new perspectives

on reading and writing

processes
Emilia Ferreiro and Margarita Gomez
Palacio (compilers)
twenty-first century publishers, sa of cv
WATER HILL 248, EGACIN COYOACAN, 04310. MEXICO CITY, DF

21st century publishers argentina, sa

LAVALLE 1634 FLOOR 11-A C-1048AAN, BUENOS APES. ARGENTINA
Cover by Marta Luisa Martinez Passarge

First edition, 1982 Seventeenth edition, 2002

C century 21st editors, sa of cv
isbn 968-23-1600-6 rights reserved according to law
printed and made in mexico/printed and made in mexico
 INDEX

PRESENTATION

1THE READING PROCESSES

LANGUAGES AND DEVELOPMENT, by KENNETH S. GOODMAN

(University of Arizona) 1
15 NON-ACQMQDATI-COGNITIVE GYAS

READING, by WALTBB H. McGINITIE, KATHERINE MARIA and

SUSAN KIMMEL (University of Victoria, Canada)29
PREDICTABILITY: A I 20’23 RSAL IN READING AND WRITING, by
JEROME v. HARSTE and Carolyn L. burke (University of
Indiana)50 SCHOOL DYSLEXIA AND EXPERIMENTAL DYSLEXIA, by MARIA
A.
CARBONELL DE GROMPONE (Uruguay) 68
SUMMARY OF THE DISCUSSION 85

II THE WRITING PROCESSES

THE DEVELOPMENT OF WRITING: ADVANCES. PROBLEMS AND
PERSPECTIVES, by HERMINE SINCLAIR (UNIVERSITY OF GI
nebra)93
WRITING DEVELOPMENT IN VERY YOUNG CHILDREN, by YETTA GOODMAN
(UNIVERSITY of Arizona) 107
OR
TURA, by EMILIA FERREiRO (Research and Education Center)
$,30ta..........................................................................................................................................................................................................................................................................53
2/-KlVLLLlLLLC<%r: c,.......................................................................................................................................................................................................................................53
428>Mp>.- •J,bAax Wh -....................................................................................................................................................................................................................................53
'(or......................................................................................................................................................................................................................................................................................55
' AEG/.............................................................................................................................................................................................................................................................................55

III LITERACY IN ITS SOCIAL AND SCHOOL CONTEXT

[7]
 8 INDEX

WRITTEN LANGUAGE IN SCHOOL CONTEXTS, by COURTNEY B. CAZDEN (Harvard

University)207
THE INFLUENCE OF SCHOOL ON LEARNING TO READ. by JOHN DOWNING
(University of Victoria, Canada)
WRITING THE SUNDAY LANGUAGE.by CLAIRE BLAN- CHE-BENVENISTE
(University of Provence)____________________________________________
READING AND WRITING AS A CULTURAL PRACTICE.by ALONZO B.
230
ANDERSON and WILLIAM h. TEALE (University of California, San Diego)
SCHOOL USES OF WRITTEN LANGUAGE, by ELSIE ROCKWELL (Center for
247
Research and Advanced Studies, IPN, Mexico)
READING AND WRITING IN INTERACTIONS: A SEARCH
OF THE DIMENSIONS AND MEANINGS IN THE SOCIAL CONTEXT, BY 271
CLAIRE A. WOODS (University of Pennsylvania) DISCUSSION SUMMARY
296

321
346
PRESENTATION

What kind of activity is this we call reading? How do we come to

understand the system of marks that make up our alphabetic writing?
How can we make literacy closer to the real process and more -
effective? These are some of the questions on which the works
collected in this volume focus. They were presented in Mexico City, at
the International Symposium on New Perspectives in Reading and
Writing Processes, organized by the General Directorate of Education
Special session of the Secretariat of Public Education, which took
place from July 1 to 4, 1981.
The publication and dissemination of these works is fully
justified, since literacy remains a central educational issue in Latin
America, and it is necessary to deepen our understanding of the nature
of the processes involved in order to clarify our teaching practice. On
the other hand, it is precisely a topic that has acquired new dimensions
in recent years, thanks to the
a group of researchers who have completely renewed our vision
on the subject, some of whom were present at that Symposium.
An interdisciplinary field par excellence, literacy requires being
viewed from different angles. That is why we brought together
psychologists, psycholinguists, anthropologists and linguists at this
Symposium, who spoke about learning processes, about social
practices, and not about teaching methods. Despite the differences in
perspectives, everyone agrees that it is not possible to reduce the
acquisition of written language—at any point in the process—to
perceptual-motor skills; that the intervention of the subject's linguistic
competence and cognitive competence are determining factors.
The discussions among the participants were as interesting as the
presentations themselves. We cannot publish them in full, but we
decided to summarize them, because they add precious elements to the
reflection on the subject.
Yo

READING PROCESSES

19]
THE READING PROCESS: CONSIDERATIONS ACROSS
LANGUAGES AND DEVELOPMENT

KENNETH S. GOODMAN

INTRODUCTION

For several years, one of the main goals of my work has been the
development of a theory of reading processes. This theory is based on
my research on errors in oral reading in English. From my studies with
many readers of all ages, I have characterized reading as a
“psycholinguistic guessing game” (Goodman, 1967). It is a process in
which thought and language are involved in continuous transactions *
when the reader tries to obtain
meaning from the printed text. I currently think that

not only to English but to reading in all languages. In this tra

Below I will defend the point of view of reading from a universal,
multilingual perspective, and considering development.
Before doing so, however, it may be useful to put my theoretical
perspective into the historical context of American education. Huey, a
pioneer of psychology, recognized the complexity of reading in
psychological terms at the beginning of the 20th century (Huey, 1908).
He viewed reading as being essentially a search for meaning and as
being constructive. His work influenced thinking about reading in the
first quarter of the present century but fell into disregard as attention
shifted to the development of a reading technology centered around the
beginning reader.
This reading technology also had an impetus to
beginning of the century in the work of the Committee on Economy of
Time, a national committee dedicated to efficient show off more
elementary and secondary education at a time
when education in the United States was progressing toward becoming
truly universal (nsse, 1915-19). For the first time in a nation's history,
virtually every school-age child would be in school. The committee's
goal was to combine a time-efficient curriculum with a scientifically -
based methodology. In the 1920s, there was great optimism that

* See clarification of this term in the discussion on this topic (pp. 86 et

seq.).

[13]
14 KENNETH S. GOODMAN

science would provide solutions to all educational problems, including

universal literacy. This “scientism” coincided with the rise of
behaviorism in psychology and logical empiricism in philosophy.
Specificity, sequence, and quantification became the keywords of
American education.
In the field of reading, William S. Gray played a decisive role in
the work that resulted from the committee's efforts and in the
development of the initial reading (Gray, 1925, 39). The technology
that centered around initial reading had these main characteristics:
1] Production of graded materials, beginning with preparatory
exercises in preschool and first grade and continuing through sixth or
eighth grade.
2] Focus on controlled vocabulary. Word lists based on word
frequency studies such as the Thorndike study were used. First grade
texts, and later preschool texts, introduced only a few words at a time
and repeated them frequently. Each subsequent book in the series
repeated the
use of the words previously introduced a lucid and added to the
few others. Editors competed with each
other to see who could produce more material with fewer words. Each
book ended with a list of the words used.
Teachers judged progress by students' ability to recognize words out of
context.
3| Production of workbooks to practice skills and provide word
exercises. Millions of copies of these workbooks were sold each year.
4| Development of skill hierarchies, starting with phonic skills, the
relationships between spelling and phonology. Word attack skills were
systematically cultivated through exercise books. Selected readings in
the student books focused the children's attention on the skills and -
words being taught.
THE READING PROCESS 15

5] The reading program was divided into developmental readers,

designed to take students through a normal progression, and remedial
readers, designed to teach skills to those who had not been successful
in the developmental program.
6] Testing became an important and eventually dominant aspect
of the technology. They were developed to measure success and
diagnostic tests to determine deficiencies for remedial programs. As
technology developed, tests increasingly became tests of skill mastery,
measured outside of reading meaningful texts. Less and less
importance was given to reading meaningful texts in basic educational
programs while skill-drills acquired central importance. This was all
the more true as public interest and school policy centered around
average scores on reading achievement tests.
tests of 00 hardness to predict success, performance tests
We can summarise this by saying that a systematic technology for
teaching reading was growing, based on a controlled vocabulary and
the development of a hierarchy of skills. Reading instruction became a
central theme of the curriculum from elementary grades through high
school. The use of tests with a strong emphasis on component skills
was dominant.
If there was a theory of reading processes implicit in this
technology, it was this: reading is identifying words and putting them
together to create meaningful texts. Learning to read was considered as
mastering the ability to recognize words and acquire a vocabulary of
sight words, words known by sight. The main controversy—phonic -
approaches. global— is located within this perspective, since the
importance of words is not discussed, but rather the best way to
identify them.
My own work began at the time when linguistic science was
gradually turning its attention from sounds to syntax and when
linguistic theory took up linguistic competence as its subject.
Psycholinguistics was emerging as an interdisciplinary bridge between
cognitive psychology and linguistics, whose purpose was the study of
the interaction between thought and language. It had become
necessary, in my opinion, to understand the reading process and its
development so that we could examine our educational technology; I
thought we needed a psycholinguistic perspective.
These are some of the key convictions underlying this perspective.
In a literate society there are two forms of language—oral and written
—that run parallel to each other. Both are fully capable of achieving
communication. Both forms have the same underlying grammar and
use the same rules to relate their underlying structure to surface
representation, oral or written. (In some cases, as in Arabic, the written
language is based on a classical form or dialect and therefore its
16 KENNETH S. GOODMAN

grammar is not identical to the spoken dialect of any reader of that -

language.) What differentiates the spoken language from the written
language are mainly the circumstances of use. We use oral language
primarily for immediate face-to-face communication, and written
language to communicate across time and space.
Each form has a productive process and a receptive process.
Speaking and writing are productive or expressive. Reading and -
listening are receptive. But both processes are processes in which
meaning is actively exchanged. When using language productively or
receptively, trans-actions between thought and language take place. So
to speak, speaking, writing, listening and reading are psycholinguistic
processes.
Linguistic processes are both personal and social. They are
personal because they are used to satisfy personal needs. They are
social because they are used to communicate between people. To the
extent that languages are social, they are all constrained by the same
need to be comprehensible to others, even though they may have basic
differences from each other. There are only limited ways in which
languages can vary and still be comprehensible. Therefore, they are
similar in their purposes and limitations.
Furthermore, although written forms of languages may relate to
the oral form of the same language in different ways, they must all
fully represent meaning in some comprehensible way that does not
depend on the possibility of converting it to its oral counterpart. Some
use alphabetic writing. Some do not represent vowels. Some use
symbols to represent syllables. Some represent ideas directly. But
written languages are not modes of representation of oral language;
they are alternative and parallel forms of oral language as modes of
representing meaning. If they could only be understood by conversion
to oral language, then they would not serve the special purposes for
which written language is needed, namely communication across time
and space. Silent reading is much faster than speech because readers
understand the meaning directly from the written text.

A SINGLE READING PROCESS

Written language, like oral language, is a social invention. When a

society needs to communicate across time and space and when it needs
to remember its heritage of ideas and knowledge, it creates a written
language. This occurs when societies reach a certain level of
complexity and size. Since the purposes of written language are
basically the same across languages, and the need to be understood by
others is universal across languages, I believe there is one and only
THE READING PROCESS 17

one reading process for all languages, regardless of differences in

spelling. There are not many ways to make sense of a text, only one.
For the same reasons, I believe that there is only one reading
process, regardless of the level of ability with which this process is
used. The difference between a capable reader and a non-capable
reader, or a beginner, does not lie in the process by which they derive
meaning from the text. There is no different way in which poor readers
make sense of text, when compared to good readers. The difference
lies in how well each reader uses this single process.
I also believe that there is a single reading process for reading any
type of text, regardless of its structure and the reader's purpose at the
time of reading. This single reading process must be flexible enough to
allow for differences in the structures of languages that differ in their
orthographies, in the characteristics of different types of texts, and in
the abilities and purposes of readers.
We can compare reading to driving a car or truck. There are small
cars, large cars, old cars, new cars; trucks, buses; heavy traffic, light
traffic. All of these differences require flexibility on the part of the
driver. And yet there is only one way to drive. You can drive well or
badly but you cannot drive without using the accelerator, brakes and
steering wheel. Somehow you have to make the car move forward,
stop, and go where you want to go.
From the In this way, although flexibility is needed in reading,
the process has essential characteristics that cannot
vary. It must begin with a text in some graphic form; the text must be
processed as language, and the process must end with the construction
of meaning. Without meaning there is no reading, and readers cannot
achieve meaning without using the process.

THE READING PROCESS

To understand the reading process, we must understand how the

reader, the writer, and the text contribute to it.
Since, as we have said, reading involves a transaction between the
reader and the text, the characteristics of the reader are as important
for reading as the characteristics of the text (Rosenblatt, 1978).
The relative ability of a particular reader is obviously important to
the successful use of the process. But so is the reader's purpose, social
culture, prior knowledge, linguistic control, attitudes, and conceptual
schemes. All reading is interpretation and what the reader is able to
understand and learn through reading depends strongly on what the
18 KENNETH S. GOODMAN

reader knows and believes before reading. Different people reading the
same text will vary in what they understand of it, depending on their
personal contributions to the meaning. They can only interpret based
on what they know.
We all speak at least one dialect of a language and sometimes
several registers or ways of using the language in different contexts.
The forms of language that the reader controls will strongly affect his
or her reading.
The way the text represents the writer will also affect how reader
and writer can transact meaning through it. The writer's sensitivity to
his audience and the way he has managed to convey meaning to this
audience will influence comprehensibility.
The success of the reading will also depend on the way in which
the reader and the writer agree on the ways of using language, on their
conceptual schemes, and on their life experiences. When writing a
letter to a close friend, many things can be taken for granted, whereas
in a business letter you need to be much more complete and explicit.
An important difference between oral and written language is that
in written language the two people communicating are rarely in the
presence of each other. Readers are thus left to construct meaning from
the text in the absence of the writer. We cannot turn to the writer
co II or we can do it towards the speaker and ask him “what did he
mean?” It is a long-distance transaction between the reader and
the writer. The reader must rely solely on the text to construct
meaning.

TEXT FEATURES

To understand what readers do we must understand the characteristics

of the texts with which readers are transacting. The text has a graphic
form, scattered across pages of paper. It has spatial dimensions, such
as size and directionality. English and Spanish are written from left to
right and from top to bottom. But Arabic and Hebrew are written from
right to left. Chinese and Japanese can be written vertically from top to
bottom or horizontally from left to right.
I am now going to define spelling more broadly than usual. In an
alphabetic system, the orthography consists of a system of graphemes,
usually called letters, that have a variety of shapes and styles.
Orthography also includes the spelling and punctuation rules by which
letters can be combined to represent the phonetic, morphophonemic,
morphemic, syntactic and pragmatic systems of language. In languages
THE READING PROCESS 19

that use the Roman alphabet, this orthography must have ways of
representing the particular sounds of the language. Thus the alphabet
must be modified slightly from language to language. But orthography
must also be able to represent the ways in which the sounds of
language are combined and the way they are modified by surrounding
sounds. In doing so, there are necessary choices about the level of
language to be represented. Sometimes we must choose words that
seem similar because they sound similar, or because they are related at
the grammatical level or by meaning. Spelling should indicate through
punctuation the sentence, phrase, and clause patterns that represent -
meaning.
The orthography should be useful to speakers of the language
whose dialects differ in their phonology. This means that you must
deviate from a consistent representation of the sounds in any of the
dialects. Many languages standardize orthography across dialects so
that the writing is understandable to speakers of a wide range of
dialects. Spellings do not fit all dialects equally (e.g., street, tortilla, -
horse, much, in Spanish).
The phonological and orthographic systems have different
limitations such that there can never be a complete one-to-one
correspondence between their units. The written text has a syntactic
structure. It must represent the syntax of the language to be
understandable. The score, the order of
sentences, grammatical suffixes, are what readers the indices that
use to get to the syntax. Written language
sometimes tends toward more formal syntactic structures than oral
language, perhaps because it is preserved after being produced. On the
other hand, the sentence structure of written language is often less
complex than in spoken language because punctuation is a less
comprehensive system than intonation for avoiding syntactic
ambiguity. Thus, writers use less complex structures to avoid
ambiguity.
Written texts also have semantic structures.
cas. Although stories vary considerably, there are only a limited
number of structures they can have. This is partly because of the
nature of the meaning; partly it is a cultural fact, and partly it is a
custom. These semantic structures make them predictable to readers. A
common story structure has a series of events that constitute a
problem, a central event, and then a resolution to the problem. Texts
also have cohesive resources that link the text and provide it with
unity. For example, the repetition of the same word and its synonyms
form a cohesive chain. All of these text features are used by the reader
when making predictions and inferences in the construction of
20 KENNETH S. GOODMAN

meaning.
READING STRATEGIES

The reading process employs a series of strategies. A strategy is a

broad framework for obtaining, evaluating, and using information.
Reading, like any human activity, is intelligent behavior. People do not
simply respond to environmental stimuli. They find order and
structure in the world so that they can learn from, anticipate, and
understand their experiences. Readers develop strategies for engaging
with text in such a way that they can construct meaning, or understand
it. Strategies are used in reading but strategies are also developed and
modified during reading. In fact, there is no way to develop reading
strategies except through reading. Readers develop sampling
strategies. The text provides redundant indexes that are not equally
useful. The reader should select from these indexes only those that are
most useful. If readers used all available indices, the perceptual
apparatus would be overloaded with unnecessary, useless or irrelevant
information. But the reader can choose only the most productive
indexes because of the schema-based strategies the reader develops for
text features and meaning.
Because texts have recurring patterns and structures, and because
people construct schemata as they try to understand the order of things
they experience, readers are able to anticipate the text. They can use
prediction strategies to predict the ending of a story, the logic of an
explanation, the structure of a complex sentence, and the ending of a
word. Readers use all their available knowledge and schemas to
predict what will come next in the text and what its meaning will be.
Typical silent reading speeds demonstrate that readers are predicting
and sampling as they read. They couldn't work with so much
information as efficiently if they had to process all of the information.
They predict based on the indices from their sampling of the text and
sample based on their predictions.
Inference is a powerful means by which people supplement
available information by using the conceptual and linguistic
knowledge and schemas they already possess. Readers use inferential
strategies to infer what is not explicit in the text. But they also infer
things that will become explicit later. Inference is used to decide on
the antecedent of a pronoun, on the relationship between characters, on
the author's preferences, among many other things. Inference can even
be used to decide what the text should say when there is a printing
error. Inference strategies are so widely used that readers rarely
remember exactly whether a given aspect of the text was explicit or -
implicit.
Because sampling, predictions, and inferences are basic reading
THE READING PROCESS 21

strategies, readers are constantly checking their own reading to make

sure it makes sense. I think readers actively control the process as they
read. There are risks involved in sampling, predictions, and inferences.
Sometimes we make promising predictions that later turn out to be
false, or we discover that we have made unfounded inferences.
Therefore the reader has strategies to confirm or reject his previous
predictions. This process of self-monitoring through the use of
strategies and confirmation is the way in which the reader shows his or
her concern for comprehension. But it is also used by the reader to test
and modify his strategies. Readers learn to read through self-
monitoring of their own reading.
The process I am describing is very efficient. The same indices
used to make new predictions and inferences are used to confirm
previous inferences and predictions. Effective reading gives meaning
to written texts. But efficient reading uses as little time, effort and
energy as possible to be effective. Use only the information you need
from the text to get meaning, and no more.
If readers are successful and self-confident, they take big risks and
increase their efficiency. If they find the text difficult to understand,
they proceed more cautiously but less efficiently. Readers must also
have self-correcting strategies to reconsider the information they have
or obtain more information when it cannot confirm their expectations.
Sometimes this involves rethinking and coming up with an alternative
hypothesis. But sometimes it requires going back to earlier parts of the
text looking for additional useful indexes. Self-correction is also a
form of learning, as it is a response to a point of imbalance in the
reading process.
READING IS A CYCLIC PROCESS

We can think of reading as consisting of four cycles, starting with an

optical cycle, moving to a perceptual cycle, then to a grammatical
cycle, and finally ending with a meaning cycle. But as the reading -
progresses, another series of cycles follows, and then another and
another. Thus, each cycle follows and precedes another cycle until the
reader stops or until reading has come to an end.
The reader is always focused on getting meaning from the text.
Attention is focused on meaning, and
the rest (such as or letters, words or grammar) only receives
mindfulness when the reader has difficulty
obtaining meaning. Each cycle is tentative and may not be completed
if the reader goes directly to the meaning. In a truly efficient reading, it
takes only a few cycles to complete before the reader gets meaning.
But in retrospect, the reader will know what the structure of the
22 KENNETH S. GOODMAN

sentence is and what the words and letters are because the reader will
know the meaning, and this will create the impression that the words
were known before the meaning. In a real sense, the reader is
constantly jumping to conclusions.
Reading is an intelligent behavior and the brain is the center of
human intellectual activity and information processing. The brain
controls the eye and directs it to look for what it expects to find. Thus,
even in the optical cycle, the reader actively controls the process. We
know what is the most useful information to look for, where to find it,
and what information to ignore. The human eye is an optical
instrument. It has a lens with a focal length that can only collect clear
information from a small part of the text. But the eye also has a less
clear peripheral field and the brain can make use of what has been seen
blurredly if it has expectations to guide it.
Our ability to predict language patterns is so strong that what we
think we see is mostly what we expect to see. As long as what we see
is sufficiently consistent with our predictions and as long as it makes
sense, we are satisfied. Once we have made sense of the text we have
the illusion that we have seen all the graphic details of the text. This
makes the perceptual cycle very efficient. We can do very well with
very few indexes, if we are dealing with meaningful and predictable
texts. Speed reading is associated with high comprehension not only
because good readers can process perceptual input more quickly, but
also because they are efficient at using the smallest amount of visual
cues necessary. They are not distracted by paying attention to
irrelevant information in the text; they use minimal perceptual indices
to activate their schemas.
The syntactic cycle strongly requires the use of prediction and
inference strategies. Readers should be able to use key elements of
sentence patterns, grammatical linkers and suffixes, and punctuation to
predict the pace of a sentence.
syntactic tasks when they begin to be processed. Otherwise they eit
cannot give their correct value to each syntactic element and know
where to look for the most useful information. Just as one needs to
know where one wants to go when beginning a journey, one must also
know the pattern when beginning to read a sentence.
The clause is the most important unit of language because
meaning is organized and presented through clauses and their
interrelationships. Readers must be able to predict the patterns of
clauses and their interrelationships (independent, coordinate, or
subordinate) to classify clauses and derive meaning.
The search for meaning is the most important characteristic of the
reading process, and it is in the semantic cycle that everything takes its
value. Meaning is constructed as we read, but it is also reconstructed
THE READING PROCESS 23

as we must continually accommodate new information and adapt our

forming sense of meaning. Throughout the reading of a text, and even
afterwards, the reader is continually re-evaluating the meaning and
reconstructing it as new insights are gained. Reading is, therefore, a
very active dynamic process. Readers use all their conceptual
frameworks when trying to understand.

THE READING PROCESS IN DIFFERENT LANGUAGES

I began this work by saying that I believe that the reading process is
essentially universal across languages and orthographies. In all
languages, readers have the same essential purpose: to obtain meaning
from the text. To do this they must enter into trans-actions with the
text created by a writer to express meaning. In all languages, readers
must use the same psycholinguistic indices and the same strategies.
They must sample, predict, infer, confirm and correct. They must go
through the same optical, perceptual, syntactic and semantic cycles.
These are universal across languages.
Of course, the reading process is flexible to account for
differences in languages and their orthographies. English readers
expect adjectives to appear before nouns. Readers of French or
Spanish expect adjectives to appear after nouns. English syntax relies
heavily on a fixed word order rather than on grammatical word
endings. That's why English readers base their predictions heavily on
word order. Spanish readers will look for important indices in suffixes
and base predictions on them.
Although European languages are written alphabetically,
punctuation varies in European orthographies to accommodate
phonological variations and conventions. Spanish puts question marks
and exclamation marks on both sides of the sentence. English only
uses them at the end. Both systems work because readers of each learn
what to expect and base their predictions on their scoring schemes.
This is a difference of convention. German capitalizes many more
nouns than most European languages, and this is another difference in
convention. But some spelling differences reflect linguistic
differences. Russian has two systems of consonants, strong and weak,
and that is why it places a “weak sign” on some weak consonants.
Spanish has some word pairs that are identical except for accentuation,
and so it includes an accent mark as part of the spelling.
Even within a language there are sometimes variations in spelling.
American spelling differs from that used in England and other English-
speaking countries in some respects: labor/labour, jewelry/jewellery,
center/centre, for example. This difference was deliberately created by
24 KENNETH S. GOODMAN

Noah Webster in his first major American dictionary. It was part of a

movement to make literature created in North America look different
from British literature.
In the United States, many people believe that some languages are
more difficult to read than others, and that English,
THE READING PROCESS 25

because of the complexity of its spelling, is one of the most

difficult languages to read and to learn to read. It is true that English -
spelling is complex. This is especially because of the multiple roots of
English. Some sets of English spelling rules are Germanic, while
others are based on Latin. Others have a French influence and there are
some Danish rules in some of our words that begin with “kn” like
know, knight, knee. There is no evidence that this spelling complexity
makes reading English more difficult. Readers can tolerate a great deal
of complexity. They learn to ignore some indexes and use others since
they have to obtain samples of the text anyway. What to look for will
vary from language to language but it remains a universal fact that in
any language some indices are more useful and reliable than others.
Readers construct schemata for the complexities and variabilities of
orthographies and use them in their reading.
Non-alphabetic orthographies relate phonology through meaning
primarily, although part of each Chinese character may indicate the
sound.
sible. But Chinese characters are basic 09 ideographic

cos. They directly represent the meaning. They are similar to eg

the numerical system of European languages. 3 + 3 = 6 is
written the same way in any language, although I read it as "three and
three are six" and a Spanish speaker reads it as "three and three are
six." It is very useful for us to use this ideographic system for
arithmetic in science and commerce, as it is significant across
languages.
It is also helpful in China that they have a writing system that can
be read with understanding by people who speak a wide variety of
Chinese dialects. Chinese readers will place much more emphasis on
the semantic and syntactic systems of their reading. But the process of
deriving meaning from text is essentially the same for them as for
other readers, once adjustments have been made for spelling
differences.

READING DEVELOPMENT

Generally schools have operated on the principle that reading and

writing should be taught in school. Traditional reading instruction
relies on teaching spelling features, letter names, letter-sound
relationships, and so on. It is usually focused on learning to identify
letters, syllables and words.
Such traditions are not based on an understanding of how the
reading process operates. They are not developmental considerations
based on an understanding of how and why people learn a language.
26 KENNETH S. GOODMAN

They do not put learning to read in the context of increasing control

over the process.
From my point of view, learning to read begins with developing a
sense of the functions of written language. Reading is searching for
meaning and the reader must have a purpose for searching for meaning
in the text.
Learning to read involves developing strategies to make sense of
the text. It involves the development of schemes about the information
that is represented in the texts. This can only happen if beginning
readers are responding to meaningful texts that are interesting and
make sense to them. In this sense, the development of oral and written
language are not really very different. Both depend on the
development of the process through its functional use.
For the same reason, I believe that you only learn to read and
write once. If you are literate in one language, it is very easy to
become literate in another. Furthermore, an important way to learn a
second language is through reading. Often the first language processes
a second language that is going to be acquired by people who are
already literate in another language.
It is no more difficult to learn to read and write than to learn oral
language. But instructional programs must move away from traditions
of treating written language as a school subject to be mastered. Rather,
they should be based on an understanding of the process and the
child's natural growth within the written language.

LITERATURE

Huey, EB, The psychology and pedagogy of reading (re-publication.

Cambridge, MiT Press, 1968).
National Society for the Study of Education, Fourteenth yearbook, part
i, 1915: Minimum essentials in elementary school subjects.
National Society for the Study of Education, Sixteenth yearbook, part

I, 1917: Second report of the Committee on Minimum Essentials

in Elementary School Subjects.
National Society for the Study of Education, Seventeenth yearbook,
part i, 1918: Third report of the Committee on Economy ofl^me in
Education.
National Society for the Study of Education, Eighteenth yearbook, part
ii, 1919: Fourth report of the Committee on Economy of Time in
Education.
National Society for the Study of Education, Twenty-fourth yearbook,
part i, 1925: Report of the National Committee on Reading (WS
THE READING PROCESS 27

Gray, chairman).
National Society for the Study of Education, Thirty-sixth yearbook,
part i, 1937, The teaching of reading (WS Gray, chairman).
Rosenblatt, Louise, The reader, the text, the poem, Carbondale,
Southern Illinois University Press, 1978.
THE ROLE OF NON-ACCOMMODATIVE COGNITIVE
STRATEGIES IN CERTAIN READING COMPREHENSION
DIFFICULTIES†

WALTER H. McGINITlE
KATHERINE MARIA
SUSAN KIMMEL

Yo

One of the fascinating aspects of the field of learning disabilities is that

the child whose problem is apparently so limiting in specific tasks may
prove to be extremely competent in all others. Indeed, this contrast is
so dramatic that many are inclined to limit the term “learning disabled”
to children with average or above-average intelligence, while offering
no argument for excluding those whose intellectual performance is
lower but who suffer from the same defects, whatever they may be.
Thus, many children with learning disabilities are children who -
perform well in daily life activities and achieve good scores on
intelligence tests, but who encounter serious difficulties in specific
tasks that are important for their academic performance. And both the
school and the society it serves are often uncharitable in their
assessment of the child who happens to have the wrong type of
disability for the school's requirements.

Thus it frequently happens to men |...] carrying with them the sign of a single
defect that nature or chance has imprinted on them, even if their virtues were as
many as are granted to a mortal, and as pure as heavenly goodness, they will
nevertheless be tarnished in the public opinion by that single vice that -
accompanies them (...)
(Hamlet, Act One, Scene X)

The best-known contrast between apparent general ability and

difficulty with a specific school task is the problem that many
otherwise skilled children have in learning to read. It is not surprising,
of course, that reading is so widely recognized as the greatest obstacle,
since reading is a vital, if complex, skill with many perceptual and
cognitive components.
† This work was funded by contract no. 300-77-0492 conducted between
the United States Department of Education, Office of Special Education and the
Research Institute for the Study of Learning Disabilities. Teachers College,
Columbia University.
|28|
COGNITIVE STRATEGIES 29

Previously, researchers in the field of learning disabilities focused

primarily on determining the factors responsible for decoding
difficulties. There is now growing interest in the relationship between
learning disabilities and difficulties in reading comprehension (Weaver,
1978; Weaver & Dickinson, 1979). Research carried out several years
ago on learning problems has sought to define these problems as
deficiencies in basic processes (Johnson and Myklebust, 1967) or
neurological differences (Mattis, French and Rapin, 1975). More recent
work has begun to focus on the role of word processing strategies
(Maria & McGinitie, 1980). This focus on strategies has been
stimulated by the growing appreciation of the complexity and
interactive nature of the reading process.
Several theorists (Rumelhart, 1977; Kintsch, 1979) have proposed
models of the interactive process of reading. The model described by
Kenneth Goodman is also an interactive model. Previous models, such
as that of Gough (1972), suggested that comprehension of a text results
from the reader's progression through a hierarchy of processes from
identifying certain features to recognizing letters and words and,
finally, processing sentences and text. Interactive models view the -
reader as engaging in parallel processing at many levels and at the
same time. In these models, processing proceeds in two directions:
from bottom to top, as in the previous models, and also from top to
bottom, so that knowing what the story is about and identifying the
letters in a word simultaneously contribute to the identification of a
particular word. It is important to note that this more sophisticated
understanding of reading blurs the distinction between comprehension
and decoding, since each of these interacts with the other.
Kintsch's (1979) interactive model assumes that the processes of
word identification, access to word meaning, and syntactic analysis are
bottom-up processes acting at the same time as top-down processes.
Top-down processes are based on factors such as the reader's purpose
in reading, his or her knowledge of the world, and the patterns that
structure the text. These top-down processes are very important
because usually a higher process will decide on the particular meaning
to encode based on the context long before the lower-level analyses are
completed (Kintsch, 1977; Marslen-Wilson & Welsh, 1978).
The interactive nature of the reading process and the limited -
capacity of memory force the reader to use all the resources at hand,
that is, to program his or her processing. This programming includes
the use of particular strategies; thus, a reading problem may be the
result of the use of an inappropriate strategy in a particular reading
task. It is possible that in some children a deficit in basic processing
may lead them to use an inappropriate strategy. On the other hand, a
reader may use a strategy that is not generally optimal for a particular
task but that compensates, to some extent, for a deficit in basic
30 WH McGINITIE/K. MARIA/S. KIMMEL

processing.
The ultimate goal of identifying the causes of reading performance
problems is to be able to suggest remedial methods that improve such
performance. Remedial programs designed to ameliorate deficits in
basic processing (eg, Frostig & Home, 1964) have generally improved
performance on tasks that are thought to require that particular type of
processing (e.g., e.g., figure-ground perception) but have not improved
reading performance (Hammill & Bartel, 1975). The strategies used by
the reader are more directly related to the reading process. Therefore, a
search for causes at this closer level may be more useful in designing
procedures to improve reading performance.
The work of Spiro and his associates at the Center for the Study of
Reading (Spiro, 1979, 1980; Spiro & Smith, in press; Spiro & Tirre,
1980) has included the classification of poor readers in terms of the
strategies they used for reading comprehension. It is important to note
that Spiro's classification refers only to poor readers. It is very difficult
to classify good readers according to the strategies they use since there
is evidence that these are flexible and that they adjust their strategies to
suit particular reading tasks (Fredericksen, 1975). Poor readers
apparently do not have this flexibility or adjustment, but they can use a
particular strategy, whether or not it is appropriate to the task at hand.
Spiro (1979) argues for a two-level approach to individual
differences in reading comprehension. One level covers the component
skills of comprehension; the other refers to the manifestation of
deficiencies in skills related to the style of reading comprehension.
When faced with poor reading skills, a reader has two options. You
may persevere in the problem area, or you may modify your processing
resources in an effort to compensate for the problem. For example,
there are readers who decode laboriously but who persevere in their
decoding efforts. Given the limitations in information processing -
capacity, this behavior can cause a “bottleneck” in the system (Perfetti
and Lesgold, 1978). The result may be that higher-level, more
knowledge-based processes are not used. On the other hand, readers
who decode laboriously may prefer to avoid the decoding task and rely
on their prior knowledge to guess what the text is likely to contain. In
other words, the same deficiency can lead the reader to use either a
text-based comprehension style or a knowledge-based style.
Spiro (1979) found evidence to support the view that poor readers
tend to rely too heavily on either top-down or bottom-up processes.
However, the poor reader does not alternate his excessive reliance on
top-down and bottom-up processes (Spiro & Smith, in press). The poor
reader with an over-reliance on the top-down process makes a decision
from the beginning about the general topic of the text and ignores all
those details of the text that may go against his hypothesis. The poor
COGNITIVE STRATEGIES 31

reader with an over-reliance on the bottom-up process has great

difficulty in reaching the full meaning of the text, above the details.
Spiro's (1979) classification of poor readers is very useful as it is a
classification at the level of strategies that focus attention on the
different ways in which ascending and descending processes can affect
comprehension. However, based on extensive informal work with
children with written language comprehension problems (Maria &
McGinitie, 1980), there are other important differences within Spiro's
general categories.
Our studies of children in the later elementary grades lead us to
believe that poor readers who rely excessively on top-down processing
employ a variety of unproductive strategies that are often distortions of
the strategies used by good readers. One could argue that they are
failures to maintain an appropriate balance between the processes
described by Kenneth Goodman. Any particular child's reading
performance may be related to one or more of these unproductive
strategies. In our work to date we have studied two groups of poor
readers, each characterized by an excessive reliance on a particular
strategy.
One of these groups consists of children who exaggerately expand
a general frame of reference drawn from their prior knowledge. They
use a small number of words from the text to evoke related knowledge
from their experiences, but are not limited to the information present in
the text. They read as if the text simply said what they already know.
These children “assimilate” the text into their schemes. When the
source of information is written text, these children cannot use the data
to modify their schemas. These children apparently show a form of
excessive dependence on the top-down process of interpreting the text.
The strategy used by them will be called non-accommodative.
The other group we studied consists of children who inflexibly
apply an initial hypothesis and schemes based on the
text. These children are ran an interpretation based on a
or more of the first sentences of the text and
try to interpret the rest of the text according to this initial interpretation.
Unlike children who use a non-accommodative strategy, these children
are aware of contradictions. In fact, they frequently give wild
interpretations to later portions of the text in order to make them fit the
initial interpretation. Apparently, these children show another form of
over-reliance on top-down word processing. In some ways, your
strategy too
It is non-accommodative since it does not an to a is- k
accommodate any information structure that does not
begin with a statement about the topic. However, since the
fundamental characteristic of their strategy is, from a clinical point of
 32 WH McGINITIE/K. MARIA/S. KIMMEL

view, a failure to test and modify an initial hypothesis about the

meaning of the text, we will refer to the strategy used by this second
group as the fixed hypothesis strategy.
The main part of this paper will describe a study of children who
rely on the fixed hypothesis strategy.* We will then discuss the
theoretical significance of the non-accommodative strategy and
describe our preliminary work with children who use this strategy.**

ii

The recent view of language comprehension considers the interlocutor

or reader to be a much more active participant than previous
researchers thought.
Meaning is not contained in the words, sentences, or paragraphs
themselves. What language offers
It is a skeleton, a l sketch for the creation of meaning.
In order to derive meaning from a text,
the reader undertakes an active construction process based on the -
formulation and testing of various hypotheses. The efficient reader
builds tentative hypotheses about the meaning of the text read and
about the content of what follows. Hypotheses remain tentative until all
related information has been taken into account. The good reader
builds and reconstructs a plausible model that takes into consideration
all the details of the text.
A tip Particularly bad reader creates hypothesis but fails
by evaluating and modifying them appropriately based on
the subsequent text. Rather than checking this interpretation against all
the new information gained from reading, these readers may actually
misperceive details in the text by fitting them into the original
interpretation rather than changing the interpretation. These poor
readers are those we have characterized as those who use a fixed
hypothesis strategy for interpreting texts.
In informal work with some of these poor readers it became
evident that they had greater difficulties

♦ This work is presented in more detail in Kimmel (1980).

•• A more extensive discussion of this non-accommodative strategy can be
found in Maria and McGinitie (1980).

to understand paragraphs organized in particular ways. The types of

paragraph structures that presented difficulty for these readers were,
naturally, those that contained structures that did not begin with the
topic or most important idea.
COGNITIVE STRATEGIES 33

A careful reading of many English school textbooks for 3rd

graders. to 6th. grade showed that these materials contained many
paragraphs with this characteristic. Particularly in expository materials
for children, some paragraph structures that did not begin with the
topic or main idea were very frequently presented, as if the author were
writing according to a certain structural formula. These grouped
structures or "formulas" will be informally called inductive structures,
since they share the characteristic of having initial sentences that lead
to the central idea of the paragraph, leaving the main idea located later
within the paragraph. Various types of these inductive formulas were
recognized. These types include:
1J Negation paragraphs, in which a belief or idea is presented at
the beginning of the paragraph, which is later said to be false within the
paragraph. For example, one paragraph began: “Many people believe
that deserts are made of sand. .and then he went on to say that this was
not true.
2] Paragraphs of analogies, in which the subject (a thing, a fact or
an idea) is explained by comparison with another thing, fact or idea: its
analogue. If the analogue is mentioned early in the paragraph, the
inflexible reader is very likely to construct a meaning centered on the
analogue and skip the central topic or simply use it as an adjunct.
3J Paragraphs with topics-examples (explicit), are paragraphs in
which examples of a topic are presented, followed by a statement that
concludes the topic, and which is supposed to link the paragraph.
41 Paragraphs with implicit topic-examples are paragraphs in
which examples of a general topic are presented, but the unifying
theme is not stated explicitly.
This is an example of a paragraph with an example-topic (explicit)
organization:

In a cloudy forest, far away, on the other side of the world, huge
elephants are pushing heavy logs. At the end of the world, dogs run
through deep snow, pulling loaded sleds. And in distant deserts, camels
swing by, carrying folded tents and goods to be traded. Animals all
over the world carry cargo for people (Fay, Ross, and LaPray, 1978, p.
20).

Paragraphs that do begin with the topic or central point—

paragraphs with “deductively” organized structures—are also common
in children's texts. These deductive structures seem to be much easier
to understand for children who use a fixed hypothesis strategy for
interpreting texts (and probably also for many good readers), since it is
more likely that the reader's first hypothesis about the central theme of
the paragraph will be correct.
In our main study on the fixed hypothesis strategy (hf) 255
34 WH McGINITIE/K. MARIA/S. KIMMEL

children in 5th grade. and 6th. Grade 12 students were tested with an
instrument consisting of 48 short paragraphs taken from published
children's reading texts at a 4th grade reading level. degree
approximately. Half of the paragraphs were organized deductively and
the other half inductively. After each paragraph there was a multiple
choice question referring to the central idea of the paragraph.
Correct responses on both the deductive and inductive paragraphs
were tallied for each child and converted to a regressive standard score.
The 16 children with the largest differences between scores, that is, the
children for whom the inductive paragraphs were particularly difficult,
were selected as those most likely to be using the hf strategy. They
constituted the group that we called the fixed hypothesis group (hf
group) in later work.
An equal number of children, matched for their overall
performance on the text test, constituted the comparison group. The 16
children selected for the comparison group had regressive standard
scores as similar as possible on the inductive and deductive paragraphs.
The comparison group thus had a general ability to read the test -
materials equal to that of the children in the hf group, but they did not
show any tendency to have special difficulties with inductively
structured texts.
Various tasks were used, administered individually, to compare the
performance of the hf group with that of the comparison group. First,
twelve paragraphs were constructed with
COGNITIVE STRATEGIES 35

the intention of disorienting those readers who will tend to

persevere in applying an unconfirmed initial interpretation of the text.
Each paragraph was designed to evoke a plausible hypothesis at the
beginning of the text and, eventually, to clearly show the need for
another interpretation. After each paragraph, a multiple-choice question
was included, testing whether the incorrect interpretation induced at the
beginning had been adequately revised by the time the entire paragraph
was finished reading. The hf group scored significantly lower than the
comparison group on this task. Reading time was analyzed to
determine whether children in the hf group were reading the passages
thoroughly. Reading times for the two groups were virtually identical.
Subsequent studies of reading time with additional paragraphs
structured both inductively and deductively showed that the groups
differed little in the time taken to read the two types of paragraphs.
(The effects for groups, structures, and interactions were all
(non-significant.) Children in the group HF EVIDENTLY HAS
understand something about inductively
structured paragraphs, but they apparently cannot learn well when they
read such paragraphs.
An adaptation of a “children's word search test” developed by
Pajurkova, Orr, Rourke, and Finlayson (1976) was also administered to
all children. In the adaptation constructed for the purposes of this work,
each word-search item consisted of a paragraph of four sentences, each
of which described a person or object that was always referred to with
the nonsense word “grob-nick.” The task the child had to perform was
to find the real word that could replace the nonsense word. 24
paragraphs of this type were constructed. Twelve of these items were
written in an “inductive” version in which the statement with the most
criterion attribute was placed at the end; and the other twelve were
written in a “deductive” version where the aforementioned statement
was placed at the beginning. This is an example of a reagent written in
an inductive version.

Every school has grobnicks. You can learn from a grobnick. A

grobnick can make you a better reader. A grobnick can give you tasks.
And this is the 0 or reagent written in its deductive version:
‡
Asame:
PL grobnick to 0give you tasks. Every school has grobnicks. You can
‡ The procedure for rating children's recall was as follows: Before
examining children's recall, each paragraph was analyzed to determine the
words or phrases that appeared to be major clues to the paragraph's structure.
A score was awarded based on the number of these key words or phrases that -
occurred in the response (synonyms or equivalent phrases were accepted). An
analysis of covariance was performed on the number of keywords present in the
responses with the total number of words in the response as covariates. The
results indicated that children in the comparison group included a significantly
36 WH McGINITIE/K. MARIA/S. KIMMEL

learn from a grobnick. A grobnick can make you a better reader.

In the task, each item was included alone or in an inductive or

deductive version, and in each item the word grobnick replaced a
different person or object. The examiner read each paragraph to the
child. Then the boy would tell him what he thought was a grobnick.
Analysis of the responses showed that children in the hf group had
relatively greater difficulty with inductively structured items than with
deductively structured items. (There was a significant interaction of
groups by structures.)
Finally, each child was given eight additional inductively
structured paragraphs to read and remember. After each paragraph the
child was asked to tell the examiner everything he could remember
about the paragraph. He
analysis of these responses II other than the children in the group
HF TENDED TO OMIT KEY WORDS
THAT WERE CLUES TO THE STRUCTURE OF THE PARAGRAPH.* IN
PARAGRAPHS BASED ON A NEGATION FORMULA, FOR EXAMPLE, KEY
WORDS WOULD INCLUDE ANY ADVERSATIVES SUCH AS: but, although,
however. (“MANY PEOPLE THINK THAT DESERTS ARE MADE OF SAND,
but...”) THE RESPONSES OF THE CHILDREN IN THE FIXED HYPOTHESIS
GROUP DID NOT CONTAIN THESE ADVERSARIAL STATEMENTS.
The results of the study support the hypothesis that there are
readers who use the hf strategy to interpret texts and that inductively
structured paragraphs are more difficult to understand for this type of
readers. The 16 children who had the largest score differences on the
initial test performed on all subsequent tasks as if they were
consistently using that strategy. The results of the word search task

greater number of target words in their responses.

 COGNITIVE STRATEGIES 37

indicated that the use of a fixed hypothesis interpretive strategy

can go beyond the reading process. The word search task was read to
the children, yet children in the fixed hypothesis group performed
better on the deductive items than on the inductive items.

The results of the reading recall task were similar.

particularly informative. Children in the hf group remembered the
comparison group, but their responses were qualitatively different.
Subjects in the comparison group tended to include key words that
were guides to the inductive structure of the paragraph and tended to
organize their responses around these structural cues. One of the
paragraphs with an inductive structure was the following (note the
importance of the word "hay" as a clue to the structure of this
paragraph):

approximately the same: 0 or number of words that children

The tiger's fur is reddish tan4 above. Below, it turns half white. It is
covered with black stripes. There is a plant with orange flowers and
black stripes like tiger skin. This flower
It is called “tigriolia”.

Subjects in the comparison group tended to organize their responses

around the word there is. For example, one child said:

There is a flower that looks exactly like tiger skin. It is orange and has
black stripes. The flower is called tigriolia.

Subjects in the hf group, on the other hand, generally omitted key

words with greater structural importance. For example, one child
recalled:
The tiger has reddish fur. He is tanned on top. Below is white. It has
black stripes all over its body.
Furthermore, what the children in the hf group frequently recalled were
elaborations of initial misinterpretations based on the central point of
the paragraph. This is another paragraph from the same assignment:

People have written stories about snakes that crush and squeeze their
enemies into pulp. These tales of snakes crushing their enemies with
tremendous force until they change shape are not true. Snakes only
squeeze hard enough to prevent their enemies from breathing.

The following is a response from a child in the fixed hypothesis group,

which appears to be an elaboration of the child's initial -
misinterpretation:
38 WH McGINITIE/K. MARIA/S. KIMMEL

Snakes crush and squeeze until their shape changes completely. They
are crushed and turned into pulp. People write stories about snakes.

It is very important to consider that inductively structured texts—

which frequently mislead children who use the fixed hypothesis
interpretive strategy—are commonly found in school materials (at least
in English). It would be important to know whether similar formulas
are commonly used in school textbooks in other languages.
It seems likely that many students who use the fixed hypothesis
interpretive strategy can be taught to evaluate their initial hypothesis as
they continue reading. One possible approach to such instruction is to
provide the child with guided practice in recognizing various
“formulas” for inductive organization of texts that present particular
difficulty for him. Informal work with some students suggests that
some can easily learn to recognize and respond appropriately to
negation statements beginning with “Some people think that. . or some
similar form. Another technique for developing a more flexible
approach to the reading task may be to have children read the same
stories from different perspectives. For example, they would be asked
to read “Cinderella” from Cinderella’s point of view and then from the
point of view of the evil stepmother. Children could then discuss how
these different points of view would affect our understanding of the
story and our memory of it. Instruction and practice in the use of words
with multiple meanings may help children who tend to use the hf
strategy for reading develop a more flexible approach to the task of
comprehension of a text. Opportunities could be provided to practice
accessing possible meanings of a word based on clues from the
beginning of a story and then given the opportunity to confirm or reject
these tentative hypotheses based on the rest of the text.
If children with fixed hypothesis interpretive strategies can learn
more flexible reading strategies, the work we now present is very
promising since we have an evaluative instrument to identify such
children. The instrument has only modest reliability at present, but can
easily be completed and refined based on current reagent information.
One possibility that needs to be explored is that the various inductive
formulas define distinct subtests within the assessment instrument; it is
possible that some children with a fixed hypothesis interpretive
strategy may find one or more of the formulas to be particularly
difficult relative to the others. It is also plausible that children
identified with this instrument do not form a uniform group.
It is possible that the same results obtained in the evaluation may
originate from various specific non-productive strategies.
Although the hf group performed significantly worse than the
COGNITIVE STRATEGIES 39

comparison group on both the misleading stories and the “inductive”

word-search items, there was considerable variability within the fixed -
hypothesis group. One aspect that requires further exploration is the
range of tasks to which children apply hf strategies. Our work suggests
that at least some of the children in the hf group apply such a strategy
when listening to written language read to them, as demonstrated by
the results of the word search task. Some children may extend their use
of the hf strategy to problem solving in other domains. The nature of
the hf strategy requires further exploration leading to the development
of a better theory on this type of behavior. It is also important to realize
that there were also some children who performed relatively well on
the inductively structured paragraphs of the initial instrument. The
reading strategies of these children have not yet been studied. Clearly
these are children who reserve their judgments and who are able to
remember and maintain their possible interpretations until they manage
to construct a model of the complete paragraph that best fits a set of
possible interpretations. Again, it is important
40 WH McGlNITIE/K. MARIA/S. KIMMEL

you know what the range of tasks might be to which this strategy could
be extended.

III

We will now turn to the examination of the second unproductive

reading strategy that we have studied: the non-accommodative strategy
in which the poor reader depends excessively on his or her prior
knowledge and adjusts the information in the text to already
established schemes. We currently have non-systematic data on this
non-accommodative strategy. However, the non-accommodative
strategy for word processing is potentially very interesting, since
reliance on this strategy when reading results in performance that has
all the appearance of a classic case of a learning disability: a child who
shows good verbal intelligence, who proves to be bright and competent
in activities of daily living but whose reading performance is poor.
Although we all must use what we already know to help us
understand what we read, Bobrow and Norman (1975) have suggested
that the efficient processor is one who focuses on the unexpected and
superficially processes the expected. There is evidence that efficient
readers superficially process expected information (Spiro & Esposito,
1977), but that poor readers who use a non-accommodative strategy do
not. The latter focus on what they already know.
Children who use a non-accommodative strategy do not learn
written language well. They are unable to understand written language
when it contains new or unknown information. And yet, they learn.
The fact that their intelligence is normal and that they can learn the -
information that teachers present verbally in class is evidence of this.
Furthermore, they do not appear to have any problems with expressive
language in a conversational situation. Children who use the above
strategy frequently perform very well on verbal intelligence tests. All
of the non-accommodative children we have studied (Maria &
McGinitie, 1980) scored average or above on the verbal subtests of the
Wisc-R (Kaufman, 1979).
In intelligence tests such as the Wisc-R, previously acquired -
knowledge and the ability to apply that knowledge are shown through
short questions to which the child must also answer verbally. In the
information subtest, the child is asked questions such as, “Why does oil
float on water?”; in the analogies subtest, questions such as, “How are
beer and wine alike?”; in the vocabulary subtest, questions such as,
“What does still mean?”; and in the comprehension subtest, questions
such as, “Why are criminals imprisoned?” Most of these questions
could be answered by any child who observes the world around him,
who is able to acquire knowledge through language and who is able to
COGNITIVE STRATEGIES 41

make use of that knowledge. Children who use a non-accommodative

reading strategy appear to be attentive observers of the world around
them and are able to acquire knowledge through some types of
language. The same strategy they seem to use to process written texts
serves them very well in many intelligence tests since the strategy
involves making use of what they already know. We suggest that these
children have learned, and they have learned by listening, but they have
not learned by listening to written texts. They have learned by listening
to oral language, particularly conversations, television and teacher
explanations. Within the broad categories of oral and written language
there are many forms that differ from each other according to the
various dimensions discussed by Rubin (1978). For the purposes of this
paper, however, we can focus on the differences between expository -
written text, on the one hand, and teacher conversation and
explanation on the other.
Although most recent research (eg, Kavanagh & Mattingly, 1972)
has focused on similarities between oral and written language, there are
also very real differences (Adams, 1979; Olson, 1977; Halliday, 1980).
Schallert, Kleiman, and Rubin (1977) have reviewed the existing
literature on the differences between oral and written language. Some
of these differences may explain why poor readers with a non-
accommodative strategy for understanding written texts learn well
through oral language and not from written texts, even if the latter are
presented to them orally.
The language children hear is specifically designed for them. The
people speaking to them are usually in the same place as they are, so
there is a shared linguistic context and participants can interact with
each other. But writers do not prepare their texts with a particular type
of reader in mind. According to Cazden (1972): “The written text is the
end point in the dimension of development towards independence from
non-linguistic context” (p. 199). This lack of shared context suppresses
cues such as pointing and gestures, which may be important for these
children. It also makes deictic terms—that is, words whose
interpretation depends on the context of use—more difficult to
understand. Deictic terms include pronouns, temporal terms such as
now or a week ago, and adverbs of place such as here and there. In
order to understand these terms in a written text, the child must take
into account the frame of reference given by the text. We know that
young children have difficulty understanding another person's point of
view (Piaget and Inhelder, 1956). This subgroup of poor readers may
present this same type of difficulty with written texts. There is also
evidence that the perspective the reader selects influences what the
reader understands from the text (Pichert and Anderson, 1977).
Therefore, if they fail to adopt the perspective that the text indicates,
understanding will suffer.
42 WH McGINITIE/K. MARIA/S. KIMMEL

Lack of interaction can pose a difficulty.

special for these children. There is mixed evidence that speakers
modify their language to suit their interlocutor (Gelman and Shatz,
1976; Gleason, 1973; Snow,
1972). People who speak to these children help them laugh
understand by observing their reactions or asking them
questions. In a speaking situation, these children may receive further
clarification through repetition.
One difference between much of oral language and expository
written texts is the greater redundancy of oral language. Speakers tend
to repeat what they say and use more words than they would use to
communicate the same message in written form (Wilkinson, 1971).
Halliday (1980) makes the same point slightly differently, when he
suggests that spoken and written language differ in the amount of
lexical density found in each of these forms. Lexical density is defined
as the number of content words per clause or sentence. Apparently,
lexical density increases as a function of the amount of control that
occurs in the production of a discourse. As a result. The written text,
which is more “edited” than the spoken discourse, has greater lexical
density. The oral instruction that children receive is, therefore,
probably less dense, lexically speaking, than the texts they read. This
may be particularly true with respect to the language with which
teachers address children. Intuitively, it appears that teachers focus
attention on new information in a variety of ways. Some ways to
accomplish this may include repeating the new information and
specifically stating the importance or novelty of the new information.
As Schallert, Kleiman, and Rubin (1977) suggest, one way to test these
intuitions would be to collect samples of the language used by teachers
with these children.
Another difference is that writing tends to be syntactically more
complex and more detailed and precise than speech (Horowitz &
Berkowitz, 1967; Wilkinson, 1971). There is some evidence that
certain types of complex discourse structures or organizations may
appear more natural in writing (Danks, 1974). Early work by Adams
(1979) indicates that children have several ways of avoiding certain
complexities in speech that characterize the written materials they are
required to read.
Written text also tends to differ from oral language in the type of
information it communicates. He met-
II Common sense transmitted through oral language is related to

particular and concrete actions and events. Knowledge communicated

through written language is usually more abstract, general and logical
(Olson, 1977). In written language, there is a more complex and
COGNITIVE STRATEGIES 43

extensive use of words that indicate relationships between sentences

and even between paragraphs, and that in this way unite the text. These
words include pronouns, conjunctions, temporal terms such as then or
later, adverbs of place such as here and there, as well as many others.
In English, the use of these words, called “cohesive markers,” has been
outlined by Halliday and Hasan (1976). These cohesive markers are
also used in spoken discourse, but, according to Halliday (1980), the
type of cohesion used is primarily related to the rhetorical function of
the discourse. Understanding the meaning of these cohesive markers
may be of particular importance for written language. Even in
children's texts, these markers are often crucial to understanding a text.
We have previously pointed out the advantages of classifying
learning problems according to the strategies used by the subjects in
question. However, once the learning problem has been linked to a
particular strategy, it may be possible for the researcher to understand
what circumstances influenced the adoption of that particular strategy.
In the case
of a non-acom strategy dative, which represents an over-
reliance on prior knowledge, there are
several plausible etiologies that can be fruitfully explored once children
whose written language comprehension learning problem involves this
strategy have been identified.
A general reason for adopting a strategy that relies on prior
knowledge might be its usefulness. In a text that is congruent with prior
knowledge, such a strategy will often be successful. During text
processing, such a strategy can also compensate
impairments such as a lack of deciphering skills or a limited ification
immediate memory capacity, which reduces processing
ability to more basic levels. Likewise, many of the texts read by -
children in the early grades and many of the questions asked by
teachers about those texts are likely to be structured in such a way that
the strategy of relying on prior knowledge is reinforced (McGinitie,
1979). Poor readers who use non-accommodative strategies, such as
the poor readers studied by Weaver and Dickinson (1979), may tend to
show poor verbal memory of what they read. Relying on prior
knowledge can be a retrieval strategy that allows them to reconstruct
what the text should have said when they have forgotten what it said
(Bartlett, 1932).
In summary, within the group of children with problems in
understanding written language, there is evidence of a subgroup of
children who have good verbal intelligence but who depend
excessively on their prior knowledge when processing written
language, assimilating the information presented in the text to their
44 WH McGINITIE/K. MARIA/S. KIMMEL

schemata, but without accommodating their schemata to the

information, with the result of a misinterpretation of the text that does
not agree with their prior knowledge and with the resulting difficulty of
not being able to obtain new information from what is written. The fact
that these subjects can learn from oral language can be explained by the
differences between oral and written language (basically, differences in
redundancy, complexity, cohesiveness and situational support). Their
particular difficulty with written language may be explained by the fact
that it provides different types of cues and they cannot perceive or
attend to the cues it does provide. The success of these children in tasks
related to verbal intelligence tests is explained by the fact that these
tasks measure prior knowledge and the ability to apply it. These tasks
require precisely the strategy that, when used excessively, causes some
children problems in understanding written language.
We are continuing our work with children who rely excessively on
particular strategies for word processing. We hope, of course, that by
discovering the nature of the strategy that governs the child's inefficient
performance, we will be able to work with him to modify his use of the
strategy in question and enable him to experience more effective
performance.
We assume that there are many and diverse unproductive strategies that
cause difficulties in learning. We hope that many teachers and
researchers will explore these possibilities.

LITERATURE

Adams, M., “Studies on oral conversation in relation to written texts”,

in B. Stan (ed.), Progress report 2 on the BBN group of the Center
for the Study of Reading (BBN Report No. 4106), Cambridge, Bolt
Beranek, and Newman, 1979.
Bartlett, F. C., Remembering, Cambridge, Cambridge University Press,
1931.
Bobrow, EG, and Norman, DA, “Some principles of memory -
schemata”, in DG Bobrow and A. Collins (eds.), Representation
and undentandig: Studies in cognitive science. New York,
Academic Press, 1975.
Cazden, CB, Child language and education, New York, Holt, Rinehart
and Winston, 1972.
Danks, JM, “Comprehension in listening ai.d reading: same or -
different?”, Report of the Interdisciplinary Institute in Reading
and Child Development, University of Delaware, 1974.
Fay, L., Ross, R.R., and LaPray, M., Telephones and tangerines (Level
10), Chicago, Rand McNally, 1978.
Fredericksen, CH, “Effects of context-induced processing operations
on semantic information acquired from discourse”, Cognitive -
Psychology, 1975, 7, 139-166.
COGNITIVE STRATEGIES 45

Frostig, M., and Horne, D., The Frostig program for the development
of visual perception, Chicago, Follett, 1964.
Gelman, R., and Shatz, M., “Appropriate speech adjustments: the
operation of conversational constraints on talk to two-year-olds,”
in M. Lewis and L. Rosenblum (eds.), Conversation, interaction,
and the development of language, New York: Wiley, 1976.
Gleason, JB, “Code-switching in children's language”, in TE Moore
(ed.), Cognitive development and the acquisition of language, New
York, Academic Press, 1973.
Gough, P.B., “One second of reading,” in J.F. Kavanagh and IG
Mattingly (eds.), Language by ear and by eye, Cambridge, mit
Press, 1972.
Halliday, MAK, and Hasan, R., Cohesion in English, London:
Longman, 1976.
Halliday, M. “Cohesion and register”, paper presented at the American
Educational Research Association meetings, Boston, April 1980.
Hammill, D., and Bartel, N., Teaching children with learning and
behavior problems, Boston, Allyn and Bacon, 1975.
Horowitz, MW, and Berkowitz, A., “Listening and reading, speaking
and writing: An experimental investigation of differential
acquisition and reproduction of memory”, Perceptual and Motor
Skills, 1967, 24, 207-215.
Johnson, D., and Myklebust, H., Learning disabilities: Educational -
principles and practices, New York, Grune and Stratton, 1967.
Kaufman, A^„ Intelligent testing with the Wisc-R, New York, Wiley,
1979.
Kavanagh, JF, and Mattingly, I.G. (comps.). Language by ear and by
eye, Cambridge, mit Press, 1972.
Kimmel, S., “Children with a perseverative text interpretation strategy:
The effect of text organization,” unpublished doctoral dissertation,
Teachers College, Columbia University, 1981.
Kintsch, W., “On understanding stories”, in M. Just and P. Carpenter
(eds.), Cognitive processes in comprehension, Hillsdale, Erlbaum,
1977.
Kintsch, W., “On modeling comprehension”, Educational
Psychologist, 1979.14,3-14.
McGinitie, WH, “What do published comprehension lessons teach?”,
in Proceedings, Reading '79, Toronto, York University, 1979.
Maria, K., and McGinitie, W., “Prior knowledge as a handicapping -
condition”, paper presented at the Fourth International Colloquium
in School Psychology, Jerusalem, July 1980.
Marslen-Wilson, W., and Welsh, A., “Processing interactions and
lexical access during word recognition in continuous speech.”
Cognitive Psychology, 1978, 10, 29-63.
Mattis, S., French, JH, and Rapin, I., “Dyslexia in children and young
adults: Three independent neuropsychological syndromes”, -
Developmental Medicine and Child Neurology, 1975, 17, 150-163.
Olson, D.R., “From utterance to text: the bias of language in speech
and writing*', Harvard Educational Review, 1977, 47, 257-281.
46 WH McGINITIE/K. MARIA/S. KIMMEL

Pajurkova, E., On, R., Rourke, BP, and Finlaysen, MAJ, “Children's
word-finding test: A verbal problem solving task.” Perceptual and
Motor Skills, 1976,42,851-858.
Perfetti, CA, and Lesgold, AM., “Discourse comprehension and
sources of individual differences”, in M. Just and P. Carpenter
(eds.), Cognitive processes in comprehension, Hillsdale, Erlbaum,
1977.
Piaget, J., and Inhelder, B., The child's conception of space, London,
Routledge and Kegan Paul, 1956.
Pichert, JW, and Anderson, RC, “Taking different perspectives on a
story”, Journal of Educational Psychology, 1977, 69, 309-315.
Rubin, A., A theoretical taxonomy of the differences between oral and
written language (Tech. Rep. No. 35), Urbana, University of -
Illinois, Center for the Study of Reading, 1978.
Rumelhart, DE, “Toward and interactive model of reading”, in S.
Domic (comp.), Attention and performance (vol. 6), Hillsdale,
Erlbaum, 1977.
Schallert, DL, Kleiman, GM, and Rubin, AD, Differences between oral
and written language (Tech. Rep. No. 29), Urbana, University of
Illinois, Center for the Study of Reading, 1977.
Snow, CE, “Mothers' speech to children learning language”, Child
Development, 1972,43,549-565.
Spiro, RJ, “Constructive processes in prose comprehension and recall”,
in RJ Spiro, BC Bruce, and WF Brewer (eds.), Theoretical issues
in reading comprehension, Hillsdale, Erlbaum. 1980.
Spiro, RJ, Etiology of reading comprehension style (Tech. Rep. No.
124), Urbana, University of Illinois, Center for the Study of
Reading, 1979.
Spiro, RJ, and Esposito, J. Superficial processing of explicit inference
in text (Tech. Rep. No. 60), Urbana, University of Illinois, Center
for the Study of Reading, 1977.
Spiro, RJ, and Smith, D., Patterns of overreliance on bottom-up and
top-down processes in children, in press.
Spiro, RJ, and Tirre, WC, “Individual differences in schema utilization
during discourse processing”, Journal of Educational Psychology,
1980, 72.204-208.
Weaver, P., “Comprehension, recall and dyslexia: a proposal for the -
application of schema theory.” Bulletin of the Orton Society, 1978,
28,92-113.
Weaver, P., and Dickinson, D., "Story comprehension and recall in -
dyslexic students", Bulletin of the Orton Society, 1979, 29, 157-
171.
Wilkinson, A.M., The foundation of language: talking and reading in
young children, London, Oxford University Press, 1971.
PREDICTABILITY: A UNIVERSAL
IN READING AND WRITING

JEROME C. HARSTE
CAROLYN L. BURKE

We asked Dawn, age 4, to write a story and she delivered this piece of
art (Figure 1). Then, using his hand

FIGURE 1. (DAWN, 4 YEARS OLD)

To mark what was written, she read it this way: “My name is Dawn. I
go to college. My brother Tommy goes to the Kids' Corner. He used to
go to college too” (making sure that when he said “also” his finger was
on the bottom left side of his writing). Dawn knows a lot about what
writing is in her cultural environment. He knows that it goes from left
to right and from top to bottom. You know what it is

150
)
PREDICTABILITY IN READING AND WRITING 51

functional. Understand that it is used to convey meanings.

Najeeba, also 4 years old, a little girl from Saudi Arabia, wrote this
page (Figure 2). It should be noted that in

FIGURE 2. (Najeeba, 4 years old)

In contrast to Dawn's writing, Najeeba's writing appears decidedly

Arabic. Najeeba said: “Here is my story, but you can’t read it because
it is in Arabic.” And then, taking his pen and making the dots that can
be seen on the paper, he added: “and in Arabic we use many more dots
than you use in English.”
Although Najeeba’s writing, like Dawn’s, is far from
conventional—if we are to base it on the perception of other speakers
of the culture in which she lives—it is important to note that Arabic
scribbling looks different than English scribbling.
Dalia is a 4-year-old Israeli girl and her story clearly
it looks like something written in Hebrew (figure 3).
From a sociolinguistic perspective p we read to
observe
 52 J. HARSTE/C. BURKE
$,30ta
2/-KlVLLLlLLLC<%r: c,
2 IR4C, $ *,23*20,,
ee 17 • ' L •2 UX,, < Gv
Ui .v LLbL- G10.3 (S2
•(Lw 6 2} 9, $)
21.584450 -6~ , (•—2-62
A4A, • 9,41,2,42 L,
192246*44352 $ 2
33 * ti I 912 • \$4.0511}r, (hLLi
428>Mp>.- •J,bAax Wh -
FIGURE 3. (DALIA, 4 YEARS OLD)
SEVERAL IMPORTANT ASPECTS. SOCIOLINGUISTICS IS THE STUDY
of the normative rules of language use that operate across situational
contexts within a culture and how these rules differ from one cultural
group to another from a conceptual standpoint, such as the study of -
language across situational contexts within the same culture (story
writing as opposed to letter writing) and also how these rules differ
from one cultural group to another (Dawn's story writing as opposed to
Najeeba's story writing). This is a very fruitful area of study as we have
found that cross-cultural studies often clarify and illuminate key
elements of the process and thus reveal our cultural perceptions.
similar situational texts. P Let's look at sociolinguistics,
To illustrate our notion of sociolinguistics within a culture, let us
examine Megan's writing in two situations: writing a story and writing
a letter. Megan—age 4—produced this when we asked her to write a
story (Figure 4). He read it like this: “Once upon a time there was a
PREDICTABILITY IN READING AND WRITING 53

ghost. A family of three ghosts. One day they went for a walk. They
honked their horn because they saw Mr. Wood and said, "Hello!", then
they went back to Mrs. Comer's house and honked their horn and said,
"Hello!" End". While this is clearly not the best story we've ever heard,
it's important to note that it really does sound like a story (“Once upon
a time... It is structured like a story (“Once upon a time there was a
family of three ghosts” = the theme; “They went for a walk” = the
initiating event; and so on) and is even organized on paper like a story.
This fact is evident when we compare it with other writings that Megan
has done. When he completed his story, we gave him another blank
sheet of paper and asked him to write a letter. She wrote and read: “-
Dear Mary, I wish you would bring me to this place every day. End.
Megan” (figure 5). We can notice that his letter sounds like such, is
structured like such, and, in contrast to the writing of his story, is
presented graphically like such. Megan knows what we seem to forget:
that written language is both culturally and contextually specific.
 54 J. HARSTE/C. BURKE

'(or
' AEG/
FIGURE 5. (Megan, 4 years old, letter)

Through ongoing experiences with stories in their culture (some

direct experiences, others indirect, but all of them real linguistic
experiences) Megan and other children in her cultural environment
learn how story texts work and not only that: they also learn how they
are organized semantically, syntactically and symbolically. Similarly,
through ongoing experiences with letters and letter writing in their
cultural environment (lived directly or indirectly, but all real linguistic
experiences) Megan and other children in her cultural environment
learn not only how writing works in this letter-writing context, but they
also learn how it is organized semantically, syntactically, and even
symbolically in contrast to other available contexts of written material.
These are not pseudo forms of writing. They represent the
raw material of real writing. T true users of These children are ver
written language. All the decisions they
made are the same ones we and our ancestors had to make. They, like
us and our cultures, had to invent written language for themselves,
from the inside out. They, like us, have the only advantage of a social
group with which they can share, construct and expand meanings.
Each of these children has developed key understandings of how
writing operates through ongoing natural experiences with diverse
written materials in their respective written language-using cultures.
PREDICTABILITY IN READING AND WRITING 55

These natural experiences allowed each child to develop a set of

anticipatory, cultural and situational schemas about how writing
operates in these particular situations. These cultural and situational
anticipatory schemata direct the strategies used and, consequently, also
direct the selection (in writing) or sampling (in reading) of what is
represented in the written material.
Looking at the four young informants and their productions, we
can notice some universals:
• Each one hoped that what was written would be meaningful.
• Everyone understood that the language of the stories was
functional.
• Each hoped that the marks made on the paper were
representing meaning.
• Each had developed a perception of the form of language in
his own culture.
• Each saw writing as organized and systematic.
These perceptions formed an anticipatory schema that was both
culturally and contextually specific and that directed not only the
strategies used, but also the preparation.
chosen sitting. It is imp It is important to note that understanding
The knowledge that Megan, Dawn,
Najeeba and Dalia have gained about how written language works in
their own cultures will be useful to them in learning to read and write
in other cultures.
With this broad framework in place, let's take a closer look at this
model. To this end, we have selected a particular situational context
(children's reading of middle grade scripts), in a particular culture (in
the United States of America). We believe that a deep analysis of
children's sophisticated and orchestrated decisions, in the face of a
particular situational context, allows us not only to clarify the reading-
writing process, but also to identify those motives that young children
bring to formal reading situations and on which we can build.
56 J. HARSTE C. BURKE

We showed 3-, 4-, 5-, and 6-year-olds samples of writing from the
environment and asked them what they thought it said. Let's look at a
set of typical responses we got when faced with a box of Crest (a well-
known brand of toothpaste):
Boyd (3 years old): “brush teeth.”
Michelle (3 years old): “toothbrush”.
Michele M. (3 years): “It says Florida pasta.”
Jeremy (4 years old): “It says toothpaste.”
Charles (5 years old): “toothpaste”.
Heather (age 6): “It’s Crest toothpaste.”
Justin (6 years old): “Crest”
One of the interesting aspects of all the answers is the assumption
that what they were taught would necessarily make sense.
We have found in our research with young children that. When
asked what they think is said in about 20 different samples of written
material taken from their environment, 97% of the responses reflect the
fact that children approach print with the expectation that it will be
personally meaningful (like Boyd's). Yetta Goodman, who also uses
this test and whose research we have relied on, showed a little boy a
box of Crest toothpaste and asked him what he thought it said. His
response was classic: “CC- CRRREEESSSTTT”, and then back:
“TTTSS- SEEER RRCCC. . . “Toothpaste”. What makes this answer
interesting is that while it shows us that he knows the names of the
letters and can answer in that manner, ultimately the writing has to
make sense and so he chooses to place importance on meaning rather
than deciphering the letters.
Answers like this are what support our vision of what language is.
Meaning is the core system. This system of meaning is circumscribed
in a syntactic system and both in turn in a graphemic or phonemic
system. A second aspect we observe is that Boyd's response seems (has
the structure of) a response from written language rather than from
spoken language. What becomes evident when contrasting Boyd's
response with that of other children when we ask them, "What do you
think it says here?" (as opposed to the question where they are required
to say what they know about the object we present), is that from the
PREDICTABILITY IN READING AND WRITING 57

From a syntactical point of view, the answers we received to our

question about reading are more similar to written language than the
answers they give when we ask them to talk about the objects
presented. For example:

What does it Tell me what you know about this:

say here
Toothpaste. My mom buys Colgate toothpaste at the
supermarket.
Toothbrush. We brush our teeth with that.
Brush your teeth.
Crest I have seen that toothpaste in the
Colgate. supermarket.
Toothpaste. My brother got in trouble because he
Crest. spilled all the toothpaste in the
bathroom.

This is very interesting because it tells us that the children's

responses demonstrate their sensitivity to the requirements of oral or
written forms of language. And we are talking about children who are
only 3 years old.
When Boyd was asked to formulate an oral response, he used the
usual form of oral language; when asked about written material, he
paid attention to the written material and used a form of written
language. It is very important to understand this “give and take”
relationship that exists between language users in a given linguistic
event. You can't really understand what's going on if you study a
speaker in isolation. It is always necessary to study the speaker in
interaction with another. The reason for this is that language is
eminently social and developed precisely because there was more than
one user. In order for a reader to exist, a writer is needed. To have a
receiver, a speaker is required. At any given time there is more than
one user of the language, even if there appears to be only one.
Anticipatory schemes for language are not only cultural and
situational, but eminently social (directed towards others, sensitive
towards others, “situated” with respect to others). There cannot be a
linguistic situation where the interlocutor can be suppressed, even
when we are supposedly writing for ourselves, such as when we write a
shopping list. In this case, we first act as writers and then as readers of
the same when using it in the market. Thus, in any case, two users of
the language appear and not just one.
The goal of language is communication; therefore, the central
system in both is meaning. In this sense, language is functional. It
provides a vehicle through which the two users can signify and share
meanings. From a functional perspective, Boyd's production of “brush
58 J. HARSTE/C. BURKE

teeth” represents the type of language that might be used to direct,

remind, or inform a subordinate, represented in this case by a child,
what to do in a particular situational context. In Boyd's case, while it is
true that "brush your teeth" as a text can be found in a dentist's office or
on the sink as a reminder from Mom to get the kid to do it, the
important thing is that we could also find these words on the side of a
toothpaste box; pragmatically speaking, this is the kind of written
language we would use to inform a non-active speaker about a desired
social activity that should be carried out in relation to this writing
context.
Pragmatics is defined as the rules of language use that are
operative in a particular situational context. It is the system that links
speakers. Children demonstrate an understanding of pragmatics when
they produce responses that indicate their sensitivity to the relationship
they have with another language user or users who are involved in the
context in which they are responding. In reading, this sensitivity is
often referred to as “stance,” which affects anticipatory schemas (and
specifically, intention), strategy use, and selection (within subsequent -
attention).
Context is an integral part of the text in any written material. The
situational context in which the writing is found, together with the
writing itself, act as a complex system of signals. And it is not a matter
of the context appearing first, but rather, having identified the context,
we can identify what is written. Actually both are simultaneous (figure
6).
To further clarify this situation, let us examine a set of responses
that evidence this fact. We showed 4-year-old Kibi a box of sticky
fabric and asked him what he thought it said. He replied: “Fabric. Let
me see. No; no fabric. What is that thing called? Adhesive fabric-
PREDICTABILITY IN READING AND WRITING 59

FIGURE 6. Context of situation

go Writing in relation to context conveys meaning to us.

es. We showed 6-year-old Emily a sign advertising a house for
sale. He replied: “I don’t know. I haven't seen it anywhere. It looks like
he's on the beach. Wait! “Yes, your, a boat is for sale.” We showed 6-
year-old Justin a box of Fisher Price toys and asked him what they
said. He replied: “What is that?” “Fisher Price cereal?”§
When we asked 4-year-old Alison what it said on a piece of paper
we showed her where we had typed the name of Kraft Cottage Cheese
using upper and lower case letters, she replied, “It looks like the
alphabet, except it doesn’t start with A.”
Justin and Alison's responses are especially valuable as they
illustrate the relationships between meaning and context, even if their
attempts may seem unsuccessful. Knowing the context allows us to
know a range of possible meanings. We have concluded, therefore,
through the analysis of texts from the medium, that in these writing
contexts there is only a limited universe of semantic options.
If we examine the writing printed on a box of Jell-O, for example,
we can see the following:

§ In all these cases, children use some letters of the text as indexes, to arrive at
a meaning in accordance with the context. |T.|
60 J. HARSTE/C. BURKE

Gelatin: refers us to the referent

Dessert: refers to the class of the referent
Strawberry: It tells us an attribute of the referent
Jell-O: is the particular name given to this reference
Test the quality: It is a functional description of how the -
manufacturers would like us to react to this
product.
Recipe: refers to something related that is included in
the box.
Net It is a contextual description of this particular
weight: instance of the referent.

Once this analysis of how writing functioned on printed material

from the medium was completed, we found that all instances of writing
from the medium fit neatly into this limited set of semantic features. In
other words, any label on any product (Com Flakes, Rice Krispies,
Cornstarch, etc.) .), any street signs (Stop, No Parking, etc. . .), any
store logo, any writing sample from the medium we could think of,
could be classified using this limited number of semantic options.
Equally interesting was the analysis of the responses given by
children aged 3, 4, 5 and 6 to questions about the media's printed
matter. We found that a large majority (approx. 86%) also fit into this
set of options.
This means that the children's semantic choices were far from
random. They fully understood the semantic options available to them
in these scriptures.
For example, the response “Box of pudding” tags the referrer.
“Dessert” identifies the class of the referent. “It tastes delicious” is a
functional description of how the product relates to the language user.
“Pie filling” is a related concept. “Strawberry”, an attribute that
describes a quality of the referent. "Jell-O" is its name. “Contains a
little sugar” is a description of this particular box or instance of the
referent.
These data are showing us that the written language of
environmental writing is predictable for children aged 3, 4, 5 and 6
years.
However, predictability is not an inherent characteristic of writing,
even though it may seem that way in most writing situations with
which we are familiar. Predictability is a semantic transaction that
occurs between language users in familiar linguistic situations. Its
genesis is social. In any literate society, children have many
opportunities to encounter language and its users.
From these relationships with language and its users in particular
situational contexts, such as writing situations, the child is able to
identify the way in which writing means, in relation to the context; he
PREDICTABILITY IN READING AND WRITING 61

or she is also able to know which features of writing serve as signifiers

and what role writing plays in the processes of signification.
The letters IRA mean something very different depending on the
context in which they are found, whether on Kenneth Goodman's
(International Reading Association) tie or on the wall of a building in
Northern Ireland. In either of these two situations the writing means
according to the context. However, in a given culture and in a given
situational context, one of these relationships between context and
writing
is more predictable than the other. In this sense, it, including the au
the absence of a predictable context-writing
relationship would also be significant simply because it does not have
the expected relationship. In other words, the predictable relationship
between context and writing is what is marking the meaning for the
reader, even in unpredictable contexts.
From this perspective, predictability can be considered a central
characteristic of the linguistic process.
All of this background leads me directly to the main point I want
to make: we have been thinking
doing in the reading and writing process with an approach that has E
so far been wrong. We think that writing is providing the meaning;
that writing directly provides the meaning. However, literacy must be
redefined now that we have evidence that the language user is using
script-context relationships to mark meaning. We must redefine
reading now that we have clear evidence that the language user, when
faced with a writing situation, makes decisions that he or she would not
have made if this material were not present.
Note that we said “writing situation” and not just writing. One of
my colleagues in Indiana asked me, “Well, Dr. Harste, if you were to
take the writing off the Crest toothpaste box and show it to a three-
year-old and ask, “What do you think it says here?” and the child
responded, “Brush your teeth,” would that be evidence of
62 J. HARSTE/C. BURKE

literacy?” I replied, “That would be evidence of reading and not

necessarily literacy.” The child would have read, in the same way that
we can “read” the sky to forecast the weather.
My colleague's question is, however, problematic, somewhat like
asking about "reading" clouds in a cloudless sky. In a reading material,
writing is an integral part. If it is removed - as in the phenomenon of
the sky without
clouds—may still be F meaning-bearer, but it turns out
I find it strange to talk about it like
reading clouds.
There is, however, a parallel, and that is that the child would still
have to read the available symbols or cues in order to reach a decision,
and of course there is always the possibility that there are more
symbols or cues available than are actually needed, given the desired
level of semantic refinement.
Reading is a phenomenon of sign use and not a phenomenon of
printed marks per se. When reading material in a writing situation,
print is an integral component of the situation.
This is true both in reading written material
from a book as in reading written material from street advertisements.
In both cases what provides the meaning is the writing in relation to the
context in which it is written.
is found. Without this writing-context sign we therefore have no way to
delimit the semantic options: it could be anything, any option.
notion of the anticipatory scheme to which p We will access a
The act of reading always involves the multiple use of many
indexes and writing is only one part of them. When I performed the
experiment of removing the “Crest” writing from its corresponding box
and asked a 3-year-old, “What do you think it says here?” the child
responded, “Well, it should say toothpaste, but you took the words
out!”
Stevie Hoffman, a professor at the University of Missouri in
Columbia, has been making videotapes of some preschool children
reading under different circumstances. In one of his interviews, one of
his informants was pretending to read a story, as he had been asked to
do. While reading he came to a page that had a drawing but no writing.
The girl scanned the page quickly and then said with disgust, “Hey, I
wish there were some words here.”
These anecdotes show us that the presence of writing is important
even when the response does not show a term-to-term correspondence,
as in the case of Boyd anticipating “Brush Your Teeth” compared to
Crest’s writing. When we remove writing from its context, the absence
of writing itself calls attention; the absence is itself a sign. If we
remove context, the absence of context draws attention to itself and -
PREDICTABILITY IN READING AND WRITING 63

disrupts, if not destroys, the process.

It is for this reason that we have needed to distinguish between
reading that occurs in a reading class and real reading. In the vast
majority of classrooms there is very little relationship between actual
reading and reading class. Obviously, they should be related. But this is
only possible when reading instruction is seen as a supportive process
that does not interfere with or destroy the natural signaling process that
exists between writing and its situational context.
The absence of natural context clues, as often happens in the way
writing is handled in the classroom, becomes its own signal. For this
reason, children become subjects oriented exclusively from the school
point of view; instead of understanding texts by themselves, they do so
through the teacher or for the teacher. The teacher creates, in relation to
the text, a new signal for knowing how to read and process the text.
Literacy, including school-based literacy, occurs when we have
evidence that elements within a writing situation operate as cues.
Progress in learning to read and write can be viewed from this
perspective as the orchestration of multimodal systems of signals,
many of which are found in the predictable or unpredictable
relationships between writing and context.
Yet it is 2-year-old David who naturally picks up the Sunday
paper, goes to his dad's couch, puts a toy in his mouth to pretend he's
smoking, adjusts the cushion, picks up the comics section of the paper
and announces, “I'm reading the paper.”
It's also Joe, 9 years old, son of our best friends who, when we
showed him a sign shaped like a stop sign, but with the letters ban and
asked him to read what it said, replied: “I bet it says STOP in German
or something.” “This is just the kind of junk you carry around and pull
out and question unsuspecting children.”
Predictability, as a perspective for considering reading, allows us
to appreciate the visual and non-visual elements of literacy. It allows us
to understand not only the natural relationships of writing signage in
context, but also the way in which new signs and meanings are created
personally and culturally.
Predictability as a perspective has broad implications for both
research and teaching. Suggests that written language learning is
handled contextually.
If our goal is to give children universal literacy (that is, if we want
them to acquire the ability to read stories, narrative materials, reading
materials in certain areas, recipes, manuals, tax forms, poetry, etc.),
then we need to provide them with multiple opportunities to have
contact with all of these types of materials in the classroom. Our
approach to classroom organization should be to alternate these
materials. Our goal, predictable relationships between writing and
64 J. HARSTE/C. BURKE

context.
This perspective does not suggest that there are multiple reading
processes, but it does suggest that, in specific detail, the processes of
prediction, confirmation, and integration will differ in their intent and
attention across situational and cultural contexts. It also suggests that
teaching activities should be functional: that is, they should be real and
natural instances of language uses and not extracted from their natural
situational context. What we do with language in the name of teaching
must be predictable given the extracurricular perspective of language
use. It also suggests that reading and writing should be fused in
teaching programs, just as they are in the natural use of language. The
dividing lines between reading and writing that we draw in the
curricula are unpredictable from the point of view of the natural use of
language. It also suggests that we can take as a starting point what the
child already knows about language and that we can begin with those
writing contexts that are most familiar to him or her. It suggests that
our role as reading teachers will be better served by helping children
discover the predictability of print as it operates in a variety of print
materials. Too often, when children arrive at school, the teacher starts
from scratch, as if the child knew nothing about written language.
Often, too, we approach language in the abstract; we take it out of
context, and instead of taking our role in helping the child discover the
predictability of writing in alternate contexts, we teach them that
writing is unpredictable.
We asked Jeff, a first grader, to write a story. Jeff drew instead of
writing, but he shared his story about what he liked to do at home:
playing Star Wars with his brother and going to the zoo. “But I wanted
you to write down your story so I could remember it and tell it to my
friends,” I said. He replied: “But I don’t know how to spell.” I told
him, “Don’t worry about it, just do the best you can.” He hesitated and
then said, “I just can’t spell like you.” I told him, “Do the best you can.
If you could write like me there would be no reason for you to be in
first grade.” Jeff thought for a moment and then asked, “Would it be
okay if I just wrote what the things are in my drawing? “That way,” he
said with great conviction, “you could remember what the story was
about.” I told him it was okay (see figure 7).
PREDICTABILITY IN READING AND WRITING 65

FIGURE 7. (JEFF, FIRST YEAR)

Jeff wrote “house,” spelling it like this: hos, and then said, adding
the e. “I bet there’s a silent e at the end of that word.” Above the ship
FIGHTER he did the same: "T-[-FOFOR," and then added the e
saying: “I bet there's a silent e at the end of this word.” For “cloud” he
wrote kld, then added an L to the end and said, “I bet there’s a silent L
at the end of that word.”
Can we have any doubt about the type of teaching Jeff received?
My presence in this environment had the meaning of “master” and
therefore indicated a particular way of processing writing. Teaching
has made writing unpredictable, full of silent words instead of
predictable.
The way we visualize and perceive reading and writing is of great
importance. Our approach is truly paramount because, as Jeff
demonstrates, children's view of language and literacy will be
influenced by how we view it. If our research with young children is
correct, Jeff had a better perception of written language before school
instruction than after it. This model, then, is important as a heuristic
tool for the study and research of literacy learning.
This model allows us to frame important questions regarding how
the context of book reading at school might differ from the same
activity at home. It allows us to identify key transactions between
variables that operate to make schools and school teaching rare or
funny instances for language and language comprehension.
Its heuristic value does not lie solely in the study of initial literacy.
We have found it to be a useful model for understanding the problems
66 J. HARSTE/C. BURKE

faced by international students in their classes at our University.

In Saudi Arabia, students are taught to take very different
pragmatic positions in relation to books than we expect them to take in
our university classrooms. The same relationships between writing and
context have different meanings in different cultures. Our Arab
students have been helped and found it beneficial to have them “re-
read” the signs that their university professors provide regarding the -
position they should take and the responsibility they should assume
when reading and interpreting written university texts.
This model has, in itself, some value as well as first diagnosis
explanatory value. This is equally true for our
understanding of data from children from different cultures. Think how
smart one of the first-grade students Carol Edelsky investigated was.
He was in a bilingual program in the United States. In English culture
you will find a store called “K Mart” (Pronounced: Kei-mart).
Although he had experiences with other similar stores in his early
culture, none were called K Mart. When asked to write a story in
Spanish, instead of writing K Mart, which is the writing he had seen
and which would be pronounceable in his native language, he translates
and writes “Cei-mart.”
Although he had several options available to him as to how to
write it, in this situation his understanding of the rules of language
usage in his original culture allowed him to do what any good linguist
or translator would do, that is, he made a change appropriate to the
culture in which he was writing.
And this is probably a good example to finish with, as this fact
shows that being bilingual means knowing twice as much.
I am hopeful that a cross-cultural view of language and the literacy
acquisition process—which this conference aims to encourage—can
lead all of us, teachers and researchers, to the same results.
SCHOOL DYSLEXIA AND EXPERIMENTAL DYSLEXIA

MARIA A. CARBONELL OF GROMPONE

I. SCHOOL DYSLEXIA

For many years we have been dedicated to the study of school dyslexia,
which we include, for practical reasons, any significant delay in reading.
Children with specific learning disorders, as they are now preferred to be
called, which includes dyslexics, have been defined as "children who present
an alteration in one or more of the basic psychological processes involved in
the understanding and use of spoken or written language, alteration that may
manifest itself in an imperfect capacity to pay attention, think, speak, read,
write, spell or perform mathematical calculations. These disorders include
conditions such as perceptual defects, brain injury, minimal brain
dysfunction, dyslexia, and developmental aphasia. The term does not,
however, include children with learning problems resulting primarily from
visual, auditory or motor acts, from mental weakness, emotional disturbances
or environmental problems (Gearheart).
Of all the students we have examined for reading deficits, some have
been considered true dyslexics and others simply delayed in reading. The key
to such a distinction, given the intricate and obscure nature of the etiology of
dyslexia, was based on whether or not the student presented the deficiencies
classically attributed as the cause of such a condition. But for the purposes of
this paper, we cover under the general name “dyslexia” all reading
difficulties when the student shows at least two years of delay between his
grade level and the level at which he reads.
We believe that we have been able to make several important
observations in the years dedicated to the study of dyslexic children. They
refer to reading and other concomitant modalities that occur in them. Are:
1] Students with reading or reading and writing difficulties have IQs for
which no criteria can be specified. It is known that a limit is set towards the
lower limit, to exclude the mentally retarded.

168
1
 SCHOOL AND EXPERIMENTAL DYSLEXIA 69

2] Most of these students also have a deficiency in other fields of

language, namely: a] in the field of
the ort challenges, since they always learn to write words with
complex spelling correctly with a delay, thus confirming the
famous saying of Mme. Borel: “First dyslexic, then dyslexic-dysortographic
and finally dysorthographic, over time”; b] an enormous poverty in writing
that makes it so that when they start writing on their own they are unable to
write more than a few sentences. Their written language is poor both because
of the simplicity of its structure, because of the repetitions, and because of the
obscurity regarding what they want to say, because of the ambiguous way in
which they say it, because of the syntax errors, out of place for the school
course they are taking. In many of these students, the poverty of language is
more marked in writing than in speaking.
3| In many of these cases, but not all or even the majority, there are also
problems in learning arithmetic: difficulty in remembering tables, or errors in
fundamental operations, or in writing and placing quantities. In other cases
we have found that learning arithmetic is unscathed and even the way it is
done can serve as a witness to the presence of dyslexia. Of course, we do not
count among the arithmetic difficulties those that arise from poor reading, as
we consider it secondary to the main problem.
4] There are no errors in reading or writing that are typical of dyslexics;
the errors that these students make are the same as those of those who are
learning to read or who are already fluent readers, the only difference being
that they last longer or are made in greater numbers when compared to poor
readers.
5] You cannot judge a poor reader's understanding of the text when he is
reading aloud, based on how difficult it is for the listener to understand. There
is no correlation between poor reading aloud and poor comprehension; some
of these students have understood well enough to perform fairly well on a
multiple-choice comprehension test.
6] Many poor readers, when they are able to read with their eyes,
without articulating, understand the text better than if they read it out loud.
7] Students whose native language is Spanish, even when they are poor
readers, do not make many mistakes
 70 MARIA A. CARBONELL OF GROMPONE

8] in reading, but rather they read slowly, failing to rise to a level of

fast reading, as they should do for the year they are in.
After confirming the peculiarity of poor readers in Spanish, their inability
to read fluently at a speed
acceptable dad, we have studied this fact in 1 difficulty. cases of
Each subject was placed, regardless of the grade they were
in, in the one that corresponded to their reading speed. Table 1 shows the
distribution of these students according to their reading speed.
TABLE i
Reading grade in relation to the grade students are currently in
Grade they are studying
Primary Education Grade Reading 3rd. Secondary education Totals
4th. 5th. 6th. you in the pedagogical it. 2nd.
definition of dyslexia. It also tells us
that these students spend more than one school year advancing from one
grade to another in reading. Although we do not know what happened to
them when they were in their first two grades, it is legitimate to assume that
they started off late and were never able to catch up with their classmates.
it. 1 3 2 6
lo./2o.* 9 6 4 4 1 2 26
2nd. 3 4 4 5 3 19
2nd/3rd. 7 4 6 5 3 1 26
3rd. 3 3 4 2 3 15
3rd/4th. 11 1 3 6
4th. 2 2

Totals 20 20 20 21 7 12 100
If these data were confirmed with larger samples, we would be entitled to
conclude that students with reading difficulties do not manage to surpass the
level of 3rd grade students. school grade in terms of speed. It is clear that
students in the last grades of our sample, that is, those in the higher grades,
are lagging behind in reading by a year and a half, up to more than four years.
It seems that the higher they advance in school grades, the smaller the rate of
improvement in speed is, or, to put it another way, that poor readers only
manage to reach a very mediocre level which they then do not surpass.
We are here at the point that we were interested in highlighting: that poor
readers are not so much so because they have not learned the grapheme-
phoneme correspondence and make mistakes when reading (from which it
would follow that in Spanish such correspondence is relatively easy, which
we agree with), but rather what characterizes them is the difficulty or
impossibility of being able to achieve fluent reading, of surpassing the pace at
which a schoolchild who has only three years of reading experience reads.
The marked slowness of reading continues and therefore creates an
impossibility of breaking out of a certain reading rhythm, no matter how
 SCHOOL AND EXPERIMENTAL DYSLEXIA 71

many years of schooling they have. This reading is slow, because it has been
done in a combination of two ways: a series of words (generally
monosyllabic and sometimes disyllabic and very common in written
language) are read at first sight, but words that are less found in the written -
vocabulary or long ones can only be read by syllabifying. Therefore,
qualitatively, the reading is short, inexpressive, without adequate rhythm,
monotonous, as if the reader could not anticipate what is coming, based on
the context before him or his knowledge of the language; the reader stops at
almost every word and to someone who heard it offhand, without paying
attention to the meaning, it could very well give the impression that he is not
reading a text, but a list of words. This lack of a sense of anticipation, of a
guiding thread through the meaning that is gradually emerging from the
context, which is essential for the speed we call “normal” reading, is
noticeably absent, and what is most notable, for the witness of this reading
aloud, is the almost visible struggle to recompose the word through the -
syllables or even the phonemes that are being isolated. Word-by-word
reading is therefore dominant, as opposed to the fluid reading that could be
characterized in this respect.
as “reading prayers.” In reading! For a fluent reader, whether spoken aloud or
silently, the sentence, or at least an important part of the sentence—one in
which the words that make it up are united, because the whole supports the
meaning—is the true unit of reading. In the case of the dyslexic, this stage
has not been reached and the word, still divided into its components -
syllables or sounds - continues to remain in the foreground, in isolation, each
one of them rising before the eye and the mental process behind the work of
the eye, like a mountain to be climbed individually.
Foucambert (cited by Ferreiro and Teberosky, 1979) rightly says:
“Dyslexias are not disturbances of
reading but of decoding.” It could not be defined perhaps
In fewer and more precise words, this
“illness” of someone who is trying to learn to read.

II. EXPERIMENTAL DYSLEXIA

We had been planning this experimental situation thinking about the

possibility of finding a circumstance that would put adult readers in a
situation similar to that of dyslexics, who cannot get out of the decoding
phase.
Despite our attempts, we were unable to construct an experimental model
that would replicate the act of reading in its entirety, that is, of facing a text.
The good adult readers we were going to work with have fully or almost fully
mastered the two conditions of linguistic competence and cognitive abilities
necessary for reading. The experimental situation we were to find tended to
elude the two advantages that these two conditions provide to non-dyslexic
readers.
 72 MARIA A. CARBONELL OF GROMPONE

Much to our regret, this led us to dispense with reading a text - an almost
universal reading situation for adults - and to resort to reading words, which
is - and we recognize this - a very artificial and arbitrary reading situation.
Since our work with dyslexic children had led us to the conclusion that they
read by words and not by sentences or parts of sentences based on the thread
of comprehension, the situation that we would propose to good readers would
be artificial, of course, but it would put them in the situation of reading
words, which is what the dyslexic does even when faced with a text that
makes sense.
In this word reading situation, would good readers have an advantage
over poor readers? We knew that a fundamental aspect of fast or fluent
reading is that the adult does not have to decipher the words and does not
read them by analyzing the graphemes that compose them one by one to later
synthesize them. We know that the fluent reader attacks words in a global
way and it is certain aspects of them that allow him to read them, in no way
the spelling or syllabification and then the reconstruction. Therefore,
fundamental elements of reading such as semantic and syntactical
information that is derived from the context for its understanding and
anticipation, had to be abandoned, much to our regret, and we were forced to
reduce the experience, to put it bluntly, to checking whether, when faced with
isolated words, a good reader has or does not have an advantage over one
who does so with difficulty.
Our hypothesis when planning the experience was that good readers,
based on their previous reading experience and, above all, their familiarity
with the written word, would not encounter obstacles in this task.
After several projects, we finally decided to choose, on the one hand,
words from the Spanish language and, on the other, using the computer,
artificially created words that appeared to belong to our language.*

We made 44 lists of 15 words each, made up as follows:

Lists a (named 5a to 15a) composed of five-letter words
the first had six graphemes, the second had six, etc., up to 15, which were typewritten,
in lower case; a total of 11 lists.
Lists b (named by the same procedure above from 5b to 15b) written entirely in
capital letters, constituting a total of another 11 lists.
Lists c (from 5c to 15c) made up of words that do not exist in Spanish, written in
lowercase; there are 11 in total.
Lists d (5d to 15d) written in capital letters and constituting the last series of 11.

♦ See Appendix. For further details on this experience, see Carbonell de

Grompone, 1980.

The grand total was 660 words across the 44 lists. This material would
allow us to:
 SCHOOL AND EXPERIMENTAL DYSLEXIA 73

a] compare the errors made in the two types of lists, namely real words
and invented words.
b] compare the reading speed of words made up of the same number of
graphemes, when some were authentic words of the language and others
invented.
cj the influence that the fact that the word was written entirely in capital
letters had on both factors, accuracy and especially speed.
The situation was presented to the subjects informally, individually, and
the following words were spoken out loud and never read: "I am going to
give you a list of words to read out loud. I will take your time but you read
them at the speed that is most comfortable for you.”
The subject read the lists in order, that is, all the lists.
of five-letter words (5a, b, c, and d): t, etc. The reading the sixes,
time was recorded and, if possible, the mistakes made.
No work was done with a recorder.
Once the experiment was over, the subject was talked to in the most
informal way possible and was asked to describe how he or she had felt when
reading lists c and d (of invented words) and any other information that could
be provided.
We worked with 18 adult subjects, good readers and for whom reading
was among their professional activities (teachers, professors, university
professionals, etc.).

Qualitative aspects of reading. The interesting aspects of reading occurred in

the lists of invented words.
When faced with long and unfamiliar words (because they were made
up), the subjects resorted to some of the following strategies:
a] spell or spell out the entire word, breaking it down into its elements.
b] spell or spell only the central part of the word.
c] follow any of the above procedures but mentally, and say the word
out loud, once it has been deciphered.
d| divide the word into two or three pieces, as if it were two or three
words ("diendo-paterzos", "alcarta-des- todel", for example).
e] point to the syllables with your finger when reading.
Some subjects told us about the difficulty they found in identifying
syllables, especially in words that contained syllables with a complex
structure, that is, syllables formed by a vowel between two consonants
immediately followed by another syllable that began with a consonant or
those that contained liquid /1/ and /r/: "cadorplidortas", “dendela- morre".
The errors were mainly made in the lists of meaningless words. These
errors were:
 74 MARIA A. CARBONELL OF GROMPONE

a] deletion of part of a word, especially the final piece, especially if it

was preceded by syllables that constituted Spanish words.
b] deletion of syllables and graphemes that occupied the middle
positions of words.
c] transpositions of graphemes and syllables.
d] deletion of diphthongs.
e] additions of graphemes (much more frequently than deletions).
fj errors in reading vowels and consonants, especially in digrams: /rr/
read /r/, /ch/ read /c/.
g] errors in complex syllables, particularly when there is a
conjunction of consonants between the terminal of one syllable and the initial
of the next.
h] errors in the accentuation of words that did not have an accent.
Although words ending in a vowel, /n/ or /s/ have been read as acute and the
rest as grave, there was a tendency to read most of the words as grave,
perhaps because this is the tendency in Spanish accentuation.
While everything we have said has occurred mainly in lists of made-up
words, there have been cases where these same errors have appeared in real
words. When questioning these subjects, we discovered that, despite their
culture, many real words were unknown to them and, therefore, presented
them with the same difficulties as the invented ones.
As for fabricated words, their reading generally caused a lot of
discomfort to good readers
 SCHOOL AND EXPERIMENTAL DYSLEXIA 75

who unexpectedly found that they had difficulty reading. In many

subjects this was openly appreciated, because while they were reading they
would interject out loud expressions such as “How horrible! I have never
seen this word”, “How difficult!”, “It seems that I cannot read”, etc.; in these
cases it was difficult for us to measure time. When they were in
When questioned, the subjects almost the reading section,
unanimously admitted at the end of the
session that they had felt uncomfortable. This situation was attributed to
words formed by 8 or 9 graphemes.
The slowness of the reading process required by the material was a
conscious fact, so depressing to some subjects that, when moving from lists c
and d (invented words) to lists a and b (real words) in the next series, they
proceeded at a rapid pace as if to compensate for the bad impression they -
suspected they had made on the experimenter.
We find it interesting to note, because it shows the success in the
formation of the invented words, that the subjects never stated that they had
suspected that they were dealing with fabricated words. Many of them did
confess that they had never seen them and some were curious to find out
where we had gotten them from.
We then considered at the end of the experiment and after what the
subjects had told us that we had managed to create in these subjects a kind of
experimental dyslexia, at least as far as reading lists of words was concerned.

Quantitative aspects. The times taken by the subjects to read the four types of
lists are shown in the following table:

TABLE 2
Stockings and DS of the reading times of the four lists
Average time in
seconds DS

Lists to (real words) 12.18 5.51

Lists b (real words written in capital letters)
Listase (forged words) 13.26 6.44
Lists d (contrived words written in capital letters) 23.65 11.67

24.08 11.26
In the lists, the vast majority of subjects
doubled their time when moving from words made up of 5 graphemes to
those made up of 15. The time doubled from words with 12-13 graphemes
and did not increase from then on.
In lists b, the time doubled when reaching words with 10, 11 and 12
graphemes, according to the people, but they did more than just double the
time, since a word made up of 15 graphemes always takes more than double
 76 MARIA A. CARBONELL OF GROMPONE

the time of a word with 5, but without reaching triple.

In lists C, the time doubled when words made up of 8 graphemes
appeared, it tripled in lists 10-11, and it quadrupled in lists from then on. As
you will remember, these were forged words.
In series d the same thing happened as in the previous series, but in a
more pronounced way: the time doubled when the words contained 9
graphemes, tripled when they contained 10-11, quadrupled when they
contained 12, and continued to increase from then on.
These data clearly inform us of the difficulty in reading words that do not
exist in the language and also of the difficulty, although lesser, in reading
words written entirely in capital letters. In Table 2, the growth in times for the
series aa to d and the important jump that occurs between b and c, that is,
when the forged words begin to appear, is evident: reading time has
practically doubled.
To verify the validity of the differences in reading times, we obtained the
value of the differences in means, from which it resulted that the only
significant difference is between a (real words) and c (fabricated words). This
indicates that although the words have the same length, the fact that some are
invented and others are not, that is to say, some are unknown and others are
known to the subject, significantly alters the reading time and produces, in
fact, a situation that without much distortion, we can consider to be
"experimental dyslexia."

III. CONCLUSIONS: HYPOTHESIS ABOUT READING

Even with the limitations of these two experiences, we believe that valid
results can be drawn from them. The two situations compared here have
important differences that cannot be ignored, because they shed light on the
reading process.
 SCHOOL AND EXPERIMENTAL DYSLEXIA 79

res, as they indeed are in the definitions, but not so much in the concrete
facts. The most typical example of “guessing” is the so-called riddles that are
asked to children. Let's take one that is very common in our latitudes:

A small barrel full of hazelnuts that gather during the day and scatter at
night.

Let us now recall the musical game called “What is this?” The subjects
gathered together are made to listen to a piece of a record on the record player
that is not announced. Here, by conjecture, one cannot deduce what is
unknown: if one of those present does not know the author of the piece
because he has never heard anything about him, he has no way of finding the
solution; if he knows the author but not the composition, it is possible that he
will be able to name him, but not to identify the piece; if the listener knows
the author and the composition, he may be able to name him, that is, he was
able to recognize him, to “discover his identity by carefully examining the
thing.”
In the case of divination, identification can be achieved with very little
knowledge, as happens to children with riddles; in the riddle of the stars, you
do not need to know almost anything about the stars, only that they are visible
at night but not during the day.
In the case of recognition, a fairly deep or at least much deeper -
knowledge is involved; discovering by ear in the case of “What is this”
implies a very intimate knowledge of music or of certain authors and certain
musical works. When it comes to reading, it involves an intimate knowledge
of the written appearance of the word.
In riddles, what is to be discovered is disguised or concealed by
something else or even symbolized by something else; in the case of
recognition, the object to be recognized is exposed as it is and it is through a
close and intimate examination of the object that its identity is recognized.
Good readers had a hard time reading the forged words
because they did not know them previously; they had ios data
perceptive t for identification, but they were not able to
quickly reach identification, as happens with real and supposedly known
words. When the word is unknown, when we have never seen it before, the “-
partial impression” that is being received by the current method of reading
words is insufficient, and then the reader is left with two paths:
a] go back in time to when, in order to read, he was forced to decipher,
because the data he was receiving was not sufficient and he resorted to
analyzing the word, inspecting all its parts (it can be said that by doing this,
this reader is not truly reading, but rather taking very ancient steps prior to
mastering reading).
bj risk making a “closure” with the insufficient data that he has extracted
visually and producing the wrong word.
 80 MARIA A. CARBONELL OF GROMPONE

Whether the good reader, now turned into a “dyslexic”, risks one
solution or another, his discomfort is evident, because he notices that his
effectiveness has diminished, in one case, and
because he notices that he can only show off this “old” process
approach the reading of the word, if he
wants to be sure of reading correctly, by going back to much lower levels. In
our hypothesis, it is because the word is unknown that the subject cannot
“guess” it; the data that comes in does not allow him to arrive at the exact
solution and he is forced to decipher. The subject tries to make use of “the
keys,” but finds that this is of as little use to him as if someone not versed in
classical music were to be shown Beethoven’s Seventh Symphony to “guess”
what it is. The good Beethoven connoisseur does not guess that it is the
Seventh Symphony; he recognizes it.
We could say, then, that in the authentic poor reader or in the
experimental dyslexic there is a failure in recognition
instant recognition, although in both cases the reasons are uzea by
different: in the case of poor readers it is because they have not
managed to acquire “instant recognition”; in the case of experimental readers,
it is because, although they have mastered it, it is not useful for a word that
does not appear in the language and which, therefore, they have never known.
Those who have not mastered the system of instant recognition, those who
have not been able to achieve it - the dyslexic - will always be slow readers,
even if they read a text, because the words are obscure to them and it is
imperative for them to decipher them, although in many cases they can come
to understand it.
effectively while the process we describe the laborious procedure
here takes place. Those who suffer from
experimental dyslexia caused by the material given to them feel that their
recognition system also fails, even though they capture the perceptual data.
Despite the culture and profession of many of our subjects, they (like most
adults, not excluding teachers) have so little awareness that reading words is a
very different task from reading a text—their usual way of reading—that they
were all embarrassed by how they read and not one of them thought of
thinking and expressing it, as a justification: “It's because I never read single
words, but rather I read books (i.e. text).”
The poor reader never becomes familiar with words, or at least he
becomes familiar with very few; the rest always seem new to him and
therefore he has to scrutinize them, analyze them, break them down in
order to reconstruct them later.
This hypothesis of the difficulty or impossibility that dyslexics have
in recognizing words accounts for a fact that all teachers know only too
well, even if they cannot explain it. A dyslexic who is already
SCHOOL AND EXPERIMENTAL DYSLEXIA 81

dysorthographic often writes the same word three times on a page, in

three different ways, and it is not unusual for one of the three to be
correct. Faced with this fact, parents and teachers say to the student:
"Why did you write it wrong if you know how to do it well? “It is well
written here.” What those who reason in this way fail to understand is that
the dysorthographic person, due to failures in the word recognition
mechanism, has no way of knowing that one of his versions of the word is
the exact one.
From the combination of the two works that we have brought
together in this communication, we believe that we can formulate the
following conclusions as a hypothesis:
1. In reading texts, the semantic and syntactical aids that come from
the text itself and are projected and complemented by the reading
experience and the total experience of the subject are fundamental; we
would almost dare to say that they are even more important than the so-
called "reading keys", since these refer only to elements of the printed
material. Given the data provided to us by the syntax and semantics of the
text, as well as what we know about the subject we are reading about, it is
not necessary for us to read each word in its entirety, or the entire passage
in detail.
mind, but we can skip words, just as we skip
part of the word when making its identification.
This process, that of reading a text with the corresponding
understanding, is the only one that deserves to be called “reading” and
until we master it we do not know how to read. It follows from this that
reading is a continuous process throughout our entire life, which is
continually being perfected.
What kind of activity is this we call reading? How do we come to understand
the system of marks that make up our alphabetic writing? How can we make
literacy closer to the real process and more effective? These are some of the
questions that are the focus of the works compiled in this volume and that were
presented at the International Symposium on New Perspectives in Reading
and Writing Processes, organized by the General Directorate of Special
Education of the Secretariat of Public Education, which took place in Mexico
City from July 1 to 4, 1981.
The publication and dissemination of these works is fully justified, since
literacy remains a central educational issue in Latin America and it is
necessary to deepen our understanding of the nature of the processes in order
to clarify teaching practice.
An interdisciplinary field par excellence, literacy requires being viewed from
different angles. That is why psychologists, psycholinguists, anthropologists
and linguists met at this symposium, where they discussed learning processes
and social practices, rather than teaching methods. Despite the differences in
perspective, everyone agreed that it is not possible to reduce the acquisition of
written language to perceptual-motor skills, and that the intervention of the
subject's linguistic competence and cognitive competence are determining
factors.

ISBN 968-23-1600-6

twenty-one
editors
• Due to the standards of the tests, it is necessary to locate
certain degrees of speed between two school grades.
 As can be seen, if we leave aside all the students who are reading
beyond the third grade, the sum of all the others reaches 92%, but if we
take all the students who do not reach the 4th. school grade comprise
98% of the sample.
The same table allows us to observe the ranges between which the
students belonging to the different school grades have moved. Thus we
see that the students who are in 3rd year. degree are located between lo.
and 2nd/3rd; those in 4th, between 10 and 14 years. and 3rd; those of
5th. between it. and 3rd/4th; those in 6th. between the./ 2nd. and 4th, and
secondary school students between 1st/2nd. and 3rd/4th. degrees.
The study of this table shows us, first of all, that the students do not
move in the school grades maintaining the distance of two, as is generally
accepted.