills through to the fluent word recognition and skilled text comprehension characteristic of expert readers. We explain why phonics instruction is so central to learning in a writing system such as English. But we also move beyond phonics, reviewing research on what else children need to learn to become expert readers and considering how this might be translated into effective classroom practice. We call for an end to the reading wars and recommend an agenda for instruction and research in reading acquisition that is balanced, developmentally informed, and based on a deep understanding of how language and writing systems work.
Article
Programs for teaching English reading, especially for students with dyslexia, and educational practice standards often recommend instruction on dividing polysyllabic words into syllables. Syllable division is effort intensive and could inhibit fluency when reading in text. The division strategies might still be useful if they work so consistently that they will help students decode most unfamiliar polysyllabic words. No study of the English lexicon has confirmed that the pattern is highly consistent. This study addresses this gap in the literature. The utility of the two most frequently taught patterns was examined in a corpus analysis of 14,844 words from texts used in grades 1–8. The VC|CV pattern involves a single vowel (V), two consonants (CC), and another vowel. According to the expected pattern, the first vowel should have a short (lax) sound, such as the a in rabbit. This was true of 70.6% of instances in VCCV words in the corpus. For the V|CV pattern, the first vowel is expected have a long (tense) sound, such as the a in mason. This was true in 30.5% of instances in VCV words in the corpus. The patterns were more consistent for bisyllabic words than longer words (78.8% vs. 62.5% for VCCV words and 47.3% vs. 18.8% for VCV words, respectively). When comparing only short‐ and long‐vowel pronunciations (ignoring other sounds such as schwa), the first vowel followed the expected pattern in 94.3% instances of VCCV words and 53.3% of VCV words. The unreliability of VCV may not justify the effort required to use the strategy. There are implications for the debate about the science of reading.
Article
The overarching goal of the new Florida State University/Haskins Laboratory/University of Connecticut Learning Disability (LD) Hub project is to align computational and behavioral theories of individual word reading development more closely with the challenges of learning to read a quasi‐regular orthography (i.e., English) for both typically developing (TD) children and, more specifically, children with dyslexia. Our LD Hub adopts an integrated approach to better understand the neurocognitive bases of individual differences in word reading development by specifically examining the experiential (exogenous) and child‐specific (endogenous) factors that determine acquisition of orthographic–phonological knowledge at different subword granularities using behavioral and computational modeling. Findings are intended to enrich understanding of the processes that influence individual differences in word reading development in TD and dyslexic children and significantly inform issues of practice (e.g., curriculum, instruction, diagnosis, and intervention). Here, we briefly provide the rationale for the Hub and present findings from the initial behavioral and computational modeling studies.
Article
In English, gains in decoding skill do not map directly onto increases in word reading. However, beyond the Self-Teaching Hypothesis, little is known about the transfer of decoding skills to word reading. In this study, we offer a new approach to testing specific decoding elements on transfer to word reading. To illustrate, we modeled word-reading gains among children with reading disability enrolled in Phonological and Strategy Training (PHAST) or Phonics for Reading (PFR). Conditions differed in sublexical training with PHAST stressing multilevel connections and PFR emphasizing simple grapheme-phoneme correspondences. Thirty-seven children with reading disability, 3rd to 6th grade, were randomly assigned 60 lessons of PHAST or PFR. Crossed random-effects models allowed us to identify specific intervention elements that differentially impacted word-reading performance at posttest, with children in PHAST better able to read words with variant vowel pronunciations. Results suggest that sublexical emphasis influences transfer gains to word reading.