Haskins Laboratories Status Report on Speech Research

1990, SR-101 /102, 1-13

The Alphabetic Principle and Learning to Read*

Isabelle Y. Liberman, t Donald Shankweiler, t and Alvin M. Liberman tt

Proper application of the alphabetic principle rests on an awareness of the internal

phonological (and morphophonological) structure of words that the alphabet represents.
Unfortunately for the would-be reader-writer, such awareness is not an automatic
consequence of speaking a language, because the biological specialization for speech
manages the production and perception of these structures below the level of
consciousness. Not surprisingly, then, awareness of phonological structures is normally
lacking in preliterate children and adults; the degree to which it does exist is the best
single predictor of success in learning to read; lack of awareness usually yields to
appropriate instruction; and such instruction makes for better readers. That some children
have particular difficulty in developing phonological awareness (and in learning to read) is
apparently to be attributed to a general deficiency in the phonological component of their
natural capacity for language. Thus, these children are also relatively poor in short-term
memory for verbal information, in perceiving speech in noise, in producing complex speech
patterns, and in finding the words that name objects. All children will benefit from
instruction that is intelligently designed to show them what the alphabet is about.

For some 15 years, we have been exploring the Beyond letter identification, reading requires
sources of the problems beginners encounter in mastery of a system that maps letter shapes to
learning to read. Since children are quite fluent in units of speech. However, as we noted years ago
their native language when first encount~ring the (Liberman, 1973), there is no evidence that chil-
language in print, we began by asking what dren of normal intelligence, given proper instruc-
seemed to us the obvious question: what is re- tion, have difficulty in associating individual let-
quired of the child in reading a language but not ters of the alphabet with their appropriate speech
in speaking or listening to it? The first answer equivalents. Perhaps, then, they are defeated by
that came to mind, of course, was the discrimina- the often complex and irregular relations in
tion of the visual shapes of the letters. But inves- English between spelling and language. Surely,
tigators who had done comprehensive studies of the complexities of English spelling do create
many different aspects of the reading process (see some problems. But even when the items to be
Doehring, 1968), or who had compiled exhaustive read include only those words that map the sound
reviews of the visual factors involved in reading in a simple, consistent way, many children still
(Benton & Pearl, 1978; Stanovich, 1982; Vellutino, fail (Savin, 1972).
1979; Vernon, 1957), were all in agreement that Learning to identify the letters, learning to as-
beginners who were making little progress in sociate them with consonant and vowel sounds,
learning to read generally showed no significant learning to cope with the irregularities of English
difficulty in the visual identification of letters. spelling-none of these is the primary obstacle in
learning to read. What is it then that makes read-
ing so hard while speech is relatively so easy? In
Parts of this paper were adapted from "Phonology and the the seventies, we (Liberman, 1971; Shankweiler &
problems oflearning to read and write." Remedial and Special
Education, 1985,6,8-17. This research was supported in part
Liberman, 1972) and other investigators (Elkonin,
by grant HD-01994 to Haskins Laboratories and by grant 1973; Gleitman & Rozin, 1977; Klima, 1972;
NllI·21888 to Yale UniversitylHaskins Laboratories from the Mattingly, 1972) proposed another possible source
National Institute of Child Health and Human Development. of difficulty in reading that is not present in

1
2 Liberman et ai.

speech. Although both reading and speech require the underlying phonology, not in the surface
some degree of mastery of language, reading re- appearances of the sound (Liberman, Cooper,
quires, in addition, a mastery of the alphabetic Shankweiler, & Studdert-Kennedy, 1967).
principle. This entails an awareness of the inter- To emphasize that letters stand for sounds also
nal phonological structure of the words of the lan- risks making it that much harder for a reader to
guage, an awareness that must be more explicit understand perfectly reasonable aspects of
than is ever demanded in the ordinary course of English spelling-for example, cats and dogs, in-
listening and responding to speech. If this is so, it stead of cats and dogz; or bat and batter, instead of
should follow that beginning learners with a bat and badder; or an apple, instead of uh napple;
weakness in phonological awareness would be at and so on. As for those aspects of English spelling
risk. that are most egregiously unreasonable-for ex-
We here first set forth the considerations that ample, through, rough, and the like-we confess
led us to that view, followed by the evidence that that even a proper understanding of the alpha-
supports it. Then we say why we should consider betic principle is not likely to be of much help, but
that deficits in awareness of the phonological then neither will anything else, short of learning
structure may be only one symptom of a more about how the language has changed since the
general underlying deficiency in the phonological orthography was developed. (Such spellings need
component of the beginning reader's capacity for not be a hindrance if they are introduced only
language. Finally we consider the implications for after the more systematic aspects of the orthogra-
instruction. phy have been understood.)
The identification of letters with sounds pro-
PHONOLOGY AND THE ALPHABETIC motes yet another misunderstanding, this one
PRINCIPLE about the nature of words and how they are per-
We begin, then, with the assumption that ceived. For it accords all too well with the com-
reading by an alphabetic writing system requires monplace notion that it is only spoken words that
mastery of the alphabetic principle. Surprisingly, are made of phonological units. Words that corne
this assumption, which seems to us a' truism, is to us via print are incorrectly thought to be differ-
not accepted by everyone in the field, as we will ent, in that they supposedly can (and perhaps
see. But even among those who think the principle should) be perceived independently of phonology.
important, many take it-we would say mistake On this basis, some advise that the reader be
it-to mean simply an understanding by the taught from the very beginning to skip the
would-be reader that the discrete letters of the phonology (read this as skipping the sound) and
alphabet represent the discrete sounds of speech. go "direct to meaning" (Goodman, 1976; Smith,
Our view (Liberman, 1983) is different. As we see 1971). Others grant that going through the
it, the letters of the alphabet do not represent phonology-which is taken to mean "sounding it
sounds as such, but rather the more remote out"-may be useful for the beginning reader, or
phonological (and morphophonological) segments for the mature reader who encounters a strange
those sounds convey. This is not to quibble. For word, but they otherwise hold that the putatively
surely it must be somewhat confusing to children "direct" (nonphonological) route is the way to go
to be told that the word bag is spelled with three (Coltheart, 1978; Waters, Seidenberg, & Bruck,
letters, when his ears tell him plainly that it has 1984).
but one sound. The confusion is only worse con- We believe that these assumptions seriously
founded ifthe teacher insists, against the evidence misconstrue the nature of words and the processes
of what the child hears, that bag can be divided by which they are produced and perceived, in
into three sounds, and that these can then be print as in speech. Consider, in this connection, a
"blended" so as to re-form the word. For there is, critical difference between language and all other
in fact, no way, with or without the marveis of natural forms of communication. In all the nonlin-
modem technology, to divide bag into pieces of guistic systems-whether the medium is acoustic,
sound that correspond in any reasonable way to optical, electrical, or chemical-meanings are con-
the sounds of the three letters, nor is there any veyed by signals that differ holistically, one from
way to synthesize the word by somehow putting another. This is to say that there are no words.
the letter sounds together. Though bag does truly The inevitable consequence is that the number of
consist of three segments-it differs from sag in meanings that can be communicated is limited to
the first, from big in the second, and from bat in the number of holistically different signals the an-
the third-these segments are to be found only in imal can produce and perceive, a number that is
 The Alphabetic Principle and Learning to Read 3

always quite small. Even if that number can as to produce the sounds of speech, but in fact it
somehow be increased, there il) no way of doing it does not. What the reader must do is to match the
so as to guarantee that the new signals will be alphabetic transcription to the abstract phono-
immediately recognized as belonging naturally to logical structure of the word it represents. In the
a system that has a specifically communicative case of a familiar word, this structure and its
function. associated meaning(s) are available in the reader's
Language is different in a most important way. lexicon; in the case of an unfamiliar word, given a
Meanings are not conveyed directly by signals command of the alphabetic principle, the structure
that differ holistically, but rather by words that is easily formed and thus made ready for what-
are distinct from each other in their internal ever meanings may subsequently be attached to it.
structure. This structure is formed of a small Once the reader has the phonological form of the
number of meaningless phonological segments we word, the appropriate phonetic structure and its
know as consonants and vowels, and governed associated articulatory movements are automati-
according to a highly systematic combinatorial cally available to him for use in working memory,
scheme called phonology. The consequence is that or for reading aloud if the occasion should call for
words can (and do) number in the tens of that.
thousands. Moreover, there is a perfectly natural
basis for accommodating new words, since the
PHONOLOGICAL PROCESSING IN
phonological system, which all speakers of the lan- READING-WRITING AND IN
guage have in common, automatically recognizes a LISTENING-SPEAKING
new, but legal, structure as a word that stands Why is it normally so much harder and less
ready to have meaning attached to it. It is only natural to deal with phonological structures in
because children have this phonological system reading and writing than it is in listening and
that they are able to acquire new words and their speaking? A serious attempt to answer this ques-
meanings with such astonishing ease and rapidity tion would take us quite deeply into the phonolog-
(Studdert-Kennedy, 1987). ical system and its biology, for the answer re-
What follows, then, is that phonology govems all quires explaining, among other things, why
words, whether dead, living, or waiting to be born. speech could have evolved in the history of our
So, whatever else a word is, and regardless of species but writing systems could not, and why
whether. it is spoken _or printed, it is always a speech can develop in the child without explicit
phonological structure. If listeners or readers instruction but reading and writing typically can
correctly perceive a word, they correctly perceive not (Liberman, 1989). Here, we can offer only a
the structure that distinguishes it. They may very truncated account.
well be unsure of its meaning-indeed, they may Like all members of the animal kingdom, human
even have got the meaning wrong-but if they beings have highly specialized ways of com-
have the phonological structure, they have a municating with their fellows. In the human case,
perfectly adequate basis for ultimately getting its and only there, this specialization includes, as a
meaning properly sorted out. As for going directly critical component, the phonological system that,
to meaning-that is, independently of phonology- as we have seen, makes large vocabularies possi-
surely that is done when a person sees a picture, ble. As this system evolved in the race, and as it
for example, or hears the roar of a lion, but not develops anew in each child, it employs abstract
when one perceives a word as it is spoken or read. motor structures-let us call them gestures-that
From our point of view, then, there is no reason ultimately control the movements of the speech
to ask, as some do, whether readers must, or organs (Browman & Goldstein, 1985; Liberman &
should get to meaning via the phonology. To make Mattingly, 1985; Liberman & Mattingly, 1989).
sense of this question, one must make three false These gestures are adapted for one purpose and
assumptions. The first is that the meanings can be for one purpose only: the production of strings of
communicated in language independently of consonants and vowels at rates many times more
words (that is, phonological structures), but in fact rapid than could otherwise be achieved. These
they cannot. The second is that the phonological rates, which run about eight to ten per second on
units that form all words are equivalent to the average, are managed by precisely overlapping
sounds of speech, but in fact they are not. And the and merging the articulatory movements that
third is that an alphabetic transcription specifies, produce the phonologically significant aspects of
on a segment by segment basis, how the speech the speech sounds. This coarticulation, as it is
organs are to be articulated and coarticulated so called, is a most complex process, but it does not
 4 Liberman et al.

appear so in the normal exercise of speech func- read and write all the words of the language-
tions, because it is done automatically and natu- those they were already familiar with and those
rally by this aspect of the specialization for lan- they had yet to encounter. But they could exploit
guage. A consequence is that a neurologically this wonderful invention only if they understood
normal child, put in a speech environment, can the discovery on which it was based.
hardly be prevented from learning to form phono-
logical structures and to exploit coarticulation for AWARENESS OF PHONOLOGICAL
that purpose. (Lacking this specialization, non- STRUCTURE AND READING
human primates do not, and cannot, learn to pro- Development of phonological awareness
duce these structures;' this is to say that they can-
not produce words.) A more important conse- in children
quence for our purposes is that to speak a word Considerations such as these led us at the very
one need not know how it is spelled. The speaker outset of our research on reading to suppose that
need only think of the word; the phonological preliterate children would not naturally have
component of his grammar "spells" it for him. made the discovery that underlay the invention of
Indeed, the automaticity of this specialization the alphabet, from which it would follow that they
makes it that much harder to be aware of how the would not be prepared to understand and apply
word is spelled, or even to know that such a thing the alphabetic principle. So we began to examine
as spelling exists. developmental trends in phonological awareness
Perception of the speech signal is correspond- by testing the ability of young children to segment
ingly complex and automatic. Given coarticula- words into their constituent elements. We inves-
tion, there is no direct correspondence between tigated the children's segmentation of spoken
the phonological structure intended by the words both by syllable and phoneme (Liberman,
speaker and the surface properties of the sound. Shankweiler, Fischer, & Carter, 1974). (The latter
Most relevant to our concerns is the fact that, as class of units of the phonological representation
we have so often pointed out, the number of seg- comprise the consonants and vowels. Heretofore
ments in the sound is not equal to the number of we have r-eferred to these only by the general
segments in the phonological structure it conveys term, phonological segments. From now on we will
, (Liberman et al., 1967). Thus, the three conso- call then phonemes to distinguish them from sylla-
nants and vowels of a word like bag are so thor- bles). We found that normal preschool children
oughly coarticulated as to produce a single seg- performed rather poorly, but that the phonemes
ment of sound. But this is no problem for listen- presented the greater difficulty by far.
ers, for they have only to rely on their phonologi- It was clear from these results that awareness of
cal specialization to automatically process the phoneme segments, the basic units of the
speech signal and recover the coarticulated ges- alphabetic orthography, is initially harder to
tures that caused it (Liberman & Mattingly, achieve than awareness of syllable segments, and
1989). It is a problem for would-be readers, how- develops later, if at all. More relevant to our
ever, because, given the complex relation between present purposes, it was also apparent that a
phonological structure and sound, and the auto- large number of children, about 30 percent of our
maticity with which this relation is dealt with in sample, had not attained an understanding of the
speech, they find it just that much harder to be internal phonemic structure of words, even at the
aware that the word does have an internal struc- end of a full year in school. Surely, they are the
ture and thus to appreciate why an alphabetic ones we need to worry about, because they are the
transcription makes sense. ones who are deficient in the linguistic awareness
Small wonder then that an alphabetic writing that may provide entry into the alphabetic system.
system is such a comparatively recent devel-
opment in the history of our species. In contrast to
Lack of phonological awareness and
the naturally evolved phonological structures it reading failure in children
represents, it is an artifact. The development of Is lack of phonological awareness in fact related
this artifact had to wait on the discovery-and it to failure in reading and writing? That the answer
was a discovery-that words have an internal is yes is strongly supported by studies in a
structure. Once that discovery was available, number of languages. In English, the relation has
someone could and did invent the notion that, by been found, for example, in studies by Blachman
representing the units of that structure with (1984); Bradley and Bryant (1983); Fox and Routh
arbitrarily chosen optical shapes, people could (1980); Goldstein (1976); Helfgott (1976); Treiman
 The Alphabetic Principle and Learning to Read 5

and Baron (1981); Vellutino and Scanlon, (1987). the third grade and the other with good and poor
Their findings have been supported by studies in readers in adult education classes (Pratt, 1985).
Swedish by Lundberg and associates (Lundberg, All the subjects were given three linguistic
Olofsson, & Wall, 1980) and Magnusson and awareness tests and one nonspeech control task
Naucler (1987), in Spanish by de Manrique and identical in format to one of the linguistic
Gramigna (1984), in French by a group of Belgian measures. Significant differences were found
researchers (Alegria, Pignot, & Morais, 1982), and between the good and poor readers at both age
recently in Italian by Cossu and associates (Cossu, levels on all three linguistic measures, but not on
Shankweiler, Liberman, Tola, & Katz, 1988). the nonspeech control task. Thus the poor readers,
The study carried out by Lundberg and his asso- whether young or old, had no more difficulty in
ciates in Sweden (Lundberg et aI., 1980) is worthy segmental analysis than the good readers when
of special mention on two counts. It provides one the task was nonlinguistic; their problem was
of the most intensive examinations of the linguis- limited to the segmental analysis of speech.
tic abilities of the kindergartners. It is noteworthy Not only reading, but early spelling proficiency
as well because it also addresses the question of has also been found to be closely related to
whether the children's deficiency is, in fact, lin- analytic phonological skills. In a study of kinder-
guistic or whether it might be attributable to a gartners (Liberman, Rubin, Duques, & Carlisle,
deficiency in general analytic ability. Their 1985), the children's ability to produce invented
battery of 11 tests given to 200 kindergartners spellings, given only some knowledge of letter
included both linguistic and nonlinguistic tasks. names (see Read, 1986), was related to their
In the linguistic set were: (1) word synthesis tasks performance on a series of language-based tasks.
that varied in two dimensions of two levels each- It was found that the children's proficiency in
with or without memory load and using either spelling was more closely tied to phonological
phoneme or syllable units, and (2) word analysis awareness than to any of the other aspects of
tasks analogous to those for synthesis and, in language development tested. Of the eight tasks
addition, three others demanding analysis of in the study, only the three which unquestionably
phoneme position in words, reversal of phoneme tapped phonological analysis skills made a
segments in words, and rhyming. Since the lin- difference statistically. They combined to account
guistic tasks required the child to shift attention for 93 percent of the variance in proficiency in
from meaning to abstract form, thus possibly re- invented spelling. A phoneme analysis test
flecting a general cognitive function not exclu- patterned after Lundberg et aI. (1980) made the
sively limited to linguistic material, nonlinguistic largest contribution-67 percent ofth~ variance. A
control tasks that simulated those cognitive de- test of the ability to write letters to phoneme
mands were also included. The most powerful dictation accounted for 20 percent more, and a
predictors of later reading and writing skills in phoneme deletion task ("Say milk without the m")
the entire battery turned out to be those requiring added another 6 percent. (A fourth task, picture
phonological awareness, specifically the analytic naming, added 1 percent, but did not reach
ability to manipulate phonemes in words. In significance in the correlation. As we will note
contrast, the poor readers showed no particular later, naming can, however, be a subtle indicator
deficiency in the non-linguistic tasks. of more general phonological difficulties.)
These findings that tasks of linguistic, rather Among the four language-based tests that did
than nonlinguistic analysis, and specifically not contribute to the invented spelling perfor-
phonemic analysis, were predictive of reading mance were three that are frequently included in
failure have since received support from other clinical evaluations-receptive vocabulary, articu-
studies here and abroad. For example, in a study lation as measured by the repetition of simple
of six to nine-year-olds with severe reading words, and letter naming or writing. The fourth
disability (Morais, Cluytens, & Alegria, 1984) it was a syllable deletion test ("Say bookcase without
was found that these children were poorer on the book"). Being able to segment words by sylla-
segmenting words into their constituent 'parts but ble was, as we would expect, not enough to equip
performed just as well as normal readers in a the child to produce alphabetically written words.
matched task that required them to deal analyti-
cally with musical tone sequences instead of The remedial effect of training in
words. The question of a possible general analytic awareness
deficit was also addressed in two complementary Given the abundant evidence that phonological
experiments, one with good and poor readers in awareness is predictive of success in reading, it is
6 Liberman et al.

of interest to know that such awareness can be implications not only for preschool instruction but
trained even in preschool and kindergarten also for the design of literacy teaching programs
(Bradley & Bryant, 1983; Content, Morais, geared to adolescents and adults. This question
Alegria, & Bertelson, 1982; Lundberg, Frost, & was explored in an unusual investigation by a
Peterson, 1988; Olofsson & Lundberg, 1983). It is Belgian research group who examined the phono-
of special interest to find, moreover, that the logical awareness of illiterate adults in a rural
training can have a salutary effect on future area of Portugal (Morais, Cary, Alegria, &
reading skill. Impressive evidence for the efficacy Bertelson, 1979). They found that the illiterate
of the training comes from a pair of experiments adults could neither delete nor add phonemes at
by Bradley and Bryant (1983). The first experi- the beginning of nonsense words, whereas others
ment confirmed the high correlations found by from the same community who had received
others between preschoolers' phonological aware- reading instruction in an adult literacy class
ness and later reading skill. This was done by a succeeded in performing those tasks. The authors
comparison of children's performance on rhyming concluded that awareness of phoneme
tasks and their achievement in reading and segmentation does not develop spontaneously
spelling several years later. The second experi- even by adulthood but arises as a concomitant of
ment was directed to an examination of the effect reading instruction and experience.
of various kinds of early training on the later aca- In view of these findings, we believed it would
demic achievements of children considered to be prove of value to test the phonological awareness
at risk for failure. To this end, the children who of adults who had had reading instruction but
had been found in the first experiment to have a were nonetheless poor readers. To this end, our
low level of phonological awareness were divided reading research group (Liberman, et al., 1985)
into four groups. One group was trained to sort tested the members of a community literacy class,
pictured words on cards by phonological cate- all of whom were having serious decoding
gories. A second received the same training except problems despite years of schooling. What we
that letters corresponding to the phonemic cate- found was that these adults performed with
gories were added. A third group was trained to difficulty on a very simple task -in which the
sort by semantic categories. A control group was subjects were required only to identify the initial,
given equal time and unrelated card play. The two medial, or final sound in monosyllabic words.
phonologically trained groups were fourid to be Though this is an exercise that one might expect a
superior to the others in subsequent tests of read- first grader to be able to perform, our adults
ing and spelling. Moreover, in follow-up studies, managed to produce correct responses on only 58
they continued to maintain their advantage. percent of the items. Moreover, they clearly found
Further evidence for the positive effect of early it to be singularly frustrating and unpleasant.
training in phonological awareness is found in a This inability of adults with literacy problems to
longitudinal study by Lundberg and associates perform well on tasks demanding explicit
(Lundberg, et al., 1988). An experimental group of understanding of phonological structure has also
kindergartners who had participated in a variety been found by other investigators-Byrne and
of analytic word games was found at year-end to Ledez (1983) in Australia; Marcel (1980) in
be superior in phoneme awareness to a matched England; and Read and Ruyter (1985) in a prison
control group. When -compared in academic population in the U.S.
achievement in the first grade, the experimental
group was slightly below the controls in math and A BROADER PHONOLOGICAL
IQ but significantly superior in both reading and DEFICIENCY AND READING
spelling. Moreover their advantage was main- Why do some people have difficulty in achieving
tained when the children were retested in the sec- the understanding of phonological structure that
ond grade. application of the alphabetic principle requires?
One possibility, as we noted before, is that they
Lack of phonological awareness and adult may suffer from a general deficiency in the ability
literacy to divide objects of all kinds into their constituent
What about phonological awareness in adult elements. But as we pointed out, the results of
nonreaders? Is it still a problem for them? The several studies suggest that the difficulty is
question as to whether phonological awareness specifically linguistic. Another possibility-one
improves spontaneously with age or requires some that we and others have pursued-is that the poor
form of instruction is a crucial one, with obvious reader's difficulty with analyzing words into their
 The Alphabetic Principle and Learning to Read 7

constituent units is one among several symptoms message. In keeping with current usage (see
of a general deficiency in the phonological Baddeley & Hitch, 1974; also Shankweiler &
component of the child's natural capacity for Crain, 1986), we call this form of memory working
language. If the underlying biology tends to set up memory, a term which is used throughout this
phonological structures weakly, then it should book. A phonological deficiency would understand-
follow that these structures would be that much ably impair the functions of working memory and
harder for the child to bring to a level of explicit could be expected, in turn, to have repercussions
awareness. But there would be other conse- on comprehension, whether of spoken discourse or
quences for the processing of language, and these printed text. For example, in sentence processing,
we consider below. the parsing of phonological segments into lexical
units and the grouping of these units into higher-
Problems in short-term memory and level phrasal structures requires phased control of
sentence comprehension the flow of linguistic information through the
Because short-term memory depends on the language apparatus.
ability to gain access to phonological structure and We could therefore expect that children with
to use it to hold linguistic information (Conrad, reading disability would sometimes comprehend
1964; Liberman, Mattingly, & Turvey, 1972), we sentences poorly because of their difficulties -in
might expect people who have underlying setting up and retaining phonological structures.
phonological deficiencies to show various The difficulty should be especially acute in
limitations on verbal tasks that tap short-term reading, where the problem of decoding from print
memory. This expectation is amply borne out. would create an additional processing load in an
The research literature contains many reports unskilled reader who decodes poorly. The impor-
that young children who are poor readers are de- tant insight that the lower-level and higher-level
ficient in short-term memory. Typically they reading problems of the poor reader are causally
retain fewer items from a set of fixed size than connected through constriction of working memo-
age-matched good readers (see Mann, Liberman, ry was contributed by Perfetti and Lesgold (1979).
& Shankweiler, 1980; Shankweiler, Liberman, In their terms, poor decoding skills coupled with
Mark, Fowler, & Fischer, 1979; Shankweiler, the limitations ofworking memory create in the
Smith, & Mann, 1984; Wagner & Torgesen, 1987). poor reader a "bottleneck" in information flow
However, memory difficulties for poor readers with severe repercussions for comprehension.
appear to arise only under specific conditions; But, as Crain and Shankweiler discuss (Crain,
chiefly, they occur when the items to be retained 1989; Shankweiler, 1989), comprehension
are words and nameable objects. When the test difficulties of poor readers are not limited to read-
materials do not lend themselves to verbal (i.e., ing. From our working memory perspective,
phonological) encoding, as in memory for nonsense difficulties should also arise in spoken language
shapes or unfamiliar faces, memory testing does processing, especially if the sentence material con-
not find poor readers at a disadvantage (Katz, tains remote dependencies or structural ambigui-
Shankweiler, & Liberman, 1981; Liberman, ties that necessitate reanalysis, or if the compre-
Mann, Shankweiler, & Werfelman, 1982). The hension task presents additional complexities that
problem seems, therefore, to be a material-specific further dilute memory resources. Several reports
one, not an all-embracing memory impairment. in the literature indicate that disabled readers do
It is noteworthy, in addition, that memory have problems in comprehending such sentences
differences between good and poor readers may in spoken form as well as in reading (Byrne, 1981;
also depend on other demands of the task-tasks Mann, Shankweiler, & Smith, 1984; Stein, Cairns,
that require rote recall of a list of unstructured & Zurif, 1984). More recent findings (Shankweiler,
items may be less differentiating than tasks that 1989) indicate that the poor readers fail not
require both storage and further processing of the because they lag behind their good reading peers
incoming material as in sentence processing in comprehension of grammar as such, but
(Daneman & Carpenter, 1980; Perfetti & because working memory is overloaded due to
Goldman, 1976). Since language structures are deficient phonological processing.
hierarchically organized and sequentially trans- By changing the task in various ways to reduce
mitted, comprehension of language, either by ear the demands on memory while testing the same
or by eye, depends on a short-term memory grammatical structures, it has been shown that
system that transiently stores and continuously poor readers can succeed as well as good readers
processes the incoming segments of the linguistic in comprehending complex grammatical struc-
8 Liberman et al.
-----------------_-=.:..:.~=..:..:....:::.:.-_----------------

tures (Crain, Shankweiler, Macaruso, & Bar- mals on three different tasks in which their
Shalom, 1990; Smith, Macaruso, Shankweiler, & speech production was stressed. The author
Crain, 1989). Thus, a memory impairment concluded, as we would, that their difficultie s in
stemming from a weakness in phonological pro- speech production may be an extension of deft cits
cessing can masquerade as a grammatical or in the phonological realm.
semantic deficit. More evidence for a broad phonological defici tin
poor readers was provided by a study of the Per-
Other language-related problems formance of second graders on a naming test
Thus far we have discussed difficulties involving (Katz, 1986). This study confirmed what others
the phonological components of language that had shown-that poor readers named more Words
directly affect reading. Reading is affected both by incorrectly. But it went further to show that their
the difficulties of accessing and mentally difficulties are often phonological and not seman-
manipulating phonemic segments and by the tic, as might be assumed. Three kinds of evidence
limitations on use of the working memory that we were presented in this regard. first, when quizzed
have just discussed (though in the case of working about the characteristics of the object they had
memory the consequences are not confined to named incorrectly (e.g., "tornado" for volcano), the
reading). We now turn to other deficits displayed poor readers were often able to describe it accu-
by poor readers that are phonological in nature, rately. They clearly knew what the object Was.
but do not affect reading directly. These are worth That is, they described a volcano, not a tornado.
mentioning, both for their diagnostic value, and Second, given the name of the item, they could
because they add to the weight of the evidence select it from a group of pictured objects. That is,
that all the elements of the syndrome of many they could identify it correctly. And third, their
poor readers may stem ultimately from a naming errors were often related to the phonolog_
deficiency in phonological processing. ical and not the semantic aspect of the word. For
One such deficit is suggested by some example, though the name given to the picture of
preliminary research into the speech perception of a volcano was incorrect, it shared syllable count,
poor readers that was carried out by Brady and stress pattern, and vowels with the target word.
associates at Haskins Laboratories (Brady, Distorted production of the word for an item
Shankweiler, & Mann, 1983). In their experi- that had been correctly identified could stem
ments, good and poor readers were tested on two either from deficient specification of the phono-
auditory perception tasks, one involving words logical structure in the lexicon, or from deficient
and the other nonspeech environmental sounds. retrieval and processing of the stored phonological
The identification tasks were' presented under two information. In either case, the source of the
conditions-with favorable and unfavorable noise difficulty relates to the phonological structure of
ratios. The findings were that the poor readers did the words and not to their meanings.
show a deficit, but it was specific to the speech
stimuli and occurred only in the noise-masked
PHONOLOGY AND THE SUCCESSFUL
condition. They did not differ from the good DEAF READER
readers in the perception of nonspeech environ- The congenitally deaf constitute a population
mental sounds, whether the sounds were noise- with a phonological impairment arising from an
masked or not. Note that the poor readers entirely different source. Surprisingly, this group
apparently needed a higher quality of signal than represents some of the most compelling evidence
the good readers for error-free performance in for the importance of phonological abilities for
speech, but not for nonspeech environmental reading. It is well known that profound deafness
sounds. These results suggest that the minor from birth or early life usually results in attain_
deficit displayed by the poor readers may derive ment of a low level of reading skill. The hearing
from phonological structures that are set up more impaired of all ages tend to read far below grade
weakly than in good readers, or are more difficult expectations. But, nonetheless, differences in
to activate. reading achievement are related to differences in
Additional evidence for a broader phonological phonological abilities even in deaf populations.
deficit in poor readers is provided by a study of Moreover, a few congenitally, profoundly deaf
speech production, specifically, the errors ofjunior individuals can read well, even up to the college
high school students (Catts, 1986). The critical level.
finding was that the reading disabled students How are these successful deaf readers different
made significantly more errors than matched nor- from the majority? Vicki Hanson and her
 · The Alphabetic Principle and Learning to Read 9

associates at Haskins Laboratories asked that derstand how the internal phonological structure
question in a series of experiments (Hanson, 1982; of words relates to the orthography, or why it is
Hanson & Fowler, 1987; Hanson, Liberman, & hard for children to achieve this understanding.
Shankweiler, 1984) reported on by Hanson (1989). In fact, teachers are all too often being provided
Briefly, they found that the successful deaf with an instructional procedure that directs them
readers were not limited to reading English words specifically not to trouble the learner with details
as if they were logographs; that is, they were not, of how the alphabet works. Instead, they are told
as one might assume, dependent on a limited store to view reading as a "guessing game" (Goodman,
of words learned in paired associate fashion as 1976) in which the general import of the message,
visual designs. The results, which are discussed in and not the actual words of the text, is to be
detail by Hanson (1989) showed that unlike their emphasized. Beginners are encouraged to memo-
poor reading peers among the deaf, these subjects, rize the appearance of words as visual patterns by
despite so little exposure to sound, were able to whatever means they can muster and to use their
access that knowledge both in reading and in store of memorized words and their "whole
retaining verbal material in short-term memory. language" capability as a basis for guessing the
In reading, the good readers among the deaf rest of the message from picture cues and context.
displayed their phonological sensitivity by re- Thus, they are not to be corrected when reading
sponding differentially to rhyming and non- "kids" for children in a story about a playground,
rhyming pairs of words (save I wave vs. have Icave) "Crest" for toothpaste in a story about dental
and by being able to name the real word hygiene, and "cats" for dogs in a story about pets,
equivalents of nonwords (flame for f-l-a-i-m; tall Fortunately, many children-the lucky 75
for t-a-u-l). In a short-term memory experiment, percent or so who learn to read whatever the
the successful deaf readers were more affected by method-manage to pick up the alphabetic
phonetically confusing words than by those that principle without much explicit instruction, if any.
were orthographically confusing or whose signs That is, given experience with printed material,
were formation ally confusing. These results they begin to discover for themselves the common-
certainly suggest that successful deaf readers are alities between similarly spoken and written
using phonological processing, a conclusion also words. When tested in kindergarten, these
reached by Conrad with a less severely impaired children turn out to be the ones with strengths in
population of deaf readers (Conrad, 1979). the phonological domain. For the large group of
The question ofhow the congenitally, profoundly children with phonological deficiencies who do not
deaf might develop phonological sensitivity with- understand that the spoken word has segments,
out being able to hear the sounds of speech is ex- and who have not discovered on their own that
plored by Hanson (1989). She identifies several there is a correspondence between those segments
sources of information that may be helpful. The and the segments of the printed word, the current
orthography itself tells them something about the vogue for the so-called (and from our point of view,
systematic phonological forms of words. In addi- misnamed) psycholinguistic guessing game and its
tion, oral training when available supplies infor- offshoots, the "whole language" and "language
mation about the gestures used to produce speech. experience" approaches, are likely to be disas-
Lipreading also provides considerable useable in- trous. Many children taught this way are likely to
formation, and the deaf individual's own attempts join the ranks of the millions of functional illit-
at speech may reflect more phonological sensitiv- erates in our country who stumble along, guessing
ity than is apparent to the hearing listener. at the printed message from their inadequate
store of memorized words, unable to decipher a
IMPLICAnONS FOR INSTRUCfION new word they have never seen before.
In view of all the evidence that has accumulated For those beginners who do not discover the
in the past 15 years to support the critical impor- alphabetic principle unaided, an introductory
tance of phonological sensitivity for the attain- method that provides them with direct instruction
ment of literacy in an alphabetic system, one in what they need to know is critical (Liberman,
would surely expect teacher training to reflect 1985;· Liberman & Shankweiler, 1979). Direct
these findings. Unfortunately, all too often it does instruction could begin with language analysis
not. Many teachers are being trained to teach activities that are incorporated into the daily
reading without themselves ever having learned reading lesson. These activities can take many
how an alphabetic orthography represents the different forms, limited in number and variety
language, why it is important for beginners to un- only by the creativity of the teacher. The Auditory
10 Liberman et aI.

Discrimination in Depth Program of Lindamood reading skills even in an inner city school with a
and Lindamood (1975) is an ingenious method for high incidence of reading failure. They are out-
helping the student to apprehend the internal lined in Figure 1. In the first procedure, one origi-
phonological structure of words It does this by nally devised by the Soviet educator, Elkonin
calling the student's attention to the perceived (1973), Blachman presents the child with a simple
distinctiveness of the articulatory gestures for the line drawing representing the word to be ana-
various phonemic constituents of spoken words lyzed. A rectangle under the drawing is divided
and then demonstrating their sequences in into squares equal in number to the phonemes in
syllables with variously colored blocks. The the picture word. The children are taught to say
method was originally developed for individual the word slowly, placing a counter in the appro-
reading remediation, but is currently being adapt- priate squares of the diagram as the word is being
ed for classroom use. slowly articulated. The words selected must begin
Adaptations of three exercises that we advocated with a fricative, liquid, or nasal rather than a stop
some years ago (Liberman, Shankweiler, consonant in order to permit their component
Blachman, Camp, & Werfelman, 1980) have re- phonemes to be accessed readily. Later, as the
cently been shown by one of our colleagues child progresses, the counters are color-coded-
(Blachman, 1987) to be effective in improving one color for vowels, another for consonants.

ELKONIN (1973)

I ~I
ITO. ITO
ENGELMANN (1969) aI

SLINGERLAND (1971)

Figure 1. Language analysis activities (after Blachman, 1987 and Liberman et aL, 1980).
 The AlphJ:lbetu: Principle and Learning to Read 11

Letter symbols can be added as well. In another those few words they have already memorized.
activity, this one adapted from Engelmann (1969), They will not learn to decode new words-the
the children are taught how to read as a single essence of true reading skill-unless the method
unit the combination of a consonant followed by a initially includes more intensive, direct, and
vowel. For example, the teacher writes a conso- systematic training in phonological structure and
nant on the blackboard (preferably a fricative, demonstrates how it relates to the way words are
nasal, or liquid}-the letter s, for example-and written. Research support for this view has been
pronounces it, holding it over time until she writes available for at least twenty years (see Chall, 1967
the vowel and pronounces that. The length of time or Pflaum, Walberg, Karegianes, & Rasher, 1980).
between the pronunciations of the initial It is surely time to put the research into practice.
consonant and the vowel (as well as a line drawn
between them on the board) is then reduced step REFERENCES
by step until the two phonemes are pronounced as Alegria, J., Pignot,E., & Morais, J. (1982). Phonetic analysis of
a single unit-"sa." By adding stop consonants in speech and memory codes in beginning readers. Memory and
the final position and pronouncing the resultant Cognition, 10, 451-456.
Baddeley, A D., & Hitch, G. (1974). Working memory. In G. H.
words, the children can begin to accumulate a pool Bower (Ed.), The psychology of learning and activation (Vol. 4).
of real words (sag, sat, sad, etc.). Thereafter, new New York: Academic Press.
vowels and new consonants can be introduced in Benton, A. L., & Pearl, D. (1978). Dyslexia: An appraisal of current
the same way, and built into new words which are knowledge. New York: Oxford University Press.
incorporated into stories to be read and written. Blachman, B. (1984). The relationships of rapid naming ability and
language analysis skills to kindergarten and first grade reading
A similar effect can be produced by a third achievement. Journal of Educational Psychology, 76, 610-622.
procedure, adapted from Slingerland (1971), in Blachman, B. (1987). An alternative classroom reading program
which a small pocket chart is used by the child at for learning disabled and other low-achieving children. In W.
each desk to manipulate individual letters to form Ellis (Ed.), Intimacy with language: A forgotten basic in teacher
new words and learn new phonemes. The words education. Baltimore: The Orton Dyslexia Society.
Bradley, L., & Bryant, P. E. (1983). Categorizing sounds and
thus constructed, along with a few nonphonetic learning to read-a cauSal connection. Nature, 301, 419-421.
"sight" words, can be used in stories and poems to Brady, S. A, Shankweiler, D., & Mann, V. A (1983). Speech
be read and written by the child. Note that the perception and memory coding in relation to reading ability.
child is now reading and writing words the Journal of Experimental Child Psychology, 35, 345-367.
structure of which is no longer a mystery and the Browman, C. P., & Goldstein, L. M. (1985). Dynamic modeling of
phonetic structure. In V. Fromkin (Ed.), Phonetic linguistics.
understanding of which can be used productively New York: Academic Press.
to form related words (bag, bat, bad, big, bit, bid, Byrne, B. (1981). Deficient syntactic control in poor readers: Is a
etc.). weak phonetic memory code responsible? Applied
All these language analysis activities and others Psycholinguistics,2,201-212.
like them can be played as games in which the Byrne, B., & Ledez, J. (1983). Phonological awareness in reading-
disabled adults. Australian Journal of Psychology, 35,185-197.
introduction of each new element not only informs Catts, H. W. (1986). Speech production/phonological deficits in
but delights. Beginning readers with adequate reading disordered children. Journal of Learning Disabilities,
phonological ability will require only a relatively 19(8), 504-508.
brief exposure to such activities. They will soon Chall, J. (1967). Learning to read: The great debate. New York:
develop skills that will enable them to decode the McGraw Hill.
Coltheart, M. (1978). Lexical access in simple reading tasks. In G.
new words of the text and to go from them to the Underwood (Ed.), Strategies in information processing. London:
meaning of the passage. For such readers, Academic Press.
language analysis can be quickly followed, or even Conrad, R. (1964). Acoustic confusions in immediate memory.
accompanied, by practice with interesting reading British Journal of Psychology, 55,75-84.
materials from other sources. These children will Conrad, R. (1979). The deaf child. London: Harper & Row.
Content, A., Morais, J., Alegria, J., & Bertelson, P. (1982).
benefit from the added skill that comes from Accelerating the development of phonetic segmentation skills
increased reading practice and the further in kindergartners. Cahiers de Psychologie Cognitive, 2, 259-269.
enhancement of vocabulary and knowledge that Cossu, G., Shankweiler, D., Liberman, I. Y., Tola, G., & Katz, L.
comes with expanded reading and life experience. (1988). Awareness of phonological segments and reading
. But unless they receive extra assistance, the ability in Italian children. Applied Psycholinguistics, 9,1-16.
Crain, S. (1989). Why poor readers misunderstand spoken
many beginners with weakness in phonological sentences. In D. Shankweiler & I. Y. Liberman (Eds.), Phonology
skills, who may include as many as 20 to 25 and reading disability: Solving the reading puzzle (pp. 133-165).
percent of the children, will remain locked into a !ARLD Research Monograph Series. Ann Arbor: University of
sight-word stage of reading, able to cope only with Michigan Press.
12 Liberman et ai.

Crain,S., Shankweiler, D., Macaruso, P., & Bar-Sholom, E. (1990). Liberman, A. M., & Mattingly, I. G. (1989) A specialization for
Working memory and sentence comprehension: Investigations speech perception. Science, 243, 489-494.
of children with reading disorder. In G. Vallar & T. Shallice Liberman, A. M., Mattingly, I. G., & Turvey, M. T. (1972).
(Eds.), The neuropsychology of memory disorder. Cambridge, Language codes and memory codes. In A. W. Melton & E.
England: Cambridge University Press. Martin (Eds.), Coding processes in hUTniln memory. New York:
Daneman, M., & Carpenter, P. A. (1980). Individual differences in Winston.
working memory and reading. Journal of Verbal Learning and Liberman, I. Y. (1971). Basic research in speech and lateralization
Vmal Behavior, 19, 450-466. of language: Some implications for reading disability. Bulletin of
Doehring, D. G. (1968). Patterns of imptlirment in specific reading the Orton Society, 21, 71-87.
disability. Bloomington: Indiana University Press. Liberman, I. Y. (1973). Segmentation of the spoken word and
E1konin, D. B. (1973). U. S. S. R. In J. Downing (Ed), Comparative reading acquisition. Bulletin of the Orton Society, 23, 65-77.
reading. New York: MacMillan. Liberman, I. Y. (1983). A language-oriented view of reading and
Engelmann, S. (1969). PreTJenting failure in the priTnilry grades. its disabilities. In H. Myklebust (Ed.), Progress in learning
Olicago: Science Research Associates. disabilities (Vol. 5). New York: Grune & Stratton.
Fox, B., & Routh, D. K. (1980). Phonetic analysis and severe Liberman, I. Y. (1985). Should so-called modality preferences
reading disability in children. Journal of Psycholinguistic determine the nature of instruction for children with learning
Research, 9, 115-119. disabilities? In F. H. Duffy & N. Geschwind (Eds.), Dyslexia: A
Gleitman, L. R, & Rozin, P. (1977). The structure and acquisition neuroscientific approach to clinical eTJaluation. Boston: Little,
of reading. Relations between orthographies and structure of Brown.
language. In A. S. Reber & D. L. Scarborough (Eds.), Toward a Liberman, I. Y., Mann, V., Shankweiler, D., & Werfelman, M.
psychology of reading Hillsdale, NJ: Erlbaum. (1982). Children's memory for recurring linguistic and
Goldstein, D. M. (1976). Cognitive-linguistic functioning and nonlinguistic material in relation to reading ability. Cortex,18,
learning to read in preschoolers. Journal of Experimental 367-375.
Psychology, 68, 680-688. Liberman, I. Y., Rubin, H., Duques, 5., & Carlisle, J. (1985).
Goodman, K. S. (1976). Reading: A psycholinguistic guessing Linguistic abilities and spelling proficiency in kindergartners
game. In H. Singer & R B. Ruddell (Eds.), Theoretical models and and adult poor spellers. In J. Kavanagh & D. Gray (Eds.),
processes of reading. Newark, DE: International Reading Biobehavioral Measures of Dyslexia. Parkton, MD: York Press.
Association. Liberman, I. Y., & Shankweiler, D. (1979). Speech, the alphabet
Hanson, V. L. (1989). Phonology and reading: Evidence from and teaching to read. In L. B. Resnik & P. A. Weaver (Eds.),
profoundly deaf readers. In D. Shankweiler & I. Y. Liberman Theory and practice ofearly reading. Hillsdale, NJ: Erlbaum.
(Eds.), Phonology and reading disability: Solving the reading puzzle Liberman, I. Y., & Shankweiler, D. (1985). Phonology and the
(pp. 69-89). IARLD Research Monograph Series. Ann Arbor: problems of learning to read and write. Remedial and Special
University of Michigan Press. Education, 6, 8-17.
Hanson, V. L (1982). Short-term recall by deaf signers of Liberman, I. Y., Shankweiler, D., Blachman, B., Camp, L., &
American Sign Language: Implications of encoding strategy for Werfelman, M. (1980). Steps toward literacy. In P. Levinson &
order recall. Journal of Experimental Psychology: Learning, C. Sloan (Eds.), Auditory processing and language: Clinical and
Memory, and Cognition, 8, 572-583. research perspectiTJes. New York: Grune and Stratton.
Hanson, V. L., & Fowler, C. A. (1987). Phonological coding in Liberman, I. Y., Shankweiler, D., Fischer, F. W., & Carter, B.
word reading: Evidence from hearing and deaf readers. (1974). Explicit syllable and phoneme segmentation in the
Memory and Cognition, 15(3), 199-207. young child. Journal of Experimental Child Psychology, 18,201-
Hanson, V. L., Liberman, I. Y., & Shankweiler, D. (1984). 212.
Linguistic coding by deaf children in relation to beginning Lindamood, C. H., & Lindamood, P. C. (1975). DIM Teaching
reading success. Journal of Experimental Child Psychology, 37, Resources. Allen, Texas.
398-393. Lundberg, I., Olofsson, A., & Wall, S. (1980). Reading and spelling
Helfgott, J. (1976). Phoneme segmentation and blending skills of skills in the first school years, predicted from phonemic
kindergarten children: Implications for beginning reading awareness skills in kindergarten. Scandinavian Journal of
acquisition. Contemporary Educational Psychology, 1, 157-169. Psychology, 21, 159-173.
Katz, R. B. (1986). Phonological deficiencies in children with Lundberg, I., Frost, J., & Petersen, O-P. (1988). Effects of an
reading disability: Evidence from an object-naming task. extensive program for stimulating phonological awareness in
Cognition, 22, 25-257. preschool children. Reading Research Quarterly, 23/3, 263-284.
Katz, R B., Shankweiler, D., & Liberman, I. Y. (1981). Memory for Marcel, A. (1980). Phonological awareness and phonological
item order and phonetic recoding in the beginning reader. representation: Investigation of a specific spelling problem. In
Journal of Experimental Child Psychology, 32, 474-484. U. Frith (Ed.), Cognitive processes in spelling. London: Academic
Klima, E. S. (1972). How alphabets might reflect language. In J. F. Press.
Kavanagh & I. G. Mattingly (Eds.), Language by ear and by eye: Magnusson, E., & Nauder, K. (1987). Language disordered and
The relationships between speech and reading. Cambridge, MA: normally speaking children's development of spoken and
MIT Press written language: Preliminary results from a longitudinal
Liberman, A. M. (1989). Reading is hard just because listening is study. Reports from Uppsala University, Linguistics Department,
easy. In C. von Euler, I. Lundberg, & G. Lennerstrand (Eds.), 16,35-63.
Wenner-Gren International Symposium Series 54: Brain and Mann, V., Liberman, I. Y., & Shankweiler, D. (1980). Children's
Reading. Hampshire, England: Macmillan. memory for sentences and word strings in relation to reading
Liberman, A. M., Cooper, F. 5., Shankweiler, D. P., & Studdert- ability. Memory and Cognition, 8, 329-335.
Kennedy, M. (1967). Perception of the speech code. Mann, V., Shankweiler, D., & Smith, S. (1984). The association
Psychological Review, 74, 431-461. between comprehension of spoken sentences and early reading
Liberman, A. M., & Mattingly, I. G. (1985). The motor theory of ability: The role of phonetic representation. Journal of Child
speech perception revised. Cognition, 21,1-36. Language, 11, 627-643.
 The Alphabetic Principle and Learning to Read 13

deManrique, A. M. B., & Gramigna, S. (1984). La segmentacion Journal of Experimental Psychology: Human Learning and Memory,
fonologica y silabica en ninos de preescolar y primer grado. 5,531-545.
Lectura y Vida,S, 4-13. . Shankweiler, D., Smith, S. T., & Mann, V. (1984). Repetition and
Mattingly, 1. G. (1972). Reading, the linguistic process, and comprehension of spoken sentences by reading disabled
linguistic awareness. In J. F. Kavanagh & I. G. Mattingly (Eds.), children. Brain and Language, 12, 241-257.
Language by ear and by eye: The relationships between speech and Slingerland, B. H. (1971). A Multisensory Approach to Language Arts
reading. Cambridge, MA: MIT Press. for Specific Language Disability Children: A Guide for Primary
Morais, J., Cary, L., Alegria, J., & Bertelson, P. (1979). Does Teachers. Cambridge, MA: Educators Publishing Service.
awareness of speech arise spontaneously? Cognition, 7, Smith, F. (1971). Understanding reading: A psycholinguistic analysis
323-331. of reading and learning to read. New York; Holt, Rinehart, &
Morais, J., Ouytens, M., & Alegria, J. (1984). Segmentation Winston.
abilities of dyslexics and normal readers. Perceptual and Motor Smith, S. T., Macaruso, P., ShankweiIer, D., & Crain, S. (1989).
Skills, 58, 221-222. Syntactic comprehension in young poor readers. Applied
Olofsson, A., & Lundberg, I. (1983). Can phonemic awareness be Psycholinguistics, 10, 429-454.
trained in kindergarten? Scandinavian Journal of Psychology, 24, Stanovich, K. E. (1982). Individual differences in the cognitive
35-44. processes of reading: I. Word coding. Journal of Learning
Perfetti, C. A., & Goldman, S. R. (1976). Discourse memory and Disabilities, 15, 449-572.
reading comprehension skill. Journal of Verbal Learning and Stein, C. L., Cairns, H. 5., & Zurif, E. B. (1984). Sentence
Verbal Behavior, 14, 33-42. comprehension limitations related to syntactic deficits in
Perfetti, C. A., & Lesgold, A. M. (1979). Coding and reading disabled children. Applied Psycholinguistics, 5, 305-322.
comprehension in skilled reading and implications for reading Studdert-Kennedy, M. (1987). The phoneme as a perceptuo-motor
instruction. In L. B. Resnick & P. A. Weaver (Eds.), Theory and structure. In A. Allport, D. MacKay, W. Prinz, & E. Sheerer
practice ofearly reading, Volume 1. (Eds.), Language perception and production. London: Academic
Pflaum, S. W., Walberg, H. J., Karegianes, M. 1., & Rasher, S. P. Press.
(1980). Reading instruction: A quantitative analysis. Educational Treiman, R. A., & Baron, J. (1981). Segmental analysis ability:
Research, 9, 12-18. Development arid relation to reading ability. In G. E.
Pratt, A. (1985). The relationship of lingUistic awareness to reading MacKinnon & T. G. Walker (Eds.), Reading research: Adt'llnces in
skill in children and adults. Unpublished doctoral dissertation, theory and practice, 3. New York: Academic Press.
University of Rhode Island. Vellutino, F. R. (1979). Dyslexia: Theory and research. Cambridge,
Read, C. (1986). Children's creatiue spelling. London: Routledge & MA: MIT Press.
Kegan Paul. Vellutino, F. R., & Scanlon, D. (1987). Phonological coding and
Read, c., & Ruyter, 1. (1985). Reading and spelling skills in adults phonological awareness and reading ability: Evidence from a
of low literacy. Reading and Special EducatiofJ, 6, 43-52. longitudinal and experimental study. Merrill-Palmer Quarterly,
Savin, H. (1972). What the child knows about speech when he 33/3,321-363.
starts to learn to read. In J. F. Kavanagh & 1. G. Mattingly Vernon, M. D. (1957). Backwardness in reading, a study of its nature
(Eds.), Language by ear and by eye: The relationships between speech and origin. Cambridge, England: Cambridge University Press.
and reading. Cambridge, MA: MIT Press. Wagner, R. K., & Torgesen, J. K. (1987). The nature of
ShankweiIer, D. (1989). How problems of comprehension are phonological processing in the acquisition of reading skills.
related to difficulties in decoding. In D. Shankweiler & 1. Y. Psychological Bulletin, 101, 192-212.
Liberman (Eds.), Phonology and reading disability: Soluing the Waters, G. 5., Seidenberg, M.S., & Bruck, M. (1984). Children's
reading puzzle (pp. 35-68). IARLD Research Monograph Series. and adult's use of spelling-sound information in three reading
Ann Arbor: University of Michigan Press. tasks. Memory & Cognition, 12, 293-305.
ShankweiIer, D., & Crain, S. (1986). Language mechanisms and
reading disorders: a modular approach. Cognition, 24,139-168. FOOTNOTES
ShankweiIer, D., & Liberman, 1. Y. (1972). Misreading: A search 'In D. Shankweiler & 1. Y. Liberman (Eds.), Phonology and
for causes. In J. F. Kavanagh & 1. G. Mattingly (Eds.), Language Reading Disability: Soluing the Reading Puzzle. IARLD Research
by ear and by eye: The relationships between speech and reading. Monograph Series. Ann Arbor: University of Michigan Press.
Cambridge, MA: MIT Press. (1989).
Shankweiler, D., Liberman, I. Y., Mark, L. 5., Fowler, C. A., & t Also University of ConnE:<:ticut:Storrs.
Fischer, F. W. (1979). The speech code and learning to read. ttAlso Yale University and University of Connecticut.