(Download PDF) Language and Reading Disabilities 3Rd Edition Allyn Bacon Communication Full Chapter PDF

Language and Reading Disabilities (3rd
Edition ) (Allyn & Bacon

Communication
Visit to download the full and correct content document:
https://ebookmass.com/product/language-and-reading-disabilities-3rd-edition-allyn-ba
con-communication/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
eTextbook 978-0137021062 Clinical Phonetics (The Allyn

& Bacon Communication Sciences and Disorders Series)
https://ebookmass.com/product/etextbook-978-0137021062-clinical-
phonetics-the-allyn-bacon-communication-sciences-and-disorders-
series/
Teaching Reading to Students Who Are At Risk or Have

Disabilities, Enhanced
https://ebookmass.com/product/teaching-reading-to-students-who-
are-at-risk-or-have-disabilities-enhanced/
Top 50 SAT Reading, Writing, and Language Skills, 3rd

Edition Brian Leaf
https://ebookmass.com/product/top-50-sat-reading-writing-and-
language-skills-3rd-edition-brian-leaf/
Reading Picture Books with Infants and Toddlers:

Learning Through Language Jane Torr
https://ebookmass.com/product/reading-picture-books-with-infants-
and-toddlers-learning-through-language-jane-torr/
Vagueness and Thought Andrew Bacon
https://ebookmass.com/product/vagueness-and-thought-andrew-bacon/
500 SAT Reading, Writing and Language Questions to Know

by Test Day, Third Edition Inc.
https://ebookmass.com/product/500-sat-reading-writing-and-
language-questions-to-know-by-test-day-third-edition-inc/
Elsevier Weekblad - Week 26 - 2022 Gebruiker
https://ebookmass.com/product/elsevier-weekblad-
week-26-2022-gebruiker/
Critical Care EEG Basics-Rapid Bedside EEG Reading for

Acute Care Providers (Feb 29,
2024)_(1009261169)_(Cambridge University Press) Jadeja
https://ebookmass.com/product/critical-care-eeg-basics-rapid-
bedside-eeg-reading-for-acute-care-providers-
feb-29-2024_1009261169_cambridge-university-press-jadeja/
Language Intervention Strategies in Aphasia and Related

Neurogenic Communication Disorders 5th Edition – Ebook
PDF Version
https://ebookmass.com/product/language-intervention-strategies-
in-aphasia-and-related-neurogenic-communication-disorders-5th-
edition-ebook-pdf-version/
Language and Reading Disabilities
Alan G. Kamhi Hugh W. Catts
Third Edition
Pearson Education Limited
Edinburgh Gate
Harlow
Essex CM20 2JE
England and Associated Companies throughout the world
Visit us on the World Wide Web at: www.pearsoned.co.uk
© Pearson Education Limited 2014
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted
in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without either the
prior written permission of the publisher or a licence permitting restricted copying in the United Kingdom
issued by the Copyright Licensing Agency Ltd, Saffron House, 6–10 Kirby Street, London EC1N 8TS.
All trademarks used herein are the property of their respective owners. The use of any trademark
in this text does not vest in the author or publisher any trademark ownership rights in such
trademarks, nor does the use of such trademarks imply any afﬁliation with or endorsement of this
book by such owners.
ISBN 10: 1-292-02198-5

ISBN 13: 978-1-292-02198-0
British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library
Printed in the United States of America

P E A R S O N C U S T O M L I B R A R Y
Table of Contents
1. Language and Reading: Convergences and Divergences

Alan G. Kamhi/Hugh W. Catts 1
2. Reading Development
3. Defining and Classifying Reading Disabilities
4. Causes of Reading Disabilities
5. Assessment and Instruction for Phonemic Awareness and Word Recognition Skills
6. Spelling Assessment and Intervention: A Multiple Linguistic Approach to Improving Literacy Outcomes
7. Perspectives on Assessing and Improving Reading Comprehension
8. Assessing and Remediating Text Comprehension Problems
9. Learning to Write
10. Developing Knowledge and Skills for Writing
Index 309
I
This page intentionally left blank
Language and Reading:
Convergences and Divergences
Alan G. Kamhi and Hugh W. Catts
I
t is now well accepted that reading is a language-based skill. At that time, the idea that most reading
disabilities were best viewed as a developmental language disorder was an emerging one. A develop-
mental language perspective of reading disabilities was the major theme of our original book and
continues to be the major theme of the present text. This view rests, in part, on the fact that there are nu-
merous similarities between spoken and written language. Reading shares many of the same processes
and knowledge bases as talking and listening. Reading, however, is not a simple derivative of spoken lan-
guage. Although spoken language and reading have much in common in terms of the knowledge and
processes they tap, there are also fundamental, nontrivial differences between the two. Knowledge of the
similarities and differences between spoken language and reading is critical for understanding how chil-
dren learn to read and why some children have difficulty learning to read. In this chapter, we begin by
defining language and reading. This is followed by an in-depth comparison of the processes and knowl-
edge involved in understanding spoken and written language. Other differences between spoken and
written language are then discussed.
DEFINING LANGUAGE
Definitions of language are broad based and highly integrative. An example of such a definition is of-
fered by the American Speech-Language-Hearing Association (ASHA, 1983):
Language is a complex and dynamic system of conventional symbols that is used in various modes
for thought and communication. Contemporary views of human language hold that: (a) language
evolves within specific historical, social, and cultural contexts; (b) language, as rule-governed be-
havior, is described by at least five parameters—phonologic, morphologic, syntactic, semantic, and
pragmatic; (c) language learning and use are determined by the interaction of biological, cognitive,
psychosocial, and environmental factors; and (d) effective use of language for communication
requires a broad understanding of human interaction including such associated factors as nonverbal
cues, motivation, and sociocultural roles. (p. 44)
From Chapter 1 of Language and Reading Disabilities, Third Edition. Alan G. Kamhi, Hugh W. Catts.
Copyright © 2012 by Pearson Education, Inc. All rights reserved.
1
Language and Reading: Convergences and Divergences
As reflected in the definition, it is generally agreed that there are five parameters of lan-
guage. These parameters are described briefly in the next section.
Phonology
Phonology is the aspect of language concerned with the rules that govern the distribution and
sequencing of speech sounds. It includes a description of what the sounds are and their compo-
nent features (phonetics), as well as the distributional rules that govern how the sounds can be
used in various word positions and the sequence rules that describe which sounds may be com-
bined. For example, the // sound that occurs in the word measure is never used to begin an
English word. Distributional rules are different in different languages. In French, for example,
the // sound can occur in the word-initial position, as in je and jouer. An example of a sequence
rule in English would be that /r/ can follow /t/ or /d/ in an initial consonant cluster (e.g., truck,
draw), but /l/ cannot.
Semantics
Semantics is the aspect of language that governs the meaning of words and word combinations.
Sometimes semantics is divided into lexical and relational semantics. Lexical semantics involves
the meaning conveyed by individual words. Words have both intensional and extensional mean-
ings. Intensional meanings refer to the defining characteristics or criterial features of a word. A
dog is a dog because it has four legs, barks, and licks people’s faces. The extension of a word is
the set of objects, entities, or events to which a word might apply in the world. The set of all real
or imaginary dogs that fit the intensional criteria becomes the extension of the entity dog.
Relational semantics refers to the relationships that exist between words. For example, in
the sentence The Panda bear is eating bamboo, the word bear not only has a lexical meaning, but
it also is the agent engaged in the activity of eating. Bamboo is referred to as the “patient”
(Chafe, 1970) because its state is being changed by the action of the verb. Words are thus seen as
expressing abstract relational meanings in addition to their lexical meanings.
Morphology
In addition to the content words that refer to objects, entities, and events, there is a group of
words and inflections that conveys subtle meaning and serves specific grammatical and prag-
matic functions. These words have been referred to as grammatical morphemes. Grammatical
morphemes modulate meaning. Consider the sentences Dave is playing tennis, Dave plays ten-
nis, Dave played tennis, and Dave has played tennis. The major elements of meaning are similar
in each of these sentences. The first sentence describes an action currently in progress, whereas
the next sentence depicts a habitual occurrence. The last two sentences describe actions that have
taken place sometime in the past. What differentiates these sentences are the grammatical mor-
phemes (inflections and auxiliary forms) that change the tense and aspect (e.g., durative or per-
fective) of the sentences.
Syntax
Syntax refers to the rule system that governs how words are combined into larger meaningful
units of phrases, clauses, and sentences. Syntactic rules specify word order, sentence organiza-
tion, and the relationships between words, word classes, and sentence constituents, such as noun
phrases and verb phrases. Knowledge of syntax enables an individual to make judgments of
2
well-formedness or grammaticality. For example, all mature English speakers would judge the
sentence The boy hit the ball as well formed and grammatical. In contrast, the sentence Hit the
boy ball the would be judged as ungrammatical. It should be apparent that knowledge of syntax
plays an important role in understanding language.
Pragmatics
Pragmatics concerns the use of language in context. Language does not occur in a vacuum. It is
used to serve a variety of communication functions, such as declaring, greeting, requesting infor-
mation, and answering questions. Communicative intentions are best achieved by being sensitive
to the listener’s communicative needs and nonlinguistic context. Speakers must take into account
what the listener knows and does not know about a topic. Pragmatics thus encompasses rules of
conversation or discourse. Speakers must learn how to initiate conversations, take turns, maintain
and change topics, and provide the appropriate amount of information in a clear manner.
Different kinds of discourse contexts involve different sets of rules (Lund & Duchan, 1993;
Schiffrin, 1994). The most frequent kinds of discourses children encounter are conversational,
classroom, narrative, and event discourses.
DEFINING READING
Reading, like spoken language, is a complex cognitive activity. Gates (1949), for example, de-
fined reading as “a complex organization of patterns of higher mental processes . . . [that] . . .
can and should embrace all types of thinking, evaluating, judging, imagining, reasoning, and
problem-solving” (p. 3). A view of reading that emphasizes higher-level thinking processes is a
broad view of reading (Perfetti, 1986). Thinking guided by print is another way to characterize
a broad view of reading. Reading ability defined in this way is associated with skill in compre-
hending texts. Although this is a widely accepted view of reading, particularly among practition-
ers, there are both practical and theoretical problems with this broad definition.
The fundamental problem with the broad view of reading is that it conflates two very dif-
ferent abilities—word recognition (word-level reading) and comprehension. Word recognition
involves a well-defined scope of knowledge (e.g., letters, sounds, words) and processes (decod-
ing) that can be systematically taught. Comprehension, in contrast, is not a skill with a well-
defined scope of knowledge; it is a complex of higher-level mental processes that includes
thinking, reasoning, imagining, and interpreting (see Kamhi, 2009a). With a broad definition of
reading, a theory of reading necessarily becomes a theory of inferencing, a theory of schemata,
and a theory of learning (Perfetti, 1986). The problems with the broad view of reading led Gough
and his colleagues (Gough & Tunmer, 1986; Hoover & Gough, 1990) to propose the Simple
View of Reading. The central claim of the Simple View is that reading consists of two compo-
nents: decoding and linguistic comprehension. Decoding refers to word recognition processes
that transform print into words. Linguistic comprehension (i.e., listening comprehension) is de-
fined as the process by which words, sentences, and discourses are interpreted (Gough &
Tunmer, 1986).
The Simple View of Reading has appealed to many researchers and practitioners. Some re-
searchers, however, prefer restricting the definition of reading to just the decoding component
(e.g., Crowder, 1982). One advantage of a narrow view of reading is that it delineates a restricted
set of processes to be examined (Perfetti, 1986). Crowder (1982), who advocates a narrow view
of reading, made the following analogy between the “psychology of reading” and the “psychology
3
of braille.” The psychology of braille does not include such topics as inferences and schema ap-
plication. These abilities involve broad-based cognitive-linguistic processes. Crowder argued
that it was superfluous to make the study of these higher level processes part of the study of
braille. The study of braille is necessarily restricted to the decoding process, or how a reader de-
codes braille to language. By analogy, the study of reading should also be restricted to the decod-
ing process.
Kamhi (2009a, b) recently suggested that embracing the narrow view may provide a solu-
tion to the reading crisis in the United States. The basic argument was that it is possible to elim-
inate reading failure if reading is defined narrowly as decoding abilities. Reading proficiency
levels should reach 90 percent, at a minimum, given the numerous research-supported instruc-
tional programs that have been shown to effectively teach word-level reading (National Reading
Panel [NRP], 2000; Simmons et al., 2007). As Catts (2009) pointed out, a narrow view of read-
ing promotes a broad view of comprehension that recognizes its complexity. Not only are there
different levels of understanding (e.g., literal, analytic, creative), but comprehension also de-
pends on thinking and reasoning processes that are domain and content specific rather than do-
main general (cf. Kintsch, 1998). This is why the best predictor of comprehension is often
familiarity with content knowledge domains (Hirsch, 2006; Willingham, 2006).
It should be apparent that the way one defines reading will have a significant impact on
how reading is measured and taught. We encourage educators to embrace a view of reading that
clearly distinguishes word recognition processes from the reasoning and thinking processes in-
volved in comprehension.
MODELS OF SPOKEN AND WRITTEN LANGUAGE COMPREHENSION

In a book about language and reading, an understanding of the similarities and differences be-
tween spoken and written language is crucial. The sections that follow compare the specific
processes and knowledge involved in comprehending spoken and written language. To set the
stage for these comparison, a brief overview of models of language and reading is provided.
Models of spoken and written language comprehension have often been divided into
three general classes: bottom-up, top-down, and interactive. Bottom-up models view spoken
and written language comprehension as a step-by-step process that begins with the initial de-
tection of an auditory or visual stimulus. The initial input goes through a series of stages in
which it is “chunked” in progressively larger and more meaningful units. Top-down models,
in contrast, emphasize the importance of scripts, schemata, and inferences that allow one to
make hypotheses and predictions about the information being processed. Familiarity with the
content, structure, and function of the different kinds of spoken and written discourse enables
the listener and the reader to be less dependent on low-level perceptual information to con-
struct meanings.
Reliance on top-down versus bottom-up processes varies with the material being processed
and the skill of the reader. Bottom-up processes are presumed to be necessary when reading iso-
lated, decontextualized words, whereas top-down processes facilitate not only word recognition
but also discourse-level comprehension. Top-down processes are especially important when
reading partially illegible material, such as cursive writing.
Many language and reading theorists (Perfetti, 1985; Rumelhart, 1977; Stanovich, 1985)
have advocated interactive models in which both bottom-up and top-down processes contribute
to reading and language comprehension. An interactive model of reading comprehension, for
example, would acknowledge that individuals must have proficient word recognition skills as
4
well as higher-level linguistic and conceptual knowledge to be good readers. Whereas bottom-up
and top-down models emphasize sequential processing, interactive models allow for parallel or
simultaneous processing to occur. Later stages could thus begin before earlier stages have been
completed. Although more complex than serial processing models, parallel processing models
better reflect the types of processing that occur in complex tasks such as reading.
Connectionist models have also been used to explain how children learn to recognize
words (e.g., Seidenberg, 1995; Seidenberg & McClelland, 1989). With this approach, the lexicon
is viewed as an interactive network of connections among different layers of processing. Instead
of depicting different routes (top-down or bottom-up) to access meaning, Seidenberg and
McClelland (1989) propose two different layers of units, orthographic and phonological, that
connect with each other and another layer of units that represents meaning. Because activation
levels are input driven, word frequency has a significant impact on word recognition because the
more often a particular set of units is activated together (e.g., phonological, orthographic, con-
ceptual), the greater the strength of the pathway associated with the particular word (cf. Whitney,
1998). A detailed review of parallel processing models of spoken and written language process-
ing is beyond the scope. For our purposes, it is sufficient to note that simplistic serial processing
models, whether bottom-up or top-down, cannot adequately capture the complex interactions
that occur within and between different processing levels.
COMPREHENDING SPOKEN AND WRITTEN LANGUAGE

We have found that the model depicted in Figure 1 provides a useful framework for comparing
the processes and knowledge involved in comprehending spoken and written language. This
model, though unique, shares components with other processing models (Gough & Tunmer,
1986; Thomson, 1984). Although the components of the model will be discussed in a linear, bot-
tom-up fashion, the model should be viewed as an interactive one that allows for parallel pro-
cessing within and between levels.
Auditory Auditory
Input Analysis
(speech) Phonological
Representation
Word Sentence/Text
Comprehension
Meaning Processing
Visual Visual
Visual
Analysis Representation
Input
(print)
Lexicon
Perceptual Analysis Word Recognition Discourse-Level Processes
FIGURE 1 A Model of Spoken and Written Language Comprehension
5
Perceptual Analyses
The input to the perceptual analysis is speech or print. For this input to be recognized, it must be
detected and analyzed. The sensory mechanisms involved in the detection of speech and print are
distinctive; the ear is used to detect speech, and the eye is used to detect print. Sensory deficits
involving hearing or vision place a child at risk for spoken and written language problems.
Children born deaf cannot detect the speech signal through the auditory modality and, as a result,
have considerable difficulty developing intelligible speech. Individuals who are blind cannot de-
tect print through the visual modality. Braille, which relies on the tactile modality, is one way to
bypass the visual deficit. An intact auditory system provides the blind another avenue to access
text material by way of tape recordings.
Once the input has been detected, the segmental and suprasegmental features of spoken
and written words are analyzed. In speech, the processes underlying phonetic discrimination
and phonemic identification are involved. Phonetic discrimination refers to the ability to hear
the difference between two sounds that differ acoustically and phonetically. For example, the
initial t in the word tap is phonetically different from the final t in the word bat. Phonetic differ-
ences that do not affect meaning are often referred to as allophonic variations. If the t sounds in
the preceding words were changed to k sounds, this would change the meaning of the words.
Tap would become cap, and bat would become back. The phonetic differences between /t/ and
/k/ are thus also phonemic differences because they change the meaning of the word. The task
for the young child learning language is to determine which differences between sounds make a
difference in meaning.
The language a child is learning determines which phonetic differences are phonemic. In
Japanese, for example, the differences between /r/ and /l/ are allophonic. In English, however, the
phonetic differences between /r/ and /1/ make a difference in meaning. In French, the front
rounded vowel /y/ is phonemically different from the back rounded /u/. An American who does
not make this distinction will not be able to differentiate between the words tout (all) and tu
(you). These examples are meant to illustrate that learning phonemic categories requires knowl-
edge of the language being learned. The acquisition of phonological knowledge about language
necessarily involves higher-level conceptual processes. Low-level perceptual processes, such as
detection and discrimination, do not lead to knowledge about phonemic categories. In light of
these points, it is important to note that in most listening situations, individuals seldom have to
make distinctions between minimal phoneme pairs (e.g., p/b in the words pin and bin) that are
common stimuli on tests of discrimination. In many instances, lexical and higher-level language
knowledge often eliminate the need for phonemic-level identification.
In reading, just as with speech, discrimination and identification processes are involved. In
reading, discrimination refers to the ability to see the visual differences between letters.
Identification requires knowledge of the correspondences between letters and phonemes. For ex-
ample, the child who confuses the letters b and d in words such as bad and dad is often said to
have a visual discrimination problem. It is more likely, however, that the child can perceive the
visual differences between the letters b and d but has not learned that the letter b is associated
with the phoneme /b/ and the letter d is associated with the phoneme /d/. In other words, the child
has not learned the phoneme–letter correspondences for these two sounds.
To illustrate the difference between low-level visual discrimination ability and higher-
level conceptual (identification) ability, consider the following analogy. In teaching large
classes, it is common to confuse students. The first author once called a girl named Aimee,
Anna. Although Aimee and Anna were both 20-something female graduate students, they could
6
be easily differentiated by their physical characteristics, personalities, clothes, and so forth. He

had no difficulty differentiating between the two students. The problem he had was associating a
particular characteristic or a set of characteristics with a name. The similarity between the two
names makes it more difficult to consistently use the right name with the right student. This is
similar to the problem children have associating the features of a particular phoneme with the
features of a particular letter. When letters and sounds are similar, as is the case for “b” and “d,”
it is particularly difficult to learn the correct correspondences.
These examples are meant to show that sound or letter confusions are not necessarily
caused by phonetic or visual discrimination problems. With respect to spoken language, the dif-
ficulty is learning which phonetic differences make a difference in meaning. With respect to
reading, the difficulty is learning which sounds are associated with which letters. In both cases,
what often appear to be discrimination problems are in fact identification problems.
Word Recognition
Reading and spoken language begin to share similar knowledge domains and processes in the
word recognition stage. Until this point, the processing of print and speech involves different
sensory and perceptual processes. In the word recognition stage, the features identified in the
previous perceptual stage are used to access the mental lexicon. The words heard or seen must
activate or be associated with previously stored concepts in the individual’s mental lexicon.
These stored concepts in the mental lexicon represent one’s vocabulary. Importantly, the content
and structure of the mental lexicon is essentially the same for both reading and spoken language.
The content of the lexicon includes information about the word’s phonological or visual form as
well as information about the word’s meaning and how the word relates to other words.
Consider, for example, the kind of conceptual information that might appear in the mental lexi-
con for the word pencil.
It refers to an instrument used for writing or drawing; it is a manmade physical object, usually
cylindrical in shape; and it functions by leaving a trail of graphite along a writing surface. . . .
A pencil is one of a class of writing instruments and a close relative of the pen, eraser, and
sharpener. (Just & Carpenter, 1987, p. 62)
The mental lexicon also includes syntactic and semantic information that indicates part of speech
(e.g., noun, verb, or adjective) and possible syntactic and semantic roles. For example, the syn-
tactic information about pencil might indicate that it is a noun that functions semantically as an
instrument (“She wrote the letter with a pencil”) or as a patient (“Peggy bought a pencil”).
The structure of the mental lexicon has received considerable research attention during the
past 30 years. Network models consisting of nodes corresponding to concepts and features have
been a popular way to depict the structure of the lexicon (Collins & Loftus, 1975; Collins &
Quillian, 1969). Early network models were hierarchical in nature, with the ordering in the hier-
archy defined by set inclusion relations. For example, higher-order concepts such as animal in-
cluded lower-order concepts such as bird and sparrow. Other network models have been referred
to as heterarchical, reflecting concepts from ill-structured domains (Just & Carpenter, 1987).
Although theorists might differ in their portrayal of the content and structure of the mental lexi-
con, they generally agree that the mental lexicon is the same for language and reading. The way
in which word meanings are accessed can differ, however, in spoken language and reading.
In processing speech, word meaning is accessed through a word’s phonological represen-
tation. The output of the perceptual analysis is a representation of a word’s acoustic and phonetic
7
features. These acoustic–phonetic representations of speech input are used by the listener to
activate or instantiate a word’s phonological representation in the lexicon. This may involve the
listener in attempting to match acoustic–phonetic representations with phonological representa-
tions. Phonological representations are directly linked to a word’s meaning because this informa-
tion is stored together for each word in the mental lexicon.
Phonological representations of words stored in the mental lexicon can take one of sev-
eral forms. Words may contain discrete phonetic and phonemic segments or syllable segments,
or be represented as whole words or short phrases (e.g., “it’s a” as “itsa” and “did you know” as
[dIdəno]). Although young children’s phonological representations begin to contain more dis-
crete phonetic and phonemic information as they progress through the preschool years, the abil-
ity to access this information may not develop until age 5 or later, depending on early literacy
experiences and formal instruction. Studies of young children’s speech perception (e.g.,
Nittrouer, Manning, & Meyer, 1993) have found that there is a gradual shift in the acoustic cues
used to make phonological decisions. Nittrouer and colleagues hypothesize that as children
gain experience with a native language, they become more sensitive to phonetic structure. In a
more recent study, Nittrouer (1996) showed that this shift is related to children’s developing
phonemic awareness. It seems that early exposure to reading as well as developmental changes
in speech perception both contribute to young children’s ability to represent speech as discrete
phonemic segments.
In contrast to speech, in which there is only one way to access a word’s meaning, in read-
ing there are two ways: indirectly, by way of a phonological representation, or directly, by way of
a visual representation (see Figure 1). Use of a visual representation to access the lexicon is var-
iously referred to as the direct, visual, look-and-say, or whole-word approach. In accessing the
lexicon in this way, the reader locates the word in the lexicon whose visual representation con-
tains the same segmental and/or visual features as those identified in the previous perceptual
analysis stage. In other words, a match is made between the perceived visual configuration and a
visual representation that is part of the mental lexicon for the particular word.
Word meaning can also be accessed through a phonological representation. With this
indirect or phonological approach, the reader uses knowledge of phoneme–letter correspondence
rules to recode the visually perceived letters into their corresponding phonemes. Individual
phonemes are then blended together to form a phonological sequence that is matched to a simi-
lar sequence in the lexicon. The phonological approach is particularly important in the develop-
ment of reading. The ability to decode printed words phonologically allows children to read
words they know but have never seen in print. Reading by the phonological approach also causes
the child to attend to the letter sequences within words. The knowledge gained about letter se-
quence makes the child’s visual representations more precise.
Reading by the phonological route is thus similar to speech recognition in that a word is
recognized by way of its phonological representation. There is one important difference, how-
ever, in using phonological representations to access meaning in comprehending spoken and
written language. To successfully use the phonological route in reading, one must have explicit
awareness of the phonological structure of words, specifically, the knowledge that words consist
of discrete phonemic segments (Liberman, 1983). These segments are not readily apparent to
young children because the sound segments of speech are blended together in the acoustic signal.
For example, the word cat is one acoustic event; its sound segments do not correspond exactly to
its three written symbols. Although preschool children might show some phonological aware-
ness, much explicit instruction and practice is usually required for a child to become efficient in
using the phonological approach.
8
The recognition that there were two possible routes to word recognition led to the popular-
ity of dual-route models of word recognition (cf. Stanovich, 1991). Although early proponents of
dual-route models agreed that there were two routes to word recognition, they differed in as-
sumptions about the various speeds of the two access mechanisms and how conflicting informa-
tion was resolved. The size of the sound–letter correspondences in the phonological route also
differed from model to model (e.g., sound-by-sound, syllables, word level). Discussions of the
different variations of these models can be found in Coltheart, Curtis, Atkins, and Haller (1993),
Patterson and Coltheart (1987), and Rayner and Pollatsek (1989).
Questions about the nature of the print-to-sound conversion have recently taken a new turn
due to the increasing popularity of parallel-distributed processing models that contain no word-
level representations or lexicon in the network (Share & Stanovich, 1995). Regardless of how the
print-to-sound conversion takes place, there is recent evidence that this conversion is essential
for the large numbers of low-frequency words that cannot be recognized on a visual basis (Share
& Stanovich, 1995). In contrast, high-frequency words seem to be recognized visually with min-
imal phonological recoding even in the very earliest stages of reading acquisition (Reitsma,
1990). The more exposure a child has to a word, the more likely a visual approach will be used.
The use of a visual versus a phonological approach to word recognition depends on the fre-
quency of the word rather than the particular reading stage a child is in.
Discourse-Level Processes
Up to this point, we have considered the processes involved in recognizing words. Spoken and
written language, however, consists of longer discourse units, such as sentences, conversations,
lectures, stories, and expository texts. Psycholinguistic studies carried out in the 1960s and
1970s (cf. Carroll, 1994; Clark & Clark, 1977) explored the role that syntactic, semantic, and
world knowledge played in comprehending larger units of spoken and written discourse. By
focusing on the independent contribution these different types of knowledge made toward mean-
ing, these early studies were limited in what they could tell us about the interaction of different
types of knowledge and whether different discourse types are processed the same way by listen-
ers and readers. Despite these limitations, it is useful to consider how structural, propositional,
and situation or world knowledge can be used to construct meaning.
STRUCTURAL KNOWLEDGE. A variety of structural cues are used by listeners and readers in
comprehending speech and text. These cues include word order, grammatical morphemes, and
function words such as relative pronouns, conjunctions, and modals. Listeners and readers often
use syntactic and morphologic cues to figure out the meaning of unknown words. Grammatical
morphemes, for example, provide information about word classes. Adverbs are signaled by the
inflections -ly and -y, whereas adjectives are marked by the suffixes -able and -al. Verbs are sig-
naled by the inflections -ed, -ing, and -en. Nouns are marked by definite and indefinite articles,
plural and possessive markers, and suffixes such as -ment and -ness. The reason why readers are
able to make any sense at all out of a sentence like “Twas brillig and the slithy toves did gyre and
gimble in the wabe” is that inflections (y and s) and syntactic markers (the and did) provide cues
about grammatical form class.
Clark and Clark (1977) provide an excellent review of studies that demonstrate the influ-
ence syntactic and morphologic knowledge have on sentence comprehension. It has been shown,
for example, that listeners use function words to segment sentences into constituents, classify the
9
constituents, and construct meanings from them (e.g., Bever, 1970; Fodor & Garrett, 1967).
Consider the following two sentences, one with relative pronouns and one without:
1. The pen that the author whom the editor liked used was new.
2. The pen the author the editor liked used was new.
Fodor and Garrett (1967) found that listeners had more difficulty paraphrasing sentences like (2)
than sentences like (1). More recent studies have continued to attempt to prove that the initial
segmentation of a sentence (i.e., parsing) is performed by a syntactic module that is not influ-
enced by other kinds of knowledge (e.g., Frazier, 1987).
PROPOSITIONAL KNOWLEDGE. Although structural knowledge may play an important role in

understanding sentences, memory for extended discourse rarely maintains structural informa-
tion. The fact that we generally store and remember the gist of what we hear or read suggests that
processing resources must be devoted primarily to constructing meaningful propositions. A
proposition is an idea-unit that consists of a predicate and its related arguments. It is generally
agreed that listeners and readers use their knowledge of predicates and their inherent arguments
to construct propositions. The predicate give, for example, requires three noun phrases or argu-
ments: an agent to do the giving, an object to be given, and a recipient of the object. When listen-
ers hear a sentence like Alison gave the book to Franne, they look for the three arguments
entailed by the predicate give.
A simple semantic strategy suggested years ago by Bever (1970) is that listeners and read-
ers might use content words alone to build propositions that make sense. For example, if the
words pile, raked, girl, leaves were presented without any other syntactic information, it would
be apparent that two propositions were involved: The girl raked the leaves and the leaves were in
a pile. To show that listeners used content words to build propositions, researchers (e.g., Stolz,
1967) showed that semantically constrained sentences (3) were much easier to paraphrase than
semantically unconstrained sentences (4).
3. The vase that the maid that the agency hired dropped broke on the floor.
4. The dog that the cat that the girl fought scolded approached the colt.
It has also been shown that propositional complexity influences processing time. Kintsch
and Keenan (1973), for example, showed that sentence 5, which contains eight propositions,
took significantly more time to read than sentence 6, which contained only four propositions.
Note that the two sentences have about the same number of words.
5. Cleopatra’s downfall lay in her foolish trust in the fickle political figures of the Roman world.
6. Romulus, the legendary founder of Rome, took the women of the Sabine by force.
Subsequent studies have examined the hierarchical networks of propositions that listeners and
readers construct to link propositions within spoken discourse and text. Not surprisingly, re-
searchers have found that the propositions listeners and readers construct are affected by many
factors, such as nature of the discourse/text, knowledge of the world, processing capacity, inter-
est level, and so forth.
WORLD KNOWLEDGE. Structural and propositional knowledge are crucial for constructing
meaning, but an individual’s knowledge of the world or what has come to be called situation
model representations also plays an important role in comprehension. Consider, for example,
how world knowledge makes the sentence Jake ate the ice cream with relish unambiguous,
10
whereas a similar sentence, Jake ate the sausage with relish, is ambiguous (Just & Carpenter,
1987). We know that relish is normally not eaten with ice cream. Such information is not specific
to language; instead, it reflects general knowledge about the tastes of foods to assign with relish.
World knowledge can be divided into knowledge of specific content domains and knowl-
edge of interpersonal relations. Specific content domains would include academic subjects such as
history, geography, mathematics, and English literature; procedural knowledge such as how to fix
a car, tie a shoelace, and play tennis; and scriptlike knowledge of familiar events. Interpersonal
knowledge involves such things as knowledge of human needs, motivations, attitudes, emotions,
values, behavior, personality traits, and relationships. It should be evident how these kinds of
world knowledge play an important role in processing spoken and written language.
Because world knowledge can be so broad, psychologists have focused attention on the sit-
uation-specific world knowledge that listeners and readers use to construct meaning (e.g., van
Dijk & Kintsch, 1983). The assumption is that as we process discourse, we construct a mental or
situational model of the world as described by the discourses.
MODELS OF DISCOURSE PROCESSING. To understand larger units of spoken and written dis-
course, it is necessary not only to construct representations that consider structural, proposi-
tional, and situational information, but also to relate these representations to one another. One
must also use this information to make inferences about meaning and make decisions about
which information should be remembered. Given the variety of knowledge types and cognitive
processes involved in discourse processing, no one model can expect to capture all these facets of
discourse processing. It is useful, however, to consider the kinds of models that have been pro-
posed. Although these models deal primarily with how readers construct meaning from texts,
their basic principles can be applied to spoken language discourse as well.
Kintsch and van Dijk’s initial model of text comprehension proposed that multiple levels
of representation were needed to construct meanings based on different kinds of knowledge
(Kintsch & van Dijk, 1978; van Dijk & Kintsch, 1983). Three levels of representation correspond
to the three knowledge types: structural, propositional, and knowledge of the situation/world.
This initial model of comprehension relied on schema-driven, top-down processing to
build the knowledge of the world (i.e., situation model) representation. Kintsch (1988), however,
felt that these notions were not adaptive to new contexts, were too inflexible, and could not ac-
count for how schemas were initially constructed. His most recent theory, called construction in-
tegration theory, acknowledges that many elements enter into the comprehension process
(Kintsch, 1998). These include perceptions, concepts, ideas, images, or emotions. A crucial con-
sideration in the theory is where these elements come from—from the world via the perceptual
system or from the individual in the form of memories, knowledge, beliefs, body states, or goals.
For Kintsch, the heart of the theory “is a specific mechanism that describes how elements from
these two sources are combined into a stable mental product in the process of comprehension”
(Kintsch, 1998, p. 4).
Kintsch goes on to provide a brief synopsis of the theory: One starts with a reader who has
specific goals, a given background of knowledge and experience, and a given perceptual situa-
tion, such as printed words on a page of text. The propositional idea units created from these
words are then linked to the reader’s goals, knowledge, and experiences to create an interrelated
network of idea units. Unlike Kintsch’s earlier schema-driven models in which context was used
to construct meaning, the construction of the network of idea units is viewed as an entirely
bottom-up process, unguided by the larger discourse context. The initial context-insensitive con-
struction process is followed by “a constraint-satisfaction, or integration, process that yields if all
11
goes well, an orderly mental structure out of initial chaos” (Kintsch, 1998, p. 5). The constraint-
satisfaction process involves selectively activating those elements from the network of idea-units
that fit together and deactivating the rest.
If it all sounds pretty complicated, it is because complicated models and theories are
needed to explain how text information is integrated with a reader’s background knowledge and
experiences to construct meaning. Simplistic bottom-up and top-down models are too general to
explain how meaning is actually constructed, but some of the notions from these models, such as
scripts and schemas, still work well for understanding how children construct meaning for cer-
tain prototype forms of discourse such as familiar events and stories. A schema is generally
thought of as a structure in memory that specifies a general or expected arrangement of a body of
information. Familiar events, for example, are well captured by scripts, which are a particular
type of schema. Scripts contain slots for the components of an event, such as the main actions,
participants, goals, and typical position of each action. Scripts make it easier to process familiar
events by providing individuals with a coherent structure into which they can insert new informa-
tion. Scripts also allow individuals to add necessary information that might be omitted in spoken
or written discourse. For example, familiarity with a restaurant script allows listeners and readers
to anticipate some mention of the menu. If no mention of the menu is made, but information
about the kind of restaurant is given (e.g., Italian), one can infer the contents of the menu.
Certain types of discourse, such as stories, seem to have a consistent structure or grammar.
This was recognized years ago when researchers proposed that stories had a common story
grammar or schema. A story schema can be viewed as a mental framework that contains slots for
each story component, such as a setting, goal, obstacle, and resolution. Story grammars represent
a slightly different characterization of the knowledge of story structures. Story grammars specify
the hierarchical relations among the components more directly than a story schema (Mandler &
Johnson, 1977; Stein & Glenn, 1979). Story grammars attempt to specify the structural organiza-
tion of stories in the same way that syntactic grammars specify the structural organization of sen-
tences (Just & Carpenter, 1987, p. 231). The main structural components of a story are a setting
and an episode. The setting introduces the characters and the context of the story. Episodes can
be further divided into an initiating event, internal response, attempt, consequence, and reaction.
Knowledge of the structure and function of stories, like knowledge of scripts, can facilitate com-
prehension of spoken and written language (Just & Carpenter, 1987; Perfetti, 1985).
OTHER METAPHORS FOR DISCOURSE AND TEXT UNDERSTANDING. The notion that multi-
ple sources of knowledge or representations are involved in processing discourse and text is an
important one for understanding what is involved in comprehension. Other notions about compre-
hension, however, are important as well. Graesser and Britton (1996) have found that five
metaphors capture the essence of the various ways of thinking about text comprehension. The first
metaphor, understanding is the assembly of a multileveled representation, has already been dis-
cussed. Speech-language pathologists and other educators are familiar with at least two of the
other metaphors: understanding is the process of managing working memory and understanding is
inference generation. The two metaphors we may not be familiar with are understanding is the
construction of a coherent representation and understanding is a complex dynamical system. To
these five metaphors, we will add a sixth: understanding is a metacognitive ability. Although
Graesser and Britton apply these metaphors to text understanding, in most cases they can be ap-
plied to spoken discourse as well. Each of these last five metaphors will be discussed briefly next.
Understanding Is the Management of Working Memory. Most psychologists and educators
are comfortable with the assumption that comprehension is managed in a limited-capacity working
12
memory. Every educator has had firsthand experience with this metaphor. For example, when the
demands of comprehension exceed the limitations of working memory, students’ comprehension
decreases dramatically. Students with low working memory spans often experience difficulty
when comprehension components tax working memory. Poor comprehenders have also been
shown to have problems suppressing irrelevant information from working memory
(Gernsbacher, 1996).
Understanding Is Inference Generation. The ability to construct meaning requires more
than interpreting explicit propositions. It involves accessing relevant world knowledge and gen-
erating inferences that are needed to make sentences cohere (local coherence) and to relate text
to world knowledge (global coherence). A number of different systems exist to classify infer-
ences. In several studies comparing inferencing abilities in good and poor readers (e.g., Kucan &
Beck, 1997; Laing & Kamhi, 2002; Trabasso & Magliano, 1996), inferences were classified as
either predictions, associations, or explanations. A predictive inference speculates about events
or actions that may occur based on what has already occurred in a story or text. For example, a
predictive inference for the sentence She played hard everyday would be She probably will be in
good physical shape. An associative inference is a statement that makes generalizations about
characters, actions, objects, or events in a story or text. Associative inferences can also be speci-
fications of procedures or responses to wh-questions. An associative inference for the sentence
He ate ice cream would be that He likes ice cream or He was hungry. An explanatory inference
provides causal connections between actions and events in a story or text. They are usually re-
sponses to “why” questions that provide explanations for a state, event, or action. For example,
in a story about a child who wants a faster computer, an explanatory inference might be He was
not very happy because he wanted a new computer. The proportion of explanatory inferences
generated has been found to be significantly related to comprehension performance (e.g.,
Trabasso & Magliano, 1996). This is not surprising because explanatory inferences require re-
trieving and remembering causal information that serves to unite propositions in a story.
Inferences can also be distinguished according to whether they are derived from the con-
tent of activated world knowledge structures (e.g., scripts and schemas) or whether they are
novel constructions that are needed to construct the situation model. Inferences that are gener-
ated from existing world knowledge tend to be generated “online.” Graesser and Britton (1996)
argue that a satisfactory model of text understanding should be able to accurately predict infer-
ences that are quickly or automatically made during comprehension as well as those that are time
consuming. Inferences generated online include those that address readers’ goals, assist in estab-
lishing local or global coherence, and are highly activated from multiple information sources
(e.g., Long, Seely, Oppy, & Golding, 1996). Inferences that are more time consuming may be
caused by minimal world knowledge about the topic or by contradictions, anomalies, or irrele-
vant propositions in the text. Readers attempt to generate explanations and justifications to re-
solve the contradictions and anomalies. The process of generating these “elaborative inferences”
is necessarily time consuming and may not be used by readers with low motivation (Graesser &
Britton, 1996, p. 350).
Understanding Is the Construction of Coherent Representations. The basic notion with this
metaphor is that the more coherent the discourse or text, the easier it is to understand. A text is
fully connected if every proposition is conceptually connected to one or more other propositions.
Some theorists, following Kintsch (1974), believe that noun-phrase arguments are critical for
connecting propositions and establishing coherence. More recent research, however, has shown
that argument overlap is neither a necessary nor sufficient condition for establishing coherence;
13
instead, it is merely one type of connection (cf. Graesser & Britton, 1996). Other types of con-
nections that have been considered include the connections between predicates of propositions
(Turner, Britton, Andraessen, & McCutchen, 1996), causal connections and goals of story char-
acters (van den Broek, Risden, Fletcher, & Thurlow, 1996), and the connections that tie deep
metaphors to lexical items and explicit expressions (Gibbs, 1996).
Despite the challenge of identifying the specific types of connections that tie texts together,
the “understanding-as-coherence” metaphor makes a large number of predictions about compre-
hension performance. Most of these predictions are generally intuitive. For example, a proposition
has a greater likelihood of being recalled when it has more connections to other propositions in the
text, and reading time increases when there is a break in coherence. However, some are counterin-
tuitive. For example, Mannes and St. George (1996) found that there are more connections (or
stronger ones) between text and world knowledge if there is a discrepancy between an outline and
text content. The discrepancy causes improved problem solving, though recall for the text suffers.
Understanding Is a Complex Dynamic System. As mentioned earlier, static, linear models of

spoken and written language may be useful to identify specific processes and knowledge
domains, but they do not have the flexibility to handle complex dynamic systems such as com-
prehension. A detailed description of a dynamic text comprehension model is beyond the scope
(cf. Graesser & Britton, 1996). It is interesting to note, however, that even researchers committed
to these models recognize the difficulty involved in testing their psychological plausibility
(Graesser & Britton, 1996, p. 347). Despite the difficulty in determining which dynamic model
provides the best “goodness of fit,” no cognitive theorist has rejected the “understanding is a
complex dynamic system” metaphor.
Understanding Is a Metacognitive Ability. Metacognition refers to one’s knowledge and

control of one’s cognitive system (Brown, 1987). Metacognitive abilities have been associated
with several aspects of reading, including establishing the purpose for reading, identifying im-
portant ideas, activating prior knowledge, evaluating the text for clarity, compensating for failure
to understand, and assessing one’s level of comprehension (Brown, 1987). Brown added that it is
not clear whether all or just certain components of these activities are metacognitive.
The ability to monitor comprehension plays an important role in both spoken and written
language comprehension (e.g., Dollaghan & Kaston, 1986). When faced with a word, sentence,
paragraph, or other text element that is not understood, it is necessary to do something to aid un-
derstanding, such as ask for clarification or reread the text in question. Individuals who are adept
at monitoring their comprehension are more proficient processors of spoken and written language.
Summary
We have attempted in this section to provide a way of thinking about the knowledge and processes
involved in understanding spoken and written language. Although the emphasis has been primarily
on the similarity of knowledge and processes, some important differences in the word recognition
processes were acknowledged. In our discussion of discourse comprehension processes, we tended
to treat research as if it applied both to spoken and written language comprehension when, in fact,
it rarely did. Our assumption here was that a model of comprehension that is sufficiently dynamic,
flexible, and multifaceted would apply equally well to both spoken and written discourse. Although
the six metaphors discussed were meant to illuminate the different aspects of comprehension, per-
haps they made a complete muddle of comprehension for some. Graesser and Britton (1996)
thought that after reading through their book on text understanding with all its different models and
14
views of comprehension, someone might ask, “What is text understanding?” Readers might won-
der the same thing about our view of comprehension. With a slight modification to include
discourse as well as text comprehension, the definition of comprehension Graesser and Britton
suggest provides a good answer to the question:
Text [and discourse] understanding is the dynamic process of constructing coherent represen-
tations and inferences at multiple levels of text and context, within the bottleneck of a limited-
capacity working memory. (p. 350)
Having emphasized the similarities between spoken and written language up to this point, in the
next section we consider some of the differences between the two.
DIFFERENCES BETWEEN SPOKEN AND WRITTEN LANGUAGE

Delineating the similarities and differences in the processes and knowledge involved in spoken
and written language comprehension only begins to capture the complex relationship that exists
between language and reading. Consider, for example, the following question posed by Gleitman
and Rozin (1977, p. 2): Why is the more general and complex task of learning to speak and un-
derstand less difficult and less variable than what appears to be a trivial derivative of this (i.e.,
learning to read and write)? These authors proceed to point out two major differences between
learning to talk and learning to read. We add a third important difference.
The first major difference is that learning to read requires explicit knowledge of the phono-
logical aspects of speech. To become an efficient reader, one must learn the various correspon-
dences between phonemes and letters. The knowledge that words consist of discrete phonemes is
crucial for constructing phoneme–grapheme correspondence rules. Spoken language compre-
hension also requires analysis of utterances into smaller phonological units. But the analysis of
the speech stream by the listener is carried out below the level of consciousness by evolutionar-
ily old and highly adapted auditory perceptual processes (Lieberman, 1973). The human percep-
tual system is thus biologically adapted to process speech. In contrast, the human visual system
is not biologically adapted to process written text. This introduces the second major difference
between learning to talk and learning to read: Reading is a comparatively new and arbitrary
human ability for which specific biological adaptations do not yet exist.
A third important difference is that almost all humans are reared in environments in which
spoken language is the principal means of communication. Thus, not only are we biologically
endowed to learn language, but we are also socialized to use spoken language to communicate.
This is not true for reading. In 2008, there were 796 million illiterate adults worldwide and
two-thirds of them (64%) were women (UNESCO Institute for Statistics, 2010). More than half
(412 million) of these individuals live in Southern Asia and about 20% (176 million) reside in
sub-Saharan Africa. These are regions with limited educational opportunities and cultures that
place little value on reading.
Perera (1984) points out additional differences between spoken and written language. An
understanding of these differences helps to further explain why reading is not a simple derivative
of spoken language. The differences discussed in the following sections, however, in no way di-
minish the language bases of reading and reading disabilities.
To emphasize the contrasts between written and spoken language, Perera (1984) compared
prototypical speech (conversation) to prototypical written language (literature or informative
prose). She acknowledged, however, that there is a full range of spoken and written discourse
types. Certain discourse types have some characteristics of written language and vice versa. For
15
example, speeches and lectures can be planned much like writing, radio talk lacks a visual di-
mension and contextual support, and tape recordings are durable.
Physical Differences
Whereas speech consists of temporally ordered sounds, writing consists of marks made on a sur-
face (e.g., paper) in a two-dimensional space. As such, writing is relatively durable; it can be read
and reread. Speech, unless it is recorded, is ephemeral. It has no existence independent of the
speaker. The durability of writing gives the reader control over how fast or slow to read. Certain
texts can be savored, whereas others can be skimmed. The listener, in contrast, is tied to the fleet-
ing speech of the speaker. Missed words or sentences will be lost if clarification is not requested.
Perera (1984, p. 161) noted that readers often have the benefit of a whole range of visual
cues, such as running headlines, different-size type, color, and summaries or abstracts. In addi-
tion, a device such as the footnote allows the writer to provide additional information without in-
terrupting the main thread of the text. Such devices allow the reader to decide the level at which
he or she will read. The listener, in contrast, is completely dependent on the speaker’s selection
of material. Note, however, that the listener could choose not to listen to the speaker’s message.
Situational Differences
The most frequent type of spoken language is face-to-face communication. Conversations are
often interactive exchanges between two or more individuals. Questions are followed by an-
swers, requests by responses, and statements by acknowledgments. When a listener does not un-
derstand something, a clarification is requested. Careful planning is not the rule in conversational
discourse. When speakers pause too long before talking, they will usually be interrupted. Despite
this time pressure to speak, misunderstandings are infrequent; when they occur, they are easily
resolved by repeating or rephrasing the message. Nonverbal communication acts, such as ges-
tures, facial expressions, and body postures, can help to clarify messages. Speakers and listeners
also share the same nonlinguistic setting. People and objects that are visible can be referred to by
pronouns rather than by noun phrases (even without prior reference), and many adverbials and
prepositions can be expressed by here, there, and like this.
In contrast, writing and reading are often individual endeavors. The writer receives no
prompting about what to write and no immediate feedback on the clarity of the writing. But the
writer is generally under less severe time constraints and can thus take more time to search for
the best way to express a message. The writer can also correct and revise a text until a final copy
is produced. Such care and precision is necessary in writing because there are no contextual and
nonverbal cues to aid comprehension. The written text thus has to bear the whole burden of com-
munication, which is one reason why writing is usually more precise than talking.
Functional Differences
One of the earliest needs to generate a writing system was to retain accurate records of property,
commercial transactions, and legal judgments. A Chinese proverb holds that “The palest ink is
better than the best memory.” Writing has enabled the knowledge of centuries to accumulate,
thus allowing each new generation to build on the ideas, discoveries, and inventions of the gen-
eration or generations before. Many academic subjects, such as history, geography, the physical
sciences, and social sciences, owe their very existence to writing (Perera, 1984, p. l64). Another
function not served by speech is labeling. Although speech is used to label objects in a referential
16
sense, written labels serve more of an information function. Consider such labels as street
names, signposts, nameplates on theaters and public buildings, brand labels, and danger warnings.
Written language can also serve a variety of communicative functions, such as relating stories,
events, and experiences or sharing information and making requests. Finally, a specialized function
of writing is found in literature. Societies have oral literatures, but oral literatures are restricted to
a few types, such as ballads, epic poetry, drama, folk stories, and myths. Essays, novels, diaries,
and memoirs are some of the genres that are particular to writing.
Perera (1984) has suggested that the most basic uses of writing involve the recording of
facts, ideas, and information. Although speech also has an informative function, an equally im-
portant function of speech is the role it plays in establishing and maintaining human relation-
ships. A large part of everyday speech with friends, acquaintances, and other individuals serves
social-interpersonal functions rather than intellectual ones. E-mail and instant messaging now
serve this role as well.
One advantage writing has over speech, according to Perera (1984, p. l65), is that it allows
ideas to be explored at leisure and in private. Writing can thus become a means of extending and
clarifying one’s thinking and ideas. Often in conversation when a controversial topic is raised,
there is a tendency for opinions to polarize. Someone who tries to take both sides of a issue might
be pressed to select one particular view. In writing, however, one can take time to develop a line of
thought, weigh opposing arguments, notice errors in reasoning, and develop new lines of thinking.
Form Differences
The most obvious difference in form is that speech consists of sounds, whereas written language
consists of letters. As indicated earlier, this would not be so much of a problem if speech sounds
(i.e., phonemes) stood in one-to-one correspondence with written letters. Form differences be-
tween spoken and written language are not limited to the discrete segments (i.e., phonemes and
letters) that make up speech and text. Spoken and written language also differ in how they repre-
sent suprasegmental, paralinguistic, and prosodic features. Paralinguistic features include pitch
and timbre differences that distinguish male and female voices; general voice quality, such as
breathiness, hoarseness, or nasality; and the general manner of how an utterance is produced,
such as shouted, whispered, or spoken. Perera has pointed out that these features do not usually
affect the actual meaning of an utterance; however, they may reflect the speaker’s attitude about
what is being said.
Prosodic features include intonation, stress, and rhythm. Perera (1984) presented four
functions of prosodic features: (1) to enable the communicative intent of an utterance to differ
from its grammatical form (e.g., He’s lost it versus He’s lost it?), (2) to group words into infor-
mation units, (3) to place emphasis, and (4) to convey the speaker’s attitude. These functions dif-
fer in the extent to which they can be reflected in writing. Whereas punctuation effectively
changes the communicative intent of an utterance, it is not so effective in signaling which words
belong together in information units. Italics, underlining, and the use of capital letters are some
ways to distribute emphasis throughout a written utterance. But heavy use of these devices in for-
mal writing is usually discouraged. Expressing attitudes in writing is clearly difficult. Perera
(1984, p. l78) provided an example of how much attitudinal information is conveyed by prosodic
features in the following quote of a journalist who listened to one of the Watergate tapes:
Once you hear the tapes, and the tone in which he (Nixon) uttered the comments which previ-
ously have only been available in a neutral transcript, any last shred of doubt about his guilt
must disappear.
17
Perera (1984) goes on to consider the extent to which the writing system represents the
segmental and suprasegmental aspects of speech. Among other things, she pointed out that
graphemes represent the “citation” (well-spoken) form of words rather than the degraded
productions that often occur in fast speech (e.g., compare “did you know” to [dIdəno]).
Punctuation can signal the grammatical function of a sentence and mark some prosodic bound-
aries. The writer, however, has no conventional way to express voice quality, volume, rate of
speech, rhythm, and into national patterns.
Vocabulary Differences
One would expect that there would be differences in the vocabulary used in spoken and written
language because writing provides more time and, therefore, more resources to select words. The
additional time allows writers to choose words that communicate meanings clearly. Clear, unam-
biguous writing is necessary to ensure that the author’s intended meaning is derived. Readers,
unlike listeners, don’t have the luxury of requesting clarification when the message is unclear. In
contrast, conversational speech provides little opportunity to consider alternative word choices,
and though one can revise a word choice once it is spoken, too much fumbling detracts from ef-
fective communication (Chafe & Danielewicz, 1987). With writing, time is often not a factor;
one can spend seconds, minutes, hours, or days finding the appropriate word or expression, and
even after a selection is made, the writer is free to revise without anyone noticing. Word process-
ing programs do not keep track of earlier drafts.
The consequence of these differences is that the vocabulary of spoken language tends to be
more limited in variety. A simple way to demonstrate this is by calculating type-token ratios
(TTRs) for spoken and written language. TTR is calculated by dividing the number of different
words by the total number of words in a sample. Chafe and Danielewicz (1987) show that TTRs
for spoken language are consistently lower than for written language. Interestingly, the ratio for
academic lectures is about the same as in conversations (.19/.18), indicating that rapid production
of spoken language produces less varied vocabulary, regardless of the kind of speaking involved.
The frequent use of nonspecific terms (e.g., thing, whatever, “doohickey,” “whatchamacallit”),
hedges (sort of, kind of), and maze behaviors (interjections, disfluencies, false starts, repetitions)
are all indications of the processing demands of spoken language. Chafe and Danielewicz also
show how spoken language has less referential explicitness than written language. Nonspecific
third person pronouns (it, this, that) are used frequently in spoken language and are one of the
factors that differentiate good from poor writing.
The lexicon that speakers and writers choose from is also not the same for writing and
speaking. There is a literate lexicon (Nippold, 2007) that writers draw from in formal writing.
For example, conjunctive adverbs like thus, therefore, hence, and accordingly are rarely used
in spoken language. Chafe and Danielewicz (1987) suggest that spoken language compensates
for its restricted lexical variety by assigning a premium to freshness. Speakers must stay cur-
rent. Cool may have been cool 10 or 20 years ago, but not now. Freshness of vocabulary is less
important in writing, where there is more of a premium on choosing the right word to convey
a particular meaning. Not surprisingly, conversations and academic papers differ considerably
in their use of literary and colloquial vocabulary (Chafe & Danielewicz, 1987). Academic pa-
pers had only 1 instance per 1,000 words of colloquial vocabulary (e.g., kid bike, figure out),
whereas conversations had 27. Lectures and letters fall somewhere in the middle, indicating
that there is nothing in the nature of speaking that prevents a speaker from using literary
vocabulary and nothing in the nature of writing that prevents a writer from using colloquial
18
vocabulary. Lectures are thus more literary than conversations, whereas letters are more con-
versational than academic papers.
Grammatical Differences
Samples of spoken language uncover relatively high frequencies of coordination, repetition, and
rephrasing. Conversational discourse is typically low in lexical density and high in redundancy.
Lexical items are spaced out and separated by grammatical words, and a high number of total
words are used to convey a relatively small amount of information. Written language, in contrast,
is high in lexical density and low in redundancy. This results from the use of grammatical struc-
tures that decrease redundancy and increase lexical density.
Studies (e.g., Chafe & Danielewicz, 1987) have shown that in conversation it is more com-
mon to provide smaller amounts of information at a time. Most written language, by contrast, is
denser lexically as well as propositionally. Conversations, because of their interactive nature, are
generally less coherent than writing. Speakers are free to change the subject at almost any point
in a conversation. Topics need not be related in any logical way. In writing, however, an overall
theme is necessary. Topic changes must be justified and explicitly made. Writing also has pre-
scribed rules for organizing content. These rules cover the use of topic sentences, paragraph
structure, and introductory and concluding statements.
Processing Differences
Earlier in this chapter, we talked about top-down processing models, discourse-level comprehen-
sion processes, and the higher-level knowledge schemas that contribute to comprehension of
spoken and written language. The focus in these sections was on the commonalities between un-
derstanding speech and text. There are very important differences, however, in the contribution
higher-level processes make to spoken and written language comprehension. The role of higher-
level processes or context effects in reading has received considerable research attention and
caused much confusion. One reason for this confusion is that researchers often fail to distinguish
between the use of context to facilitate word recognition and the use of context to facilitate text
comprehension. Context plays an important role in facilitating text comprehension; it plays a
very limited role, however, in facilitating word recognition in good readers.
Support for the limited role of higher-level processes in word recognition comes from eye-
movement experiments. Research using various eye-movement methodologies has been consistent
in finding that the vast majority of content words in text receive a direct visual fixation (Just &
Carpenter, 1987; Rayner & Pollatsek, 1989). Short function words may be skipped, but even
many of these receive a direct visual fixation. The span of effective visual information during a
fixation is thus quite small, meaning that text is sampled in a very dense manner, even when the
words are highly predictable (Balota & Chumbley, 1985).
Based in part on evidence from these eye-movement studies, most models of reading have
expectancy-based, top-down processes functioning after words have been recognized
(Seidenberg, 1985; Till, Mross, & Kintsch, 1988). Higher-level contextual information plays
more of a role in speech perception or language processing because of the well-documented am-
biguity in decontextualized speech. For example, isolated words from normal conversation are
often not recognized out of context. This is not the case, however, for written language. Fluent
readers can identify written words out of context with near-perfect accuracy. As Stanovich
(1986) has noted, the physical stimulus alone completely specifies the lexical representation in
writing, whereas this is not always true in speech. It is more important in reading, therefore, for
19
the input systems involved in word recognition to deliver a complete and accurate representation
of words to higher level processes. Paradoxically, then, poor readers who have difficulty accu-
rately decoding words must rely more on contextual information than good readers who have
proficient word recognition skills.
BASIC FACTORS IN READING AND LANGUAGE DEVELOPMENT

It should be clear that although there is considerable overlap in the processes involved in spoken
and written language, there are also many important differences between the two. These differ-
ences explain to a large extent why learning to read is not a simple derivative of learning to talk
and understand. In the definition of language given earlier in this chapter, language learning and
use were said to be determined by the interaction of biological, cognitive, psychosocial, and en-
vironmental factors. Learning to read is also determined by the interaction of these four factors.
However, the relative importance or weight of these factors for learning to read is not the same as
it is for learning spoken language.
Biological factors are crucial in learning spoken and written language. As indicated earlier,
however, one important difference between learning to talk and learning to read is that the analy-
sis of the speech stream is carried out below the level of consciousness by evolutionary old and
highly adapted auditory processes. In contrast, the human visual system is not biologically
adapted to process written text. By itself, this difference does not necessarily make learning to
read more difficult than learning to talk; it does suggest, however, that learning to read requires
more attentional resources than learning to talk.
Environmental factors play different but equally important roles in learning spoken and
written language. As noted previously, almost all humans are reared in environments in which
spoken language is the principal means of communication. The social–environmental forces to
use language to communicate are just as crucial for language learning as the physical, perceptual,
and cognitive mechanisms that make speech, hearing, and language possible. Children deprived
of early exposure to language input will eventually develop some language abilities once normal
input is provided, but they will never be normal language users, as the tragic cases of Genie and
other severely deprived children have shown (Curtiss, 1977). Although it is rare to find examples
of extreme deprivation of language input, there are still many societies in the world that place lit-
tle importance or value on literacy. These societies account for the high rates of illiteracy (40 per-
cent) in the world. Most of the individuals reared in these societies will have little exposure to
print and no formal instruction in reading.
Because the biological and social bases of reading are not as strong as they are for spoken
language, psychosocial factors, such as motivational and attentional states, often play a more im-
portant role in learning to read than in learning to talk. Unless a child has a severe emotional dis-
order, such as autism, language learning will be relatively unaffected by motivational and
attentional states. This is not the case in learning to read because reading requires a considerable
amount of motivational and attentional resources. Reading difficulties in individuals with moti-
vational and attentional problems have been well documented (e.g., Guthrie et al., 2009; Willcutt &
Pennington, 2000).
Cognitive factors play a fundamental role in learning spoken and written language because
spoken and written language are essentially cognitive achievements. Both rely on basic cognitive
processes to encode, store, and retrieve information. In addition, the same store of linguistic and
conceptual knowledge is tapped by readers as by speakers and listeners. Metacognitive abilities,
however, play a more important role in learning to read than in learning to talk and understand.
20
Another random document with
no related content on Scribd:
I hear within our house?
The ground shakes, | and the home of Gymir
Around me trembles too.”
The Serving-Maid spake:
15. “One stands without | who has leapt from his

steed,
And lets his horse loose to graze;”
. . . . . . . . . | . . . . . . . . .
. . . . . . . . . . . . . . .
Gerth spake:
16. “Bid the man come in, | and drink good mead
Here within our hall;
Though this I fear, | that there without
My brother’s slayer stands.
[113]
17. “Art thou of the elves | or the offspring of gods,

Or of the wise Wanes?
How camst thou alone | through the leaping flame
Thus to behold our home?”
Skirnir spake:
18. “I am not of the elves, | nor the offspring of

gods,
Nor of the wise Wanes;
Though I came alone | through the leaping flame
Thus to behold thy home.
19. “Eleven apples, | all of gold,

Here will I give thee, Gerth,
To buy thy troth | that Freyr shall be
Deemed to be dearest to you.”
Gerth spake:
20. “I will not take | at any man’s wish

These eleven apples ever;
Nor shall Freyr and I | one dwelling find
So long as we two live.”
Skirnir spake:
21. “Then do I bring thee | the ring that was burned

[114]
Of old with Othin’s son;
From it do eight | of like weight fall
On every ninth night.”
Gerth spake:
22. “The ring I wish not, | though burned it was

Of old with Othin’s son;
In Gymir’s home | is no lack of gold
In the wealth my father wields.”
Skirnir spake:
23. “Seest thou, maiden, | this keen, bright sword

That I hold here in my hand?
Thy head from thy neck | shall I straightway hew,
If thou wilt not do my will.”
Gerth spake:
24. “For no man’s sake | will I ever suffer

To be thus moved by might;
But gladly, methinks, | will Gymir seek
To fight if he finds thee here.”
Skirnir spake:
25. “Seest thou, maiden, | this keen, bright sword

That I hold here in my hand? [115]
Before its blade | the old giant bends,—
Thy father is doomed to die.
26. “I strike thee, maid, | with my magic staff,

To tame thee to work my will;
There shalt thou go | where never again
The sons of men shall see thee.
27. “On the eagle’s hill | shalt thou ever sit,
And gaze on the gates of Hel;
More loathsome to thee | than the light-hued
snake
To men, shall thy meat become.
28. “Fearful to see, | if thou comest forth,

Hrimnir will stand and stare,
(Men will marvel at thee;) [116]
More famed shalt thou grow | than the watchman
of the gods!
Peer forth, then, from thy prison.
29. “Rage and longing, | fetters and wrath,

Tears and torment are thine;
Where thou sittest down | my doom is on thee
Of heavy heart
And double dole.
30. “In the giants’ home | shall vile things harm

thee
Each day with evil deeds;
Grief shalt thou get | instead of gladness,
And sorrow to suffer with tears.
31. “With three-headed giants | thou shalt dwell

ever,
Or never know a husband;
(Let longing grip thee, | let wasting waste thee,—)
[117]
Be like to the thistle | that in the loft
Was cast and there was crushed.
32. “I go to the wood, | and to the wet forest,

To win a magic wand;
. . . . . . . . | . . . . . . . .
I won a magic wand.
33. “Othin grows angry, | angered is the best of the

gods,
Freyr shall be thy foe,
Most evil maid, | who the magic wrath
Of gods hast got for thyself.
34. “Give heed, frost-rulers, | hear it, giants,

Sons of Suttung,
And gods, ye too,
How I forbid | and how I ban
The meeting of men with the maid,
(The joy of men with the maid.)
[118]
35. “Hrimgrimnir is he, | the giant who shall have

thee
In the depth by the doors of Hel;
To the frost-giants’ halls | each day shalt thou fare,
Crawling and craving in vain,
(Crawling and having no hope.)
36. “Base wretches there | by the root of the tree

Will hold for thee horns of filth;
A fairer drink | shalt thou never find,
Maid, to meet thy wish,
(Maid, to meet my wish.)
37. “I write thee a charm | and three runes

therewith,
Longing and madness and lust;
But what I have writ | I may yet unwrite
If I find a need therefor.”
[119]
Gerth spake:
38. “Find welcome rather, | and with it take

The frost-cup filled with mead;
Though I did not believe | that I should so love
Ever one of the Wanes.”
Skirnir spake:
39. “My tidings all | must I truly learn

Ere homeward hence I ride:
How soon thou wilt | with the mighty son
Of Njorth a meeting make.”
Gerth spake:
40. “Barri there is, | which we both know well,

A forest fair and still;
And nine nights hence | to the son of Njorth
Will Gerth there grant delight.”
Then Skirnir rode home. Freyr stood without, and

spoke to him, and asked for tidings:
41. “Tell me, Skirnir, | ere thou take off the saddle,
Or farest forward a step:
What hast thou done | in the giants’ dwelling
To make glad thee or me?”
[120]
Skirnir spake:
42. “Barri there is, | which we both know well,

A forest fair and still;
And nine nights hence | to the son of Njorth
Will Gerth there grant delight.”
Freyr spake:
43. “Long is one night, | longer are two;
How then shall I bear three?
Often to me | has a month seemed less
Than now half a night of desire.”
[107]
[Contents]
NOTES
[108]
Prose. Freyr: concerning his father, Njorth, and the race of the
Wanes in general, cf. Voluspo, 21 and note. Snorri thus describes
Njorth’s family: “Njorth begat two children in Noatun; the son was
named Freyr, and the daughter Freyja; they were fair of aspect and
mighty. Freyr is the noblest of the gods; he rules over rain and
sunshine, and therewith the fruitfulness of the earth; it is well to call
upon him for plenty and welfare, for he rules over wealth for
mankind. Freyja is the noblest of the goddesses. When she rides to
the fight, she has one-half of the slain, and Othin has half. When she
goes on a journey, she drives her two cats, and sits in a cart. Love-
songs please her well, and it is good to call on her in love-matters.”
Hlithskjolf: Othin’s watch-tower; cf. Grimnismol, introductory prose.
He said: both manuscripts have “Then Skathi said:” (Skathi was
Njorth’s wife), but Bugge’s emendation, based on Snorri’s version, is
doubtless correct.
1. My son: both manuscripts, and many editors, have “our son,”
which, of course, goes with the introduction of Skathi in the prose. As
the stanza is clearly addressed to Skirnir, the change of pronouns
seems justified. The same confusion occurs in stanza 2, where
Skirnir in the manuscripts is made to speak of Freyr as [109]“your
son” (plural). The plural pronoun in the original involves a metrical
error, which is corrected by the emendation.
4. Elfbeam: the sun, so called because its rays were fatal to elves
and dwarfs; cf. Alvissmol, 35.
6. Gymir: a mountain-giant, husband of Aurbotha, and father of

Gerth, fairest among women. This is all Snorri tells of him in his
paraphrase of the story.
7. Snorri’s paraphrase of the poem is sufficiently close so that his

addition of another sentence to Freyr’s speech makes it probable
[110]that a stanza has dropped out between 7 and 8. This has been
tentatively reconstructed, thus: “Hither to me | shalt thou bring the
maid, / And home shalt thou lead her here, / If her father wills it |
or wills it not, / And good reward shalt thou get.” Finn Magnusen
detected the probable omission of a stanza here as early as 1821.
8. The sword: Freyr’s gift of his sword to Skirnir eventually proves

fatal, for at the last battle, when Freyr is attacked by Beli, whom he
kills bare-handed, and later when the fire-demon, Surt, slays him in
turn, he is weaponless; cf. Voluspo, 53 and note. Against the giants
grim: the condition of this line makes it seem like an error in copying,
and it is possible that it should be identical with the fourth line of the
next stanza. [111]
10. Some editors reject line 3 as spurious.
12. Line 2 is in neither manuscript, and no gap is indicated. I have

followed Grundtvig’s conjectural emendation.
13. This stanza is almost exactly like many in the first part of [112]the
Hovamol, and may well have been a separate proverb. After this
stanza the scene shifts to the interior of the house.
15. No gap indicated in either manuscript. Bugge and Niedner have

attempted emendations, while Hildebrand suggests that the last two
lines of stanza 14 are spurious, 14, 1–2, and 15 thus forming a
single stanza, which seems doubtful.
16. Brother’s slayer: perhaps the brother is Beli, slain by Freyr; the
only other references are in Voluspo, 53, and in Snorri’s paraphrase
of the Skirnismol, which merely says that Freyr’s gift of his sword to
Skirnir “was the reason why he was weaponless when he met Beli,
and he killed him bare-handed.” Skirnir himself seems never to have
killed anybody. [113]
17. Wise Wanes: cf. Voluspo, 21 and note.
18. The Arnamagnæan Codex omits this stanza.
19. Apples: the apple was the symbol of fruitfulness, and also of
eternal youth. According to Snorri, the goddess Ithun had charge of
the apples which the gods ate whenever they felt themselves
growing old. [114]
21. Ring: the ring Draupnir (“Dropper”) was made by the dwarfs for
Othin, who laid it on Baldr’s pyre when the latter’s corpse was
burned (cf. Voluspo, 32 and note, and Baldrs Draumar). Baldr,
however, sent the ring back to Othin from hell. How Freyr obtained it
is nowhere stated. Andvari’s ring (Andvaranaut) had a similar power
of creating gold; cf. Reginsmol, prose [115]after stanza 4 and note.
Lines 3 and 4 of this stanza, and the first two of stanza 22, are
missing in the Arnamagnæan Codex.
25. The first two lines are abbreviated in both manuscripts.

26. With this stanza, bribes and threats having failed, Skirnir begins
a curse which, by the power of his magic staff, is to fall on Gerth if
she refuses Freyr.
27. Eagle’s hill: the hill at the end of heaven, and consequently
overlooking hell, where the giant Hræsvelg sits “in an eagle’s guise,”
and makes the winds with his wings; cf. Vafthruthnismol, 37, also
Voluspo, 50. The second line is faulty in both manuscripts;
Hildebrand’s emendation corrects the error, but omits an effective
touch; the manuscript line may be rendered “And look and hanker for
hell.” The Arnamagnæan Codex breaks off with the fourth line of this
stanza.
28. Hrimnir: a frost-giant, mentioned elsewhere only in Hyndluljoth,

33. Line 3 is probably spurious. Watchman of the gods: Heimdall; cf.
Voluspo, 46. [116]
29. Three nouns of doubtful meaning, which I have rendered rage,

longing, and heart respectively, make the precise force of this stanza
obscure. Niedner and Sijmons mark the entire stanza as
interpolated, and Jonsson rejects line 5.
30. In Regius and in nearly all the editions the first two lines of this
stanza are followed by lines 3–5 of stanza 35. I have followed
Niedner, Sijmons, and Gering. The two words here translated vile
things are obscure; Gering renders the phrase simply “Kobolde.”
31. The confusion noted as to the preceding stanza, and a metrical

error in the third line, have led to various rearrangements and
emendations; line 3 certainly looks like an interpolation. Three-
headed giants: concerning giants with numerous heads, cf.
Vafthruthnismol, 33, and Hymiskvitha, 8. [117]
32. No gap indicated in the manuscript; Niedner makes the line here
given as 4 the first half of line 3, and fills out the stanza thus: “with
which I will tame you, / Maid, to work my will.” The whole stanza
seems to be either interpolated or out of place; it would fit better after
stanza 25.
33. Jonsson marks this stanza as interpolated. The word translated

most evil is another case of guesswork.
34. Most editors reject line 3 as spurious, and some also reject line
6. Lines 2 and 3 may have been expanded out of a single line
running approximately “Ye gods and Suttung’s sons.” Suttung:
concerning this giant cf. Hovamol, 104 and note. [118]
35. Most editors combine lines 1–2 with stanza 36 (either with the
first two lines thereof or the whole stanza), as lines 3–5 stand in the
manuscript after line 2 of stanza 30. Hrimgrimnir (“The Frost-
Shrouded”): a giant not elsewhere mentioned. Line 5, as a repetition
of line 4, is probably a later addition.
36. For the combination of this stanza with the preceding one, cf.
note on stanza 35. The scribe clearly did not consider that the stanza
began with line 1, as the first word thereof in the manuscript does not
begin with a capital letter and has no period before it. The first word
of line 3, however, is so marked. Line 5 may well be spurious.
37. Again the scribe seems to have been uncertain as to the stanza
divisions. This time the first line is preceded by a period, but begins
with a small letter. Many editors have made line 2 [119]into two half-
lines. A charm: literally, the rune Thurs (þ); the runic letters all had
magic attributes; cf. Sigrdrifumol, 6–7 and notes.
40. Barri: “The Leafy.” [120]
42. Abbreviated to initial letters in the manuscript.
43. The superscription is lacking in Regius. Snorri quotes this one

stanza in his prose paraphrase, Gylfaginning, chapter 37. The two
versions are substantially the same, except that Snorri makes the
first line read, “Long is one night, | long is the second.” [121]
[Contents]
HARBARTHSLJOTH
The Poem of Harbarth
[Contents]
Introductory Note
The Harbarthsljoth is found complete in the Codex Regius, where it
follows the Skirnismol, and from the fourth line of stanza 19 to the
end of the poem in the Arnamagnæan Codex, of which it occupies
the first page and a half.
The poem differs sharply from those which precede it in the Codex
Regius, both in metrical form and in spirit. It is, indeed, the most
nearly formless of all the Eddic poems. The normal metre is the
Malahattr (cf. Introduction, where an example is given). The name of
this verse-form means “in the manner of conversation,” and the
Harbarthsljoth’s verse fully justifies the term. The Atli poems
exemplify the conventional use of Malahattr, but in the Harbarthsljoth
the form is used with extraordinary freedom, and other metrical
forms are frequently employed. A few of the speeches of which the
poem is composed cannot be twisted into any known Old Norse
metre, and appear to be simply prose.
How far this confusion is due to interpolations and faulty

transmission of the original poem is uncertain. Finnur Jonsson has
attempted a wholesale purification of the poem, but his arbitrary
condemnation of words, lines, and entire stanzas as spurious is quite
unjustified by any positive evidence. I have accepted Mogk’s theory
that the author was “a first-rate psychologist, but a poor poet,” and
have translated the poem as it stands in the manuscripts. I have
preserved the metrical confusion of the original by keeping
throughout so far as possible to the metres found in the poem; if the
rhythm of the translation is often hard to catch, the difficulty is no
less with the original Norse.
The poem is simply a contest of abuse, such as the early Norwegian

and Icelander delighted in, the opposing figures being Thor and
Othin, the latter appearing in the disguise of the ferryman Harbarth.
Such billingsgate lent itself readily to changes, interpolations and
omissions, and it is little wonder that the poem is chaotic. It consists
mainly of boasting and of references, often luckily obscure, to
disreputable events in the life of one or the other of the disputants.
Some editors have sought to read a complex symbolism into it,
particularly by representing [122]it as a contest between the noble or
warrior class (Othin) and the peasant (Thor). But it seems a pity to
take such a vigorous piece of broad farce too seriously.
Verse-form, substance, and certain linguistic peculiarities, notably

the suffixed articles, point to a relatively late date (eleventh century)
for the poem in its present form. Probably it had its origin in the early
days, but its colloquial nature and its vulgarity made it readily
susceptible to changes.
Owing to the chaotic state of the text, and the fact that none of the
editors or commentators have succeeded in improving it much, I
have not in this case attempted to give all the important emendations
and suggestions. The stanza-divisions are largely arbitrary.
[Contents]
Thor was on his way back from a journey in the East,

and came to a sound; on the other side of the sound
was a ferryman with a boat. Thor called out:
1. “Who is the fellow yonder, | on the farther shore
of the sound?”
[123]
The ferryman spake:
2. “What kind of a peasant is yon, | that calls o’er

the bay?”
Thor spake:
3. “Ferry me over the sound; | I will feed thee

therefor in the morning;
A basket I have on my back, | and food therein,
none better;
At leisure I ate, | ere the house I left,
Of herrings and porridge, | so plenty I had.”
The ferryman spake:
4. “Of thy morning feats art thou proud, | but the

future thou knowest not wholly;
Doleful thine home-coming is: | thy mother,
methinks, is dead.”
Thor spake:
5. “Now hast thou said | what to each must seem

The mightiest grief, | that my mother is dead.”
[124]
The ferryman spake:
6. “Three good dwellings, | methinks, thou hast

not;
Barefoot thou standest, | and wearest a beggar’s
dress;
Not even hose dost thou have.”
Thor spake:
7. “Steer thou hither the boat; | the landing here

shall I show thee;
But whose the craft | that thou keepest on the
shore?”
The ferryman spake:
8. “Hildolf is he | who bade me have it,

A hero wise; | his home is at Rathsey’s sound.
He bade me no robbers to steer, | nor stealers of
steeds,
But worthy men, | and those whom well do I know.
Say now thy name, | if over the sound thou wilt
fare.”
Thor spake:
9. “My name indeed shall I tell, | though in danger I

am, [125]
And all my race; | I am Othin’s son,
Meili’s brother, | and Magni’s father,
The strong one of the gods; | with Thor now
speech canst thou get.
And now would I know | what name thou hast.”
The ferryman spake:
10. “Harbarth am I, | and seldom I hide my name.”
Thor spake:
11. “Why shouldst thou hide thy name, | if quarrel

thou hast not?”
Harbarth spake:
12. “And though I had a quarrel, | from such as

thou art
Yet none the less | my life would I guard,
Unless I be doomed to die.”
[126]
Thor spake:
13. “Great trouble, methinks, | would it be to come
to thee,
To wade the waters across, | and wet my middle;
Weakling, well shall I pay | thy mocking words,
If across the sound I come.”
Harbarth spake:
14. “Here shall I stand | and await thee here;

Thou hast found since Hrungnir died | no fiercer
man.”
Thor spake:
15. “Fain art thou to tell | how with Hrungnir I

fought,
The haughty giant, | whose head of stone was
made;
And yet I felled him, | and stretched him before
me.
What, Harbarth, didst thou the while?”
[127]
Harbarth spake:
16. “Five full winters | with Fjolvar was I,

And dwelt in the isle | that is Algrön called;
There could we fight, | and fell the slain,

(Download PDF) Language and Reading Disabilities 3Rd Edition Allyn Bacon Communication Full Chapter PDF

Uploaded by

Copyright:

Available Formats

(Download PDF) Language and Reading Disabilities 3Rd Edition Allyn Bacon Communication Full Chapter PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Download PDF) Language and Reading Disabilities 3Rd Edition Allyn Bacon Communication Full Chapter PDF

Uploaded by

Copyright:

Available Formats

Language and Reading Disabilities (3rd

Edition ) (Allyn & Bacon

eTextbook 978-0137021062 Clinical Phonetics (The Allyn

Teaching Reading to Students Who Are At Risk or Have

Top 50 SAT Reading, Writing, and Language Skills, 3rd

Reading Picture Books with Infants and Toddlers:

500 SAT Reading, Writing and Language Questions to Know

Elsevier Weekblad - Week 26 - 2022 Gebruiker

Critical Care EEG Basics-Rapid Bedside EEG Reading for

Language Intervention Strategies in Aphasia and Related

Visit us on the World Wide Web at: www.pearsoned.co.uk

© Pearson Education Limited 2014

ISBN 10: 1-292-02198-5

British Library Cataloguing-in-Publication Data

Printed in the United States of America

1. Language and Reading: Convergences and Divergences

Alan G. Kamhi and Hugh W. Catts

MODELS OF SPOKEN AND WRITTEN LANGUAGE COMPREHENSION

COMPREHENDING SPOKEN AND WRITTEN LANGUAGE

be easily differentiated by their physical characteristics, personalities, clothes, and so forth. He

PROPOSITIONAL KNOWLEDGE. Although structural knowledge may play an important role in

Understanding Is a Complex Dynamic System. As mentioned earlier, static, linear models of

Understanding Is a Metacognitive Ability. Metacognition refers to one’s knowledge and

DIFFERENCES BETWEEN SPOKEN AND WRITTEN LANGUAGE

BASIC FACTORS IN READING AND LANGUAGE DEVELOPMENT

The Serving-Maid spake:

15. “One stands without | who has leapt from his

17. “Art thou of the elves | or the offspring of gods,

18. “I am not of the elves, | nor the offspring of

19. “Eleven apples, | all of gold,

20. “I will not take | at any man’s wish

21. “Then do I bring thee | the ring that was burned

22. “The ring I wish not, | though burned it was

23. “Seest thou, maiden, | this keen, bright sword

24. “For no man’s sake | will I ever suffer

25. “Seest thou, maiden, | this keen, bright sword

26. “I strike thee, maid, | with my magic staff,

28. “Fearful to see, | if thou comest forth,

29. “Rage and longing, | fetters and wrath,

30. “In the giants’ home | shall vile things harm

31. “With three-headed giants | thou shalt dwell

32. “I go to the wood, | and to the wet forest,

33. “Othin grows angry, | angered is the best of the

34. “Give heed, frost-rulers, | hear it, giants,

35. “Hrimgrimnir is he, | the giant who shall have

36. “Base wretches there | by the root of the tree

37. “I write thee a charm | and three runes

38. “Find welcome rather, | and with it take

39. “My tidings all | must I truly learn

40. “Barri there is, | which we both know well,

Then Skirnir rode home. Freyr stood without, and

42. “Barri there is, | which we both know well,

6. Gymir: a mountain-giant, husband of Aurbotha, and father of

7. Snorri’s paraphrase of the poem is sufficiently close so that his

8. The sword: Freyr’s gift of his sword to Skirnir eventually proves

10. Some editors reject line 3 as spurious.

12. Line 2 is in neither manuscript, and no gap is indicated. I have

15. No gap indicated in either manuscript. Bugge and Niedner have

17. Wise Wanes: cf. Voluspo, 21 and note.