An

Introduction to
Linguistics
L. Ben Crane - Edward Yeager
Randal L.Whitman
 AN
INTRODUCTION
TO
LINGUISTICS
 *

; 7 ae
.

» MOR IVIGORTHE 2
ae 5 de a 7
OFT Arba 2 ees *
 AN
INTRODUCTION

LINGUISTICS
TO

L. Ben Crane
Edward Yeager _
Randal L. Whitman

fp
LITTLE, BROWN AND COMPANY
Boston Toronto
To Marilyn, Cindy, and Leanna

Copyright © 1981 by L. Ben Crane, Edward Yeager,

and Randal L. Whitman

All rights reserved. No part of this book may be re-

produced in any form or by any electronic or mechan-
ical means including information storage and retrieval
systems without permission in writing from the pub-
lisher, except by a reviewer whe may quote brief pas-
sages in a review.

Library of Congress Catalog Card No. 80-83800

ISBN 0-316-160156

9876543 2

ALP
Published simultaneously in Canada
by Little, Brown & Company (Canada) Limited

Printed in the United States of America

 PREFACE

Not very long ago, an introductory course in linguistics consisted

largely of plowing through a set of English phrase-structure rules and
memorizing the details of a great number of transformations. The re-
mainder of the course typically included some sort of treatment of
phonetics and phonemics and perhaps a sprinkling of comments on
the nature of language and linguistic theory. Since the formulation of
phrase-structure and transformational rules changed every few years,
students learned something about the way syntax was done at a par-
ticular time and very little else about linguistics.
The emphasis on syntax certainly was understandable. Chomsky’s
Syntactic Structures and Aspects of the Theory of Syntax were the
landmarks of transformational-generative grammar, and the over-
whelming enthusiasm of linguists for Chomsky’s work led to an early
tendency to equate linguistics with transformational-generative
syntax — at least in the classroom.
Times have changed. Because of advances in the past decade in
such areas as phonology, semantics, psycholinguistics, sociolinguis-
tics, and neurolinguistics, linguists have come to realize that their
subject demands a broader view of language than they might have
previously suspected.
Textbooks, too, have changed. A trend has begun to provide more
balanced coverage of the subareas of linguistics, and syntax in par-
ticular has been given a more reasonable emphasis. We hope to con-
tinue this trend by covering such areas as the history of linguistics,
psycholinguistic experiments, recent approaches to semantics, and
acoustic phonetics — areas that have been virtually ignored in most
other texts.
We have designed this book to be an introduction to linguistics for
undergraduates and beginning graduate students, although we hope
Vv
vi Preface

that it will be of some value to anyone who wishes to know something

about linguistics. It is divided into three major parts.
In the first part of the book, ‘‘Language and Linguistics: An Over-
view,’ the first two chapters consider the so-called communicative
acts of animals and the particular type of communication used by hu-
mans. Language is discussed in terms of its three basic subsystems:
speech, writing, and gesture. Chapters 3 and 4 provide a historical
perspective on linguistics and the English language. Chapter 3, ‘‘His-
tory of Linguistics’? serves a twofold purpose. It introduces the sub-
ject, allowing meaningful references in other chapters to particular
approaches to grammar; and it impresses upon the student the linguis-
tic tradition that has served as the basis for such modern ideas as
generativity, case grammar, and universal grammar. The ‘‘History of
English’’ discusses the beginnings and development of English — the
language used for analysis in the text and the primary language of
most readers of the text.
Part Two, ‘Linguistics: Grammar,’’ will undoubtedly be a major
b]

topic in an introductory course in linguistics. Chapters 5-7, on pho-

netics and phonology, deal with the articulatory basis of sound, how
sound is structured, and how sounds function in language. The next
chapter, “‘Morphology’’ serves somewhat as a bridge between the
chapters on sound and those on meaning and deals with word forma-
tion. The remaining two topics, syntax and semantics, deal with
meaningful units in language and how they occur in patterns. Phonol-
ogy, morphology, syntax, and semantics are all approached primarily
from the transformational-generative point of view, with introductory
sections on previous approaches. Possible modifications of the stan-
dard transformational-generative approach also are indicated where
pertinent.
In the final part of the text, ‘Linguistics: A Broader View,’’ we
examine aspects of linguistics that go beyond grammar. The chapters
on psycholinguistics and sociolinguistics are of great importance in
this respect, since the consideration of psycholinguistic experimenta-
tion and language variation has recently re-established the view of
language as a fundamentally psychological and sociological phenome-
non. The chapters on anthropological linguistics and_historical-
comparative linguistics view language across cultures and through
time — two. essential perspectives on language and linguistics. The
final chapters on neurolinguistics and acoustic phonetics discuss the
initial and terminal points of language: the brain and physical sound.
Each chapter includes a summary that recapitulates the main points
of the chapter and a list of sources for further study. Questions for
discussion and exercises, ranging from the somewhat simple to the
somewhat difficult, also are provided; answers to the exercises appear
at the end of the text.
 NOTE
TO THE
STUDENT

Linguistics is one of the most unusual and interesting subjects in your

studies. It is a broad-based field, ranging from abstract mathematical
and philosophical theory, to the relationship of the behavioral sci-
ences to language, to simple, fun-and-games word play.
In this text, we have attempted to provide an introduction to the
various aspects of linguistics. Some of the material presented may
appear difficult on a first reading, but much of it will be readily acces-
sible. The student is encouraged to find his or her particular interest
in this large field and pursue it in more detail through the suggested
readings that appear at the end of each chapter.
Throughout this text we have been guided by a theory of language
known as transformational-generative grammar. We have found this a
useful way to organize and present the material in the text. Further-
more, it has been the dominant approach to linguistics in the past
twenty-five years. However, if students continue in linguistics beyond
this introductory course, they will surely encounter opposing view-
points and theories. In fact, we have tried to incorporate many of
these in the text. Thus, the student is encouraged to keep an open
mind concerning the subject and to question those statements which
may appear somewhat dogmatic.

Vil
 i

4 oe. ae ai (Aan sirai? fe v, Ra an

et er a ee envib ery: * Sew; Cte

ee ee ees .
7 obian’ Som). id in TTY ews so tn rene

ta ee ee Laas ile,
0 oCyie ere howe Sy tra aed ccmabienaeieees
Oo are Fg Qt ee eee nila
SUA NE hie i Zum Fae storia
; do ‘ Sepheatel ives Bi sine wh
: 'yeeee geen ice? evo Ve
7 ee Sn a press ;
pra, Hea (ratne ieneoa oe aiae
POR O of wit ol Cite: le
aia Ch galt sp oelhengo ead
nae yp ae peeither ahha
i, We Wc? Maher tan) aieinte tieggta
Seee oe a papas A eabaly
wey eo tee 1 geldy t -

ne re ;

same
 CONTENTS

PART I
LANGUAGE AND LINGUISTICS: AN OVERVIEW/1

1. Animal Communication/3
Communication and Animals/3
Characteristics of Animal Communication/6
The Messages of Animals/6
Learned and Innate Communication/8
Language, Apes, and Humans/9
Defining Language/10
The *‘Talking Apes’’ and Language/1 1
Summary/12
Sources for Further Reading/12
Questions for Discussion/13

2. Human Communication: Three Systems/14

Speech/14
Sound and Meaning/15
Syntax and Semantics/16
Writing/17
Logographic Writing/17
Syllabic Writing/18
Alphabetic Writing/19
Mixed Writing Systems/21
Gesture/22
Summary/26
Sources for Further Reading/26
Questions for Discussion/27
X Contents

3. History of Linguistics/28
The Ancient Grammarians/28
The Middle Ages, the Renaissance, and the Age of
Enlightenment/30
The Nineteenth Century: Linguistics Becomes a Science/32
Traditional and Structural Linguistics in the Twentieth
Century/33
Transformational-Generative Grammar/34
Summary/36
Sources for Further Reading/36
Questions for Discussion/38
. History of English/39
The Establishment of Anglo-Saxon/39
Old English/43
Middle English/44
Early Modern English/47
Modern English/48
American English/50
Summary/52
Sources for Further Reading/5S3
Questions for Discussion/54

PART II
LINGUISTICS: GRAMMAR/S55
. Phonetics/57
The Production of Speech Sounds/57
Consonants/59
Stops/59
Fricatives/62
Affricates/63
Nasals/63
Liquids/63
Glides/64
Vowels/64
Distinctive Features/66
Suprasegmentals/68
Stress/69
Intonation/69
Summary/70
Sources for Further Reading/70
Exercises/7 |
 Contents XI

6. Phonology 1/72
Structuralist Phonemics/72) |
Minimal Pairs and Phonemes/73
Morphophonemes/74
Structuralist Phonology versus Transformational-Generative
Phonology/75
Phonological Rules/76
Underlying Representations and Related Issues/78
Dealing with Exceptions/8 1
The Status of the Concept/82
Summary/83
Sources for Further Reading/83
Exercises/84
. Phonology II/85
Introduction/85
Stress in Transformational-Generative Phonology/86
Redundancy/88
Markedness and Natural Phonology/91
Summary/93
Sources for Further Reading/94
Exercises/94

. Morphology/96
Structuralist Morphology/96
Types of Morphemes/97
Allomorphs/98
Morphology and Transformational-Generative Grammar/99
Summary/100
Sources for Further Reading/100
Exercises/101

. Syntax 1/102
Defining a Sentence/102
The Traditional Approach to Syntax/103
The Structuralist Approach/104
The Transformational-Generative Approach/106
Performance and Competence/106
A Model of Competence/108
Phrase-Structure Grammar/109
Lexical Insertion/114
Summary/115
Sources for Further Reading/115
Exercises/1 16
xii Contents

10. Syntax II/117

Transformations/117
Deep Structure, Surface Structure, and Meaning/118
Two Obligatory Transformations/120
Imperatives, Questions, and Negations/121
Transformations with Embedded Sentences/122
The Ordering of Transformations/124
Constraints in Syntax/125
Summary/127
Sources for Further Reading/127
Exercises/127
11. Semantics I/129
Words and Word Relationships/130
Dictionaries and Lexicons/131
The Semantics of Sentences/134
Summary/ 135
Sources for Further Reading/136
Exercises/136
12. Semantics II/138
Interpretive Semantics/138
The Semantic Component/139
Extended Standard Theory/141
Generative Semantics/142
Relating Semantic Structure and Surface Structure/142
Comparing Generative and Interpretive Semantics/143
Case Grammar and Relational Grammar/144
Advantages of Case Grammar/146
Relational Grammar/146
Linguistics and Logic/148
Speech Acts/150
Summary/152
Sources for Further Reading/152
Exercises/153

PART II
LINGUISTICS: A BROADER VIEW/155
13. Psycholinguistics/157
The Psychological Reality of Transformations and Constituent
Structure/158
Transformations/159
Constituents/160
 Contents Xiil

The Psychological Reality of Phonological and Semantic

Representations/161
Distinctive Features and Markedness/162
The Lexicon and Memory/162
Other Approaches/163
Language Acquisition/164
Phonological Acquisition/164
The Two-Word Stage/165
Syntactic Acquisition/166
An Alternate View: The Behaviorist Approach/166
An Assessment/168
Perception and Production of Language/169
The Analysis-by-Synthesis Model/169
The Speech-Perception Grammar/170
Speech Production/171
Summary/172
Sources for Further Reading/172
Exercises/174

14. Sociolinguistics/175
Diversity in Language and Society/175
Linguistic Geography/177
Dialect Differences/179
Lingua Francas, Pidgins, and Creoles/181
Slang/183
The Correlation of Social Patterns and Linguistic Data/183
Language Variation and Grammatical Theory/186
Summary/188
Sources for Further Reading/189
Exercises/190

15. Anthropological Linguistics/192

Language, Thought, and Culture/192
The Whorfian Hypothesis/193
Culture and Language/194
Language Families and Language Types/194
The Indo-European Family/195
Other Language Families/196
Structural Types/198
Language Universals/199
Summary/201
Sources for Further Reading/202
Exercises/203
XIV Contents

16. Historical-Comparative Linguistics/204

Language Change/204
Phonological Change/206
The Causes of Phonological Change/206
The Process of Phonological Change/208
Describing Phonological Change/208
Morphological, Syntactic, and Semantic Change/211
Simplification and Elaboration in Syntactic Change/212
Aspects of Semantic Change/213
Methods of Reconstruction/214
The Comparative Method/214
Internal Reconstruction/217
Summary/218
Sources for Further Reading/219
Exercises/220
17. Neurolinguistics/222
Aphasia and Brain Mapping/222
Types of Aphasia/223
Brain Areas and Language/224
Speech and Species/225
Language and Consciousness/226
Summary/229
Sources for Further Reading/229
Questions for Discussion/229
18. Acoustic Phonetics/230
The Physical Basis of Speech Sounds/230
Vowels/233
Consonants/236
Some Applications/237
Summary/238
Sources for Further Reading/238
Questions for Discussion/239

ANSWERS TO EXERCISES/240
GLOSSARY/249
BIBLIOGRAPHY/262
INDEX/273
 AN
INTRODUCTION
TO
LINGUISTICS
 : ro Loar

as Fiche: 7
a _ (d=

sie 2
4
a 1

ie

ei
_ =
_
7 -

a a
_
-
 PART

LANGUAGE
AND
LINGUISTICS:
AN
OVERVIEW
fe

¥ re
<= ~~

f a. j =
+

a ¢

ir
7

: =e
5
e

Pe) «

OAUOVAL z
a,
-

tyee
7
=
a
"|
a

‘ Pa
 CHAPTER

ANIMAL
COMMUNICATION

When Darwin presented his explanation of how living things evolve,

he challenged the belief that human beings have a very special, indeed
unique place in the universe. Since his time, people have searched for
proof of their superiority to other animals. This search has usually
focused on the powers of the large human brain, and especially on the
ability to use language. ‘‘All animals can communicate, but only
human beings have language’”’ is a statement that has been made re-
peatedly in the twentieth century.
This claim raises a host of questions. What, for example, is com-
munication, and how does animal communication differ from human
communication? What makes language different from the cries of
birds? Can apes be taught a language — and, if so, what does that say
about apes and about humans? By looking first at the nature of com-
munication and then at how animals communicate, we can clarify
what language is and why the ability to use language is very signifi-
cant in defining human nature.
t
Communication and Animals

anitmenes uses onl: or srmboles fea cepa ants to transmit

3
4 Animal Communication

the information ‘‘I am thirsty,’’ he or she must first put that infor-
mation into the symbolic system of language — in this case, En-
ish.
ymmunication. Next, he or she may choose
to Verbalize this message, as conics to writing it or miming it.
3. Delivering the symbols through a medium. A medium is the physi-
cal basis for communication, for example light, air, or ink. In this
case, the medium is the air which conveys the sound waves of a
verbalized message.
4. Perceptual processing
munication is to occur, a receiver must perceive the symbols; the
receiver must see or hear or feel the symbols sent. In this example,
the AED ear receives the gouud waves.
Di ig yr iformation. Even if the
receiver woueshes the symbols, nothin is communicated unless
the receiver is able to decode the message contained in the sound
waves. One assumes that the receiver knows the individual words
and grammar of the language, and thus comprehends the message.

According to this definition of communication, how do we know

when communication has taken place? This question might arise even
when considering human communication, but it is especially interest-
ing — and perplexing — when it concerns animal communication.
Let us say we observe two dogs being walked in opposite directions.
As they approach and then pass, one, a Pekinese, gives out a fury of
yaps with much straining at his leash; the other, a larger nondescript
pooch, looks at the yapping Pekinese but otherwise does not respond.
Has communication occurred? Even if we stipulate that the Pekin-
ese’s noises are intended as communication — that they do encode
some type of information—we cannot then assume that com-
munication has occurred; for communication requires that the goal, or
the pooch, process and decode the message. How do we know that
the pooch has done so? Quite simply, we do not.
Despite such problems in analyzing the behavior afanimals, animal
psychologists generally agree on the conditio
sense to as abet ene Salient

Ste coreantarehus, if swan A ruffles itseu

ore and swan B promptly dives, then it seems reasonable to con- ‘
sider the question of what the ruffling of feathers by a swan ‘‘says’’ to
another swan. We are not obliged to conclude that this behavior is
communicative; in fact, such a conclusion on the basis of one obser-
vation would be unwarranted. If, however, we observe time after
time that when one swan ruffles its feathers, another swan dives, then
 Communication and Animals
. ‘a . .

the hypothesis that ruffling of feathers by swans is a communicative

event becomes increasingly attractive.
Establishing such a connection between the behavior of one animal
and another, however, is far from easy. Karl v n Frisch’s studies
5 mmunication among bees providea model of how
eos relationships can be s udied. Von Frisch spent decades of his life
observing the behavior of bees. Eventually, he came to suspect that
certain bee behavior is communicative. To study this behavior, he
carefully organized the environment of his bees to elicit the same be-
havior repeatedly and to evaluate the response of other bees. He es-
tablished a new source of nectar within the foraging radius of his bee
hive and waited for a bee to discover it. After finding the nectar, the
bee would return to the hive and go into a ‘‘dance.’’ As an apparent
result, other bees would then fly directly to the new source of nectar.
This sequence of events happened often enough to permit the tenta-
tive conclusion that there was something about the_bee’s dance that
communicated the location of the source of the nectar.
Von Frisch went on to decipher the nature of this communication
among bees, working out the code carried by the bee’s dance. Thus
he learned that a specific dance translated into English might say,
‘“‘This quality food is located at a distance of 140 meters, 300° to the
left of the sun’s bearing.’’ Later in the chapter we will examine how
bees communicate this information.
The more general significance of von Frisch’s studies is that he
showed that bees have a spontaneous nonhuman communication
system that can be reliably translated. We may suspect that the howls
and twitters of many animals are communicative, but only through
studies and findings such as von Frisch’s can we demonstrate beyond
a shadow of a doubt that they are in fact communicating.
What happens if, for one reason or another, we are unable to es-
tablish a translation or a function for the call of an animal? What if we
cannot establish clearly a relationship between the call of one animal
and the behavior of another? Such a case appears to exist when we
consider the sounds made by cetaceans — marine mammals such as
dolphins and whales. They have the ability to make an incredible vari-
ety of sounds — far more than any other animal, including humans,
can make. Their low-pitched sounds are believed to function as ce-
tacean sonar; that is, like bats they use echos to locate objects and to
navigate. But even if we subtract these low-pitched sounds, the in-
ventcry of sounds that cetaceans might use for communication is
staggering. The impulse to conclude that they do communicate with
these sounds is very nearly overwhelming. As oceanographer Jacques
Cousteau (1972) put it: ‘‘My friends and I have perhaps spent too
much time with whales; we may be the victims of an illusion. But how
6 Animal Communication

can we explain these alternating voices, and such a diversity of modu-

lation, except by concluding that it is actually conversation?’’’ No
direct evidence that communication is involved has been found; the
conclusion that cetaceans communicate is defended instead by ask-
ing, What else can these sounds be? For now, at least, scientists can
say little more than this about many forms of animal behavior.

Characteristics of Animal Communication

Much has been learned, however, about many systems of animal

communication. A vast array of creatures has been observed — ants
A
and bees, sticklebacks and whales, chimpanzees and baboons — and, e 6
not surprisingly, many types of communication have been found: __
analyzing these varied systems and comparing them with human com-
munication, two questions are especially interesting: What kinds of
messages can particular animals convey, and are their systems of
communication genetically set or learned?

THE MESSAGES OF ANIMALS

Many animals howl, twitter, or sing. Some birds even mimic human
words and phrases. But what, if anything, are they saying? Lorenz
(1952) noted the case of a parrot who surprised everyone by calling,
appropriately, ‘‘The chimney sweep is coming!’’ But Lorenz denied
that the bird knew what it was saying. The parrot, Papagallo by name,
had a terrible fear of the chimney sweep; his visits were times of
severe trauma for Papagallo, who learned to associate them with the
cook’s cry, “‘The chimney sweep is coming.’’ If the cook had cried,
‘“‘The fat’s burning’’ at exactly those same instances, it may be pre-
sumed that this is what Papagallo would finally have said.
Like Papagallo, many animals vocalize in response to stress. Some
cries and actions, however, have more specific meanings that are con-
veyed to other members of the species. Scientists have been able to
assign rough translations to many bird and mammal calls. The famed
animal behaviorist Konrad Lorenz (1952), for example, has discussed
a number of calls of the jackdaw: the kia (‘‘fly with me’’), kiaw (‘‘let’s
go home’’), zick (a nesting call), and others. But Lorenz noted that
these calls are ‘‘not comparable to our spoken language, but only to
those expressions such as yawning, wrinkling the brow, and smiling.
. . . The ‘words’ of the various animal ‘languages’ are merely inter-
jections.’’?
Some animal calls or signals indicate a readiness to mate; others,

1 Jacques-Yves Cousteau, The Whale: Mighty Monarch of the Sea, p. 116.

? Konrad Lorenz, King Solomon's Ring, p. 78.
 wate re . °
Characteristics of Animal Communication

ownership of a territory; others, that danger is near. African vervet -

monkeys, for example, use four or five alarm calls to identify a spe-
cific enemy: a chatter for snakes, uh! or nyow! for a minor bird or
mammal, and rraup when a large predatory bird is seen. The response
of other vervets depends on which signal is given (Wilson 1980).
Many animal calls and actions occur in response to Nely specific stim-
uli and call ian a very ea response. 1e animal cant
; it only slightly — and cadiOoeD
als. For example, Wilson (1980) noted
that nonhuman vertebrates seem to have no more than thirty or forty
specialized patterns to convey information. Thus, animal com-
munication seems to be stereotyped and limited.
But some animals can communicate sophisticated messages, as the
behavior of bees indicates. Von Frisch (1967) found that the dance
performed by many kinds of bees after they have soy a nichessource

source of food deh but not where the source of oe is to be

found. If the food is farther from the hive, however, dee returning
ro ch source of food goes into a waggle dance immediately on —
ts returr he hive. Other bees, perhaps cued by sensing the
eye of rs food adhering to the dancing bee’s body, follow the
dancer and so perceive certain aspects of the dance that communicate
the location of the food source. One aspect of the dance indicates the
direction to the food; another, the distance to it.

FIGURE I-1
Round Dance (left) and Waggle Dance

eS
a 7
—— ue pi og Roe

Source: Reprinted from Karl von Frisch, Bees: Their Vision, Chemical Senses, and
Language, p. 70. Copyright © 1950, © 1971 by Cornell University. Used by permission
of the publisher, Cornell University Press.
8 Animal Communication

The dance, which proceeds as a repetitive figure eight, includes a

straight portion. The angle that this straight portion makes with re-
spect to the vertical is the same (to within about three degrees of arc)
as the angle lying between the direction to the food source and a hori-
zontal line in the direction of the sun. (This description applies to ver-
tical hives, which is the normal hive arrangement.) Thus, if the
straight portion of the dance is exactly vertical, the food source lies in
exactly the direction of the sun.
The frequency with which the dancer accomplishes the straight por-
tion of the dance is inversely related to the distance of the food
source; the faster the dance, the closer the food. The dance may
range from ten runs in fifteen seconds (for a food source about a
hundred meters away) down to less than five runs in fifteen seconds
(if the source is about fifteen hundred meters distant); in some cases,
bees can reliably communicate distances of five thousand meters or
more.
The information conveyed by the bee’s dance is impressive. The
dance even contains information about the quality of the food source.
The higher the sugar concentration of the food, the longer and livelier
the dance. Still, as far as we know, the bee’s vocabulary is quite lim-
ited, and its communications are stereotyped compared with human
communication.

LEARNED AND INNATE COMMUNICATION

Some animals have their communicative behavior handed to them on
a genetic platter, and it develops as they mature; other animals must
learn their communicative system. The bulk of bee communication,
for example, has been shown to be genetically built in, or innate. A
bee knows from the start how to dance and how to interpret a dance.
Interestingly enough, however, there is a crucial step that must be
learned: a young bee is a poor follower and must practice following
dancers before it can accurately perceive another bee’s dance.
Many birds have calls or songs that are completely innate. Such
birds, if raised from the chick stage to adulthood totally separated
from their species, will nevertheless develop the characteristic call of
their kind. Other birds demonstrate a mixed pattern; that is, there is
an innate component and a learned component to their song. The in-
nate and learned components of the song of the chaffinch have been
studied most carefully. William H. Thorpe (1956) has shown by ex-
perimentation that certain aspects of the characteristic chaffinch song
are undoubtedly innate: its length of approximately 2.3 seconds, a
tendency to crescendo, and a reduction of mean frequency as the
song progresses. Even if chaffinches are isolated from other chaf-
 Language, Apes, and Humans

finches at birth, their song shows these characteristics. Other aspects

of the song, however, are learned as the young develop by hearing
adult chaffinches sing. For example, if a chaffinch is raised in isola-
tion from normal adult chaffinches, it develops a song that does not
include the final flourish characteristic of normal chaffinch song. Fi-
nally, some birds, such as the skylark, must learn virtually their entire
song from others of their species.
Like the skylark’s song, human communication (at least its spoken
and written aspects) seems to be almost totally learned. However, most
linguists and psychologists assume that human language has an innate
component: an ingrained predisposition to language that causes us all
to learn language, provided we have contact with it. Many linguists
believe that there is a more important innate aspect of language that
results in a basic similarity among all the world’s languages. If this is
true, we are then a good deal like the chaffinch. Such hypotheses are,
at this moment, well beyond proof.

Language, Apes, and Humans

The gap between the calls of birds or the dance of bees and human
speech seems obvious, but in the last few decades work with chim-
panzees and gorillas has seemed to narrow the distance between the
abilities of at least some apes and humans. Recent work with these
apes, which are the animals most closely related to humans, has
brought perhaps the most significant advance in the study of animal
communication and has raised new questions about the uniqueness of
our ability to use language.
In essence, rather than try to work out the existence and nature of
a natural communicative system among apes by tedious observation,
several researchers invented codes to teach to apes. Different re-
search teams invented unrelated codes. Premacks.(1972)
The used
plastic shapes.as_“‘words.’; workers at Yerkes Primate.Center..used_a
kind_of.typewriter console keyed to.signs (Bourne 1971), as shown in
Figure 1-2; the Gardners (1969) used American Sign Language, in
which hand movements stand for concepts.
with their chimpanzee_
Washoe; and more recently, Patterson has.taught.American..Sign.Lan-
guage to a gorilla named Koko.(Hayes..1977).
In all these cases, *
the results have been similar. Within a couple of
years, the apes could communicate in a fairly restricted manner with
their human colleagues by using vocabularies that ranged from about
a hundred to three hundred symbols and beyond. According to the
researchers, the apes were not just communicating as other animals
communicate; they were using language. Thus, this work seemed to
10 Animal Communication

FIGURE 1-2
Signs Used on the Typewriter Console at Yerkes Primate Center

=| (Al[el|e
challenge the belief that using language is a uniquely human ability.
Two questions must be answered in order to evaluate the behavior of
these ‘‘talking apes’’: what is language, and what exactly can these
apes do?

DEFINING LANGUAGE
Language is clearly a form of communication, but it has never been
very well defined. Language is most commonly defined as a form of
communication that is nonstereotyped_ and nonfinite; that is, it is
learned and creative. By creative we generally mean_that language is
unlimited
in its scope. Speakers of a language are able to produce and
interpret an unlimited number of utterances that they have never be-
fore heard.
We might also say that language is the form of communication used
by humans; all other definitions seem to be attempts to sort out what
differentiates human communication from that of other species. It is
fairly easy to invoke a definition of language and then say that, for in-
stance, bee communication is not language. The poor bee’s com-
munication is limited to such basic things as the location and distance
of a source of nectar, whereas people may talk throughout their lives
without ever repeating themselves. Humans can talk about the
weather, funny religious practices, the quintessence of Shaw, and
utter nonsense. Humans may lie, play word games, and abuse their
grammar in quest of art. These are the creative, fundamentally human
sorts of things that make language language.
Language in turn is often said to help make humans human. Be-
cause we have learned the symbol system of our language, we can
refer to the past and to the future; we can develop and remember con-
cepts and think about and communicate about things we have never
seen or heard or felt for ourselves. From these concepts and com-
munications, these memories of the past and ideas of the future, we
largely develop a concept of our selves and of others and our rela-
tionship to them. Thus language seems to be a necessary basis for our
personal and social identity and for our social relationships.
 Language, Apes, and Humans

THE ‘“‘TALKING APES’? AND LANGUAGE

If language helps to make us human, where do the chimpanzee Wa-
shoe and the gorilla Koko fit in? If we define language as nonstereo-
typed and nonfinite communication, are they indeed using language?
Some researchers think so and point to several surprising achieve-
ments of these apes. They can handle not only declarative state-
ments but also interrogative forms, negation, comparison, and even
conditional statements. Furthermore, the apes not only respond to
the researchers but also routinely initiate ‘“‘conversations,’’ and they
formulate new statements that they have not learned through training.
Washoe, for example, was reported to have signalled ‘‘Open food —
drink’’ to request that the refrigerator be opened — a request she had
never been taught.
In these few apes, then, one may cautiously consider the possibility
that something rather special is going on: their communication seems
to be creative; they may be using language. Many of these interesting
results have been challenged, and the entire issue of ape language
remains quite clouded. But these few apes may be capable of doing
with their various communicative systems something very similar to
what we do — just less of it. That we may therefore relate to a chim-
panzee on a ‘‘meaningful’’ level makes the chimp a good deal more
‘‘human’’ and less ‘‘animal.’’ Interestingly, this might be a legal
point. Koko was owned by the San Francisco zoo when she was
loaned to a Stanford University psychologist for study. Koko was
taught her language, and in 1977 the zoo decided that it wanted Koko
back. Theodore S. Meth, a lawyer who specializes in animal rights,
has argued that the zoo can no longer claim to own Koko:
The gorilla doesn’t exist anymore. Under normal circumstances the only
thing this animal doesn’t have that we do is language. Now you have
changed it. When you give it the conceptual apparatus for conscious rea-
soning, for mobilizing thought, you have radically altered it. You have
given it the pernicious gift of language. If it has never been one before, it is
an individual now. It has the apparatus for the beginning of a historical
sense, for the contemplation of self... . Her right to remain in a meaning-
ful relationship with the people she has known is greater than the zoo’s
property rights. This is the whole history of jurisprudence over the last 75
years— that property rights must give way to individual rights. In this
case, you have an ape that has ascended.*

Even if it is eventually proved that human beings are not unique in

possessing language, humans do represent the first species to possess
a highly developed, creative form of language. More importantly,

3 Harold Hayes, ‘“‘The Pursuit of Reason,’ New York Times Magazine, June 12, 1977,
po 22s
12. Animal Communication

Homo sapiens may be the first species that is able to teach language
and thus foster the intellectual development of other species. In this
regard, the position and influence of humanity in the evolution of
life will be unique. Man may not be God, but humans may have the
power to bestow on other species what is perhaps the most godly gift
of all — language.

eer be

tem of enonencont animal communication that can be reliably trans-

lated into human language.
Animal communication may be learned or innate, or a combination
of the two. Human communication is assumed to have both a learned
and an innate component.
Some studies with apes have indicated that these animals may be
capable of learning language. The evidence rests on their ability to
grasp different types of statements, to initiate conversations, and to
use statements they had not learned through training. Nevertheless,
recent challenges in these claims show that our knowledge of ape
language is far from complete.

SOURCES FOR FURTHER READING

GENERAL
Akmayjian, Adrian; Demers, Richard A.; and Harnish, Robert M. Lin-
guistics: An Introduction to Language and Communication. Cam-
bridge, Mass.: MIT Press, 1979.
Lorenz, Konrad. King Solomon’s Ring. New York: Thomas Y.
Crowell, 1952.
Sebeok, Thomas, ed. How Animals Communicate. Bloomington: In-
diana University Press, 1977.
Sebeok, Thomas A., and Ramsay, Alexandra, eds. Approaches to
Animal Communication. The Hague: Mouton, 1969.
Thorpe, William H. Learning and Instinct in Animals. London:
Methuen, 1956.

BEE COMMUNICATION
Lindauer, Martin. Communication Among Social Bees. Cambridge:
Cambridge University Press, 1971.
Von Frisch, Karl. The Dance Language and Orientation of Bees.
Translated by L. E. Chadwick. Cambridge, Mass.: Belknap Press
of Harvard University Press, 1967.
 Summary 13

WHALE AND DOLPHIN COMMUNICATION

Cousteau, Jacques-Yves. The Whale: Mighty Monarch of the Sea.
Garden City, New York: Doubleday, 1972.
Lilly, John C. The Mind of the Dolphin: A Non-Human Intelligence.
Garden City, New York: Doubleday, 1967.

APE COMMUNICATION
Bourne, Geoffrey H. The Ape People. New York: Putnam, 1971.
Gardner, R. Allen, and Gardner, Beatrice T. ‘‘Teaching Sign Lan-
guage to A Chimpanzee.”’ Science 165, August 15, 1969, 664-672.
Hayes, Harold. *‘The Pursuit of Reason.’’ The New York Times
Magazine. June 12, 1977.
Linden, Eugene. Apes, Men and Language. New York: Saturday
Review Press, 1974.
Premack, David. Intelligence in Ape and Man. Hillsdale, New Jersey:
L. Erlbaum Associates, 1976.
. “‘Language in the Chimpanzee?’’ Science 172, May 1971,
808-822.
Premack, David and Premack, Ann J. ‘‘Teaching Language to an
Ape.”’ Scientific American, October, 1972, pp. 92-99.
Rumbaugh, Duane M., ed. Language Learning by a Chimpanzee: The
Lana Project. New York: Academic Press, 1977.

QUESTIONS FOR DISCUSSION

1. Define animal communication.

2. How do you know when animal communication has taken place?
3. Nearly everyone has observed animals that appear to com-
municate. Describe such an observation and evaluate the apparent
communication.
4. Give examples from human behavior of the differences in learned
and innate behaviors.
 C.HeArP
eT BR

de
HUMAN
COMMUNICATION:
THREE SYSTEMS

Through symbols humans can produce an unlimited number of mes-

sages. The power and scope of our ability to communicate are en-
hanced by the fact that we have not one but three systems of com-
munication: speech, writing, and gesture. For most people, speech is
the basic system of communication. Writing, however, allows the
preservation of communications over space and through time, and its
great importance in human history and in complex societies would be
hard to overestimate. The role of gesture is less obvious. Among the
deaf, of course, it may replace speech as the principal form of com-
munication; but for most people it is an important supplement to
speech. Together these systems allow us to communicate in a variety
of situations with subtle shades of meaning.

Speech
Speech requires the manipulation of the tongue, lips, vocal cords,
lungs, velum, and all other parts of what is commonly called the vocal
tract. Physiologically, it requires such complex integration of nerves
and muscles that it is difficult to imagine how anyone ever learns to
speak. The speech centers of the brain are physically more extensive
than the centers controlling any other form of activity. The portions
of the brain apparently involved in controlling tongue movements
alone are nearly twenty times larger than those controlling leg move-
14
 Speech

ments, despite the fact that the tongue muscles are only a fraction of
the weight of leg muscles. The large size of the speech centers reflects
the complexity of speech. When we are speaking, the tongue is in
constant motion, and its position in relation to other elements of the
vocal tract is essential to the production of appropriate sounds. In
producing speech, the brain conceives a notion to say something and
sets in motion a series of electrical impulses to all the muscles of the
vocal tract. These muscles, in turn, set up a complex sound wave,
and the result is that something is ‘‘said.’’ Thus, when humans com-
municate by means of spoken language, they express meanings that
are conveyed through sounds. Understanding the relationship be-
tween meaning and sound is the departure point for linguistic inquiry.

SOUND AND MEANING

There is no logical or necessary relationship between the sounds of

words and their meanings. Speech is arbitrary and segmentable, For
example, the word cat may be divided into at least three segments or
units of sound: a k-like sound, a vowel, and a f-like sound. There is
nothing about these three sounds that suggests felinity. These sounds,
moreover, recur regularly in English and may be recombined in dif-
ferent orders to form different words, such as act and tack. Thus, no
sound carries meaning by itself; but when sounds are put into
sequences with other sounds, these sequences may carry meaning.
Some words seem to contradict the principle that the relationship
between sound and meaning is arbitrary. Onomatopoeic words such as
bu ish, bang, and meow, for example, sound like the things they
represent. Similarly, so words contain sounds that have become
imbued, in part, with meaning, such as the s/- in slime, slippery,
slush, and slop. Similarly, the gl- in such words as glow, glimmer,
glitter, and glisten may convey a certain meaning. However, these
words make up only a tiny fraction of language; for most words, the
relationship between sound and meaning is almost completely arbi-
trary. Certain modern words like television and typewriter seem to be
predictable from other existing froms like tele (‘‘across’’), vision
(‘‘sight’’), type (‘‘print’’), and writer (‘‘source of writing’’); but the
meanings of these forms are ultimately arbitrary, too.
This arbitrary relation is also indicated by the fact that the world’s
languages use somewhat different sounds. For example, the clicking
sounds of some African languages do not occur in English; in fact,
most speakers of English would find it difficult to integrate these
sounds into speech. Certain English sounds, like the initial sounds in
judge and then, are difficult for many non-English speakers to pro-
duce in a speech context.
At any rate, whatever sounds a language employs must be strung
16 Human Communication: Three Systems

together to form a message. Sounds are strung together to form

meaning-bearing units, and these units are strung together to form
sentences. This stringing-together is accomplished according to a sys-
tem of rules called grammar. According to the most popular modern
conception of a grammar, it contains three components: a phonologi-
cal, a syntactic, and a semantic component. The phonological compo-
nent is concerned with the sounds of a language; the syntactic compo-
nent is concerned with the combining of meaning-bearing units into
the sentences of a language; and the semantic component is concerned
with the meanings of sentences. All languages are assumed to share
some basic underlying similarities in their grammars.

SYNTAX AND SEMANTICS

The relationship between syntax and semantics is close and complex.
For example, consider the three meaningful units boy, girl, and love.
Languages have two means of expressing relationships between such
units: word order and word endings. Thus, in English, word order
(structure) establishes the difference in meaning between The boy
loves the girl and The girl loves the boy. Not all languages that use
word order to specify relationships use the same order. English is
basically a subject-verb-object language; Japanese is a subject-object-
verb language; and Tagalog, an Indonesian language, is a verb-sub-
ject-object language. In Classical Latin, word order was relatively
unimportant as a means of expressing relationships; instead, word
endings (inflections) specified most relationships. For example, the
Latin roots for boy, love, and girl are puer, amo, and puella, respec-
tively. Thus, The boy loves the girl might take the form Puer amat
puellam, or Puer puellam amat, or Puellam puer amat, and so forth.
The word order may vary, but in each case the -(a)m ending on puella
(‘‘girl’’) indicates that it is the object of the verb. The form puer
(‘‘boy’’) indicates that it is the subject of the sentence. To say that the
girl loves the boy, puer would be given an -em ending. Thus The girl
loves the boy might be given the form Puella amat puerem, or Amat
puerem puella, and so on.
Most languages today use both word order and word endings to in-
dicate the relationship between words. In English, word order is ex-
tremely important for establishing these relationships, and word end-
ings are used less frequently. For instance, we distinguish between
the subject, object, and possessive forms of pronouns: he and she are
subject forms, him and her are object forms, his and hers are posses-
sive forms. We mark verbs for the third person singular present tense
by the ending -s (she runs); and for most verbs, we mark the past
tense with the ending -d. Our system of adjectives is somewhat di-
vided in that the comparative form of some adjectives is marked by
 a
Writing ty

the ending -er (bigger, prettier) whereas others require the addition of
the word more (more beautiful, more ridiculous). Thus, learning a
language involves mastering both its word endings or inflections and
its word order or structure. Only then can one understand the full
range of messages conveyed by a language.

Writing
Linguists are usually more interested in speech than in writing, and
writing has often been viewed as a reflection of speech. Although we
have no proof, we assume that speech preceded writing. This assump-
tion is supported by the observations that children can learn to speak
before they learn to write and that many of the world’s peoples do not
possess written forms of their language. The lack of a written lan-
guage need not reflect simple ignorance. Many scholars believe that
Continental Celtic, the language of much of Europe before the Roman
Empire, was unwritten as a result of a religious prohibition on any
graphic representation of speech.
At some point, however, early humans discovered that they could
communicate by making marks on material. The earliest known pic-
tures, cave drawings, date from about 20,000 B.c., but we cannot be
certain that they were intended to communicate; that is, to transmit
information from one person to another. Writing as we know it today
developed rather late — hardly more than five or six thousand years
ago — but it has become particularly important since the invention of
the printing press some five hundred years ago. Today there are three
basic types of writing: logographic, syllabic, and alphabetic.

LOGOGRAPHIC WRITING
Many languages of the world, most notably Chinese, have a writing
system in which each symbol represents a word; such a writing sys-
tem is called logographic. Thus, in Chinese the symbols hoes, \
lL mean ‘‘man,”’ ‘‘woman,’’ and ‘‘mountain,”’ respectively. This
system developed out of a much less stylized system in which, for ex-
ample, the symbol for mountain was cLh. These earlier symbols, which
provided a more pictorial representation of the concepts to be com-
municated, are called pictographs. The likelihood is strong that the
first picture messages werememory aids or personal records of
events. However, at some point they were used to convey a meaning
to another person in much the same way that one draws a map for a
friend who needs directions. The pictures eventually became formal-
ized, at which time they became logographs rather than simply pic-
tures. Many of the symbols in a logographic system bear little resem-
blance to the objects they are meant to represent.
18 Human Communication: Three Systems

The logographic system has its advantages. Speakers of the dif-

ferent dialects in China today have difficulty communicating with
each other in speech, but two Chinese may readily communicate with
each other in writing. No matter how a word may be pronounced in
the various Chinese dialects, its logograph is the same across China.
On the other hand, a disadvantage is readily apparent: one must learn
an enormous number of logographs before one can read and write
Chinese.
It is not quite correct to say that each logograph is different, how-
ever, for a logograph typically is compounded of several parts. The
Chinese word meaning ‘“‘peace,’’ for instance, is %' , which derives
from the idea of a woman ( & ) under a roof ( 1). These two symbols
recur frequently in compound logographs, including many whose
meanings bear little or no relationship to either womanhood or roofs.
Most writing systems include a number of logographic characters,
even if they use an alphabetic system otherwise. Numerals (1, 2, 3,
. . .), for example, are logographic. There is nothing about the sym-
bol 1 that suggests it should be pronounced like the word won; in
most languages of the world it is not. Similarly, the symbol & is a
logograph, as are $, %, +, and =.

SYLLABIC WRITING
In a syllabic writing system, each symbol! represents a syllable. The
Egyptian hieroglyphs originally were logographs, but eventually they
were used as syllabic symbols as well. A symbol would be taken first
to represent a word and would then be extended to represent the
sound sequence of that word. Certain hieroglyphs represented spe-
cific consonants; for instance, a hand , represented
‘the sound
‘“‘d + a vowel.’’ Vowels were not represented in the writings of the
ancient Egyptians; thus, we do not know the actual pronunciation of
the vowel. The cuneiform writing system of the ancient Mesopo-
tamians, which was wedge-shaped script, contained characters for
such syllables as tim, ma, and mut.
The kana syllabaries of Japanese are another good example of a syl-
dabic writing system. In Japanese, a syllable consists either of a vowel
sound or of a consonant plus a vowel. The word ame, for instance,
consists of the syllables -a plus -me. The Japanese writing system
includes two syllabaries, hiragana
and katakana, which are used for
different kinds of words or situations; in both syllabaries, each sylla-
ble is given its own symbol. The hiragana is shown in Figure 2-1.
The symbols for -a, -ma, and -ka are ia) . Ed ,h» , respectively. No-
tice that there is no indication in the symbols that the same vowel is
involved in each syllable; the syllables are different and so are written
differently.
 Writing 19

FIGURE 2-1
The Hiragana Syllabary

Source: Reprinted from The Japanese Language by Roy A. Miller by permission of the
University of Chicago Press. Copyright © 1967 by the University of Chicago.

ALPHABETIC WRITING
Syllabic writing served as the source for alphabetic writing. Greek
traders in the second millennium B.c. encountered numerous North
African and Near Eastern groups that wrote with syllabaries de-
20 Human Communication: Three Systems

scended from early Egyptian. From one or another of these groups,

most likely from the Phoenicians, the Greeks took the script and
adapted it to serve their own language. In the process, symbols for
vowels were established and the syllabic symbols degenerated into
symbols for consonants. This Greek alphabet was adopted and
adapted by the Romans, from whom most of the peoples of the West-
ern world borrowed their alphabets.
Generally speaking, the principle of alphabetic writing is that each
letter represents a particular spoken sound of the language. But no
language is perfect in this respect, although English is unquestionably
worse than many others. The letter c, for example, is sometimes
‘“‘hard’’ (as in cat, act) and sometimes ‘‘soft’’ (as in cider, receive),
and is thus associated with at least two distinct sounds (cello and
spacious illustrate that c is associated with other sounds as well). In
fact, only a few letters of the English alphabet —f, /, m, and a few
others — are usually related to one and only one sound. English is
notorious for exceptions, however; notice the difference, for ex-
ample, in the finalf sounds of off and of.
This inconsistent relationship between letters and sounds results
from the fact that languages are constantly changing, whereas writing
systems tend to be preserved. For example, in 1922 the Turkish gov-
ernment decided to abandon use of the Arabic alphabet. To replace it,
Turkish linguists devised a Roman-type alphabet in which each
Turkish sound was represented by its own letter. Thus, in 1922 Turk-
ish had a near-perfect one-letter/one-sound alphabet. Today, little
more than half a century later, Turkish pronunciation has changed
sufficiently for the correspondence between letters and sounds to be
less perfect, and in the future the relationship will be progressively
less consistent.
Modern English spelling is about four hundred years old, and En-
glish pronunciation has changed a good deal in that time, which is
why its spelling seems peculiar at times. Most of the European lan-
guages, while by no means perfectly consistent, are much more regu-
lar than English in the relationship they maintain between spelling and
pronunciation. (French is a notable exception.) Thus, the spelling
bee, a traditional part of English education, is largely unknown to Eu-
ropeans, to whom spelling is insufficiently challenging to warrant such
a game. In many European languages, spelling reforms have caught
up with the pronunciation changes.
Many attempts to reform English spelling have taken place in the
last hundred years or so. President Theodore Roosevelt, for instance,
tried to institute spelling reforms in White House memoranda to pro-
vide an example to the rest of the country, and the late George Ber-
nard Shaw was so disgusted with English spelling that he left a good
 Writing 21

part of his fortune to anyone who could devise a better alphabet for
English in accordance with his suggestions. Interestingly, several al-
phabetic systems commonly used in English are more efficient than
our familiar alphabet; for example, speed writing (F U cn rd ths U cn
gt a gd jb as a scrtry).

MIXED WRITING SYSTEMS

When archaeologists find a previously undiscovered writing system
on clay tablets, they can rather easily determine what kind of writing
system they are facing. If the number of symbols is fewer than forty,
they may be reasonably confident that the system is an alphabetic
one, for few languages have an alphabet of more than forty letters. On
the other hand, if there appear to be more than sixty or so symbols,
they may be confident that they have a syllabary in hand; moreover,
the language probably has a fairly simple syllabic structure similar to
that of Japanese. It would be difficult to imagine a syllabic script that
would do justice to English, which has well over two thousand possi-
ble syllables. Finally, if the number of distinct symbols is very large,
the archaeologists may conclude that the writing system is logo-
graphic.
A language may use several writing systems, however. We have
seen that English, an alphabetic system, uses some logographs. Japa-
nese writing is a good example of how different writing systems can
be mixed together to form a more or less sensible and cohesive
whole, although its complexities make it a challenge to learn. There
are basically four writing systems in Japanese.
aeonil ateTogogtaphs Each kanji symbol that was adopted from the
dor concept, but most kanji symbols have at
least two BeReOn eH. When used by itself, a kanji symbol is given
the Japanese pronunciation. The symbol , for instance, is pro-
nounced ‘‘yama’’ (‘‘mountain’’) in isolation. When used in com-
pounds, however, the Chinese pronunciation ‘‘shan,’’ derived from
Mandarin Chinese, is used. In this case, is pronounced ‘‘san’’ as
in Fujisan (never Fujiyama).
Hiragana and Katakana are the two Japanese syllabaries. Hiragana,
the more cursive of the two, is used for any word not written in kanji
__ . . . of

and for all grammatical endings. Katakana is used primarily for formal .
official writing and for words of foreign origin, such as beisubo-ru
(‘‘baseball’’).
Romaji is the Japanese system of writing that uses the Roman al-
phabet. Although it is not commonly used in ordinary writing, it is
frequently found on public signs such as street signs.
Japanese is interesting in the extent to which it combines the three
systems of writing. The combination reflects Japanese history, in
22 Human Communication: Three Systems

which China at one time and, more recently, the West have been in-
fluential. In other languages as well, the system of writing can reveal
historical processes in language. More generally, writing helps to pre-
serve history. Its role in the development and functioning of societies
is vast. In contrast, gestures are most important in personal, face-to-
face communication.

Gesture

Although speech and writing seem more systematized and pervasive

than gesture, gesture was probably the first form of communication.
The term gesture is used here to include all human communication
that involves waving of the hands and facial signals, grunts and other
vocalizations that do not make up words, what is frequently called
body language, and various manipulations of the environment that
have communicative intent such as smoke signals. A gesture, then, is
a physical manipulation that is neither verbal nor graphic but is com-
municative.
One cannot doubt that gesture was basic to the development of
human communication and that it is still much used. Even when it is
inappropriate, people gesture while they speak. If you watch people
talking on the telephone, you will see them nodding, waving their free
hands, shrugging, and otherwise complementing their speech with a
whole repertory of motions that are completely wasted on the person
at the other end of the line.
To some people, gestural communication is a necessity. The deaf,
for example, depend greatly on the use of gestures. Communication is
also conducted through gestures at distances when speech is imprac-
tical. For example, referees at football games use a system of gestures
to communicate the infraction of rules to the crowd.
The most-discussed forms of-gestural communication fall under the
heading of kinesics, which is the study of the positioning and move-
ment of the body and its parts during conversation. These gestures
range from a wink to nodding the head to thumbing the nose. Dif-
ferent cultures often employ different gestures (for example, the
difference between the handshakes of two middle-aged American
business-people and two young inner-city blacks), although certain
gestures, such as a smile, are more or less universal.
Modulations of the voice, often called paralan » also affect
communiGatonc FOrGSample thedanguase SEEHectiOn i spoken ina
soft voice whereas aggressive arguments are conducted in a loud
voice with extreme shifts in pitch.
Proxemics_is_the
study_of the spacemaintained_ between two
~speakers_—in__conversation
— another _factor__in__gestural—com-
nication. This ‘‘social space’’ varies from culture to culture. Gen-
 Gesture 23

erally, individuals from the Middle East and certain Mediterranean

countries position themselves much more closely to each other dur-
ing speech than Americans and Northern Europeans do.
The broad definition of gesture outlined here includes still other
forms of communication like lip reading, which is a combination of
speech and gesture. A deaf person may perceive certain movements
of the lips and tongue and interpret their probable sequence. The
word mope, for example, would be perceived as the sequence lip
closure, open rounded lips, lip closure. Lip readers are quite re-
stricted in what they can actually see — movements of the lips and,
to some extent, movements of the tip of the tongue. The amount of
information they can draw from these gestures is only a fraction of
what is going on. Generally, lip readers supplement this information
with data drawn from facial expressions, body gestures, and the like;
thus, lip reading is really only a part of what the deaf generally call
face reading. In fact, one of the major problems that face readers en-
counter is the well-intentioned speaker who speaks slowly and exag-
geratedly for their benefit. Slow, exaggerated speech is unnatural and
distorted, and the deaf find it far more difficult to interpret than natu-
rally paced speech.
All these forms — kinesics, paralanguage, proxemics, lip and face
reading — occur most often along with speech, not in place of it. But
gesture can also substitute for speech, as the use of systems of tones
and manual signs illustrate.
Many languages, including most of the Chinese dialects and a
number of African languages, are to nguages; that is, they use
pitch to communicate or to pce artnet As a result,
speakers of these languages develop a sensitivity to tone, a sensitivity
that some peoples have applied to gestural systems of communica-
tion. For example, the Mazatecans, a group of Mexican Indians, can
communicate with each other by whistling:
Eusebio Martinez was observed one day standing in front of his hut, whis-
tling to a man a considerable distance away. The man was passing on the
trail below, going to market to sell a load of corn leaves which he was car-
rying. The man answered Eusebio’s whistle with a whistle. The in-
terchange was repeated several times with different whistles. Finally the
man turned around, retraced his steps a short way, and came up the foot-
path to Eusebio’s hut. Without saying a word he dumped his load on the
ground. Eusebio looked the load over, went into his hut, returned with
some money, and paid the man his price. The man turned and left. Not a
word had been spoken. They had talked, bargained over the price, and
come to an agreement satisfactory to both parties — using only whistles as
a medium of conversation.!

1George M. Cowan and Florence H. Cowan, ‘‘Mazateco Locational and Directional

Morphemes,’’ Aboriginal Linguistics 1, 1947, 1-9.
24 Human Communication: Three Systems

Some African groups also use tone as the basis of distant com-
munication. Instead of whistling, however, they use a set of drums,
each of which has a different musical note; they beat out the tune of a
sentence in much the same way that the Mazatecans whistle it.
English-speaking people, who do not use pitch to differentiate
words, probably could not develop an adequate communicative sys-
tem using only whistles or drums. Both drum and whistle com-
munications depend directly on the tonal aspects of the spoken lan-
guage. Mazatecans will learn whistling as they learn their verbal
language, and each member of an African tribe is capable of under-
standing the drummed version of the tribe’s speech (although only a
few specialists learn the art of doing the drum-work).
Much more familiar to us is the Indian sign language, in which hand
movements stand for concepts. Throughout the American Plains,
from Canada to Mexico, Indian tribes whose languages were com-
pletely different used sign language. These tribes frequently met while
following the great buffalo herds, and sign communication evolved
over a period of time well before whites came on the scene. Most of

FIGURE 2-2
Selected Signs in Indian Sign Language

————s i— [ESS

A
Fragrant Grenouille
Wohlriechend Fragrant Frosch ~ Frog

Sacro ee OO aaa)
f H is ae PN No
mee AN
=S See
/
a
I
/ify
‘be LA).
he
Sees
SSS
Sat Se
yy

a
Donnez - moi Aller
GiveMe ae ee
Gieb mir
Source: Reprinted from William Tomkins, Indian Sign Language, p. 30. Copyright ©
1969 by Dover Publications, Inc. Used with permission of the publisher.
 Gesture 25

FIGURE 2-3
Sentences from Indian Sign Language

! went to the mountains with Bug Bear oe SOt ao) i

ar = eaeai
ar \e
eaie Weal
| Ep MOUNTAINS WITH BIG BEAN SSMOTOR! Pele foie S

SIPs
Fyfe &
QUES MONSYOU GOIACROSS, BIG ELK RIVER: AUNT
Source: Reprinted from William Tomkins, Indian Sign Language, p. 67. Copyright ©
1969 by Dover Publications, Inc. Used with permission of the publisher.

the signs were readily apparent because they were not arbitrary but
related to their meanings, as can be seen from Figure 2-2, a portion
of the sign dictionary given in Tomkins (1969). Sentence construction
was fairly simple, with few oddities from our point of view. Adjec-
tives followed the nouns they modified, and all questions were begun
with the question sign: an open hand held up, palm forward, waved
from side to side. Two examples of sentences are given in Figure 2-3.
The Indians are by no means the only people who have used sign
language for regular communication. Although the deaf are adept at
lip and face reading, they do not “‘speak’’ at each other. Like the In-
dians, the deaf can communicate through a sign language in which
each gesture is imbued with some standardized meaning. American
Sign Language is an example. There are more than fifteen hundred
signs that are standard among users of sign language, with a certain
amount of variation from one community to another. Although a few
correspondences exist between Indian and deaf signs (such as the sign
for ‘‘see’’), most of the deaf signs developed independently of the In-
dian system.
Like the Indian sign language, deaf signs do not represent words,
_but_concepts-Shades of meaning between words such as exhausted
and fatigued are designated not in the sign itself but in the manner
and force with which the sign is made. Because the face is important
in conveying these shades of meaning, the deaf tend to watch the face
of the person who is speaking to them. A great majority of deaf signs
are made at face level. Whereas the Indian sign for ‘‘good”’ begins at
26 Human Communication: Three Systems

heart level, the deaf sign begins at the lips. Even the feet function in
communication with the deaf. A stamp of the foot is usually in-
terpreted as a call for attention; its vibrations can be perceived readily
by a deaf person close by.
Even with these modulations, however, the signs for concepts are
relatively limited and very general in meaning. Accuracy requires
greater specification of meaning than is possible with these signs. An
alternative is finger spelling, in which gestures of the fingers represent
sounds, not concepts. Interestingly, the better-educated use finger
spelling much more than do those with less education.
Of the three methods of communication discussed, speech is by far
the quickest and the most efficient. Also, subtle nuances of meaning
can be indicated through vocal tone as well as actual words.

Summary
The three major ways that humans communicate with each other are
speech, writing, and gesture. Speech is arbitrary and segmentable. In
human language, sounds are strung together to form meaning-bearing
units, and these units are strung together to form sentences. This
stringing-together is accomplished according to a system of rules
called grammar.
Writing systems are of three basic types: logographic, syllabic, and
alphabetic. In a logographic system, each symbol represents a word;
in a syllabic system, each symbol represents a syllable; and in an al-
phabetic system, each symbol represents a sound.
The term gesture may include all forms of human communication
that are neither speech nor writing. This includes the subject matter
of kinesics, paralanguage, and proxemics and many other com-
munications systems like Mazatecan whistling, African drumming, In-
dian sign language, and sign languages of the deaf.

SOURCES FOR FURTHER READING

SPEECH

Cherry, Colin. On Human Communication. 3rd ed. Cambridge,

Mass.: MIT Press, 1957.
Hybels, Saundra and Weaver, Richard L., II. Speech Communica-
tion. New York: Van Nostrand, 1974.

WRITING

Diringer, David, Writing. New York: Praeger, 1962.

Fishman, Joshua, ed. Advances in the Creation and Revision of Writ-
ing Systems. The Hague: Mouton, 1977.
 Summary 27

Gelb, IgnaceJ. The Study of ie abl Rev. ed. Chicago: University of

Chicago Press, 1963.

GESTURE
Birdwhistell, Ray L. Kinesics and Context: Essays on Body Motion
Communication. Philadelphia: University of Pennsylvania Press,
1970.
Hall, Edward T. The Hidden Dimension. Garden City, New York:
Doubleday, 1956.
. The Silent Language. Garden City, New York: Doubleday,
1959.
Wundt, Wilhelm Max. The Language of Gestures. The Hague:
Mouton, 1973.

QUESTIONS FOR DISCUSSION

1. What do you see as the various purposes of writing?

2. Give examples of onomatopoeic words other than those in the
text. If you have trouble, check for examples in literature. Look at
Edgar Allan Poe’s ‘‘The Bells’’ and Vachel Lindsay’s ‘‘General
William Booth Enters into Heaven.”’
3. If a being from another planet visited earth, how would you ex-
plain the meaning of the words grammar, phonology, syntax, and
semantics ?
4. Make up a pictographic language and give three examples of sen-
tences using your language.
5. What words in modern English might be in the process of a spell-
ing change? What causes you to think this spelling change is in
progress?
6. What has been left out in speed writing? Why does this form of
language still communicate?
7. Observe yourself for the next few days and see how you gesture as
part of communication.
8. Notice how close or far you stand from people when you are talk-
ing. When you are angry, where do you position yourself?
 C HyAsP LER

HISTORY
OF
LINGUISTICS

Linguistics is the study of language, but unlike philology, which deals

with the analysis of written text, linguistics is most concerned with
spoken language and its anthropological, psychological, and sociologi-
cal ramifications. This study of spoken language goes back thousands
of years. During some periods, linguists have been most interested in
the changes in languages through history; during other periods, they
have concentrated on the study of languages at just one time. Some
linguists have concentrated on describing how language is used; oth-
ers, on how it should be used. Language has been seen by some
linguists as a mirror of the mind and a key to the understanding of
thought; to other linguists, mind is irrelevant to the study of language.
At one time or another, both empirical studies and philosophical anal-
yses have dominated linguistics; the earliest linguistic inquiries were
almost purely philosophical. Among the questions that early linguists
asked were What is the origin of speech? What is the relationship be-
tween the human intellect and the structure of language? How are
words related to ideas? But both the ancient Indians and the ancient
Greeks also studied how words are produced, formed, and changed.

The Ancient Grammarians

The earliest grammar of any language, as far as we know, was Pa-

nini’s grammar of Sanskrit, the classical language of India. Written in
about the fourth or fifth century B.c., Panini’s remarkable work repre-
sents a highly developed approach to linguistics,
28
 The Ancient Grammarians 29
. a .

The ancient Indian linguists identified the chief parts of the vocal
tract involved in producing sounds, and they associated them with
particular sound segments like p and m. They also studied vowel
length, tone, syllables, and other aspects of phonetics; and they in-
vestigated the rules that govern the proper combinations of sound
segments and words in sentences.
Long ago, the Indian linguists recognized the verb as central to a
sentence. They saw that all other words in a sentence bear a particu-
lar relation to the verb. But they also recognized that not all sen-
tences are grammatical. For example, a sentence like The student
reads writes buys a novel, which disobeys the rule that a sentence
cannot contain three consecutive verbs, cannot be grammatical. Nor
can a sentence like The rain ate the sun, which involves contra-
dictions in the meanings of the words, be considered grammatical.
In general, the linguistic accomplishments of the ancient Indian
grammarians were superior to those of their contemporaries, the
Greeks. But like most of the Western sciences and humanities, the
study of language in the Western world began with the ancient
Greeks. Among the very early studies of language, Plato’s Cratylus is
perhaps the best known — largely because of its naive approach in
seeking the origins of words. If Plato were living today, he might seek
the source of a word like catastrophe by noting that it seems to con-
sist of three parts, which sound like the words cat, astro, and fee. He
might point out that a cat is characterized by its ability to strike
quickly; that astro refers to something extraordinary, colossal, and
out of this world; and that fee refers to something costly, or some-
thing that is paid. Putting these meanings together, Plato might then
explain that the word catastrophe means something that is ‘‘quick,
colossal, and costly.’’ The problem with this simple and superficially
plausible method is that almost anything can be made to mean any-
thing, as just demonstrated. Despite Plato’s misguided approach to
the origin of words, he offered valuable insights into language, includ-
ing a differentiation between vowels and consonants, an appreciation
of word accent, and a division of the sentence into nominal and verbal
parts.
The study of language was more fruitful under Plato’s successors,
Aristotle and Dionysius Thrax. Aristotle classified the parts of speech
in the third century B.c., and he is often regarded as the founder of
classical European grammar. But the oldest known grammar of Greek
was written by Dionysius Thrax, who lived near the end of the second
century B.c. Thrax’s grammar, the Techne grammatike, made inroads
into the study of how sounds are produced; but it was even more
useful in its treatment of word classes and functions.
Thrax identified eight basic word classes: nouns, verbs, pronouns,
30 History of Linguistics

articles, participles, prepositions, conjunctions, and adverbs. He fur-

ther noted important properties of classes of words — such as the
properties of gender, number, and case for nouns and those of con-
jugation, tense, voice, mood, and person for verbs. To Thrax, the
principal units of grammatical description were the word and the sen-
tence. He considered the sentence to be a group of words that ex-
presses a complete thought — a definition that remained popular until
the twentieth century.
Another major force in ancient Greek linguistics was the Stoic
school of grammar, which enjoyed its greatest success in the second
century B.c. The Stoics’ most valuable contribution was their treat-
ment of case as a grammatical category of nouns expressed by word
endings, and they distinguished between the nominative and all other
cases. The Stoics, too, were the first to distinguish between proper
and common nouns.
The approach of the ancient Greeks to grammar was continued and
modified by the Romans. Writing in the first century B.c., Varro
noted that the word endings discussed by the Stoics and Thrax could
be further subdivided according to type. Some endings are used to
derive new words; others are inflectional, indicating a word’s rela-
tionship to other words in a sentence. To use a modern example, a
word like sunny contains the derivational ending -y, meaning roughly
‘tof or pertaining to.’’ Thus, the word sunny is derived from the noun
sun. The word sun can also take the inflectional ending -’s, which
indicates possession. Varro realized that derivational endings like -y
cannot be attached to all nouns (for example, they cannot be attached
to Varro) but that inflectional endings like -’s are more general and
can be applied regularly to all nouns to indicate possession (Varro’s).
Latin grammar achieved its most precise formulation in the works
of Priscian and Donatus, who wrote grammars of Latin at approxi-
mately the time of the Roman Empire’s decline. For centuries these
descriptive grammars served as the basis for learning Latin and for
learning about language in general.

The Middle Ages, the Renaissance, and the Age of Enlightenment

After the Fall of Rome and through the Dark Ages, Latin continued to
be a respected language, both in scholarly and religious circles. The
grammars of Priscian and Donatus remained in vogue for teaching
purposes, thus ensuring a continuation of linguistic awareness. The
first major new development in linguistics, however, did not take
place until the thirteenth century, when speculative, or modistic,
grammar rose to popularity.
Speculative grammar resulted from the collision of traditional Latin
descriptive grammar and scholastic philosophy, represented by Cath-
 Middle Ages, Renaissance, and Age of Enlightenment 31

olic philosophers like St. Thomas Aquinas. It sought a theoretical

basis for Latin grammar by attempting to specify the ways, or modes,
by which we perceive and signify things. Thus, the speculative gram-
marians sought to explain the differences between parts of speech
such as nouns and verbs by looking for differences in the ways the
mind perceives the things referred to-by nouns and verbs. For ex-
ample, a chair differs from the act of laughing in that it possesses an
existence or substance. This property of chairs and similar objects is
perceived by the mind, which expresses objects of this class by
means of the part of speech we call noun. The speculative gram-
marians described other modes for perceiving and signifying verbs,
pronouns, participles, and so on. Like the ancient Indians, they
recognized that words may be combined in sentences only according
to certain relationships between parts of speech and meaning.
Although the speculative grammarians concentrated on Latin, their
work led to the assumption that language has a universal basis, that
all languages are essentially the same in nature and differ only in their
surface characteristics. This was the view held by Roger Bacon
(1214?-1294), one of the first speculative grammarians, who studied
Greek, Arabic, and Hebrew, as well as Latin.
With the coming of the Renaissance and Columbus’s discovery of
the New World, several linguistic developments occurred, and views
of language became more diverse. A strong interest in literature and
the discovery of printing combined to draw attention to the inconsis-
tencies between spelling and pronunciation, thus leading to a fuller
understanding of their relationship. Furthermore, as the chief descen-
dant languages of Latin — French, Italian, and Spanish — were ele-
vated in status, linguists became more aware of changes that had
taken place since the days when Latin was spoken regularly.
Linguists thus were provided with written proof of what they had
always suspected: languages change. Finally, explorers and mis-
sionaries in the Orient and the Americas discovered new languages
for which grammars eventually were written. Thus, the fifteenth and
sixteenth centuries were marked by an increased awareness of the dif-
ferences among languages.
Then, in the seventeenth century, the idea that beneath these dif-
ferences all languages are essentially the same in nature surfaced
again. The Port Royal grammarians, under the influence of Des-
cartes, stressed the universality of thought, which they said was re-
flected in language. They also recognized that creativity is inherent in
human language, and they viewed the phrase as corresponding to a
complex idea. In many ways, their ideas are similar to Noam
Chomsky’s modern theory of transformational-generative grammar
(see pages 106 and following).
While the French grammarians following Descartes were investigat-
32 History of Linguistics

ing universal grammar, the English grammarians were studying pho-

netics, or the physical sound system of language. Much of what the
English discovered about the articulation of sounds had already been
known by the ancient Indians, but one great triumph of the English
was their elucidation of the relationship between pronunciation and
spelling. The English interest in phonetics, along with the discovery
of many new languages, led to the first attempts to create universal
phonetic alphabets. These are sets of symbols used to represent the
most common sounds in known languages. The work done by the En-
glish in universal phonetics foreshadowed some of the most important
advances in linguistics.

The Nineteenth Century: Linguistics Becomes a Science

In the nineteenth century, new developments in linguistics were stim-
ulated by close attention to concrete data and exciting philological
studies. Thus, a new approach emerged — historical-comparative lin-
guistics. The earliest significant work in comparing languages was
done by William Jones (1747-1794), an Englishman who noted certain
similarities among Greek, Latin, Gothic, and Sanskrit. Although
Jones assumed that these languages had a common origin, he some-
how missed the importance of his discovery — the seeds of the com-
parative method, which allows one to establish the relatedness of lan-
guages within a family and to reconstruct the probable forms of the
original parent language. Through the work of Franz Bopp, Rasmus
Rask, the Grimm brothers, and other nineteenth-century scholars, the
comparative method was recognized as a scientific procedure. This
procedure established the relatedness of the Indo-European lan-
guages, a family that includes Greek, Latin, Gothic, Sanskrit, En-
glish, Russian, and others. Soon linguists prouped Oriental and other
languages into similar families.
The preoccupation with the scientific investigation of concrete facts
and the success of the comparative method culminated in the late
nineteenth-century neo-grammarian school, which claimed that
‘‘sound laws operate without exception.’’ These ‘‘laws’’ were actu-
ally not laws at all, but statements of correspondences between
sounds. For example, if a certain sound in Latin was found to corre-
spond to a certain sound in Sanskrit, the correspondence was be-
lieved to hold for all cases. If it did not, another ‘‘law’’ had to exist
that would explain the apparent exceptions.
Though not a neo-grammarian, an important linguist of this period
was Wilhelm von Humboldt (1767-1835). Von Humboldt sought uni-
versal grammatical rules from the facts of particular languages. He
was particularly concerned with the relationship between language
 Traditional and Structural Linguistics 33

and thought, and he drew strong ties between the language of a peo-
ple and their spirit or culture. He also emphasized that language has
an “‘inner form’’ corresponding to the psychological structure of the
individual. The intellectual climate of nineteenth-century linguistics,
with its strong emphasis on concrete linguistic data and historical
linguistics, was at odds with von Humboldt’s more philosophical
approach to language and his interest in language at a particular time.
Hence, his contributions were not fully appreciated until the twen-
tieth century.
In conclusion, nineteenth-century linguists were challenged by the
growth of the natural sciences, and they were wise enough to turn
their attention to a scientific method of dealing with language data.
Their efforts also included many first attempts to expand the horizons
of linguistics, such as the beginnings of acoustic phonetics and the
study of dialects and bilingualism.

Traditional and Structural Linguistics in the Twentieth Century

The study of linguistics has often been divided into synchronic and
diachronic linguistics. Synchronic linguistics is concerned with the
analysis of a language at a particular time; for example, English at the
time of the American Revolution. It is also known as descriptive
linguistics. Diachronic linguistics, or historical linguistics, deals with
different states of language through time, such as the changes that
took place between Old English and Middle English. In the twentieth
century, historical-comparative studies, which preoccupied nine-
teenth-century linguists, have assumed a secondary role to syn-
chronic linguistics.
From the mid-eighteenth century until the twentieth century,
synchronic linguistics was essentially normative or prescriptive; that
is, it attempted to prescribe the norms of ‘‘proper’” language usage.
Because it did little but expand on the classifications of the ancient
Greek and Latin grammarians, it is sometimes referred to as tradi-
tional linguistics. Nevertheless, the work of some traditinal gram-
marians, particularly Otto Jespersen’s detailed grammar of English, is
valuable for its simplicity, clarity, and insight.
To put it simply, traditional grammar is the grammar that many of
us learned in elementary school. Such notions and techniques as the
parts of speech, parsing, and the Reed-Kellogg sentence diagrams
provide a simple and accurate analysis for a sentence like We do not
have any apples. The prescriptive tendencies of many traditional
grammarians, however, would hinder analysis of the semantically
equivalent sentence We ain’t got no apples, a variant that occurs with
some frequency in certain dialects. They would simply treat this sen-
34 History of Linguistics

tence as ungrammatical. Thus, the traditional approach is viewed

today as useful, but often incomplete.
The structuralist theory of language was the first major new ap-
proach to descriptive linguistics in the twentieth century. Introduced
by the Swiss linguist Ferdinand de Saussure and pioneered in the
United States by Leonard Bloomfield, structuralism flourished for
about thirty years, until the early 1960s. In its emphasis on the inves-
tigation of concrete linguistic data, structuralism logically followed
the late nineteenth-century neo-grammarian school. But structuralism
was geared toward descriptive linguistics.
Structural linguistics typically involved isolating, classifying, ana-
lyzing, and segmenting the observed language data. For example, a
structuralist encountering English for the first time would begin by
collecting samples of English utterances and would then record the
strings of sounds that constituted those utterances. The structuralist
would find that sounds like p and b in English are significant because
they distinguish words like pat and bat. Whether the p-sound is pro-
duced with an accompanying puff of air (aspiration) would be consid-
ered rather insignificant because it would not affect the meaning of
what was being said.
After isolating and classifying the significant sounds of a language,
the structuralist would notice that certain sounds appear together in
recurring segments that signal some form of meaning. Thus, the
sounds p, r, and o would often be found in succession (pro-) to signal
‘in favor of.’’ The structuralist would then classify these larger units.
For instance, book and table would be placed in the same class be-
cause they could both be found in utterances like The ________ costs
a lot of money but not in utterances like The man ______ the house.
In summary, structuralism began with concrete language data and
organized these data into levels of structure for the purpose of
description.
One criticism of structural linguistics is that it made no attempt to
deal with how humans understand and interpret the meanings of sen-
tences; that is, Bloomfield’s theory of structuralism excluded the
mind from linguistic consideration. For this reason, structuralism is
often linked with the psychological theory of behaviorism, which
similarly restricts itself to that which is concrete and observable.

Transformational-Generative Grammar

Transformational-generative grammar emphasizes that human lan-

guage is creative — that humans are able to produce and interpret an
infinitely large number of sentences that they have never before
heard. It attempts to describe what a person knows about his or her
 Transformational-Genefative Grammar 35

language, but it also claims that all languages are grounded in univer-
sal facts and principles. By postulating deep structures, which are rep-
resentations that contain the essential meanings of sentences and that
underlie actual utterances, transformational-generative grammar is in
direct opposition to structuralism. Faced with the sentence Climbing
plants can look strange, the structuralist observes one concrete utter-
ance; but the transformational-generative grammarian explains that
the utterance has two possible meanings, which can be expressed in
the form of two different deep structures. These might be paraphrased
as Plants that are climbing (plants) can look strange and The climb-
ing of plants by someone can look strange.
Historically, transformational-generative grammar appears to have
made a rather rapid break from structuralism, but the period of transi-
tion, although short in time, is worthy of study. For example, the
structuralist era produced the two respected Prague linguists, Nikolaj
Sergeevic Trubetzkoy and Roman Jakobson, whose work found use-
ful applications in transformational-generative grammar. Both
linguists were concerned with studying the basic units and contrasts
of the sound systems employed in language. Jakobson is also famous
for his work on the acquisition of speech sounds and contrasts by
young children and for his work with poetics and prosody.
Still, it was Noam Chomsky who played the most prominent part in
developing transformational-generative grammar. Chomsky studied
under the noted structuralist Zellig Harris, who introduced the con-
cept of a linguistic transformation to relate basic sentences such as
Tom read the book to other common sentence types such as negatives
(Tom did not read the book), questions (Did Tom read the book?),
and passives (The book was read by Tom). Chomsky and Harris
shared an interest in mathematical approaches to language; it was
commonly assumed in the early 1950s that the computerization of
human languages was a reasonable immediate goal. Much of the early
appeal of transformational-generative grammar was due to the mathe-
matical precision with which Chomsky formulated his theory.
But Chomsky’s structuralist schooling cannot be overlooked. For
example, one element of his grammar is phrase-structure rules. One
of these indicates that a sentence is composed of a noun phrase plus
a verb phrase (SNP + VP); but these rules are simply age-old
grammatical statements cloaked in computerlike notation. Similarly,
the tree diagrams of transformational-generative grammar are akin to
both the traditional Reed-Kellogg sentence diagrams and the analyses
of the structuralists in that all provide a pictorial representation of the
structure of sentences. Transformational-generative grammar also
adopted the notion of levels of structure from structural linguistics.
Finally, Chomsky, like the structuralists, did not provide much il-
36 History of Linguistics

lumination about how a grammar might incorporate a description of

meaning. It was not until 1963 that the transformational-generative
grammarians J. A. Fodor and J. J. Katz made the first real attempt to
deal with semantics. In the meantime, Chomsky and fellow linguist
Morris Halle had attempted to provide a transformational-generative
analysis for the sound system of English.
Linguistic theory has undergone considerable revision and modifi-
cation in the past fifteen years, particularly in its approach to mean-
ing. Linguistics has also drawn on other sciences to form a large
number of linguistic subfields, most notably psycholinguistics and
sociolinguistics. Transformational-generative grammar, the most pop-
ular current linguistic theory, has not been a panacea, but it has
inspired research and new ideas. It is this approach to language that
will be developed throughout most of this textbook.

Summary
The study of language began in ancient India and Greece. The In-
dians, in particular, dealt with linguistics in a very sophisticated way;
much of their work has been adapted, however indirectly, by modern
linguists. The earliest studies in linguistics were primarily concerned
with philosophical questions and with etymology, word formation,
and phonetics.
In the Middle Ages, the speculative grammarians sought a theoreti-
cal basis for grammar, which was developed more extensively in the
seventeenth century by the Port Royal universal grammarians. During
and after the Renaissance, views of language became more diver-
sified, largely because of increased contact among languages and cul-
tures, especially after the invention of printing.
Modern linguistics may be defined as the period of time beginning
in the late eighteenth and early nineteenth centuries with the work of
William Jones and Wilhelm von Humboldt. The major linguistic dis-
covery of the nineteenth century was the comparative method. De-
scriptive linguistics in the twentieth century has moved from tradi-
tional linguistics to structuralism to transformational-generative
grammar. One of the fundamental issues debated by linguists in the
twentieth century, and indeed earlier, has been whether the study of
mind is relevant to linguistic description.

SOURCES FOR FURTHER READING

HISTORY OF LINGUISTICS
Hymes, Dell, ed. Studies in the History of Linguistics: Traditions and
Paradigms. Bloomington: Indiana University Press, 1974.
 Summary 37

Ivic, Milka. Trends in Linguistics. The Hague: Mouton Press, 1965.

Parret, Herman, ed. History of Linguistic Thought and Contemporary
Linguistics. New York: Walter De Gruyter, 1976.
Robins, Robert H. A Short History of Linguistics. Bloomington: In-
diana University Press, 1967.
Salus, Peter H. On Language: Plato to von Humboldt. Holt,
Rinehart and Winston, 1969.

ANCIENT GRAMMAR
Belvalkar, S. K. An Account of the Different Existing Systems of
Sanskrit Grammar, Poona [the author]: 1915.
Faddegon, B. Studies on Panini's Grammar. Amsterdam: North-
Holland, 1936.
Robins, Robert H. Ancient and Medieval Grammatical Theory in
Europe. London: Bell, 1951.

MIDDLE AGES, RENAISSANCE, AND AGE OF ENLIGHTENMENT

Bursill-Hall, G. L. ‘‘Medieval Grammatical Theories.’’ Canadian
Journal of Linguistics, 1963.
Cadet, Felix. Port Royal Education. London: Swan, Sonnenschein,
1898.
Chomsky, Noam. Cartesian Linguistics. New York: Harper and
Row, 1966.
Padley, G. A. Grammatical Theory in Western Europe, 1500-1700:
The Latin Tradition. New York: Cambridge University Press, 1976.
Vorlat, Emma. Progress in English Grammar, 1585-1735. Luxem-
burg, N. D., 1963.

NINETEENTH CENTURY
Jankowsky, Kurt R. The Neogrammarians: A Re-evaluation of Their
Place in the Development of Linguistic Science. The Hague:
Mouton, 1972.
Pedersen, Holger. Linguistic Science in the Nineteenth Century.
Cambridge, Mass.: Harvard University Press, 1931.
von Humboldt, Wilhelm. Linguistic Variability and Intellectual
Development. Translated by George C. Buck and Frithsof A.
Raven. Coral Gables, Florida: University of Miami Press, 1971.

TRADITIONAL LINGUISTICS
Jesperson, Otto. Essentials of English Grammar. (Reprint) Montgom-
ery, Alabama: University of Alabama Press, 1964.
Walsh, J. M. and Walsh, A. K. Plain English Handbook. Cincinnati:
McCormick-Mathers, 1972.
38 History of Linguistics

STRUCTURAL LINGUISTICS
Bloomfield, Leonard. Language. New York: Holt, Rinehart and Win-
ston, 1933.
Gleason, Henry A. An Introduction to Descriptive Linguistics. Rev.
ed. New York: Holt, Rinehart and Winston, 1961.
Harris, Zellig. Structural Linguistics. Chicago: The University of
Chicago Press, 1951.
Hockett, Charles F. A Course in Modern Linguistics. New York:
Macmillan, 1958.
Saussure, Ferdinand de. A Course in General Linguistics. New York:
McGraw-Hill, 1966.

TRANSFORMATIONAL-GENERATIVE GRAMMAR

Chomsky, Noam. Aspects of the Theory of Syntax. Cambridge,

Mass.: MIT Press, 1965.
. Syntactic Structures. The Hague: Mouton, 1957.
Lyons, Joan. Noam Chomsky. New York: Viking, 1970.

QUESTIONS FOR DISCUSSION

1. List several early linguists who made major contributions to the

field of linguistics. Be sure to point out the contributions of each.
2. Explain the difference between synchronic and diachronic linguis-
tics.
3. Explain the structuralists’ approach to describing language.
4. What are the major differences pointed out in this chapter between
transformational-generative grammar and structural linguistics?
 CoH AvP TE R

HISTORY
OF
ENGLISH

Linguists and others have debated for years what, if anything, is uni-
versal in human language, what is common and unchanging in all lan-
guages. But for centuries they have agreed that many aspects of lan-
guage do change. Moreover, the history of every language is unique,
because each language is inherently bound to the thinking, nature,
and spirit of a people, all of which are continuously altered by the
twists and turns of events. The development of the English language
is an example of this process. It has been fashioned from the complex
history of the English people and reflects that history and those peo-
ple. This chapter reviews four periods in the history of English and
the development of American English.

The Establishment of Anglo-Saxon

Prior to the Roman invasion in 55 B.c., the language of Britain was
Celtic, an Indo-European language. The Romans left shortly after
they arrived, leaving little influence on the Celtic language. About a
century later, in A.D. 44, Roman legions returned to begin an occupa-
tion lasting almost four hundred years. The second Roman invasion
left a rich archaeological heritage but, again, little permanent linguis-
tic influence. Nevertheless, Latin was learned by Celts, primarily in
the towns, and some words derived from or influenced by Latin are
still found in Irish and Welsh today.
39
40 History of English

Roman troops in Britain were few, and the occupation proceeded

routinely until about A.D. 350, when the attention of the Rornans was
drawn to the European continent. After centuries of battle with the
Huns and migrating Germanic hordes, Rome’s forces could no longer
be spread throughout the world; Rome needed her troops at home.
Figure 4-1 sketches the migration of Germanic tribes throughout
Rome’s holdings in Europe. At least four of the barbaric tribes sacked
Rome in the fifth century. The cost of defense against these tribes
coupled with the gradual disintegration of Roman dominion led Rome
to withdraw from England completely in A.p. 410.
For a brief half-century, the British Celts were left to their own
devices. Without the support of Rome, however, they were no match
for the fierce Picts and Scots, who frequently raided from the north-
ern borders. Since British requests to Rome for help went largely
unheeded, the British Celts called on a Germanic tribe, whom
Bede identified as the Saxons, to help drive off the marauding Picts
and Scots around A.D. 450. The obliging Germanic Saxons were very

FIGURE 4-1
Germanic Migrations in Europe
 The Establishment of Anglo-Saxon 41

FIGURE 4-2
Germanic Tribes in Northern Europe

Source: Albert C. Baugh, History of the English Language, 2d ed., © 1975. Reprinted
by permission of Prentice-Hall, Inc., Englewood Cliffs, New Jersey.

successful, but they found the British countryside so inviting that

they appropriated it as their own, driving away or assimilating their
Celtic hosts. Over the next century, successive waves of Germanic
invasions established three major groups in England: Saxons, Jutes,
and Angles (see Figure 4-2). All three groups spoke Germanic dia-
lects, and the dialects were similar enough for the term Anglo-Saxon,
to describe their language.
This Anglo-Saxon language was influenced by several other lan-
guages, including Celtic. Apparently not all the British Celts had been
forced out by the Angles, Saxons, and Jutes. Some remained, either
as captives or willing citizens, and in time they were assimilated into
the Anglo-Saxon population. This mixture of Celts and Anglo-Saxons
established the origins of the modern English people. The Celts’ influ-
ence is evident today in certain English place names, such as Dun-
combe, Holcombe, and Wincombe — names rooted in the Celtic
word cumb (‘‘valley’’).
The influence of Latin came through several paths. First, the lan-
guage of the Celts had been influenced by Latin during the Roman oc-
cupation. The word castra (‘‘camp’’) in the place names of Britain,
such as Lancaster, Winchester, and Worcester, illustrates the Latin
influence. Each location had been the site of a Roman camp. In addi-
tion, Latin had colored the language of the Angles, Saxons, and Jutes
long before these tribes conquered the Celts. For hundreds of years
42 History of English

these groups attacked and traded with Roman holdings, and as a

result they adopted many Roman words. The Old English word coper
(‘‘copper’’) was derived from the Latin cuprim or cuprum as a result
of commercial or military contact. Other words similarly acquired at
this time are disc (‘‘dish’’), camp (‘‘battle’’), segn (‘‘banner’’), weall
(‘‘wall’’), stret (‘‘street’’), and mil (‘‘mile’’). A few of the commer-
cial words derived from Latin are ceap (‘‘sale’’), pund (‘‘pound’’),
win (‘‘wine’’), and mynetian (‘‘to mint a coin’’).
A more important Latin influence on Anglo-Saxon took place after
the Christianization DE iaiiniare Si arduetineaqlenion oF
about forty monks arrived in Kent, where King 42thelberht received
them. He was baptized within three months, and by A.D. 601 Canter-
bury had been established as the center of the English church. By the
end of the seventh century, Christianity was flourishing as the religion
of England. With the success of the Christian mission came influence
on the Anglo-Saxon language. Because Latin was the language of the
Church, great numbers of Latin words were incorporated into the
changing Anglo-Saxon language. Many of these words are still used to
refer to the church and its activities, including altar, anthem, candle,
deacon, hymn, martyr, mass, Pope, priest, psalm, and temple.
The next four centuries, from 600 to 1000, saw yet another impor-
tant influence on the English language: Scandinavian settlement and
absorption into the Anglo-Saxon peoples. The two cultures were akin,
and the Scandinavian dialects were similar to Anglo-Saxon. But the
Scandinavian influx was accompanied by battles and bloodshed, and
even a brief time of total Scandinavian dominion over England. The
conflict had lessened by 1042, when the Anglo-Saxon Edward the
Confessor became king.
Because of the similarity of Scandinavian to Anglo-Saxon, deter-
mining the extent of Scandinavian influence on English is not simple.
One contribution easy to identify is in modern English words with sk-
(sky, skin, skill). The Anglo-Saxon equivalent of this sound was
spelled sc, but today is spelled sh as in shall, fish, shoe. Modern En-
glish preserves some pairs of words that originally had the same
meaning in Anglo-Saxon and Scandinavian: ship and skiff, and shirt
and skirt. Not all sk- words are of Scandinavian origin, however; for
example, the origin of skipper is Dutch and skirmish is from Old
French.
All the influences discussed — Celtic, Anglo-Saxon, Latin, and
Scandinavian — were felt gradually over five to six hundred years.
No language is static, and in this era English underwent diverse
changes. The language growing and forming during these years is cal-
led Anglo-Saxon or Old English.
 ‘ Old English 43

Old English
One of the easiest ways to discuss the relationship of Old English to
modern English is to look at an example of Old English. The follow-
ing copy of the Lord’s Prayer illustrates several symbols that no
longer appear in modern English, such as p and 8. These symbols
have been replaced by th.

1. Federure, pu pe eart on heofunum

Our Father Thou who are in heaven
2. Si pin nama gehalgod. Tobecume pin rice.
Be thy name hallowed. May thy kingdom come.
3. Gewurpe din willa on eordan swa swa on heofonum.
May thy will happen on earth as in heaven.
4. Urne gedeghwamlican hlaf syle us todeg.
Our daily loaf give us today.
5. And forgyf us ure gyltas,
And forgive us our sins,
6. Swa swa we forgyfad urum gyltendum.
as weforgive our debtors.
7. And ne geled pu us on costnunge,
And not lead thou us in temptation,
8. Ac alys us of yfele. Soplice.
And release us from evil. Truly.

Fortunately, not all the words in the Old English prayer appear
foreign to the reader of current English. Although a few words have
been entirely lost, such as rice (line 2), gewurpe (line 3), and cost-
nunge (line 7), there are many that remain in modern English, even
if their forms are somewhat different. For example, the first part of
the Old English word soplice (line 8) survives. today in soothsayer,
and -lice survives as the adverbial suffix -/y. Nama (line 2) remains
almost the same, and fxder (line 1) is close enough to father to be
recognizable.
The grammar of the Old English prayer strongly reflects its Ger-
manic heritage. Nouns in Old English were inflected for case
umber; for et ee ie a
ending for the dative case. Nouns also revealed gender: costnunge
(line 7) is feminine, but yfele (line 8) is neuter, and gyltas (line 5) is
masculine. Verbs were_i on, number, tense, and
mood. The verbs si, tobecume, and gewurpe (lines 2 and 3) are all in
the third person singular present subjunctive. The -ad in forgyfad is
an example of the first person plural indicative ending.
As these few examples illustrate, Old English was inflected much
44 History of English

more than modern English. All nouns, for example, were inflected to
show gender, number, and case. The five Old English cases were
nominative, accusative (objective), dative, genitive (possessive), and
instrumental. The instrumental case was rarely used in Old English; it
expressed the agent or means by which something was done. Because
Old English word endings indicated the exact function of each word
in a sentence, word order was less fixed than in modern English.
Superficially, Old English seems more complex than modern En-
glish. The only modern part of speech to retain a high degree of
inflection is the pronoun, specifically the personal pronoun. It has
twenty-three forms if the pronomial adjectives such as mine, hers,
and ours are included. Old English also featured many more strong
verbs than modern English A strong verb isonethat changes-its form
_in_the_ principal parts (present, past, and past participle), usually by
changing a_main vowel: sing, sang, sung. Weak verbs follow a more
regular pattern, taking an -e ing_in the past tense and past par-
ticiple: look,looked, looked.
Reading ;and [i
listening to Old English reveals that all the vowels
were pronounced; there was no such thing as a “‘silent e.’’ Vowels
bore a greater similarity to the Germanic languages and to the Ro-
mance languages derived from Latin than they do today. Stress nor-
mally occurred on the root syllable of each word. Today, stress on
syllables varies with the word, although native English words still
keep their stress on the root syllable.

Middle English
In 1066 an event occurred that was of immense significance to the En-
glish language. William the Conqueror, Duke of Normandy, led a
Norman French invasion of England. He defeated King Harold of
Britain at the Battle of Hastings. After the successful invasion, the
Norman French permanently displaced the English rulers and became
the ruling aristocracy. The class distinction between the Normans and
the English was maintained, at least in part, by a linguistic distinction:
the Anglo-Norman ruling class spoke French, and all the others spoke
English.
The need for communication between the Anglo-Normans and the
English over the next 350 years led to a vastly enriched English. The
most noticeable effect of French was on the English vocabulary. A
very large portion of the modern English vocabulary consists of
French words, many of them borrowed during this period. Many of
the words taken from the Norman aristocracy were words associated
with the political and social spheres, which they dominated — words
 ‘ Middle English 45

TABLE 4-]
English Words of Norman French Origin

Government Law War Religion Leisure Fashion

govern justice peace service art apparel

state judge battle trinity joy dress
power jury arms virgin pleasure garment
crown court banner angel flowers costume
authority summons officer saint chase tailor
nation attorney sergeant preach falcon
accuse navy sermon sport
damage soldier beauty
Source: Many of these words are derived from Jesperson (1935).

of governance, law, war, religion, leisure, and fashion. Some ex-

amples are given in Table 4-1.
In some cases, these words were introduced because they named
concepts or distinctions rarely used by the native English. In many
cases, however, an Anglo-Saxon word already existed and remained
in use along with the new French word. As a result, many pairs of
words with approximately the same meaning exist in English today,
one from French and one from Old English (see Table 4-2).

TABLE 4-2
Paired Anglo-Saxon and French Words

Anglo-Saxon French

kingly royal
clothes costume
help aid
begin commence
hide conceal
hunt chase
bill beak
folk people
feed nourish
clothe dress
kin relations
look search

Source: Many of these words are derived

from Jesperson (1935).
46 History of English

In at least one instance, the class distinction between the Anglo-

Normans and the English peasantry led to an interesting set of word
doublets in modern English, in which the name of the animal is
Anglo-Saxon in origin, but the name of the meat is Norman French.
Cow, calf, sheep, and pig are all Anglo-Saxon words; beef, veal, mut-
ton, and pork come from French. The English-speaking peasants
looked after the animals. As far as they were concerned, they raised
‘‘cows’’ and, if they were lucky, on occasion they ate ‘‘cow.’’ They
could not help but notice that their Norman overlords always ate
‘‘beef.’’? Eventually, the peasants took to eating ‘‘beef’’ themselves,
but continued to raise ‘‘cows’’ as they always had.
The Anglo-Norman nobility continued to use French well into the
thirteenth century; moreover, the aristocracy maintained their French
estates and relationships. Ultimately, after lengthy disputes, the
French court forced the Anglo-Normans to divest themselves of
French possessions. The nobility became more English as its sense
of French identity waned. The use of English as the primary language
became common among the upper class, and French was not much
used in England by the end of the fourteenth century. For example, in
1362 English was ruled to be the language of legal proceedings.
While the Norman French were being assimilated into English cul-
ture and society, other changes were taking place in the language.

sie ded _ orounced to9 thyme with modern maid) became ae

ern English deed; Old English nama (which rhymes with llama) be-
came today’s name; hus (rhymes with moose) became house; riden
(rhymes with bleedin’) became ride. This change was probably com-
pleted sometime early in the seventeenth century, during the Early
Modern English period.
Another very important change in Middle English occurred with the
vowels a, o, u, and e in unstressed syllables. As unstressed vowels,
the vowels lost their_individual Teiiiciaganys aiRee naan
JoURCed ASA quick “uh.” Eventually These Soames wereTostatthe
end of most Si odern English. For example, Old English
nama became Middle English name (with the e pronounced ‘‘uh’’)
and finally our modern equivalent name.
Changes in grammar occurred as well. Many noun endings indicat-
ing grammatical function were shortened or lost. For example, the
ending -n was dropped from the possessive pronoun myn, producing
my, and from the indefinite article before a consonant, so that an be-
came a. Grammatical gender of nouns was also lost, and changes in
word order took place.
 Early’ Modern English 47

Throughout the Middle English period, English was not a unified

language. Dialects continued to diverge, as they had from the begin-
ning, with wide diversity in pronunciation, vocabulary, spelling, and
even word endings. For example, the third person plural ending of
verbs varied from no ending or -s in the Northern dialect to -en in the
Midlands and -eth in the Southern dialect.
By the close of the Middle English period, it was apparent that En-
glish had triumphed in England. It had displaced both French and
Latin as the written language of the people. Town records, parliamen-
tary records, and local ordinances were written in English. Mean-
while, French was taught as a foreign language in order that the En-
glish could communicate with their French neighbors. A national
literature, written in English, was emerging. The most memorable
writing in the ‘‘new language’’ was that of Geoffrey Chaucer, whose
Canterbury Tales is often studied today in its original, Middle English
form. William Langland, author of Piers Ploughman, and John Wy-
cliffe, translator of the Bible, made further contributions to the devel-
oping language.

Early Modern English

- Whereas Middle English was characterized by diversity, Early Mod-
ern English (circa 1500-1700) was noted for its greater uniformity. All
the written dialects began to merge as a result of three major factors:
the printing press, increased communication, and growing literacy.
In 1476 William Caxton introduced the printing press to England.
Previously, all manuscripts had been hand-copied and were limited in
number. For the first time, books and other printed materials became
widely available. Because London was the chief commercial center,
the center of government, and near the great universities, most books
were printed in the London area and written in the London dialect.
Thus, the London dialect was disseminated throughout England. It
was not superior to the others in any sense; it just happened to be the
one being printed and disseminated.
Government policy also helped to spread the London dialect. In
1516 Henry VIII established the first English postal system. The in-
creased communication that resulted reduced the isolation of the out-
lying districts, bringing them increased contact with London. For a
long time London had been the political, social, and economic center
of the country; its new accessibility allowed the city to gain even
greater importance throughout England.
During this same period, dramatic changes took place in education,
partly because of the availability of printed materials. Prior to the fif-
teenth century, only the clergy and the nobility were literate; but dur-
48 History of English

ing the fifteenth century, many middle-class people learned to read

and write. Incr iteracy, the spread of London’s dialect, and the
ing importance of London as the nu
ntry helped to establi ondon dialect as the literary standard
for the
sign one of the London dialect, uniformity followed in spell-
ing, pronunciation, and grammar. The changes in verb endings are a
good example of what was taking place. The third person singular
endings -(e)s of the Northern dialect triumphed over the -(e)th ending
of the Midlands during the fifteenth century, although both were still
common in writing at the end of the sixteenth century. Shakespeare,
who wrote during this period, used both endings, as in Portia’s line:
‘It blesseth him that gives, and him that takes’’ (Merchant of Venice,
act IV, scene 1). Of the two endings, the -(e)s proved dominant,
and it progressively displaced -(e)th in common use. As a second
example, the Middle English endings for the third person plural —
Northern -(e)s or no ending, Midlands -(e)n, and Southern -(e)th
— dropped out of use.
Many other changes occurred in English grammar during this
period. One example is the regular use of the contracted negative
forms, such as won’t, which did not begin to appear in written texts
until the Early Modern English period. Another is the regular use of
the expanded verb forms containing “‘to be,’’ particularly the ex-
panded passive as in He is being chosen.
Unlike Middle English or Old English, Early Modern English looks
familiar to today’s reader. Many of its forms appear archaic, but they
are not obscure. The word order looks like modern English, but the
diversity of spelling is substantial.

Modern English
The steps toward regularity in the written English language, which
were begun in the Early Modern English period, continued in the
eighteenth century, when many dictionaries and grammars were
printed. No doubt such works were strongly influenced by the French
Academy, which in the previous century had been charged to ‘‘labor
with all possible care and diligence to give definite rules to our lan-
guage, and to render it pure, eloquent, and capable of treating the
arts and sciences.’’! Such men as John Dryden and Jonathan Swift

1 Albert Baugh, A History of the English Language, 2nd ed., p. 317 (taken from D. M.
Robertson, A History of the French Academy, London, 1910).
 “Modern English 49

headed a movement to create an English Academy with similar func-

tions. Ultimately, the movement failed because of a growing sense of
the futility of the proposed academy’s broad purpose.
Bishop Lowth’s Short Introduction to English Grammar (1762) and
Samuel Johnson’s A Dictionary of the English Language (1755) were
probably the most important of the grammar texts and dictionaries
that appeared in the eighteenth century. These books established a
single standard for correct English. In fact, Johnson is often regarded
as the man who conferred stability on the English language. Johnson
worked alone for only seven years, yet his dictionary was a memora-
ble first for the language. Although it lacked the completeness of in-
formation that modern dictionaries possess, it displayed the English
vocabulary more fully than ever before, offering a standard for spell-
ing to which all could refer.
As Johnson’s dictionary established a uniform standard for spelling
and word use, so Bishop Lowth’s and other grammarians’ works
standardized English grammar. To grammarians in the eighteenth cen-
tury goes the credit (or blame?) for outlawing the double negative.
That is, sentences such as She didn’t go neither and Hamlet’s ‘‘Be
not too tame neither’? were declared ungrammatical. Two negatives,
it was decided, make a positive, thereby condemning a common
idiom to its doom. Which pronoun case should precede a gerund was
another question debated by grammarians of the day. Eventually, the
decision of Noah Webster prevailed, and constructions such as I
worry about his crossing the street, were accepted as grammatically
correct. The distinction between shall and will originated during this
time, as did numerous other grammatical points, such as the rule
against ending sentences with prepositions. It was a time of vigorous
activity on the part of language enthusiasts who were attempting to
codify and regulate the English language. The enthusiasts were so
eager to set prescriptive rules that they managed to amass a consider-
able number of them, some of which are retained today; others have
been abandoned.
While grammarians debated the correctness of language at home,
English colonization went on abroad. Acquisitions of new territories
meant acquisitions to the vocabulary of the mother country. Contact
with Africans, Indians, and Native Americans and with Spanish and
Portuguese colonists resulted in many additions to the language.
Thus, we have English/Mexican words such as tomato, coyote, and
chocolate. From South American regions come jaguar, poncho,
and tapioca; from India, bungalow, cheroot, juggernaut, mandarin,
and hosts of others; from Africa, chimpanzee, banana, guinea,
voodoo, and gumbo.
50 History of English

American English
When the colonists arrived in America in 1620, they brought their
Elizabethan English with them. From this beginning the development
of modern American English has depended on the unique American
experience. Rules of grammar are similar to England’s English, of
course; but other facets of the language developed with the new
country and reflect contributions from the many groups that became
part of the American ‘‘melting pot.”’
Although the colonists mingled to some extent with the Native
Americans, the Indians’ influence on American culture and language
was not extensive. But the colonists found many things in the new
world that were unfamiliar, and they adopted the Indian words for
them. Many place names, trees, plants, and animals indigenous to
America retain their Indian names today. Hickory, squash, pumpkin,
woodchuck, chipmunk, and dozens of other such words are Indian in
origin. Approximately half of the states’ names, including Connecti-
cut and Alabama, are also of Indian origin.
The French were important to the exploration of America, and their
settlement concentrated in the Midwest and around New Orleans and
Quebec. However, they had little permanent linguistic influence in the
United States except in the Creole regions of southern Louisiana. The
words they contributed to English were primarily designations for
the features of the land they explored — prairie, butte, crevasse, and
so on — although they introduced a variety of other words as well,
such as gopher, chowder, and bureau.
The Spanish, on the other hand, had extensive settlements in the
West and Southwest, and a much greater influence on American cul-
ture and language. Many words associated with life in the West are
Spanish in origin. The cowboy, one of the most popular heroes of
American culture, was Spanish in almost everything but his name. He
wore chaps and ponchos; he twirled a lariat and rode a bronco that
ate alfalfa. He lived on a ranch and fought coyotes. He may even
have been caught by vigilantes and held incommunicado in the
hoosegow.
Another linguistically important group of early American colonists
was the Dutch in New York. Although they were displaced by the
English after a short time, their influence on American vocabulary
was considerable, including such words as cole slaw, cookie, waffle,
boss, caboose, sleigh, dope, dumb, snoop, and even Santa Claus.
Later groups of immigrants — Germans, Irish, Italians, Scan-
dinavians, Africans, and Eastern European Jews — contributed a
sprinkling of familiar words to the American vocabulary, as shown in
 American English a

TABLE 4-3
Examples of Borrowed Words in American Vocabulary

German Ttalian Scandinavian African Jewish

noodle spaghetti ski gumbo kosher
delicatessen pizza smorgasbord goober goy
pretzel oregano sauna juke (box) blintz
bum ombudsman jazz Shalom
hamburger banjo oy
samba

Table 4-3. Certainly, other immigrant groups have contributed to

American English as well, although often within a limited area. For
example, Japanese and Chinese influence on American English has
largely been confined to Hawaii, where numerous indigenous Hawai-
ian words also are in common use (mauka, ‘‘seaward’’; pau, ‘‘fin-
ished’’; and, of course, aloha).
As American English developed, it did not develop uniformly
throughout the country. Instead, as the colonists from different parts
of England settled in various parts of the new country, they continued
to speak their separate British dialects. In America today, these dia-
lects are characterized most obviously by differences in pronunciation
and vocabulary.
Although American English has been relatively independent from
British English since the seventeenth century, the two have remained
similar. The difference between the New England dialect and British
speech today probably is no greater than that between the New En-
gland and the Southern dialects in America. This similarity results
from many factors. First, as a result of the printing press and the at-
tendant spread of education, the rate of change in languages generally
has slowed. Second, the availability of grammars and dictionaries in a
permanent form (or perhaps just the availability of all prose in a per-
manent form) has provided more durable models for language than
ever before. American education has traditionally stressed the study
of British literary works. Thus, British English has had a continuing
(although incalculable) effect on the American people.
Nevertheless, certain differences exist. A Briton is immediately
identifiable in America by his accent, and an American is equally dis-
tinctive in England. In other words, regular differences in pronuncia-
tion can be described. Certain differences exist in vocabulary as well,
although almost entirely for things associated with twentieth-century
inventions (see Table 4-4). Few grammatical differences are no-
52 History of English

TABLE 44
Twentieth-Century Differences in American and British Vocabulary

American British

trunk (of a car) boot

hood bonnet
gas petrol
elevator lift
truck lorry
wrench spanner
phonograph grammaphone
subway underground

ticeable, although one difference is illustrated by a sentence such as

She told me to read it, but I already had done. Americans would find
this sentence very peculiar if not unacceptable, and they would
require that the pronoun it be included at the end of the sentence.

Summary
The changes in a language parallel the history of its speakers; a lan-
guage is responsive to the social forces that shape history. The his-
tory of English is an excellent example of this principle.
When the Roman legions arrived in Britain in A.D. 44, the inhabi-
tants of Britain were loosely organized groups of Celts. During the
next several hundred years, the Romans protected the Celts from
their warlike neighbors to the north, the Picts and Scots. In 410, how-
ever, the Romans were forced to withdraw from Britain. More vulner-
able to the Picts and Scots, the British Celts called on continental
Germanic tribes for protection. The Germanic agreement to provide
help set the stage for invasions by Angles, Saxons, and Jutes in the
latter part of the fifth century. It was their political dominance that es-
tablished Anglo-Saxon, or. Old English. The language was influenced
by the Christianization of Britain at the end of the sixth century,
which added large numbers of Latin words to Anglo-Saxon, and by
several hundred years of Scandinavian invasion and colonization,
which added Scandinavian words to the growing vocabulary.
As a result of the Norman conquest of 1066, large numbers of
French settled in England and created a new phase of language devel-
opment that added vast quantities of French words. This Middle En-
glish period (1100-1500) was marked by great diversity and rapid
change. Many changes in pronunciation took place, including the
 Summary 53

Great Vowel Shift. Grammar was altered extensively as most gram-

matical endings were lost or modified. Dialects continued to diverge.
The introduction of the printing press to England brought changes
in education and literary output and firmly established the London
dialect in which most books were printed. Since London had become
the governmental and cultural hub of England, its dialect was ac-
cepted more readily. The London speech in this period (circa
1500-1700) is illustrated by Shakespeare’s writing and generally
termed Early Modern English.
By the eighteenth century, standards for the changing language ap-
peared. Dedicated grammarians compiled many dictionaries and
grammars, and the form of English became much like the language we
know today. Standards were set for sentence construction, use of
words, and spelling.
Modern American English began with the arrival of the Pilgrims
and grew with the country. It depended for its growth on the various
linguistic groups that blended their customs and vocabularies on the
American continent. From the American Indians came words for na-
tive flora and fauna, as well as place names. From the French ex-
plorers, we inherited names for the terrain; from Spanish settlers of
the West, a vocabulary describing ranch life; from later immigrant
groups, a wealth of additional words. Yet American English is not
greatly different from British English. Until quite recently, American
education and literature have been heavily dominated by British mod-
els. The most recognizable differences today are in pronunciation and
vocabulary. Dialects within the English language have not been
erased entirely.

SOURCES FOR FURTHER READING

Baugh, Albert C., and Cable, Thomas. A History of the English

Language. 3rd ed. London: Routledge, 1978.
Gordon, James Daniel. The English Language: An Historical
Introduction. New York: Thomas Y. Crowell, 1972.
Pyles, Thomas. The Origins and Development of the English Lan-
guage 2d. ed. New York: Harcourt, Brace Jovanovich, 1971.
Rigg, A. G. ed. The English Language: A Historical Reader. New
York: Appleton-Century-Crofts, 1968.
Robertson, Stuart, and Cassidy, Frederic G. The Development of
Modern English. 2d ed. Englewood Cliffs, New Jersey: Prentice-
Hall, 1954.
Williams, Joseph M. Origins of the English Language: A Social and
Linguistic History. New York: Free Press, 1975.
ey
History of English

QUESTIONS FOR DISCUSSION

ik At which point in the history of the English language did the

major Latin influence appear?
ae Explain the tremendous French influence on the vocabulary of
English, a Germanic language.
. Since the personal pronoun in modern English has twenty-three
forms, it is our most inflected part of speech. Name several of
these pronouns and explain why we need these different forms.
. What was the Great Vowel Shift that occurred in Middle English
and Early Modern English? Name some other changes that oc-
curred in English during these periods.
. What significant Middle English writings helped to establish En-
glish, rather than French, as the dominant language of the island?
. Name and discuss briefly the influences that established the Lon-
don dialect as the standard for Early Modern English in England.
. Discuss briefly the grammatical changes that took place in Early
Modern English.
. Who was responsible for the first English dictionary? What pur-
pose did this dictionary serve?
. Note the names of several early grammarians who helped codify
modern English.
10. Cite several influences on American English that help to explain
the use of the term melting pot to describe American English as
well as America.
ile Can you see any changes in progress in the grammar of modern
English? Discuss.
 PART

I]

LINGUISTICS:
GRAMMAR
 os
ic
“as
i 2 ‘

J ah,

a ——

‘ = =
- as
woe
ae ~
y
a
oe ies : _
t a i
; as
* e -
= —_—* af
‘ e

Se

5 ie

Yer 3 ”
e

a ke t

i. 7 a
= at a = a0
” re ——~ Fe, neu?

| a “aXFe
a ‘Ss tar

“i
 CHA
PT ER

PHONETICS

An obvious way of beginning an analysis of spoken language is by

studying its sounds; this study is called phonetics. We can take two
basic approaches to phonetics: the acoustic approach, which looks at
the physical properties of the sounds themselves; or the articulatory
approach, which looks at sounds in terms of how they are produced,
or articulated. The acoustic approach, discussed in Chapter 18, tells
us that speech is a continuous flow of sound. Sophisticated sound-
recording instruments reveal this continuity in speech as well as an-
other fact: no speaker ever produces two utterances that sound ex-
actly the same. In contrast, the articulatory approach looks at speech
sounds as composed of segments.
Despite its distortion of reality, we begin with the articulatory ap-
proach for two reasons: people tend to perceive speech as segmented,
and the articulatory approach gives us a way of talking about and
classifying the sounds of speech. Thus, we begin this chapter by look-
ing at how in general humans produce sounds and how we can clas-
sify sounds according to the way they are produced. Then we exam-
ine relationships among sounds and two acoustic characteristics of
speech.

The Production of Speech Sounds

Speech is an incredibly complex activity. It involves coordinated ef-
forts of all the participating parts of the vocal tract (see Figure 5-1).
One of the basic exercises in voice control is breath control, which in-
D7
58 Phonetics

FIGURE 5-1
The Vocal Tract

nasal cavity | soft palate

ees
hard palate —
(velum)

oral cavity
uvula

teeth Saas \. pharynx

(to men

volves training certain muscles involved in air expulsion, the first step
in producing sounds. Air proceeds from the lungs through the trachea
to the larynx, commonly called the voice box, which houses the vocal
cords. If the cords are slightly tensed, the passage of air sets the vocal
cords vibrating, which gives a basic sound quality to the air stream,
which continues into the pharynx, where basic voice quality is es-
tablished. Voice quality determines the unique characteristics of each
speaker’s voice, so that an individual often can be recognized by
voice alone. Above the pharynx is the uvula, which is a movable flap
that controls the passage of air through the nasal cavity. The uvula is
always open when an individual breathes through the nose, but it is
only open at certain times during the course of speech. For the most
part, the velum (or soft palate) is closed in speech, and the air moves
through the oral cavity (the mouth), the dimensions of which change
according to the interaction of the tongue and lips. These changes
 Consonants 59

result in what all speakers of a language recognize to be the sounds of

their language. The study of these speech sounds is called phonetics.
Every language has a variety of sounds, and many of the sounds
found in the languages of the world are not found in English. In this
chapter, all discussion refers to American English. The English alpha-
bet consists of twenty-six letters, but quite a few more than twenty-
six sounds are used in speaking English. To discuss these sounds, a
new alphabet of sounds, or a phonetic alphabet, is required. The pho-
netic symbols are enclosed in brackets to distinguish them from let-
ters. Thus, [p] is a phonetic symbol indicating a particular sound. Al-
though in many cases this phonetic alphabet corresponds to English
spelling, many phonetic symbols have no direct correspondent in En-
glish spelling. Thus, it is important to remember that the phonetic
alphabet describes only sounds, and each symbol represents only one
sound.

Consonants

The principal division of sounds is between vowels and conso-

nants. Every language makes this distinction. Vowels are defined as
those sounds produced with the oral cavity relatively open to the flow
of air. Consonants, on the other hand, are sounds produced with a
constriction or occlusion in the oral cavity.
The consonants may be grouped according to how the sounds are
produced. English then has six groups of consonant sounds: stops,
fricatives, affricates, nasals, liquids, and_glides (see Table 5-1). The
sounds within these groups can be further classified according to the
place of articulation, that is, the position of the lips or tongue as
the sounds are made.
1. If the lips are pressed together, the sound is a bilabial.
2. If the lower lip jspressedto the upper teeth, a labiodentalis pro-
duced.
3. The tip_of the tongue may be placed between the teeth, producing
an interdental.
4. The tip or blade of the tongue may be placed against the alveolar
ridge (the hard ridge behind the upper teeth), producing an al-
veolar. ay
5. The midsection
of the tongue-may be pressed against the hard pal-
ate |(palatal), or against the soft palate or velum (velar).

STOPS
By entirely closing off the flow of air at some point in the mouth,
stops, or plosives, are formed. Air pressure from the lungs builds up
and is suddenly released in a sharp burst of sound. In the production
60 Phonetics

KACK
TABLE 5-1
Phonetic Representation of English Consonants

‘S
5
jee
~ Interdental
Labiodental Pharyngeal

Voiceless

Stops Voiced

Voiceless
Fricatives
Voiced

Voiceless
Affricates
Voiced

Nasals Voiced

Liquids Voiced

Glides , Voiced

English Consonants Illustrated by Examples

[p] pat, map [3] then, father

[t] tin, mar [z] zinc, ties”
[k] came, pick [Z] measure, azure
[?] bottle [é] chain
[b] bud, dub [j] Jane
[d] din, mad [m] man, came
[Ig] game, flag [n] no, tin
[f] fine, leaf [n] sing, finger
[e] thin, tooth {l] /ap, fall
[s] sink, class [r] rap, tar
[S$] sure, push [w] witch, sew
{h] Hand, who [y] you, few
[v] vine, leave
 Consonants 61

of English stops, the mouth may be sealed off in three major ways. If
the lips are pressed together to seal off the air and opened in a sharp
burst, the bilabial sounds [p] and [b] are produced, as in the words
pop and Bob. In the alveolar pair of stops, the tip of the tongue is
pressed against the alveolar ridge to produce [t] and [d], as in tot and
dad. In the third pair of stops, the velars, the back of the tongue is
pressed against the soft palate, or velum, to produce [k] and [g], as in
the words kick and gag.
The difference between the members of each of these pairs lies in
the operation of the vocal cords, which are two elastic membranes
that can be moved by muscles in the larynx (see Figure 5-2). The
position of the membranes can vary from completely closed to com-
pletely open. During the act of swallowing, the vocal cords close to
prevent food or drink from going into the lungs; they are open as one
breathes. During speech, the vocal cords usually are open in varying
degrees. In the production of [p], for example, the vocal cords are
completely open and air may pass freely through the larynx. For [b],
on the other hand, the vocal cords are almost closed, leaving only a
narrow slit. As the lungs force air through this slit, the membranes
vibrate in a fashion much like a kazoo. This vibration is called voic-
ing; it can be heard by putting your hands over your ears while saying
a long, continuous [b] in which the light contact of both lips is main-
tained. Repeat this experiment, saying a continuous [p]. There will be
no vibration; [p] is a voiceless consonant, and [b] is its voiced counter-
part. Voiced sounds are produced when the vocal cords are almost
closed and air from the lungs causes them to vibrate; voiceless sounds
are produced when the vocal cords are open. Many of the consonants
appear in voiceless/voiced pairs like [p] and [b].
The voiceless stops — [p], [t], and [k] — are further characterized
by degrees of aspiration, or the amount of air that accompanies their

FIGURE 5-2
The Vocal Cords (from above)

Soh 7

open position semi-closed position

(voiceless sounds) (voiced sounds)
62 Phonetics

production. In initial position — that is, at the beginning of a sylla-

ble — the voiceless stops are aspirated, as in pin, tall, and come. If
you hold a lighted match in front of your mouth when saying such
words, the flame should go out or at least flicker. When preceded by
an s or followed by r or /, as in spin, tray, and clay, significantly less
aspiration accompanies the consonants; these sounds are called un-
aspirated. Finally, when voiceless stops occur in final position, or at
the end of a word, they frequently are unreleased or unexploded;
mop, nick, and mat are examples.
These three variations of aspiration can be distinguished in tran-
scription by using the following symbols: [p", t", k®] for aspirated
stops; [p:, t:, k-] for unaspirated stops; and [p~, t-, k~] for unreleased
stops. Because the relative aspiration of these consonants in English
is automatic, it often is not designated. For our purposes, only the
symbols [p, t, k] will be required.
A different sort of stop is the glottal stop, indicated by [?]; (it is not
shown in Table 5.1). This sound is produced by closing off the flow
of air at the glottis and suddenly releasing the air. It can be heard be-
tween the vowels in the word kitten, satin, or bottle. All these words
are spelled with t’s between vowels; but for many speakers, the words
do not contain a t-sound at all. The glottal sound can be a difficult
consonant for the novice to hear, because no part of the mouth func-
tions in its production.

FRICATIVES
By forcing air through a narrow opening in the oral cavity, a process
that creates audible turbulence in the airstream, fricatives are made.
The constriction may be made in different places of the mouth, as is
the case with stops. When the lower lip is pressed lightly against the
upper teeth and air is forced out between them, two labiodental frica-
tives, [f] and [v], may be produced. The voiceless sound [f] occurs in
the beginning of the word fine, and its voiced counterpart, [v], occurs
at the beginning of vine.
The next pair of fricatives are the interdentals, [e] and [d]. These
are spelled th in English. Some people pronounce these sounds with
their tongues between their teeth; others produce them with their
tongues pressed lightly against the upper teeth. The voiceless [e] ap-
pears in words like thin, thought, pithy, teeth, and bath. The voiced
[d] is present in then, that, bathe, and mother.
The alveolar fricatives are the voiceless [s] and the voiced [z]. They
are formed by placing the tongue in light contact with the alveolar
ridge and forcing the air out. The consonant sounds in sis and the
final sound in buzz illustrate alveolars.
The palatal fricatives, [5] and [Z], are formed by arching the center
 Consonants

of the tongue up toward the hard palate. The [§] is usually spelled sh
in English, as in ship, althoughthe sound occurs also in words like
sure, omission, and location. Its voiced equivalent, [Z], never occurs
at the beginning of native English words; it is usually found in the
middle of words — as in measure, seizure, and vision — and at the
end of a few borrowed words — as in rouge and garage — in some
dialects of English.
The last fricative, [h], is voiceless and has no voiced counterpart.
Strictly speaking, this sound does not meet the definition given for
fricatives because [h] does not usually involve turbulence caused by
constriction, although there is turbulence when [h] precedes certain
vowels, as in a word like heal. In any case, many linguists have
chosen to include [h] among the fricatives.

AFFRICATES
The affricatesare a special group of sounds that are formed by_com-_
bining a stop and a fricative. In English, only one pair of sounds
occurs in this category, [¢] as in chain and rich and [j] as in Jane and
ridge. Notice that in pronouncing [é], one seems to pronounce [t] fol-
lowed by [S]. Similarly, [j] is much like a phonetic compound, con-
sisting of [d] followed by [Z].

NASALS
In English, the three nasals, [m, n, n], are made with the lips and
tongue | respective positions as they are for [p, t, k]; how-
ever, air pressure does notbuild upasitdoes inthestops. Instead,
the_uvula (th the nasal passage) is
open, allowing the air toflow through the nose. English,
In the nasals
are always voiced. Whereas [m] and [n] may occur at the beginning as
well as at the end of a syllable in English, as in mom and nun, [py]
occurs only at the end of a syllable, as in sing.
Notice that there is no [g] pronunciation inherent in [n] for most
speakers of English. On the other hand, in transcribing the word
finger, both [pn] and [g] must be used because the second syllable
‘begins with a [g] sound. Also, the word rink would be transcribed as
tries nk):

LIQUIDS
The consonants [I] and théasa in lilt and roar, are called liquids.

to
siaihes
the tip of the tongue. >scape to
alveolar ridge and allowing air to escape
each side. The [r] sound in English is formed by curling the tip of the
tongue_up behind the alveolar_ridge and flipping itforward and up-
ward without actually touching the alveolar fs2
64 Phonetics

GLIDES
The last two consonants are the glides, [w] and [y]. A [w] is formed
with the back of the tongue arched high and the tipsin a rounded
position ——much as they are in making the sound ‘ > (as in foo).
For example, in the word woo, a word notoriously difficult for many
foreigners to learn to pronounce, the lips begin and end in the same
position. The [y] glide, much like the [w], is formed with the joneue
and lips in the same position as they are when making the sound*
(as in bee). Say the word yeast and note the position of your ior
Both [w] and [y] always appear either beforeor after a vowel in En-
glish. In both cases; the sounds ‘“‘glide”’ rapidly to or from the ar-
ticulatory position for that vowel. Since [w] and [y] possess certain
vowel-like properties — for example, they lack a definite constriction
of the oral cavity — they are not true consonants and are often called
semivowels.

Vowels

Vowels are voiced continuous sounds involving no interruption in the

flow of air through the or, al| cavity. Different vowel sounds result from
‘changing the shape of the mouth; each vowel is associated with atas
ferent configuration of the tongue and lips. For example, to say‘
represented phonetically as [i], the lips are somewhat pulled back ee
the tongue is arched up toward the palate. To say ‘‘oo,’’ as in woo
and Sue, represented as [u], the tongue is raised toward the back of
the mouth and the lips are rounded and pushed forward. For ‘‘ah,’’ as
in father and cot, represented as [a], the tongue is flattened and low-
ered.
Like the consonants, each vowel is associated witha phonetic sym-
bol. Traditionally, vowels have been described along two dimensions:
tongue height and the frontedness or backnéss of the tongue. Table
5-2, which is commonly called a vowel chart, shows the spatial re-
lationship among vowels in terms of these two dimensions.
The vowels presented in Table 5-2 generally characterize Stan-
dard English or General American speech; that is, a form of speech as
free from regional pronunciation variations as possible. English
speech sounds, especially vowels, vary considerably from speaker to
speaker; in most cases, a finer transcription than the one presented in
the vowel chart is necessary to capture these distinctions because
vowels lie along a continuum. This means that vowels may be pro-
duced slightly higher or lower or more fronted or backed than shown
in Table 5-2.
To represent these variations in vowel sounds, symbols called dia-
critics or diacritic marks may be used. Thus, arrows may show how aa
 Vowels 65

TABLE 5-2
Traditional Representation of Standard English Vowels

Front Central Back

vowel differs from the Standard English form. For example, [4] in-
dicates a slightly fronted pronunciation of [a], and [i}] indicates a
slightly lowered pronunciation of [i]. Similarly, dots may be used to
show the length of a vowel: one dot following
a vowel indicates slight
lengthening, [a‘]; and twodots_indicate greater lengthening, [a:]. A
horizontal line above a vowel also may indicate a long vowel, fil. In
this text we will rarely need diacritics.
The vowels of English are either monophthongal or diphthongal;
that is, they are made up of eithera single sound or two sounds in
sequence. The major diphthongs of English are [ay], [aw], and [oy],
as in the words ride, house, and boy (these are not shown in Table
5-2). Most linguists treat the English vowels [i], [e], [0], and [u] as
diphthongs as well, because most speakers typically pronounce them
with following glides. It
It is not uncommon to find these vowels way

5-3.

TABLE 5-3 LEX Op

English Vowels Illustrated by Examples

[i] beat, key, fee [u] book, put

{1] bit, inch {u] boot, through, suit
[e] bait, gay, fate [a] butter, rough, ratify*
[e] bet, end, head [ay] bite, fight
[ez] bat, and [aw] how, sauerkraut, about
[a] calm, father [oy] boy, hoist
[9] bought, crawl
[o] boait, snow, hoe, though

*More frequently, the symbol [a], read as ‘‘schwa,”’ is used instead of [a] in un-
stressed syllables.
66 Phonetics

Distinctive Features

So far we have discussed how sounds are produced, but little or no

phonetic theory has been involved. Sounds can be classified in a vari-
ety of ways to suit a variety of theoretical pursuits. Perhaps the most
important of the theoretical approaches is distinctive-feature theory.
Its purpose is to isolate and identify the relationships between sounds
and to provide a universal phonetic system to describe the sounds
found in all the world’s languages. Again, our application will be re-
stricted to the sounds of English. In Chapter 6, we shall see the theo-
retical implications of distinctive features; here, we are concerned
with what the features of speech sounds are.
A feature is ideally a binary attribute; that is, a phonetic aspect that
is either present or absent in a specific sound. For example, one of
the distinctive features is [voiced]. Thus, such sounds as [b], [m], [a],
and [I] can be characterized as [+ voiced] whereas such sounds as [p],
[f], and [s] are characterized as [—voiced]. In a completely accurate
phonetic description, features may be preceded by integers that de-
scribe the degree to which a feature is present; for example, [2 nasal].
The vowel [z] in the word can’t, for example, is partially nasalized
because it is affected by the following nasal [n]; however, it does not
possess as great a degree of nasality as [n]. Thus, the [n] in can’t is
[1 nasal], indicating the greatest degree of nasality, and the [z] is
[2 nasal], indicating a secondary degree of nasality. When the features
are used for classification, however, they are always either present
(+) or absent (—).
The three sounds [p], [t], and [k] all share the feature [—voiced].
Therefore, additional features must be indicated in order to distin-
guish them from each other. Two additional features are required:
[anterior] and [coronal]. The feature [+anterior] designates those
sounds formed in the front of the mouth; that is, at the alveolar ridge,
the teeth, or the lips. The feature [+coronal] specifies those sounds
made with the front half, or blade, of the tongue. With these features,
we can now differentiate the voiceless stops [p], [t], and [k]:

pet aak

Voiced - — —
Anterior + + —
Coronal -—- + -—

Thus, the feature [voiced] does not help us distinguish among [p].
[t], and [k]; but the three consonants are completely distinguished
from each other by the features [anterior] and [coronal]. Tables 5—4
and 5-S list the distinctive features that characterize English sounds.
 Distinctive Features 67

TABLE 5-4 SY (\cpi

The Distinctive Features

The following features serve to distinguish between vowels and consonants:

Syllabic — the role a sound plays inthe structure of a syllable. [+syllabic]
sounds include all_vowels, as well as cnasals and liquids when they
function_as the basis of.a syllable, as in regen, The semivowels, [w]
and [y], are [—syllabic].
Consonantal —a_ constriction or occlusion of the oral cavity.
[+consonantal] sounds include stops, fricatives, affricates, nasals,
and liquids. Note that [h] is [—consonantal], although it has tradition-
ally been classed as a consonant, because [h] does not exhibit the
constriction that is typical of consonantal sounds.
Sonorant — the resonance of a sound. Sonorant sounds are produced with
a configuration of the vocal tract cavity that allows spontaneous
voicing. [+sonorant] sounds include vowels, nasals, liquids, and
semivowels.

The following features serve to distinguish among consonants:

Continuant —- those sounds that involve a constriction or occlusion of the
_oral cavity over a period of time. The ffricatives, liquids, and glides are
[+continuant]; the stops, nasals, and affricates are [—continuant].
Strident — high-frequency noisiness. The sounds [f], [s], [S], and [¢], along
with their voiced counterparts, are [+strident]. ~
Nasal — those sounds produced with the nasal cavity open. :
Anterior — those sounds formed in the anterior or forward part of the
mouth.
Coronal — those sounds formed with the blade of the tongue.
Voiced — the presence of vocal-cord vibration.

The following features principally serve to distinguish among vowels:

High — with the blade of the tongue abovea neutral position.
Low — with the blade of the tongue below a neutral position. The
midvowels [e], [e], [A], and [o] are neither high nor low and are
marked [—high] and [—low].
Back — with the tongue further back than a neutral position. In utilizing the
distinctive features, the central vowels [a] and [a] are considered
[+back].
Round — with the lips rounded. The vowels [a] and [a] are unrounded; thus,
they are distinguished from the back vowels, which are [+round].
Tense — with the muscles of vocalization tensed. The vowels marked
[-tense] are [1], fe],[a], [a], and [u]; all other vowels are [+tense].
Vowels marked as [—tense] are frequently called lax vowels; the ar-
ticulatory positions for tense vowels are generally maintained longer
than for lax vowels. Note that consonants may also be [+tense] or
[—tense], although for consonants this feature is best defined acousti-
cally.
68 Phonetics

TABLE 5-5 ae ee
Distinctive Features A Ag \ia

English ‘
Consonants pb fem, td ec0s mh S21 Sar koe yl ae ee
Sy LEA ee = epee re re I a ee
Consonantal 72°46 eee ese ee
Sonorant Be, eae eeet +
Continuant Sad rats ie ogee Ag ean A he lpm a pti eS
Nasal allot iene cette ee dh scones igen Sk 9 gece padoghmiten tours:
Strident latsili aha 8 iaaae Maes ge a CE ee oe ee
Anterior Solin OU pars Cuepuciei = aah 5.5 eet Meee ae See +
Corolla 9 +
Voiced eV faeerie
PE ee ge ee 9i i a cn eel a 7
LOWS 7 ee Se ee ee “Re
Backes |) seats Se eee Ss Re ee ts
Round 9%) se ye ee ete 5
hela
Tense Ee ek a Ee a Se a ee ee
Note: |, r, and the nasals are [+syllabic] when they serve as the basis for a syllable.

English
Vowels iT €6é Apa 5-60

Syllabic ++4+Ht4+44 4444

Consonantal == = === = === —
Sonorant Se See ae te) oe oe Se
High +4+----—---- ee
Low 6S s Se Ae
Backs tL aiacSee aie ++ 4+ 44+
Round ek ee Ae) ae eae

Tense Lees Bicets eg so +

The descriptions are given in articulatory terms because such defini-

tions are simplest, but each feature may be defined in acoustic terms
as well (for example, see strident in Table 5-4). This feature system
makes possible the complete differentiation of all the sounds of a lan-
guage. ;

Suprasegmentals

to aspects of pronunciation that go beyond the production of indi-

vidual (segmental) sounds. Thus, they may be defined better in acous-
tic terms than in articulatory terms.
 Suprasegmentals 69

STRESS

Pitch, length, and loudness are components. of stress. The basic unit
containing stress is the syllable. A syllable generally consists of a
vowel accompanied by one or more consonants, and the most com-
mon syllable form in English is a consonant plus a vowel. In actual
pronunciation, one may stress a syllable by giving it a higher pitch,
making it louder, or making it longer — or perhaps by a combination
of all three. In English, native speakers generally recognize at least
three levels of stress: primary, secondary, and unstressed. They are
indicated with the following stress marks:
Primary: 7
Secondary: \
Unstressed: (no mark)
Alternatively, stress may be treated as a distinctive feature; thus, the
feature [+stress] may be [1 stress] (the syllable with primary stress),
[2 stress] (the syllable with secondary stress), and so on.
. The distinction between stressed and unstressed syllables in words
like cattle is easy to hear. In multisyllabic words, three levels of
stress often may be perceived. For instance, the word California has
primary (or greatest) stress on the third syllable, secondary stress on
the first syllable, and no stress on the remaining two syllables. The
word California thus may be written phonetically as [kzlofornyoa].
Notice that in some noun and verb forms of identically spelled
words — for example, the noun cénvict and the verb convict—
stress reflects a parts-of-speech categorization. This fact shows that
stress is not independent from higher-level grammatical structure. But
stress in fact is a continuum, and its division into discrete levels
beyond primary stress is, to an extent, arbitrary. As a result, levels of
stress sometimes may be difficult to perceive.

INTONATION
The other major suprasegmental is intonation, which in English is de-
scribed as a rising or falling pitch over_a group of words. Intonation
alone does not normally distinguish between any two words in En-
glish; however, it does serve to distinguish among sentences, as in the
following examples:

Mary had soup for lunch.

Mary had soup for lunch.
Mary had soup for lunch?
bh
—
WN Mary had soup for lunch?

The first sentence indicates that Mary had soup, rather than some
other food, for lunch. The second sentence indicates that Mary had
70 Phonetics

soup for lunch, and not some other meal. The third and fourth sen-
tences are the corresponding questions. It is important to note that
intonation, like stress, exists in varying degrees; it is not as simple
and uniform as the notation used in these examples suggests.

Summary
Speech sounds are formed by modifying the shape of the mouth to im-
pose characteristic modulations on the flow of air from the lungs. Far
more sounds exist in English than the English alphabet suggests; thus,
a system of phonetic symbols is required for accurate transcription.
Perhaps the most important division among sounds is that between
consonants and vowels. Consonants are defined as sounds produced
with constrictions or occlusions in the oral cavity. Consonants are
further classified according to the degree and location of constriction,
whether the nasal cavity is open, and the presence or absence of voic-
ing. Vowels are defined as sounds produced through an open oral cav-
ity. They may be classified according to the height and the fronted-
ness or backness of the tongue.
Phonetic transcription makes possibie the careful analysis of the
sounds of English, as well as any other language, and represents the
first step in any linguistic analysis. The next step is the organization
of phonetic material into a system of distinctive features. A feature is
a binary attribute that is used to describe and differentiate the sounds
of a language.
The two major suprasegmentals in English are stress and intona-
tion. Stress applies to individual syllables, whereas intonation applies
over groups of words.

SOURCES FOR FURTHER READING

Abercrombie, David. Elements of General Phonetics. Edinburgh:

Edinburgh University Press, 1967.
Catford, John C. Fundamental Problems in Phonetics. Bloomington,
Indiana: Indiana University Press, 1977.
International Phonetic Association. Principles of the International
Phonetic Alphabet. Rev. ed. London: International Phonetic As-
sociation, 1949,
Jakobson, Roman; Fant, Gunnar; and Halle, Morris. Preliminaries to
Speech Analysis. Cambridge, Mass.: MIT Press, 1963.
Ladefoged, Peter. A Course in Phonetics. New York: Harcourt
Brace Jovanovich, 1975.
. Preliminaries to Linguistic Phonetics. Chicago: University of
Chicago Press, 1971.
 Summary 71

Pike, Kenneth. Phonetics. London: Oxford University Press, 1943.

Singh, Sadand, and Singh, Kala S. Phonetics: Principles and
Practices. Baltimore: University Park Press, 1976.
Smalley, William A. Manual of Articulatory Phonetics. Rev. ed.
Committee on Missionary Personnel, Division of Foreign Missions,
NCCC, 1968.
Walsh, J. M., and Walsh, A. K. Plain English Handbook. Cincinnati:
McCormick-Mathers, 1972.

EXERCISES

1. Prepare to say the word ski and note the position of your vocal ap-
paratus. Next, do the same for the word school. What is the dif-
ference between the initial positions of your vocal apparatus for
these two words? To what would you attribute this difference?
What theoretical issue in phonetics is raised by the results of this
experiment?
2. Can the English alphabet be considered a phonetic alphabet?
3. Provide a phonetic transcription for the following words: (a)
though, (b) boxes, (c) twinkle, (d) thistle, (e) few, (f) jaw, (g) sea-
son, (h) cheese, (i) string, (j) should, (k) catastrophic.
4. Indicate the pronunciation of the following words if the speaker
pronounces them in the indicated manner.
a. all (vowel lowered and backed)
b. J (pronounced with no glide but with the vowel considerably
lengthened)
5. Why is it possible for a vowel to be [—high] and [—low], but not
{+high] and [+low]?
6. Complete the sentences by choosing one of the following features:
[+continuant], [+sonorant], [+anterior], [—coronal], [—conso-
nantal] £ & 4x
a. The interdental, alveolar, and palatal fricatives are all
Cowl NU Cit Fae
b. All labiodentals and bilabials are Ganterieft
c. The glides, nasals, and liquids are all’ = Sonera 0A/
d. [h], [w], and [y] are the only consonants that are aConsonota |
~
}
‘€.)All velar consonants are Serituart (“corome Vi
7. In what fundamental way do suprasegmentals differ from vowels
and consonants?
8. Describe phonetically the difference between the sounds rep-
resented by o in the words convict and convict.
9. Indicate the degree of stress on each syllable in the following
words: (a) resign (b) agonizing (c) petunia (d) present (as a noun)
(e) Rosy aie hy 3 Sy,
residYW Sioa >. | ]
pein’ a a C
{ ned
 CHAPTER

PHONOLOGY I

Out of all the sounds that we are physically capable of producing,

only some sounds are used in human languages, and these sounds
occur not randomly but in a limited set of patterns. Furthermore,
speakers of a language recognize as the same sound utterances that in
fact vary acoustically. What makes sounds the same, and what are
the patterns of sounds that occur in speech?
These are the kinds of questions that phonology addresses.
Whereas phonetics isthe study of sounds that occur in language,
phonology isis the study. ofhow these sounds are organized and how
they functionin language. It uses the classifications of sounds derived.
from phonetics to describe and analyze how sounds occur in speech.
According to both structuralism and _ transformational-generative
grammar, in order to explain how we understand and produce mean-
ingful sounds, we must posit the existence of abstracts units under-
lying the actual sounds of speech. These abstracts units and their
relationships to actual sounds are the main topics of this chapter.

Structuralist Phonemics

As linguists began to study sounds in fine detail, they recognized in-

creasingly complex aspects of phonetic organization. For example,
the sound we have indicated as [p] really comes in different

Wa
 Structuralist Phonemics 73

varieties in English. One of these types, indicated by {p"], occurs in

words such as pill and peace. This p-sound is produced with an ac-
companying puff of air called aspiration. In pronouncing another
major p-sound, indicated by [p:], there is little or no aspiration; this
sound occurs in a word like spill. A third major p-sound, indicated by
[p'], is the unreleased p which may occur at the end of a word like
stop. To deal with these p-sounds, the structuralists suggested the
existence of an abstract p, which they termed the phoneme /p/. (In
referring to phonemes, slashes are used.) phoneme
A was
definedby
the_structuralists as an abstract phonological unit that represents a
class_of real sounds, termed the allophones of a ‘phoneme. The
phoneme /p/ in English, then, is represented by the allophones [p"],
[p-], and [p~], representing the p-sounds in pill, spill, and sometimes
stop, if one does not release the p at the end of the word.

MINIMAL PAIRS AND PHONEMES

How do we know what these abstract units of sound called phonemes
are? In order to find the phonemes of a language, the structuralists de-
veloped the concept of the minimal pair. A minimal pair is any two
words that (1) contain the same number of segments, (2) differ in
meaning, and (3) exhibit only one phonetic difference. In such cases,
this one phonetic difference indicates that two separate phonemes are
involved. Thus, in practical terms, phonemes distinguish meanings;
and a phoneme can also be defined as the smallest meaning-
distinguishing unit of sound. For instance, the words pin and bin
mean different things, and the only difference in these words occurs
in the initial sounds. Thus, this minimal pair demonstrates the exis-
tence of the two phonemes /p/ and /b/. The pair of words pin and ban,
on the other hand, differs in more than one sound, and, therefore,
does not represent a minimal pair.
By using the concept of a minimal pair, we can determine that the
three p-sounds do not represent three phonemes. Certainly, it is pos-
sible to pronounce the word cap with either an aspirated or
unreleased p, [p"] or [p~]; however, the two forms [kzp"] and [keep]
are not a minimal pair, even though they involve different sounds,
because they are identical in meaning. p-sounds
These two are
are tl
there-
fore
said to exhibit free variation; that is, the pronunciation may vary
without_ signifi
signifying a change in meaning, One concludes that ‘the
unreleased p and the ‘aspirated p are not representations of different
phonemes in English but are, in fact, allophones of one phoneme, /p/.
Every phoneme has at least one allophone;some _have more, Jn_
English, /f/and /s/ are‘examples of phonemes having only one “allo-
phone. When fF
phonemes have more than one allophone in a language,
74 Phonology I

the allophones are said to be in complementary distribution. Comple-

mentary distribution means that the allophones of a particular phoneme
occur in different phonetic environments (that is, with different
sounds surrounding them). In other words, the phonetic environment
determines
which of the allophones is used in pronouncing a word.
For example, the three allophones of /p/—[p"], [p'], and [p” }-are in
complementary distribution. The [p"] occurs before stressed vowels,
as in pat or pot; a [p~] may occur at the end of a word, as inhip; and a
[p:] occurs in all other environments, as in spot, prom, and potato.
The fact of their complementary distribution means that these al-
lophones do not occur in minimal pairs in English.

MORPHOPHONEMES
By the structuralists’ definition of phonemes, differences in phonemes
should signify differences in meaning. But they faced a problem in
trying to explain cases of phonemic alternation, in which words with
related meanings have particular corresponding segments that differ.
The words foot and feet, for instance, are obviously related in mean-
ing, yet they differ in their phonemic representations. This relatedness
could not be expressed within the confines of phonemic theory.
If foot and feet were isolated examples of this kind of alternation,
they probably would have been cited as exceptions. But major classes
of regularly alternating sounds exist in English. For example, the al-
ternation in the vowel sounds of the related words sane and sanity
also occurs in such pairs as vain and vanity, profane and profanity,
and others. Such regularly occurring alternations of phonemes led the
structuralists to postulate an even more abstract, higher-level unit
called the morphophoneme. A morphophoneme, then, was a unit that
represented alternating phonemes. Other common alternating pho-
nemes occur in took/take, break/broke, sing/sang, and throw/threw.
The structuralists, however, were uncomfortable with the concept
of the morphophoneme. It raised several troublesome issues. First,
morphophonemes
seemed to represent only rare exceptions. No rules
could explain the possible differences between unique morphopho-
nemes (such as the one in the alternating vowel of man and men) and
the morphophonemes that occurred with some frequency (such as the
/e/ and /e/ in the sane/sanity alternation). Second, even with the more
frequently occurring alternations such as vain/vanity, the structur-
alists were unable to suggest a reason for the particular phonemes
that made up the alternating pair. Why, for example, should /e/ and /z/
alternate? What relates them phonemically to each other? The struc-
turalists never resolved these questions.
 Structuralist vs. Transformational-Generative Phonology 75

Structuralist Phonology versus Transformational-Generative

Phonology

Transformational-generative phonology is a relatively recent develop-

ment in linguistic theory. Chomsky launched transformational-genera-
tive grammar in 1957, but the earliest studies within this framework
were largely concerned with syntax. A decade passed before the first
comprehensive transformational-generative treatment of English pho-
nology appeared: Chomsky and Halle’s The Sound Pattern of En-
glish (1968). Since then, the theoretical framework that Chomsky and
Halle provided has endured, although many specific details have been
challenged and modified.
The transformational-generative phonologists made sweeping revi-
sions of the structural phonologists’ carefully constructed phonemic
theory, saying that it failed to provide an adequate understanding of
_ phonological processes. In principle, the idea of a phoneme was per-
fectly well founded; but in practice, the structuralist phoneme was in-
adequate. At root, it was just not abstract enough.
Consider the words sane and sanity. A phonemic rendering gives
/sen/ and /szniti/; and according to the principles of phonemic theory,
the different phonemes /e/ and /z/ are not related. The transformational-
generative phonologists claim that any speaker of English can per-
ceive the obvious fact that the words sane and sanity derive from the
same basic source as far as meaning is concerned. They further claim
that the large number of such /e/-/z/ pairs and the common spelling
(both sane and sanity are spelled with a) lead necessarily to the con-
clusion that there is some single abstract entity that represents the
vowel in both words. This abstract entity that accounts for both the
[e] in sane and the [z] in sanity is called an underlying representation.
Like the,structuralists’ phoneme, it is indicated in slash marks.
If the iiderlying representation were simply to ‘‘represent,’’ some-
how, the various sounds that it was assumed to underlie, it would be
no improvement at all over the structuralist morphophoneme. One
can easily suggest the existence of a morphophoneme, represented
perhaps by /A/, that is realized at the phonemic level by /e/ and /z/. But,
in fact, the transformational-generative phonologists go a great deal
further than this: they claim that there is a dynamic relationship be-
tween an underlying representation and the sounds it underlies. Spe-
cifically, they claim that sounds derive from underlying represen-
tations through a specific, describable, and coherent system of rules.
Many words that do not enter into such alternations as sane/sanity
also will have underlying representations that differ from their pho-
netic representations. In fact, according to Chomsky and Halle, all
76 Phonology I

TABLE 6-1
Contrast of Two Phonological Theories

Structuralist Phonology Transformational—Generative Phonology

Morphophonemes Underlying Representations

J
Phonemes Rule A
¢ Rule B
Phonetic Representations Rule C
etc.

Surface Phonetic Representations

words have stress indicated only in their phonetic_representations;

underlying representations are given without stress, which is pre-
dictedby aset of rules (see pages 86-88). Shas
The basic difference between the structuralists’ and the transforma-
tional-generative grammarians’ theories of phonology is outlined in
Table 6-1. Structuralist phonology recognized three major indepen-
dent phonological levels: morphophonemes, phonemes, and phonetic
representations. The phonetic representations in the structuralist and
transformational-generative theories are somewhat similar, and the
structuralist morphophoneme was a rough approximation of the trans-
formational-generative underlying representation. However, one ma-
jor difference is that in transformational-generative phonology the
underlying representations and the surface phonetic representations
are expressed in terms of the binary distinctive features discussed in
Chapter 5; these were not originally available to the structuralists.
Most importantly, the transformational-generative phonologists
strongly oppose the structuralists’ phonemic level. They replace this
level by a series of rules that directly relate underlying represen-
tations to observed phonetic representations. The central mechanisms
in transformational-generative phonology, then, are underlying rep-
resentations and phonological rules, which are discussed in the fol-
lowing sections. Transformational-generative phonology will be the
concern of the remainder of this chapter.

Phonological Rules
The idea of a rule is central to transformational-generative theory. A
rule is an operational statement in which some linguistic entity is
modified, resulting in a new linguistic entity. Rules may add elements,
remove elements, or change elements. They are theoretical state-
 Phonological Rules ae

ments on the part of the linguist, who is attempting to demonstrate

that there is order in linguistic phenomena, that linguistic patterns are
systematic. There is little proof that any specific rule actually exists in
the mind of the speaker. Instead, a system of rules may be a linguist’s
‘ hypothesis about mental operations. In any case, such a system is an
attempt to account for the abilities that speakers demonstrate.
Phonological rules may add, delete, or change phonological ele-
ments. A phonological derivation is an operation that begins with an

specific rules, yields the actual sound the speaker produces.

The representation of a phonological rule has the general appear-
ance

A> B/C
\
The A in this rule represents that which is going to be changed by the
rule. Keep in mind that a segment in both underlying and phonetic
representations is actually a bundle of distinctive features such as
[+voice], [—anterior], [—coronal]; thus, a symbol like A may repre-
sent a single phonological feature or a set of phonological features.
The arrow in the rule means ‘‘changes to.’’ The B represents the new
form. The slash mark separates the rule ‘‘A changes to B’’ from the
phonological environments or conditions under which the rule oper-
ates, here symbolized by C. This rule may thus be read: ‘‘A changes
to B under condition C.”’
Conditions are normally written before or after a horizontal line;
that is,

A—B/__ C
or =
x /

A— B/C ___

This line indicates whether the segment being changed occurs before
or after the condition C. If no C appears, the rule is understood to
operate under all conditions.
Suppose one wishes to express the fact that the word fen is pro-
nounced like the word tin in most Southern dialects. Such a fact may
be represented by the statement ‘‘ten = tin in most Southern dia-
lects.’’ However, this is not an isolated fact; for den is pronounced
like din, Ben like bin, and so on. Actually, in any case in which the
vowel /e/ precedes /n/, it is pronounced [1]. This is a general fact that
holds true even when Southerners produce words containing the com-
bination /e/+/n/ that they have never uttered before. This very general
fact can be represented by the phonological rule
78 Phonology I

/e/— [1] ——/n/

which may be read: ‘‘/e/ changes to [1] when /e/ precedes /n/.”’
The statement of phonological rules can be extremely complex, and
linguists have developed several notational devices for writing them.
Often, the following symbols will be necessary for stating the condi-
tions under which rules apply:
# indicates a word boundary
+ indicates an intraword boundary
$ indicates a syllable boundary
Other frequently used notational devices are parentheses and
braces. Parentheses indicate optionality. Thus, the rule
A B/——(C)D
means ‘“‘A changes to B before D, and A changes to B before CD’’; C
is an optional condition. This notation allows the combination of the
equivalent two rules
A— B/——D
A— B/——CD
The braces notation indicates that a rule applies under two or
more conditions. To combine the two rules
A=> B/C ee 2
Ass Bi 28s

+0 /|—
we write

which may be read: ‘‘A changes to B following C, and A changes to B

following D.”’
The combination of rules through the use of these notational de-
vices is subject to certain restrictions, but a full discussion of the
restrictions would take us too far afield. For the present, an un-
derstanding of the purpose of these devices is sufficient.

Underlying Representations and Related Issues

The transformational-generative description of phonology relates un-
derlying representations to phonetic representations by rules. We can
illustrate this process through a relatively simple example. In English
there are certain pairs of words like sign/signature and malign/malig-
nant that exhibit a regular alternation in their phonetic represen-
tations: [g] is present in the second member of the pairs but absent in
 Underlying Representations and Related Issues 79

the first member. To express the relatedness of words such as sign

and signature, we could claim that the underlying representation of
the segment in all such pairs is /g/ and that a rule operates to delete /g/
before syllable-final nasals. But how do we find the underlying repre-
“sentation of a form?
No clear-cut procedure exists for determining underlying represen-
tations, but certain considerations may help us to choose among
proposed forms. In seeking underlying representations, we prefer an
analysis that is plausible, simple, and consistent with other phonologi-
cal patterns in the language. Of course, the analysis should allow the
phonetic representations to be predictable from the underlying repre-
sentation, and not to be merely an ad hoc solution.
For example, take the case of the alternation in sign/signature.
Why should we choose /g/ rather than @ (that is, ‘‘nothing”’ or ‘‘no /g/’’)
as an underlying segment? If @ were chosen as an underlying seg-
ment, we would need a rule that inserts /g/ before syllable-final nasals
in order to yield the /g/ in words like signature. But how would we then
account for the fact that /g/ is not inserted in many words like linear
(related to line)? In other words, there would be no way to predict
which words have an inserted /g/. Moreover, since the word sign is
spelled with a g, we would be inclined to choose /g/, rather than @, as
an underlying segment. Finally, in the languages of the world, there
are cases in which /g/ is deleted before nasals in various circumstances,
but it is uncommon in the world’s languages for /g/ to be inserted
before nasals. Thus, an analysis that proposes @ as an underlying
segment in sign/signature is implausible, counterintuitive, and simply
wrong.
Actually, it is conceivable that some segment other than /g/ or @
might be the underlying representation, although in this case such an
answer would have no justification. In the past, several transfor-
mational-generative grammarians have argued for underlying rep-
resentations that do not appear in any of the surface forms of an
alternation. Indeed, this was often the approach of Chomsky and Halle
in The Sound Pattern of English. Nevertheless, a more recent trend is
to keep underlying representations as close to the phonetic represen-
tations as possible.
The rule ‘‘/g/ is deleted before syllable-final nasal’? would appear for-
mally as:
+voice
—anterior |-@ ___ [+ nasal]$
—coronal

On the left-hand side of the arrow are the features needed to uniquely
specify g among the consonants; that is, no other consonant has the
features [+voice], [—anterior], and [—coronal]. The symbols —@
 80 Phonology I

mean that ‘‘/g/ changes to nothing’’ or, more properly, ‘‘/g/ is deleted.”’
The horizontal line following the slash mark refers to the position of
g; namely, before a segment that is [+nasal]. Finally, this [+nasal]
segment occurs before a syllable boundary, as indicated by $. A less
formal way of writing this rule would be:
/g/— ©/___[+ nasal] $
Notice that this rule also helps describe such alternations as
phlegm/phlegmatic and paradigm/paradigmatic.
Consider now another example. In English, there is a prefix mean-
ing ‘‘not’’ that is added to words. This prefix alternates among the
forms [1m-], [!n-], and [m-], depending on the point of articulation of
the initial segment of the following word. If the segment begins in the
extreme front part of the mouth (labials), the form is [Im-], as in im-
proper, if it begins in the extreme back part of the mouth (velars), the
forms is [1n-], as in incomplete; and if it begins in the mid-region of the
mouth (all other sounds), the form is [In-], as in indecent.! We would
like to express the phonetic and semantic relatedness of [1m-], [1n-],
and [In-] by suggesting a single underlying representation for these
three forms.
In this case, one of the main facts governing our choice of an un-
derlying representation will be economy of description. Note that
when words beginning with vowels take the prefix meaning ‘‘not,”’
the form of the prefix is [In-], as in inoperative. If we select /In-/ as the
underlying representation, the segment /n/will be changed only before
certain consonants; thus, no rules need apply when /In-/ precedes a
word beginning with a vowel or a nonlabial or nonvelar consonant.
An economical way to state the rule describing these facts is to use
a notation that includes Greek letters. According to this notation, a
and B can mean either ‘‘+’’ or ‘‘—,’’ and —a and —6 mean the op-
posite of the original feature specification. The rule appears as fol-
lows:
ae ees i mart / (5 ae
\
B coronal B coronal
~
\ Suppose we wished to apply this rule to the underlying representation
in +proper, where + is an intraword boundary symbol. Here, the ini-
) : tial segment of proper is /p/, which is [+anterior] and [—coronal].
Thus, we let a mean ‘‘+’’ and B mean ‘‘—”’ in this case. Then the
| rule can be read: ‘“‘If [+ nasal] is followed by a segment that is [+ an-
|

\
terior] and [—coronal], then [+nasal] changes to become [ + anterior]
and [—coronal] also; these features specify [m]. Similarly, if the word

1 Actually, words beginning with [r] or [I] will take the forms [Ir-] and [il-]. This fact can
be captured in a more general version of the rule.
 Underlying Representations and Related Issues 81
a

was complete, the initial segment /k/ would be [—anterior] and >
[—coronal], and the nasal /n/ would change its point of articulation (
also to become [-—anterior] and [-—coronal], or [pn]. Notice that \
this rule also ensures the point-of-articulation agreement in other \
words; for example, the prefix con- will be affected in the same way
(com +plete, con +dense, and con +crete). Of course, few rules are
perfect, and there are exceptions, such as comfort. r

DEALING WITH EXCEPTIONS

Still another example of an alternation in English is found in pairs of
words like electric and electricity, in which the segments [k] and [s]
alternate. The two obvious choices for an underlying representation
for this segment are /k/ and /s/. Choosing /k/ as more basic, one would
need to determine the conditions under which /k/ changes to [s]. For
example, [s] appears in electricity, but [k] appears in electrical. Fur-
ther examination of such alternations implies the velar softening rule:
/k/ changes to [s] only before nonlow
onlow front vowels,
vows which may be writ-
ten informally as:
V
/ki—[s] / —— ketow |;
(back

where V indicates ‘‘vowel.”’

The velar softening rule is notorious for having a large class of ex-
ceptions; for example, the /k/ in such words as kiss and kennel does not
change to [s]. In order to account for these exceptions, Chomsky and
Halle assumed that there are two different underlying /k/s in En-
glish. These two /k/s are distinguished from each other by a new fea-
ture, [+derivable]. Those /k/s that undergo the velar softening rule are
{[+derivable] and those that do not are [—derivable]. Although
this solution accounts for the data, it is an arbitrary decision to treat
[+derivable] as a feature, since it has no intrinsic phonetic content
and serves a limited purpose.
When a rule has only a few exceptions, a different approach is
taken. For example, many pairs of words like serene/serenity and
obscene/obscenity exhibit a regular alternation between [i] and [e]. In
the second member of each pair, a rule called the trisyllabic laxing
rule has operated to lax a vowel before two syllables (that is, to
change it to a vowel that is not tensed). However, this rule does not
apply to the word obesity; both obese and obesity contain the seg-
ment [i]. To account for this fact, one may say that in the lexicon, the
speaker’s dictionary of words, the word obese is designated as an ex-
ception to the trisyllabic syllabic rule. This solution may be general-
ized to account for all exceptions: any time a word does not ‘‘obey’’ a
rule, it is designated specifically not to obey it. It is circular to suggest
82. Phonology I

that an exception is an exception because it is an exception, but for

the present this solution is the preferred one in transformational-
generative phonology.
To return to the velar softening rule, it may be pointed out that
some apparent exceptions are not exceptions at all. Consider the
words critic, criticism, critical, and criticize. In criticism, the velar
softening rule has correctly applied, because the following vowel is a
high front vowel, and thus /k/ changes to [s]. And in critical it has cor-
rectly not applied because the following vowel is a low vowel. How-
ever, the rule has applied in criticize even though the vowel following
/k/ is not a high or mid front vowel; the vowel is the diphthong ay, which
begins with a low vowel. Nevertheless, the rule has applied properly
in this case because the underlying representation of [ay] is assumed
to be /i/, which is a high front vowel. In other words, the velar soften-
ing rule applies before rules that change /i/ to [ay]. This fact illustrates
that phonological rules may be ordered with respect to each other in
their application. If the rules changing /i/ to [ay] were to apply first, the
velar softening rule would not be able to apply because its environ-
mental condition would not be met. In such a case, the ultimate
derived form would be [kritikayz], which is not an English word.
Rule ordering is thus an important device in phonological theory; a
change in the ordering of rules may lead to different results. How-
ever, not all rules need to be ordered. In fact inguists have
noted that rules do not necessarily have_tobe ordered
at all, provided
they are understood to apply simultaneously. Unfortunately, this ap-
proach may make the statement of rules or conditions more complex.
But why was /i/ chosen as the underlying representation for [ay] in
criticize? Although the reasons for this choice are complicated (see
Chomsky and Halle, 1968), it may be noted that at one point in the his-
tory of English the vowel /i/ did, in fact, change to [ay]. This change
was part of the Great Vowel Shift discussed in Chapter 4. Since we
are concerned with synchronic description, historical information
should not serve as the basis for determining underlying represen-
tations; however, we can at least feel confident that /i/—>/ay/ is a plausi-
ble change and that, consequently, /i/ is not a totally unreasonable
choice as an underlying representation.

THE STATUS OF THE CONCEPT

The underlying representation and the system of rules that relates it
to surface phonetics is a powerful tool that is, as yet, not wholly mas-
tered. Yet it accounts rather neatly for the otherwise mysterious abil-
ity that we English speakers have: we can spell most words sensibly
that we hear for the first time, and we can pronounce most words sen-
sibly that we see spelled for the first time.
The transformational-generative phonologists claim that English
 Summary 83

spelling often reflects the language’s underlying representations. They

point out, for example, that if English spelling were really as bad as it
is claimed to be, we would not know how to pronounce a word like
*bomity.” Although the word is novel, almost everyone would pro-
nounce it [bamiti] rather than [bomiti] or [bamayti] or [bomayti], and
so on. The transformational-generative phonologists account for this
near unanimity by saying that, unconsciously, we recognize the under-
lying representation indicated by the letter o in a particular environ-
ment, and we apply phonological rules to obtain the pronunciation [a].
The rules and underlying representations we have been discussing
are not universally accepted by transformational-generative phonol-
ogists. From the moment that The Sound Pattern of English ap-
peared, alternatives have been suggested for both the features of un-
derlying representations and the rules that relate them to surface
sounds. The principle of the underlying representation and the deriva-
tional rule schema, however, has not yet been modified.

Summary
Structural phonologists championed the phoneme, an abstract phono-
logical unit consisting of a class of real sounds called allophones.
Phonemes are determined primarily by investigating minimal pairs
and complementary distributions.
Because of the existence of patterned phonemic alternations, trans-
formational-generative phonologists established a more abstract level
of phonology that consists of underlying representations, thus
challenging the structuralists’ phonemic level. The underlying repre-
sentations are related to phonetic representations by a system of rules
that follow the form A— B/C. These rules may add, delete, or change
phonological elements. Transformational-generative phonology is a
universal theory of phonology that is applicable to individual lan-
guages. It attempts to characterize what speakers of a language un-
consciously know about the language’s phonological structure and
processes.

SOURCES FOR FURTHER READING

STRUCTURAL PHONOLOGY
Bloomfield, Leonard. Language. New York: Holt, Rinehart and Win-
ston, 1933.
Makkai, Valeria B., ed. Phonological Theory: Evolution and Current
Practice. New York: Holt, Rinehart and Winston, 1972.

2 An asterisk indicates an ungrammatical form.

84 Phonology I

Trubetzkoy, N. S. Principles of Phonology. Berkeley: University of

California Press, 1969.

GENERATIVE PHONOLOGY
Chomsky, Noam and Halle, Morris. The Sound Pattern of English.
New York: Harper and Row, 1968.
Harms, Robert. Introduction to Phonological Theory. Englewood
Cliffs, New Jersey: Prentice-Hall, 1968.
Hyman, Larry M. Phonology: Theory and Analysis. New York: Holt,
Rinehart and Winston, 1975.
Schane, Sanford. Generative Phonology. Englewood Cliffs, New Jer-
sey: Prentice-Hall, 1973.

EXERCISES

1. Describe briefly the difference between phonetics and oe a

2. Which of the following pairs of words are minimal pairs? (a) r ipIrib
(b)- thinithan> (c) wiitelright (d) batibats (e) salacious/sagacious
(f) back/bag ye, — S
3. What functions do underlying representations serve?
4. Some linguists transcribe rhythm and heaven as [r16m] and [hevn],
while others transcribe them as [ridom] and [hevon]. In the first
transcription, where m and n are treated as_syllabic, would the
linguist be justified in treating [m] and [n] as separate phonemes
from [m] and [n]? The latter pair, for example, would appear in
words like mother [mador] and nor [nor].
5. ‘‘Silent’’ letters, such as the ‘‘silent e’’ in reduce, are not always
completely silent. Many are silent only under certain circum-
stances and are pronounced under others. For example, the silent
g in sign is pronounced in signature.
a. Give an example of other pairs of related words in which a con-
sonant letter is silent in one word and pronounced in the other.
b. As a general principle, how do you account for their silence?
6. State in your own words the effect of the following hypothetical
rule:

B
AS Gy a= (D)E #

7. Write rules that express the following:

a. Consonants become voiced between vowels. (You may use the
symbol C to represent consonants and V to indicate vowels.)
b. A voiced consonant following m at the end of a syllable is de-
leted. (You need not write m in distinctive features.) —
 CHAPTER

PHONOLOGY II

Introduction

Now that we have covered the fundamentals of phonological theory,

we will introduce some additional concepts that lead to a more com-
plete understanding of phonology. Four important topics we will
study in this chapter are stress rules, redundancy, markedness, and
natural phonology.
Chapter 6 showed how the transformational-generativists related
underlying representations to phonetic representations by a system of
ruies. These grammarians also showed that the stress pattern of a lan-
guage is determined by a system of underlying rules. We will begin
this chapter by looking at these stress rules.
A discussion of redundancy and markedness will further expand
our understanding of phonology. The fact that language frequently
competes with other noises in the environment and yet com-
munication between individuals still occurs seems obvious in our
world of many sounds. We are able to comprehend a message even if
we do not hear every sound. This ability is associated with redun-
dancy in language. Another concept frequently employed by gram-
marians in phonological theory and analysis is markedness. We will
discuss how markedness designates what is normal, simple, expected,
and plausible and what is not.
Finally, in recent years the natural phonologists have presented a
view differing from that of the transformational-generative phonol-
85
86 Phonology II

ogists. These differences, which we will examine at the close of this

chapter, have inspired a lively debate among linguists.

Stress in Transformational-Generative Phonology

One of the earliest and most interesting claims of transformational-
generative phonology was that stress appears only in phonetic repre-
sentations, not in underlying representations; it is determined by a
series of rules. These rules operate on words that are grouped into
phrases and provided with a parts-of-speech classification. For ex-,
ample, suppose one wished to determine the stress contours for the
phrase John’s baseball cap, from a sentence such as He was wearing
John’s baseball cap. This phrase is assumed to be segmented, or
bracketed, according to the following parts-of-speech classification:

[vplp John’s]p[nInly base]y[y ball]yIniy Cap]yindne-

NP denotes a noun phrase; N indicates a noun; and D a determiner,

which is essentially a modifier. These labels fall just inside the open-
ing bracket/and just outside the closing bracket of each unit (for ex-
ample, [yp. . - ]np). Thus, this notation indicates that base, ball and
cap are nouns (N), and John’s is a determiner (D) that modifies a
noun. Furthermore, baseball is a noun, John’s baseball cap is a noun
phrase (NP), and so on.
In this representation all vowels are [—stress] originally. A series
of rules then applies to assign stress. The rules apply first to the
smallest constituents of the bracketed expression. In this example,
the smallest constituents are the words John’s, base, ball, and cap.
After all the relevant rules have applied, the rules may reapply to the
next larger constituents (for example, baseball), and so on until the
rules apply to the entire bracketed expression, which in this example
is the noun phrase John’s baseball cap. The reapplication of the rules
to progressively larger constituents is performed in cycles. Thus, the
stress rules are cyclic rules, and the principle by which they apply is
called the transformational cycle. The fact that these rules determine
stress for larger and larger constituents means that they are genera-
tive. That is, these rules generate the stress patterns for what are ul-
timately sentences. These stress patterns are part of the system of
phonological rules that determine the particular sound patterns for
each language. As with the rules discussed in the previous chapter,
these rules attempt to explain what a native speaker does naturally. —
The stress pattern characterizes the knowledge of the sound system
for a particular language that an individual has internalized in learning
the language. While one might not apply these rules in the production
 Stress in Transformational-Generative Phonology 87

of actual speech patterns, they are useful in describing the generalities

that hold within a language.
Let us look more closely at the rules involved in deriving the stress
pattern for John’s baseball cap, which English speakers pronounce
with the stress pattern

2 ey die 3
John’s baseball cap
The first rule to apply is called the stres
main rule,
sand it may be
given as
V—[I stress] f Xe Calaky
This rule states that a vowel (V) receives primary stress ({1 stress])
when three conditions are met:

1. It appears in a noun, adjective, or verb (NAV).

2. It is followed by zero or more consonants (C AR
3. The preceding element may not contain an internal word boundary
symbol. This condition is represented simply by ‘‘X.”’
The effect of the main stress rule in John’s baseball cap is as follows:
[xelp John’s ]p[nIyln base]y[y ball InInby cap Iniylne
a + Slee Cycle 1
Thus, primary stress is placed on the only vowel in John’s, base, ball,
and cap, since all three conditions of the rule are met in each case.
No other rules apply on the first cycle.
On the second cycle, a rule called the compound stress rule applies:
eT

1 ed
st [1 stress) / ey? INAV
According to this rule, any primary stressed vowel that appears to the
left of a primary stressed vowel within a noun, adjective, or verb con-
tinues to bear primary stress. By a general convention associated with
this rule, the stress values of all other vowels within the relevant
noun, adjective, or verb are reduced by one. The effect of this rule on
the noun baseball is that primary stress remains on base and the
stress on ball is decreased:
[velp John’s ]p[xiyly base]y[y ball JyInty CaP] Inne
1 1 nie Alas Cycle 1
1 Eon Cycle 2
On the third cycle, the next largest constituent, baseball cap, is
subjected to the reapplication of the compound stress rule, giving
88 Phonology II

[nplp John’s |p[ylyly basely[y ball JyIniy Caply

iyIne
1 1 tis mi Cycle 1
1 2% 0 Cycle 2
1 po 2s. Cycle 3
That is, the primary stressed vowel to the left (‘“base’’) remains the
primary stressed vowel and the other two are reduced by one each.
On the fourth cycle, rules apply to the maximal phrase John’s base-
ball cap. The relevant rule in this case is the nuclear stress rule:
1 stress 1
V jou stress} /\ yl ge mere ee)|

This rule maintains primary stress on all primary stressed vowels to the
right of primary stressed vowels. As with the compound stress rule,
all other stresses within the domain of the rule are reduced by one.
The nuclear stress rule assigns primary stress to base, and the final
result is

Inplp John’s |p[ninly base}y[y ball InInby Caplydnine

1 oe reas e.Elim cis Cycle 1!
tad a B02: we Cycle 2
errno Cycle 3
2 zie 4- ELE Cycle 4
That is, the rule assigns primary stress to base, which is to the right
of the primary stressed vowel John’s; all others are then reduced by
one. Thus, in the expression John’s baseball cap, the constituent
base bears primary stress, John’s bears secondary stress, and so on.
In considering longer, more complicated phrases, even weaker de-
grees of stress (5, 6, 7, and so on) may be predicted. Although every
speaker has some limit on the degrees of stress perceived, the predic-
tions made by the rules agree with the perceptions of nearly all
speakers, at least for the major levels of stress. The principle of the
transformational cycle and the assignment of stress by generative
rules that apply to labeled constituents were major innovations of
transformational-generative phonology, and they further challenged
the claims of structuralism.

Redundancy
Communication may take place in a subway, over the telephone, at a
rock concert, while people are eating, and even when one of the par-
ties has bad hearing. Redundancy is a property of language that assists
communication even under difficult circumstances; it means that
 ‘Redundancy 89

speech contains more information than is absolutely necessary for

communication under normal circumstances. For example, if John
calls to Fred who is walking out the door, ‘‘Don’t forget to take your
pill,’’ what prevents pill from being heard as bill? If Fred is sick and
all his bills have been paid, he will probably interpret the message
correctly. But when the context does not provide the solution, the
burden of communication falls on the phonetic properties of language.
Thus, if we consider the segments /b/ and /p/, we know that [b] is pro-
duced with a vibrating of the vocal cords, while [p] is not. We also
know from phonetic study that [p] is [+ tense] — that is, the muscles
are tensed in producing it — whereas [b] is not. But although either
voicing or tensing may be invoked as a distinctive feature that cap-
tures the distinction between /p/ and /b/, there is evidence that it is the
aspiration present in initial /p/, as opposed to initial /b/, that is most eas-
ily perceived as a distinguishing characteristic. While we would not
want to claim aspiration as a distinctive feature in English, neverthe-
less, aspiration is valuable for communication. It provides redundant
information for distinguishing /p/ and /b/.
Redundancy statements are statements in which certain information
is predictable from other information. Consider the segment /1/. The
fact that /1/ is specified as a vowel that is [+high] means that /1/ must
automatically be [—low], and the fact that /1/ is [—back] means that it
is automatically [—round]. These redundancy statements may be ex-
pressed as follows:
1. [+high] 2. [—back]
J
[—low] [—round]
Note that redundancy statements may vary from language to lan-
guage. Statement 1 holds for all languages, because it is a physical
and logical impossibility for vowels to be high and low simulta-
neously. On the other hand, statement 2 applies to English, but not to
all languages. In English, all nonback vowels are produced without a
rounding of the lips; but certain other languages, like German, have
nonback vowels that are produced with a rounding of the lips.
The_value of redundancy in communication is clear, but redun-
dancy is also useful in phonological theory also because it permits
a more efficient statement of segments in rules. Thus, if one wished
to refer to the segment /1/, it could be uniquely specified by the four
features [+syllabic], [+high], [—back], and [—tense]. The values
for the features [consonantal], [sonorant], [low], and [round] are
then predictable. The only requirement is that each redundancy-free
segment, that is, a segment containing only the essential features from °
90 Phonology I

which others can be predicted, remain completely distinguishable

from all other segments.
The concept of redundancy statements is applied to sequences of
segments by discovering which sequences of segments or words are
permissible within a given language. The grammar of every language
contains a lexicon, which may be defined for now as a dictionary or list
of the meaningful elements of a language which each speaker has in-
ternalized and stored in his or her brain. These meaningful elements,
called morphemes, include words like dog and wheel, affixes like pre-
and -able, and markers like -s (indicating plurality, possession, or
third person singular). One of the many functions of the lexicon is to
provide an underlying phonological representation for each meaning-
ful element.
But only certain combinations of segments may be used in forming
possible morphemes in a language. Thus, the speaker of English rec-
ognizes splunge as a possible English word, although it has no meaning
in present-day English. However, the combination */rtsim would
surely be rejected by the speaker of English as a possible English
word. For one thing, we are restricted in the number of initial consec-
utive consonants that we can pronounce. This is not necessarily a
physical restriction; languages differ in the number of initial consecu-
tive consonants that they will permit in a morpheme, although it is
rare for a language to have words with more than three. English per-
mits three morpheme-initial consecutive consonants, but not just any
three. We thus have the following two different kinds of redundancy
statements in English (that is, the information in one statement is pre-
dictable from the information in the other):

1. Each English morpheme may have no more than three initial con-
secutive consonants.
2. If any English morpheme begins with three consecutive conso-
nants (excluding /w/ and /y/, which are not true consonants), the first
must be /s/, the second must be a voiceless stop, and the third must
be /I/ or /r/.

The first statement may be expressed in feature notation as follows:

1. + ([—syllabic]) ({ —syllabic]) ({ — syllabic]) ({ + syllabic])

The symbol + at the beginning of the statement indicates that the fol-
lowing segments begin a morpheme. Recall that a true consonant is
[ — syllabic] and that parentheses indicate optionality. Thus, the nota-
tion says that an English morpheme may begin with 0, 1, 2, or 3 con-
secutive consonants. The second statement is expressed in feature
notation in the following way:
 Markedness and Natural Phonology 91

2. If: + [—Syllabic] [—Syllabic] [—Syllabic]

ae takhiel
Then: +Strident —Voiced +Sonorant
+ Anterior —Continuant +Consonantal
+Coronal —Strident —Nasal
— Voiced

Again, these redundancy statements are expressed in terms of fea-

tures since this is the most convenient notation for displaying what is
significant in a phonological rule.

Markedness and Natural Phonology

The concept of markedness is used frequently in transformational-
generative grammar, especially in phonological theory and analysis. It
is an attempt to characterize what is normal, simple, expected, and
plausible in language and that which is not by assigning a measure of
complexity to each. But as Hyman (1975) pointed out, the term
markedness has been used in several ways; it has been defined in
terms of the four notions of addition, frequency, neutrality, and pro-
ductivity or regularity.
When markedness is defined in terms of addition, an unmarked
term represents a basic, or normal, state; a marked term incorporates
the addition of some feature or property. Thus, /k/is unmarked and /g/ is
marked with respect to voicing, because /g/ incorporates the feature of
voicing, which is not present in /k/.
Markedness also has been defined in terms of frequency. Thus, in
the pair /s/ and /z/, the unmarked member would be /s/ because it occurs
in English words more frequently than /z/. =
The view of markedness as neutrality refers to a situation in which
one member of a pair or class of segments — the unmarked mem-
ber — is always found in a certain environment where the other(s)
cannot occur. For example, in English initial consonant clusters
beginning with /s/, the second segment may be /p/, but not /b/. Thus, one
finds words like spare, but not *sbare. Consequently, one would refer
to the ‘‘favored’’ segment /p/ as unmarked, whereas /b/ is marked.
When markedness is used to refer to productivity or regularity, a
term is marked if it is an exception to a rule. For example, obesity is
an exception to the trisyllabic laxing rule. Because obesity does not
undergo the regular rule for English vowels, it is marked in the lex-
icon as an exception to the rule.
Transformational-generative phonology uses markedness as a theo-
retical device in characterizing those segments, classes of segments,
and rules that are common and plausible in the world’s languages and
92 Phonology II

those that are not. The evidence for this type of study most frequently
comes from facts of language acquisition, language typology, and lan-
guage change. For example, studies of language acquisition have
shown that ordinarily the first vowel regularly made by an infant is /a/.
Furthermore, the vowel /a/ is virtually universal in the inventories of
the world’s languages, and it is not common for a language to change
in such a way as to eliminate /a/ from its stock of vowels. We conclude
that /a/ is the most unmarked vowel. Slightly more marked are /i/ and /u/.
The most basic or unmarked vowel system, thus, would be /i,a,u/. One
reason this system is so basic is that it allows for maximum differen-
tiation among vowels, in terms of their articulation in the vocal tract.
At the opposite extreme, front rounded vowels, as appear in German,
would be more highly marked. A similar characterization applies to
consonants; for example, /t/is unmarked and /d/ is more highly marked.
Among the most basic, or unmarked, consonants are /p, t, k, s, n/.
It is important not to assume that all features with plus specifica-
tions are always marked features. For example, /m/, /n/, and /n/ are all
[+nasal]. However, according to studies of language acquisition,
language typology, and language change, the unmarked nasal is deter-
mined to be /n/; that is, /n/is less marked than /m/ or /n/. Since /n/is the un-
marked nasal and it has the features [ + anterior] and [ + coronal], /m/ is
marked among nasals because it is [—coronal], and /n/ is marked
among nasals because it is [—anterior]. Therefore, /m/ and /n/ have
negatively specified features’ that contribute to their markedness
among nasals.
Besides identifying the most natural segments, markedness helps to
identify natural rules. From written records and linguistic observation
of the world’s languages, we know that certain types of rules are
more common or expected than others. A rule that makes word-final
stops and fricatives [ — voiced] is a natural rule; it occurs in German,
Russian, and other languages. Many natural phonologists claim that
children tend to pronounce words like tub and fizz as [tap] and [fis]
and that English-speaking children must learn to suppress these
voiceless sounds in accordance with the data provided by more so-
phisticated speakers.
As another example, a rule that causes voiced sounds to become
voiceless when occurring between voiced sounds would be highly
suspicious; that is, highly marked. Similarly, in English, when the
\, prefix in- is added to a word, the nasal segment of the prefix takes the
same point of articulation as the initial consonant of the word. When
the word begins with a vowel, the nasal is [n]; this reinforces our feel-
ings that /n/ is the unmarked nasal. However, although /m/ and /n/ are
more highly marked than /n/, it is more natural — in English and uni-
versally — for /m/ to precede labial consonants and /n/ to precede velar
 4 Summary 93

consonants. In this case, it is apparently ease of articulation that

makes this situation common in the world’s languages. Rules that
Cause segments to adopt certain features of neighboring segments are
teferred to as assimilatory, and they are usually considered quite natu-
ral or unmarked.
Natural rules generally use natural classes of segments. For ex-
ample, we think of the voiceless stops /p, t, k/ as being somewhat of a
natural class, and feature notation bears this out. If one adopts the
distinctive feature [delayed release] to distinguish the affricate /¢/ from
the voiceless stops, then only three features will be required to spec-
ify the class of voiceless stops: [— voice], [—continuant], and [—de-
layed release]. On the other hand, we would think of the segments
/p, 6, I/ as an unusual conglomeration; and the expression of this
class in terms of features is, in fact, extremely cumbersome in binary
analysis.
A theory of phonology that is heavily based on natural segments,
classes, and rules is called natural phonology; it has been developed
extensively by David Stampe, Theo Vennemann, Sanford Schane,
and many other linguists. In large part, natural phonology constitutes
a reaction to the very abstract analyses of Chomsky and Halle (1968).
For example, natural phonologists would argue against the use of fea-
tures such as [+ derivable], and they would challenge the claim that /1/
is the underlying representation of [ay] in criticize.
Unfortunately, not all natural phonologists agree on how abstract
underlying representations should be. Nevertheless, one common po-
sition among them is that for each pair of segments that alternate, the
underlying representation must be identical to one of the alternating
segments. This requirement would dictate that the underlying repre-
sentation of divine/divinity be either div/ay/n or div/i/n, but certainly
not div/i/n, as claimed by Chomsky and Halle. Of course, phonologi-
cal rules would still be required to relate the two forms.
Although the approach of the natural phonologists has generally
been applauded, some traditional transformational-generative phonol-
ogists fear that the natural phonologists’ strict concern with concrete
representations and plausible rules might represent a step back to-
ward a very empiricist methodology like structuralism. This debate
was one of the key issues in the phonology of the 1970s.

Summary
Stress contours in transformational-generative phonology are as-
signed by generative, cyclic rules according to the principle of the
transformational cycle. These rules apply to a bracketed parts-of-
speech representation in which all vowels are originally [ — stress].
94 ’ Phonology II

Redundancy is a property of language that assists communication. }

The concept is applied in phonology to allow the most economical
grammatical descriptions.
Markedness is a concept that characterizes the relative complexity
of segments and rules. Natural phonology is concerned with analyz-
ing language through the use of more concrete underlying represen-
tations and more plausible rules than those employed in traditional
transformational-generative phonology.

SOURCES FOR FURTHER READING

GENERATIVE PHONOLOGY
Chomsky, Noam, and Halle, Morris. The Sound Pattern of English.
New York: Harper and Row, 1968.
Halle, Morris. ‘‘Phonology and Generative Grammar.’’ Word, 1962.
Harms, Robert. Introduction to Phonological Theory. Englewood
Cliffs, New Jersey: Prentice-Hall, 1968.
Hyman, Larry M. Phonology: Theory and Analysis. New York: Holt,
Rinehart and Winston, 1975.
Schane, Sanford. Generative Phonology. Englewood Cliffs, New Jer-
sey: Prentice-Hall, 1973.

REDUNDANCY
Schane, Sanford. Generative Phonology. Englewood Cliffs, New Jer-
sey: Prentice-Hall, 1973.
Stanley, Richard. ‘‘Redundancy Rules in Phonology.’ Language,
1967.

NATURAL PHONOLOGY
Bruck, Anthony; Fox, Robert A.; and LaGaly, Michael W. Papers
from the Parasession on Natural Phonology. Chicago: Chicago
Linguistic Society, 1974.
Hooper, Joan B. An Introduction to Natural Generative Phonology.
New York: Academic Press, 1976.

EXERCISES

1. In which of the following pairs of words are there different pho-

netic segments for which you would suggest a single underlying
representation?
(a) electric/electricity (b) dearth/death (c) sun/son (d) drama/dra-
matic (e) finite/infinity
2. Why does one suggest an underlying /g/ in the word sign, but not in
the word gnat?
 $ Summary~95

b. What features are needed to specify uniquely the class of seg-

ments /m/, /n/, and /n/? What other features are needed to distin-
guish each of these segments from one another?
. Words like *dpove, *gbilm, and *kput are obviously not English
words. Formulate a redundancy statement (in your own words)
that expresses the fact that such words cannot be English words.
. Provide a bracketed parts-of-speech classification for the phrase
young grand-child, and describe step by step how stress is as-
signed by cyclic rules. Hint: apply the main stress rule, the com-
pound stress rule, and the nuclear stress rule—one rule per cycle.
. a. The vowel in a word such as brick receives primary stress by
what rule?
b. Would the same vowel in the word in receive primary stress by
this same rule?
. In the following pairs, decide which term is marked and which is
unmarked.
a. With respect to voicing: /f/
/v/
b. When occurring before a word-final voiced stop or fricative: /f/
/v/
c. The plural formation of truck; that is, [s].
The plural formation of fish; that is, @.
 CHA
PT ER

Morphology

Neither the phonemes of structuralism nor the underlying represen-

tations of transformational-generative grammar in themselves carry
meaning. To convey meaning, sounds are combined into words. But
in fact, words are not the smallest unit of meaning. Many words are
made of smaller units of meaning, and these units are combined in
particular ways, forming words. Morphology is the study of these
meaning-bearing units and the rules governing them, the study of the
structure of words.
How, for example, is the general relationship between such pairs of
words as refuse/refusal, deny/denial, and rebut/rebuttal to be cap-
tured? And why can the suffix -ness be added to words like happy and
silly, but not run and in? In the study of morphology, linguists try to
account for these and other facts with general rules. As in previous
chapters, we will examine the approaches of both structuralism and
transformational-generative grammar.

Structuralist Morphology

Structuralist phonemes were units that distinguished among meanings

but that did not in themselves carry meaning. But the structuralists
were interested also in defining a unit of language that did bear mean-
ing; so they proposed the concept of a morpheme. A morpheme,
which sometimes is designated with braces,{ }, is the smallest
96
 Structuralist Morphology a}

meaning-bearing unit of language. For example, re- is not a word, but

it does carry meaning. A morpheme ordinarily consists of a sequence
of one or more phonemes. The structuralists tried both to classify the
types of morphemes and to relate these abstract units to the actual
words of speech.

TYPES OF MORPHEMES
Many words are themselves morphemes, such as {dry} and {water};
they cannot be broken down into smaller units that in themselves
carry meaning. But many other words consist of more than one mor-
pheme. Most compound words, such as sandbox, are created by join-
ing together two morphemes, in this case {sand} and {box}, each of
which can be recognized as a word that carries a meaning by itself.
All the morphemes named thus far are free morphemes; that is, they
can exist as independent words. S
Another type of morpheme is the bound morpheme, which occurs
only when attached to another morpheme. This type includes prefixes
and suffixes of all kinds, such as {pre-}, {-ness}, {-ly}, {-ed}, and
many others. The word reprinted, for instance, consists of three mor-
phemes: the free morpheme {print} and the two bound morphemes
{re-} and {-ed}. Each of the three morphemes bears meaning and
contributes to the overall meaning of reprinted: {print} carries the
meaning of making an impression, {re-} signifies repetition, and {-ed}
designates the past tense.
In general, the analysis of a word into its component morphemes
requires that each morpheme occur elsewhere in the language; that is,
it must occur with the same meaning either as a free morpheme or as
a bound morpheme in other combinations. For example, the mor-
phemes {re-} and {-ed} in the word reprinted also occur in many
other words and have the same meanings in those words. For ex-
ample, in replayed, regained, and retyped, {re-} again signifies repeti-
tion and {-ed} again indicates the past tense.
A frequent problem in analyzing morphemes is the tendency to
overanalyze a word — to subdivide it into elements that do not con-
tribute to the meaning of the word. For example, the word regarded
may be divided into the morphemes {regard} and {-ed}, but it would
be a mistake to continue by dividing regard into {re-} and {gard}.
The re- in regard does not carry the meaning of repetition, as did the
{re-} in reprinted and replayed; and gard apparently does not occur
with a consistent meaning in any other combinations.
Another problem in analyzing morphemes arises in a word like
cranberry. It is apparent that {berry} can and should be considered a
free morpheme, but then {cran} also must be a morpheme. {Cran},
however, exists neither as a free morpheme nor as a bound mor-
98 Morphology

pheme in any other compound in English; thus, there is no inde-

pendent confirmation of its reality as a morpheme. According to our
earlier discussion, {cran} should therefore not be accepted as a
morpheme, and cranberry must be treated as a single morpheme — a
conclusion that violates common sense. Under what circumstances
could we accept {cran} as a morpheme? We may argue that the word
cranberry fits a pattern followed by blueberry, blackberry, and straw-
berry, each of which may be divided readily into two free morphemes
in the pattern X + berry. By analogy, {cran} should be considered a
morpheme in that pattern, even though it lacks independent status;
however, since it exists only in this one word, it must be considered a
bound morpheme.
Bound morphemes may be subdivided into derivational and inflec-
tional morphemes. A derivational morpheme is one that is added to a
root (that is, a word) to form a new word tlthat_differs, usually, in its
part--of-speech clclassification. For example, when the suffix -ness is
added to the adjective happy, the noun happiness is formed. Simi-
larly, the adjective quick becomes the adverb quickly when -ly is
added. We may also classify such bound morphemes as uwn- and pre-
as derivational morphemes, although these differ from -ness and -/y in
two basic ways: (1) un- and pre- do not change the part-of-speech
classification; for example, unambitious and predetermined remain
adjectives; (2) the addition of un- and pre- changes the meaning of the
word in a significant way, whereas the addition of -ness and -ly has
little effect on the basic meaning of the word. In English, prefixes are
usually derivational morphemes that change the meaning but not the
part-of-speech classification, whereas suffixes are usually derivational
morphemes that change the part-of-speech classification but not the
meaning. Exceptions include joy/enjoy and dear/endear.
An inflectionalal
morpheme indicates certain grammatical properties
associated with nouns and verbs, such as gender, number, case, and
_tense. Unlike highly inflected languages like Latin, English has very
few inflectional morphemes. In English, the inflectional morphemes
are all suffixes. The suffix -s, which indicates either possession or plu-
rality in nouns, is an inflectional morpheme; the past tense suffix -ed,
which is added to verbs, is another.

ALLOMORPHS
The morpheme is an abstract unit. In actual speech, one morpheme
may have several pronunciations or several phonological forms. For
example, -ed is pronounced differently in rented, employed, and
faked. Just as phonemes are abstract units realized through allo-
phones, morphemes are realized through allomorphs. Allomorphs are
any of the variant forms of a morpheme. The plural morpheme in En-
 Morphology and Transformational-Generative Grammar 99

glish, {Pl.}, is realized through many allomorphs. Consider, for ex-

ample, the words hats, dogs, and buses. Phonemically, these words
are written as /hets/, /dogz/, and /basoz/, from which it is apparent that
the plural endings are /-s/, /-z/, and /-9z/. These three allomorphs do not
occur randomly; which allomorph occurs depends on the phonetic en-
vironment. Nouns that end in one of the sibilants /s, z, 8, Z, ¢, j/ take
the /-9z/ plural allomorph, as in mazes, judges, and wishes. Nouns that
end in a voiceless consonant (other than a sibilant) form their plurals
with the voiceless allomorph /-s/, as in caps, chiefs, and wicks. All
other regular nouns end in a voiced sound and take the voiced allo-
morph /-z/, as injoys, bums, and liquids. Some irregular allomorphs of
the plural morpheme are /-an/as in oxen, /-ran/ as in children, and © as
in deer. These allomorphs help to account for the differences in pro-
nunciation of the various plural endings.

Morphology and Transformational-Generative Grammar

Until recently, transformational-generative grammar had little to say
about morphology. Instead, in transformational-generative grammar
much of what the structuralists called morphology was covered by
powerful theories of phonology and syntax. For instance, one can say
that the plural morpheme in English has a single underlying represen-
tation and that phonological, not morphological, rules determine
whether the phonetic ending is [-s], [-z], or [-az]. Exceptions such as
children and oxen are then designated in the lexicon as exceptions to
the phonological rules. Lately, however, many linguists have re-
emphasized that plurality is a morphological, not a phonological,
issue.
Other morphological phenomena were treated in transformational-
generative grammar as part of syntax. For example, the noun refusal
was assumed to derive from the verb refuse by a syntactic rule called
a transformation. Now, many transformational-generative gram-
marians consider refusal an item distinct from refuse; they do not use
syntactic rules to relate the two words. However, the relatedness
of the words is expressed in their representation in the lexicon.
Thus, the lexicon might indicate that refuse has the syntactic feature
[+verb], whereas refusal possesses the syntactic feature [+noun];
other features for the two items will be identical. (Syntactic features
will be discussed in Chapter 9.) In short, morphology was rediscov-
ered in the 1970s, and transformational-generative grammarians are re-
establishing its status as a separate component of grammar.
We have seen that transformational-generative grammarians typi-
cally express a speaker’s linguistic knowledge by means of rules. To
deal with morphological phenomena, the grammar contains rules de-
100 Morphology

scribing productive morphological processes. For example, one of

these rules, called ize-attachment, states that an adjective like com-
munal may take the suffix -ize to form the verb communalize. Such a
tule is called a word-formation rule. Each word-formation rule con-
sists of a base and an operation. The base is a set of words to which a
rule can apply; these words must be members of a major grammatical
category such as noun, verb, adjective, and adverb. The operation
specifies the effect of the rule.
For example, a word-formation rule indicates that a verb like em-
ploy may take the suffix -ee to form a noun, employee. But an adjust-
ment rule indicates that when one adds -ee to the verb nominate to
form a noun, one gets the truncated version nominee. Adjustment
rules change the shape of certain morphemes in the immediate envi-
ronment of other specific morphemes. Another type of adjustment
rule will specify that invert + -ion gives inversion. Notice that this
rule would have been considered a phonological rule (t—Z) in earlier
transformational-generative phonology.
Generally, morphological and phenological rules exhibit the follow-
ing relationship in a grammar: the word-formation rules may be sub-
ject to adjustment rules, and these, in turn, apply as a group before
the phonological rules apply. Word-formation rules are unordered,
but phonological rules may be ordered with respect to one another.

Summary
The morpheme was defined by the structuralists as the smallest unit
of meaning. The variants of a morpheme are called allomorphs. Mor-
phemes are either free or bound, and bound morphemes are sub-
divided into derivational and inflectional morphemes.
Morphology has only recently been studied intensively within
transformational-generative grammar; it has previously been sub-
sumed under the domains of phonology and syntax. There are two
major types of rules dealing with morphology: word-formation rules
and adjustment rules.

SOURCES FOR FURTHER READING

Aronoff, Mark. Word-Formation in Generative Grammar. Cam-

bridge, Mass.: MIT Press, 1976.
Bolinger, Dwight. ‘‘On Defining the Morpheme.’’ Word, 1948.
Matthews, Peter H. Morphology: An Introduction to the Theory of
Word Structure. Cambridge, England: Cambridge University Press,
1976. ;
 rs
4 {
Summary 101

Nida, Eugene A. Morphology: The Descriptive Analysis of Words. 2d

ed. Ann Arbor: University of Michigan Press, 1956.

EXERCISES

1. Identify all the morphemes in the following sentence: The older

gentleman voted wisely.
2. How many morphemes are contained in the word lawyer? Discuss
what you considered in deciding where to break the word.
3. Can the plural and possessive morphemes be grouped as one mor-
pheme since they are phonetically identical?
4. In the following sentence, indicate whether each morpheme is (a) a
free morpheme, (b) a derivational bound morpheme, or (c) an in-
flectional bound morpheme: The children skipped pope and played
games joyfully.
5. Which of the following rules are word-formation rules and which
are adjustment rules?
a. arrive + -al—arrival
b. electric + -ity—electricity
c. law + -er—lawyer
d. true + -ly-truly
6. How might one use the lexicon to show the relatedness of the
words proof and prove?
 CHAPTER

SYNTAX I

In every language, certain. patterns of words are permissible and

meaningful and others are not. Syntax is the way words are put
together to form phrases and sentences. To study syntax is to study
the patterns and relationships of words, phrases, and clauses. In 1957,
Noam Chomsky started a revolution in this age-old study with his
book Syntactic Structures. Ten years later, only a few structuralists
were left. More than twenty years later, Chomsky’s approach con-
tinues to dominate the study of syntax. In this chapter we examine
first the problems with the traditional and structuralist approaches to
syntax and then a few of the basic concepts of Chomsky’s transfor-
mational-generative grammar.

Defining a Sentence
For as long as there have been grammarians, the study of grammar
has focused on the sentence. The traditional definition of a sentence
as ‘‘a complete thought’’ goes back to the ancient Indian and Greek
grammarians, but the definition lacks persuasiveness. For example, a
person may ask the question, ‘‘Why did you climb that mountain?”’
and another person may answer, “‘Because it is there.’’ This answer
may be considered unacceptable as a sentence because it does not
represent a complete thought. In fact, it does represent a complete
thought, and we express ourselves in such terms quite frequently.
102
 The Traditional Approach to Syntax 103

Nevertheless, it is an incomplete sentence because it is a subordinate .

clause, and such clauses cannot be sentences by themselves —
according to prescriptive grammarians.
The definition of a sentence became a real problem for the struc-
tural linguists; for they discovered that a sentence, per se, had little in
the way of defining features. They rejected the easiest defini-
tion — any stretch of words between two periods — because they
believed in working with the spoken rather than the written word.
Eventually, they adopted a very rough and dubious definition: a sen-
tence was defined to be a stretch of speech between major pauses.
But everyone knew that this definition was terribly weak, just as they
knew that the likelihood of a better definition was remote.
When the transformationalists succeeded the structuralists, they
also made the sentence central to syntax; but their approach to a defi-
nition of a sentence was much simpler. In virtually the same way that
the transformational-generative grammarians defined parts of speech,
they said that a sentence was whatever native speakers knew to be a
sentence. In other words, no definition was necessary; sentences
were recognized to be sentences intuitively. Although this definition
inspired research and insight, it certainly has its shortcomings. Native
speakers’ intuitions about whether particular word strings are sen-
tences often disagree.

The Traditional Approach to Syntax

The analysis of sentences has as its ultimate purpose an under-
standing of the relationships between the words and phrases of a sen-
tence. Probably the best-known form of analysis is the Reed-Kellogg
sentence diagram, which was used in most American public schools
until the 1950s. The Reed-Kellogg diagram is reasonably simple, and
it provides a somewhat useful presentation of relationships within
sentences. For example, Figure 9-1 is a Reed-Kellogg representation
of the sentence The brown dog chased the yellow cat continuously.
To begin with, the diagram shows that a sentence consi fa syb-
ject and_a predicate, which are indi i ine divided

FIGURE 9-1
Reed-Kellogg Sentence Diagram
104. Syntax I

_by
vertical
a line. The object is also indicated on the horizontal line,
and it is separated from the predicate by a vertical line that does not
cross the horizontal line. All modifying words are placed on slanted
lines beneath the words they modify. Thus, Reed-Kellogg diagrams
provide a simple, pictorial view of sentences. Unfortunately, the
Reed-Kellogg approach is simply too simple. In effect, it cannot
match the genuine complexities of English grammar.
One result of the simplicity of traditional grammars was a contra-
diction between the rules and reality. For example, one traditional
rule held b in iV
When faced with a sentence like She was given a watch, in which
given is undeniably passive and also undeniably takes an object, the
traditionalist resolved the problem by saying that watch was a “‘re-
tained object,’ a grammatical category of objects different from
direct and indirect objects. Rather than reassessing the rule about
passive verbs, the traditionalist created an artificial distinction that
served only to preserve the rule.
An unswerving loyalty to simplistic rules sometimes led traditional
grammarians to startling conclusions. The following is taken from a
traditional grammar:
The subject of an infinitive is in the objective case:
We asked him to go. (Him is the subject of to go.)
In an infinitive clause, a predicate noun. . . used after fo be is in the ob-
jective case to agree with the subject of the infinitive:
They took me to be her.
If the infinitive to be has no subject, the predicate noun. . . following it is
in the nominative case:
He was thought to be /.
[Walsh and Walsh 1972, pp. 53-54]

These rules maintain a firm grip on logic; they demonstrate no incon-

sistencies. Why, then, does the last example sound ludicrous? Be-
cause for all their lucidity, the rules as stated are just not right. The
major flaw lies in the supposition in the second rule that the predicate
noun after to be is in the objective case in order ‘‘to agree with the
subject.’’ True, it is in the objective case; but agreement with the sub-
ject has nothing to do with it. The reasons for its being objective hold
even if the infinitive has ‘‘no subject,’’ and the last rule is simply an
instance of prescriptivism.

The Structuralist Approach

In their thorough revision of syntactic theory, the structuralists re-

placed grammatical rules with sentence patterns and the Reed-
Kellogg diagrams with immediate-constituent diagrams. Neither rep-
resented a great improvement.
 The Structuralist Approach 105

A sentence pattern was a sequence of word classes and groups. One

of the most common types of sentence patterns was the pattern:
Group A — Class I — Class II — Group A — Class I
representing such sentences as The girl kissed a boy, A woman kicked
the bucket, and A sentence has a subject. Here, girl, boy, woman,
bucket, sentence, and subject are all considered Class I words; they
all behave similarly in sentences in that they possess the properties of
nouns. In the same fashion, kissed, kicked, and has share properties
of verbs and are termed Class II words. Finally, a and the are modi-
fiers that behave similarly in sentences; they are put together in a
class of words called Group A. The business of a grammarian was to
determine the patterns of English and, if possible, to list them. This
was an exhausting chore of no great merit; for in the end the patterns
expressed nothing more than what was already known: that a variety
of sentence types existed.
diate- ituent analysis, too, was no improvement in the art
of understanding organization within sentences. Generally speaking,
immediate constituents were successive words that formed a unit.
Givetra string of three words, such as in the eye, the problem to the
structuralists was to determine which word is the more closely as-
sociated with — in or eye? The obvious answer, in this case, is eye,
and the unit formed by the eye goes with in. An immediate-constit-
uent diagram would present this information in the form shown in
Figure 9-2. A box represented a unit, or_a constituent. The smallest
boxes represented the smallest constituents, presented here as words.
An analysis of the sentence The girl kicked her younger brother fero-
ciously would appear as shown in Figure 9-3.
Certain ageless truths can be perceived in Figure 9-3. The major
division in the sentence lies between the unit the girl and the unit
kicked her younger brother ferociously — in other words, between
the traditionalist’s subject and predicate. Atticles are_closely as-
sociated with the nouns they precede, rather than other units.

though the word order of the sentence is preserved — an improve-

ment over Reed-Kellogg diagrams — such diagrams are rather dif-
ficult to read. Furthermore, th egories of the individual words are
not indicated, and the syntactic information given is often incomplete.

FIGURE 9-2
106 Syntax I

FIGURE 9-3

For example, in Figure 9-3 we are not given the complete array of
syntactic information that indicates which words can fit into which
slots. More importantly, structuralist analysis was overly concerned
with the strict surface appearance of sentences; the limitations of this
approach will become clearer in Chapter 10. In conclusion, the struc-
turalist, like the traditionalist, was unable to provide an adequate
description of the relationships between words within sentences.

The Transformational-Generative Approach

The structuralists were primarily concerned with describing the sen-
tence patterns in observed utterances. In contrast, Noam Chomsky
and the transformational-generative grammarians are concerned with
what a speaker knows about the sentences of his or her language; that
is, the unconscious ability to interpret and produce sentences that the
person may have never before heard. Once we penetrate the forbid-
ding notation of transformational-generative grammar, we find a great
deal of plain common sense.
Every day we utter and we understand sentences that we have
never heard before. Transformational-generative grammar deals with
this fact, recognizing that the number of sentences a speaker may
produce and understand is unlimited, but his or her knowledge and the
sentences actually produced are not. Its goal is to find a limited set of
rules from which an unlimited number of sentences can be generated.

erating sentences and the use of that knowledge in producing and un-
derstanding utterances, the transformationalists introduced the con-
cepts of performance and competence.

PERFORMANCE AND COMPETENCE

Performance means the_actual saying of_s l or the act_of
speech it Ideally, we would like to be able to account for the
speaker’s performance, but there are formidable obstacles. In the act
of speech, a speaker is responding to the environment, which is
 The Transformational-Generative Approach 107

usually of extreme complexity and for which we have few descriptive

tools. The speaker is responding also to a mental state; that is,
whether he or she is alert, drunk, and so on. And the speaker has lim-
itations of memory, may stutter, or may change intentions in midsen-
tence. These and other things go into a speaker’s performance at any
moment.
Competence is an essential part of performance. It is the speaker’s
knowledge of his or her language; that is, of its sound structure, its
words, and its grammatical rules. Competence is, in a way, an en-
cyclopedia of the language. Moreover, the knowledge involved in
competence is generally unconscious. For instance, if we are asked
what rules govern the passive voice of the verb, we may reply that we
have no idea. But we must know the rules because we use them in
speech. In fact, we know the rules, but we cannot verbalize them.
If competence is an encyclopedia of the language, the speaker must
also-Know the rules governing use of the encyclopedia. Such eee
Sot i Tee SG aa a as yet do little
more than guess at the forms they might take. To give only a general
idea of what these heuristics are, suppose a speaker wishes to express
the following thoughts: (1) that he or she has many bills; (2) that they
must be paid; and (3) that these bills must be paid by the end of the
month. The speaker must go-through some mental process which may
result in the sentence ‘‘I have some bills to pay before the end of the
month.’’ While we can actually write certain rules to describe the
syntactic combinations that exist in this sentence, we cannot as yet
say exactly what mental process the speaker went through, or what
performance strategy was employed, in arriving at the actual sen-
tence. But surely some such mental strategy exists. This strategy is
called heuristics.
The transformational-generative view of performance and compe-
tence is diagrammed in Figure 9-4.

FIGURE 9-4
Performance and Competence

PERFORMANCE

memory intentions speech mechanism

COMPETENCE
 108 Syntax I

FIGURE 9-5
Standard Interpretive Model of Competence
COMPETENCE

Phonological Rules

SAI. [Seman Re
| Syntactic Rules oat

A MODEL OF COMPETENCE a
Ths transtormationalsgsnes a= aulsts have taken it upon them-
se 0 give an account of competence; that is, to determine what
the rules are that underlie a language and how they work together to
create the sentences of the language. This model of competence has
three major components: syntax, phonolo and_semantics. In
Chomsky’s_view; syntax is ‘‘central,’’ which is to say that gentences
are_principally_formed
within the syntactic component. Hhonolk
He and semantics are ‘‘interpretive’’ in the sense that their rules operate
on the product of the syntactic component to clothe it with sound and
to make sense out of it. This organization of competence, which we
refer to as the standard interpretive theory, is presented in Figure 9-5.
The centrality of syntax has been something of an issue among
transformational-generative grammarians. An alternative view —
originated by such linguists as George Lakoff, James McCawley,
Paul Postal, John Ross, and others — is called generative semantics.
This view holds that semantics, ormeaning: is primary, and that syn-

tences. This view 1s eae at

in Figure 9-6.
“Tt is generally agreed that the three components probably are in-
termingled somewhat, but for practical purposes they are treated as if
they are entirely independent. For the remainder of this chapter and
the next, we will elaborate on the standard interpretive theory, al-
though much of what is said applies to generative semantics as well.
Further discussion of both theories appears in Chapter 11.
A sketch of the syntactic component in the standard interpretive
theory is shown in Figure 9-7. The syntactic component has two

FIGURE 9-6
Generative Semantics Model of Competence
COMPETENCE

|Semantic Rules |» Syntactic Rules |-~ | Phonological Rules

 Phrase-Structure Grammar 109

FIGURE 9--7
The Syntactic Component
SYNTAX

Base component

Phrase-
Structure Lexicon
rules

deep structures

Transformational
component

transformational
rules
'
surface structures

subcomponents: the base component and the transformational compo-

nent. The pase_component contains a set of rules called paras
tructure rules, which provide the constituent structure for sentences.
The constituent structure is sim i j i
combined. The base component also contains a Jexicon, a list of words
that indicates which words may be inserted into which structures.
The output of the base component is a string of terms called a dee
structure, which represents the basis for the meaning of a sentence.
This deep structure feeds into the transformational component, which
is a set of rules that turns the deep structure into a surface structure.
The surface structure basically serves as the input to the phonological
component, as well as to some semantic rules. For the time being, we
may think of the deep structure as an abstract level representing the
essential meaning of a sentence, whereas the surface structure corre-
sponds to an actual sentence, as heard in discourse. In the remainder
of this chapter, we examine the base component; the transformational
component is examined in Chapter 10.

Phrase-Structure Grammar

A phrase-structure grammar is a series of rewrite rules. These rules

pres Sow Senses establishing their basic structures, regardless
of the final form the sentences may take after transformational rules
have applied. Each rule in a phrase-structure grammar has the form:
110 Syntax I

xX- Y+Z

which means ‘‘X consists of Y and Z.’’ The first phrase-structure

rule, for instance, is:

S—NP+Aux+
VP

which means that a sentence consists of a noun phrase (NP) followed

by an auxiliary element (Aux) like tense that in turn is followed by a
verb phrase (VP). Successive phrase-structure rules indicate what
NP, Aux, VP, and other constituents consist of, until there are no
more constituents in the sentence to account for.
No complete phrase-structure grammar for English has yet been
fully accepted. It is possible that none ever will be; for the phrase-
structure rules depend in part on the theories that underlie them, and

TABLE 9-]
Partial Phrase-Structure Grammar and Lexicon

Partial Phrase-Structure Grammar Partial Lexicon

. S>NP + Aux + VP a — Lex[Det]

. NP (Det) + N girl — Lex[N]
Det— Lex[Det] can — Lex{M]
BY
. N>Lex[N] happy — Lex[Adj]
. Aux— Tense + (M) + (Perf) + (Prog) man — Lex[N]
V +(NP) + (PP) see — Lex[V]
the — Lex[Det]

. M>Lex[M] . Abbreviations
. Perf—have + -en
S — sentence
. Prog—be + -ing
NP — noun phrase
. V—>Lex[V]
Aux — auxiliary
. PP—Prep + NP
VP — verb phrase
. Prep— Lex[Prep]
Det — determiner
. Adj—Lex[Adj]
Lex — lexical item
M — modal
Notational Conventions
Perf — perfect
— means ‘‘consists of.”’ Prog — progressive
(X) means ‘‘X is optional.”’ Adj — adjective
PP — prepositional phrase
rf|means ‘‘choose between X and Y.”’
Pres — present
Lex[X] means ‘‘a lexical item marked X.”’ Prep — preposition
 Phrase-Struéture Grammar 111

these change with the linguist and the year. A partial phrase-structure
grammar and lexicon is given in Table 9-1 although it is primitive in
many ways. For example, it does not allow for conjunctions, adverbs,
and many other types of constituents that make English as rich as it
is. Moreover, the partial lexicon contains only seven words — ob-
viously far short of the thousands of words found in a typical
speaker’s vocabulary.
Despite these limitations, we can use the phrase-structure grammar
in Table 9-1 to generate the deep structure of some sentences. To il-
lustrate how the deep structure is generated, we can use a tree dia-
gram. The relationship between the phrase-structure rules an
‘diagram is quite strict; a phrase-structure rule of the form X—Y + Z
matches the tree

-
Zh hs
Y Z
Let us generate the deep structure underlying the sentence A girl
has seen the man. Rule 1 in Table 9-1, which starts with the abstract
concept of the sentence (S), generates an NP, an Aux, and a VP:

JA
NP Aux VP

The NP, according to Rule 2, consists of a noun (N) and an optional

determiner (Det); detervt ds_as and
some. In this case, the option of including a determiner is used. Thus,
we get

Rules 3 and 4 indicate that Det is rewritten as Lex[Det] and N is

rewritten as Lex[N]. Thus,
112 Syntax I

By Rule 5, Aux consists of Tense and may include other elements; in

this case, Aux is Tense plus the perfect construction (Perf).

NP
eS |e
Det N Tense Perf

Lex[Det] Lex[N]

Rule 6 looks rather complicated, but if the notational conventions of

Table 9-1 are observed, it is easily understood. Since, for this sen-
tence, we wish to generate a verb and a noun phrase, we choose the
first line of the rule: V+(NP)+(PP). From this line, we select the
obligatory V and the optional NP, and we exclude the optional PP,
since there is no prepositional phrase in our sentence. After the appli-
cation of Rule 6, the structure looks like this:

NP Aux VP

Det N_ Tense Perf V NP

Lex[Det] Lex[N]
By Rule 7, Tense is rewritten as Present; and by Rule 9, Perf is
rewritten as have and -en, where -en is the linguist’s notation for the
ending on the past participle of the verb (which may be -en, as in
broken, -ed as in typed, or nothing at all, as in put). Thus, we get

HS PS Ss seine.
Det N Tense Perf V NP

Lex[Det] Lex[N] Pres have -en

According to Rule 11, V consists of a lexical item marked Lex[V]. We

also apply Rules 2, 3, and 4 to the NP at the far right. Thus,

Lex[Det] Lex[N] Pres have -en Lex[V] Det N

Lex[Det] Lex[N]
 Phrase-Strueture Grammar 113

Next, we consult the lexicon in Table 9-1 to rewrite Lex[Det],

Lex[N], and Lex[V]. We find that the words a and girl are Lex[Det]
and Lex[N] as in the first NP; the word see meets the requirement
of Lex[V]; and the words the and man meet the requirements of
Lex[Det] and Lex[N] as in the second NP. The insertion of these
lexical items gives the following:

NP Aux

wa ae ane Ne Hes \ NP

Lex[Det] Lex[N] nih Det N

Lex[Det] Lex[N]
a girl Pres have -en — see By ihe

This final tree is the deep structure: no further phrase-structure rules

or lexical substitutions may operate. Deep structure, then, consists
both of the ordered words and terms at the bottom of the tree and the
tree itself, which is sometimes called a phrase-structure marker.
Why was girl chosen as the first NP meeting the requirements of
Lex[N], and not the noun man? And why was the determiner a cho-
sen as the first determiner meeting the requirements of Lex[Det], and
not the determiner the? The obvious answer is that we wanted to
generate A girl has seen the man. The specific set of phrase-structure
rules used in this example also will generate The man has seen a girl.
And since a speaker’s lexicon is actually very large, the same rules
would generate such sentences as A boy has petted the dog, The
witch has married the goblin, and many others.
The deep-structure string a girl Pres have -en see the man is ob-
viously not a sentence yet; it must undergo certain transformations
before it will take final form. Nevertheless, a passing resemblance to
its final form is already evident. (We will turn to the transformations
in Chapter 10.)
In fact the process described here in generating A girl has seen the
man may not be the process a person goes through in producing the
sentence. The purpose of the phrase-structure grammar is only to pro-
vide a description of the structures that underlie grammatical sen-
tences. The phrase-structure grammar, then, is part of a model of
competence that does not attempt to incorporate the speaker’s inten-
tions, important though those intentions are. What we have done is to
extract one particular sentence from the grammar for the purpose of
analysis.
114 Syntax I

Lexical Insertion

The lexicon is an extremely important part of the grammar; for no

sentence can make sense without words. Moreover, words and struc-
tures are related. Each word can enter into only certain structures
and certain relationships with other words.
In the lexicon, all relevant information is incorporated in features.
For example, the lexicon will contain a phonological representation,
given in distinctive features, for each word. The lexicon will also con-
tain two kinds of features that are important to us here: syntactic fea-
tures and semantic features.
Syntactic features are those aspects of a word that govern its use in
structures. The most important of these syntactic features is the part
of speech. It is this feature that disallows the use of destroy when a
noun is expected. That is, a tree of the following sort is not permis-
sible:
Lex[N]

destroy
because destroy has the syntactic feature [+verb] but Lex[N] calls for
a word with the feature [+ noun]. For verbs, the lexicon may desig-
nate the type of complement that a verb may take. The verb run, for
example, may take a prepositional phrase (PP) complement or no
complement, as in the sentences The athlete runs around the track
and The athlete runs. The verb run might then appear in the lexicon
something like this:
run —[V, Complement: (PP)]
where the parentheses again indicate that a feature is optional.
Through the use of syntactic features and the conventions governing
lexical insertion, the grammar will be prevented from generating the
deep structure that would underlie such sentences as *A man can see
happy. In this case, the verb see possesses a syntactic feature some-
thing like ‘‘takes an object,’’ ‘‘takes an adverb.”’ And since this is
true, it cannot be followed by an adjective like happy.
Semantic features are features that specify certain semantic require-
ments of words. For instance, it is a syntactic requirement that the
verb see have a subject NP, but it is a semantic requirement that this
subject NP must be something capable of seeing. In other words,
there is something semantically wrong with the statement *the stone
saw the boy; for it is well known that stones cannot see. See then pos-
sesses a semantic feature something like
(Subject: + vision)
indicating that the subject of see has vision.
 Summary 115

Semantic features are not unconditionally binding, as are syntactic

features. In fact many sentences of English involve semantic mis-
matches, and many of these are considered metaphors; they are per-
fectly good, even desirable, adjuncts of literary style. An example is
John Donne’s ‘‘Death, Be Not Proud’’; obviously, death is not a
higher animal capable of being proud. (Semantic features are dis-
cussed further in Chapter 11.)
Finally, the lexicon must somehow indicate the meaning of words.
Mostly, grammarians have taken it for granted that we do know the
meanings of words; generally, they have not dealt sufficiently with the
definitional part of the lexicon.

Summary
Neither the traditionalists nor the structuralists provided an adequate
account of the syntax of natural languages. Although the transforma-
tional-generative grammarians also have encountered numerous ob-
stacles in syntactic description, they recognize that language is cre-
ative, if not infinite. Therefore, grammars must be generative.
In the theory of transformational-generative grammar, a distinction
is made between competence and performance. Competence refers to
what one unconsciously knows about one’s language, whereas perfor-
mance refers to the utilization of this competence in the actual pro-
duction of speech.
A transformational-generative grammar is a model of competence,
and it contains a syntactic, a phonological, and a semantic compo-
nent. The syntactic component provides the structure for a sentence
whereas the phonological component provides the sound of the sen-
tence and the semantic component provides its meaning.
Phrase-structure rules are rewrite rules that indicate the structure
of a particular constituent. The effect of phrase-structure rules may
be represented by trees. The lexicon contains words, or morphemes,
that possess certain phonological, syntactic, and semantic features
and that are inserted into these trees. The phrase-structure rules, in
conjunction with lexical insertion rules, generate a deep structure.

SOURCES FOR FURTHER READING

TRADITIONAL AND STRUCTURAL SYNTAX

Hockett, Charles F. A Course in Modern Linguistics. New York:
Macmillan, 1958. (Structural)
Walsh, J. M., and Walsh, A. K. Plain English Handbook. Cincinnati:
McCormick-Mathers, 1972. (Traditional)
116 Syntax I

TRANSFORMATIONAL-GENERATIVE SYNTACTIC THEORY

Akmajian, Adrian, and Heny, Frank. An Introduction to the Princi-
ples of Transformational Syntax. Cambridge, Mass.: MIT Press,
1975.
Bach, Emmon. Syntactic Theory. New York: Holt, Rinehart and
Winston, 1974.
Chomsky, Noam. Aspects of the Theory of Syntax. Cambridge,
Mass.: MIT Press, 1965.
. Syntactic Structures. London: Mouton, 1957.
Katz, Jerrold J., and Postal, Paul M. An Integrated Theory of
Linguistic Descriptions. Cambridge, Mass.: MIT Press, 1964.
Lakoff, George. Irregularity in Syntax. New York: Holt, Rinehart
and Winston, 1970.

EXERCISES

1. The following sentence may have been heard in actual speech:

‘‘Uh, where is the st- st- statistics class?’’ In generating this sen-
tence, would one wish to generate wh and st- st-? Explain why or
why not.
2. Which of the phrase-structure rules shown in Table 9-1 are applied
in generating the deep structures for the following sentences?
a. Maria can be miserable.
b. The juggler has performed in the street.
c. The scholar is writing a biography.
3. Is it a syntactic or semantic feature of English grammar that pre-
vents the following sentences from being generated?
a. *Fred by run the store.
b. *Moon cow the jump the over.
c. *The water swam past the man.
4. Modify one of the phrase-structure rules to allow adverbs to be
generated.
 C-H.A
PT E.R

10.
SYNTAX II

In Chapter 9 the generating of deep structures was examined. In this

chapter we look at the transformations that a deep structure must go
through in order to take the form of the surface structure, the actual
utterance. With the addition of transformational rules, we can gener-
ate the actual sentences of a language; we can also clarify the rela-
tionship between sentences and their meanings. But these rules
should also do something else: they should avoid the generating of
ungrammatical sentences. To meet this criterion, it is necessary that
we also define certain constraints on the transformational rules.

Transformations

Transformations are operations that add, delete, or change elements

in one structure to produce another structure. According to this defi-
nition, the insertion of a lexical item into a phrase-structure tree might
be considered a transformation because it introduces (or adds) a com-
plex element into a structure, whereas the phrase-structure rules sim-
ply reanalyze, or break down, sentence structures. Nevertheless, we
will refer to the transformational _co ent as all the rules that apply
F ica] items have been inserted into tr
We need transformations for several reasons. For example, Figure
10—1 shows again the deep structure produced by the phrase-structure
grammar and lexicon in Table 9-1. It is still not an English sentence;
ig
118 Syntax I

FIGURE 10-1
Deep Structure for the Sentence A girl has seen the man.

NP Aux VP
aN Dc ONS
Det N Tense Perf i NP

Lex[Det] Lex[N] Lex[V] Det N

Lex[Det] Lex[N]

a girl Pres have -en see the man

certain other operations must be performed to yield a grammatical

English sentence. These other operations are what are called
transformations.
One of the earliest arguments for transformations concerned rela-
tions between words in sentences called co-o estrictions;
that is restrictions on the relations between words that apply to dif-
ferent types of sentences. For example, in the active sentence John
threw the ‘ball, the verb throw must take a concrete object. In the
passive sentence the ball was thrown by John, the same restriction
holds between ball and throw, although the surface structure is differ-
ent. Unless actives and passives are related by a transformation,
which we assume does not change the relationship between bail
and throw, this restriction will have to be stated twice in the
grammar— once for the active sentence and once for the passive
sentence. Therefore, transformations may simplify a grammar and in-
crease its ability to describe language.
More importantly, because they relate deep and surface structures,
transformational rules are able to express certain generalizations that
are part of the speaker’s knowledge about his or her language. The
relationship between deep and surface structures is in no way arbi-
trary. The fact that two levels of structure are posited at all results
from the principle that one of the purposes of a grammar is to account
for the relationship between sentences and their meanings.

DEEP STRUCTURE, SURFACE STRUCTURE, AND MEANING

The structuralist had been concerned strictly with the surface charac-
teristics of sentences and had, therefore, worked with only one level
of description. The transformational-generative grammarian recog-
nizes that there is no one-to-one relationship between sentences and
 a
Transformations 119

meanings. In fact, a particular meaning can be represented by several

different sentences, or paraphrases; and a particular sentence might
b i r have several meami
An example of several paraphrases Ga one basic meaning is the fol-
lowing set of sentences:
Mary gave an apple to Bob.
Mary gave Bob an apple.
. An apple was given to Bob by Mary.
Bob was given an apple by Mary.
It was Mary who gave Bob an apple.
Ne
ee Bob is the one Mary gave an apple to.
These sentences (among others) all have the same basic meaning and
are assumed to derive from one deep structure. This deep structure
may be affected by different rules in the transformational component,
and the different rules create variations in surface structure, as shown
by the different sentences that result. In the standard interpretive
theory of transformational-generative grammar, as well as in genera-
tive semantics, the application of transformational rules does not
change the meaning of a sentence. (Later versions of the standard in-
terpretive theory have relaxed this requirement, as discussed in
Chapter 12.) In general, then, deep structure represents the basis for

deep structure, as presented here, alone can provide the full meaning

tence sounds like, also is relevant for understanding such semantic

aspects ic and focus of a sentence (discussed in Chapter 12).
Another possible occurrence in natural Tanguages is that one sur-
face structure may represent several meanings, which are expressed
by different deep structures. A sentence of this type is ambiguous,
and ambiguity occurs in every language. The ambiguities of one lan-
guage, however, might not recur in the same form in any other lan-
guage. One example, from English, of the type of ambiguity we are
concerned with is the sentence Climbing plants can look strange.
This sentence, as we have seen, has two meanings, represented by
two deep structures, that are in no way related to each other. In one
deep structure, plants is the object of climbing and the meaning in-
volves a declaration about the possible strangeness of _someone’s

Subject_noun plants, and the declaration involves the possible

~Senieeness ofplants that areTogarded-av“‘ctimbing plamts” Thus,
Fe iGtet oaaicitG satis eortrectic cobicoeeretbiect rela-
tions that may distinguish sentences: To the best of our knowledge,
this particular ambiguous sentence cannot be translated into an equiv-
120 Syntax I

alently ambiguous sentence in any other language. In other words, the

ambiguity results from a peculiarity in the transformationai rules of
English.
—
Transformations, then, allow us to express the fact that one basic
meaning can have several paraphrases expressed indifferent surface
structures and that an ambiguous sentence, represented _b u

y different deep structures.

TWO OBLIGATORY TRANSFORMATIONS

One transformation applied to all sentences is the flip-flop (FF) rule,
by which certain endings are attached to the verbal forms that follow
them. The flip-flop rule may be: generalized as
FF Rule: Affix + Verbal=> Verbal + Affix
Condition: Constituents affected by FF may undergo the
rule only once.
Affix may be Tense, -en, or -ing. Verbal may be V, M, be, or have.
(Double-stemmed arrows, >, signify a transformational rule.) For an
example of how the rule operates, consider the deep structure pre-
sented in Figure 10-1. FF may operate at two points: at Pres +have
and at -en + see, which are in the correct order for application of the
rule. Thus,

Deep structure: a girl Pres have -en see the man

After FF: a girl have Pres see -en the man
Notice that we imposed the condition that the FF rule can apply to
each relevant constituent only once; otherwise, there is nothing to
prevent Pres+see from being flip-flopped. This condition thus pre-
vents the generating of such an ungrammatical sentence as *A girl
have sees -en the man.
Fo rt taetree ee in-
volves two parts: a structural description and _a structural change The
"structural description specifies the structure that must be present in
~ orderaferithe
tuleito analy (orekanipIe Bun oa
tural change indicates how thestructure is alate
ample, >Verbal + Affix). Conditions on the application of the rule
may or may not be present, and linguists have recently been working
hard to eliminate conditions wherever possible.
It is also important to note that Affix and Verbal, as stated, are
somewhat artificial categories, since they do not appear in the deep
structure tree of Figure 10-1. We use these categories because they
allow us to state the rule more easily; nevertheless, their use must be
considered a flaw in the theory, which ultimately must be corrected.
 Transformations 121

After the application of FF, only one other rule need be invoked to
transform the deep structure in Figure 10-1. This rule, which may be
called the lexical-formation LF) rule, requires thatthe lexicon be con-
final appearance of the nouns and verbs in

words of the uiiace structure; it tells us that have + Pres is has in

third
singular
person and have otherwise ‘ise and that see +-en isseen, as
follows:
By LF: a girl has seen the man
The FF and LF rules are the only transformational rules that must
be applied in order to derive all surface structures. Of course, many
other transformational rules apply under certain conditions. Ideally,
the conditions of application should be apparent in the deep structure,
although it does not always work out that well.

IMPERATIVES, QUESTIONS, AND NEGATIONS

The formation of imperatives, questions, and negations involves
operations that are signaled by constituents in the deep structure. The
first rule of the phrase-structure grammar must be altered at this point
to introduce these constituents:

Ss {|Imp
ote })+ (Neg) + NP + Aux + VP
In other words, a sentence still consists of an NP, an Aux, and a VP;
but it may also optionally be a question, an imperative, a negative, a
negative-question, or a negative-imperative. Because Imp, Ques, and
Neg are optional, it is still possible that none is realized, in which
case the sentence will be declarative and positive.
Consider how the imperative sentence Wash the car! is generated.
The underlying, or deep, structure for this sentence must include the
constituent Imp, the subject NP you, penta and the present
tense. This analysis is supported by the existence of a sentence like
Wash the car, won't you?, in which you and will ‘‘surface.’’ Further
evidence for an underlying subject you in imperatives is seen in sen-
tences such as Watch yourself. The first phrase-structure rule will
now produce the structure:

Imp
as ai age
NP Aux VP

The application of all other relevant phrase-structure and lexical in-

sertion rules will produce the deep structure given in Figure 10-2.
122 Syntax II

FIGURE 10-2
Deep Structure for the Sentence Wash the car!

Lex(N] Lex[M] Lex[V] Det N

Lex[Det] Lex[N]

you Pres will wash the car

The constituent Imp then triggers the imperative transformation,

which has the following effect: ss
Imperative
RPET transformation:
AUN CAT Imp iy
BTISTOND BUCO: ATID + you+will
ie ares> @
In other words, the constituents Imp, you, and will are deleted. The
FF rule then switches the order of the constituents Pres + wash, giv-
ing wash + Pres. The LF rule then tells us that wash + Pres is wash,
thus generating the sentence Wash the car!

TRANSFORMATIONS WITH EMBEDDED SENTENCES

The sentences discussed this far have been simple, one-clause sen-
tences, such as the following:
1. The bell rang.
2. A student entered the building.
3. The child likes the horse.
4. The horse roams on the farm.
But many sentences contain more than one clause; for example,
5. The bell rang and a student entered the building.
In order to generate sentences like this, we need the modified
phrase-structure rule:

S—S + (Conj+S)

where Conj represents and, or, and so on. Thus, the deep structure of
sentence 5 would appear as in Figure 10-3.
Figure 10-3 shows a relatively straightforward-looking deep struc-
ture; two sentences are joined by and. But in order to generate sen-
tences like
 Transformations 123

FIGURE 10-3
Deep Structure for Sentence 5: The bell rang and a student entered
the building.
S
|
ee
DS | pe
NP Aux VP Lex[Conj] NP Aux VP

| | |
Lex[Det] Lex[N] Lex[V] Lex[Det] Lex[N] Lex[V] Det N

Lex[Det] Lex[N]

The bell Past ring and a student Past enter the building

6. The child likes the horse that roams on the farm.

phrase-structure rule 2 must be revised as follows:
NP— (Det) +N +(S)
The deep structure of sentence 6 appears in Figure 10—4. In this struc-
ture, One sentence is contained within another. The sentence The
horse roams on the farm appears under the category label S, which
appears directly under the category label NP. Such a sentence is
called an embedded sentence.
In both revised phrase-structure rules, the symbol S appears to the
right of the arrow, allowing simple sentences to be contained within
more complex sentences. In theory, these rules could be applied infi-
nitely, producing sentences like

7. John has a cat and Mary has a dog and Phil has a canary and . . .
8. The student likes the professor who likes the administrator who
likes his mother who. . .

Such rules that can apply over and over again are called recursive,
and they permit an infinite number of sentences of infinite length to be
generated in a language.
The deep structure presented in Figure 10-4 is transformed into
sentence 6 through the application of the relative transformation,
which moves the second, identical occurrence of the NP the horse out
of the embedded sentence to the position of the horse in the main sen-
tence, leaving only one occurrence of the horse. Then an appropriate
relative pronoun — in this case that — is substituted into the position
from which the second occurrence of the horse was moved.
124 Syntax II

FIGURE 10-4
Deep Structure for Sentence 6: The child likes the horse that roams
on the farm.

NP ux
al VP
eS (eee
Det N Tense V NP

Lex[{Det] Lex[N] Lex[V] Det N 1 ae

Lex[Det] Lex[N] NP Aux VP
s yan
Det N Tense V PP
| | | PA
Lex[Det] Lex[N] Lex[V] Prep NP
|
Lex[Prep] Det N
|
Lex[Det] Lex[N]
|
The child Pres like the horse the horse Pres roam on the farm

THE ORDERING OF TRANSFORMATIONS

Some transformational rules, like phonological rules, may be ordered.
For example, the imperative transformation must apply before the FF
rule, as illustrated in the sentence Wash the car! (See Figure 10-2.) If
the FF rule were to apply first to reverse the order of Pres and will,
and then the imperative transformation were to apply to delete will,
we would have Pres + wash. Then we could not derive the desired
order wash + Pres, because flip-flop was already applied to Pres and
FF is permitted to apply to each constituent only once.
Furthermore, the syntactic transformations may apply in a cycle, in
pretty much the same way as the stress rules in phonology (see pages
86-88). The principle of the cycle in syntax requires that the trans-
formations be applied in order, beginning with the most deeply em-
bedded sentence. After all relevant transformations have been applied
on this first cycle, the transformations may be reapplied, again in
order, to the next highest sentence in the tree diagram. This process
continues until the final cycle, on which all relevant transformations
are applied in order to the main sentence. Transformational-genera-
tive grammarians have also considered the possibility that certain
transformations are special in that they apply either before or after
the cycle.
The transformations that we have presented are only a few of the
many that are applied in generating a language like English. The treat-
ment of others is best left to a full course in syntax.
 Constraints in Syntax 125

FIGURE 10-5

NP Aux VP
24 vee

N_ Tense
| Det
+ N
(Sona Aux VP

Lex[N] Lex[V]
|
Lex[Det] Lex[N]
| NP Tense
|
NP Conj NP
Te eee
Det N Det N Adj
| |
Lex[Det] Lex[N] Lex[Det] Lex[N] be Lex[Adj]

Constraints in Syntax
Once transformational-generative grammarians became well ac-
quainted with transformations by analyzing English and other lan-
guages, they quickly noticed that such rules are very powerful. Trans-
formations can generate a large number of sentences that speakers of
a language do not accept as proper sentences. To illustrate, consider
the structure shown in Figure 10-5. With the exception of the con-
joined noun phrases in the embedded sentence (which are generated
by a rule similar to S~S+(Conj+S)), all other rules generating this
structure have previously been presented. After the insertion of lex-
ical items, the deep-structure string in Figure 10-6 may result.

FIGURE 10-6

ieee

NP Aux VP
FOES
Vv NP ane eee We

N_ Tense Deta N sar ux VP ast

| | | neat | aba | Adj
Lex[N] Lex[V] Lex[Det] Lex[N] eK Conj i ae Tense |

Mary Pres hate the hat Det N Det N Lex[Adj]

| | |
Lex[Det] Lex[N] Lex[Det] Lex[N]
|
the hat and _ the coat Pres be dirty
126 Syntax IT

Notice that we have followed the rules perfectly in generating this

deep structure. In Figure 10-5, a sentence containing conjoined noun
phrases is embedded in another sentence, and the hat in the main sen-
tence is understood to be identical to the hat in the embedded sen-
tence. Thus, we should be able to apply the relative transformation.
However, if we apply this rule, moving the embedded occurrence of
the hat and replacing it by which, we derive the unacceptable sen-
tence *Mary hates the hat which and the coat are dirty. In order to
prevent such sentences from being generated, it is necessary to im-
pose certain constraints on the power of transformations.
One such constraint is called the constrai joined ses: NO_
phrase contained within a conjoined, or coordinate, structure may be
moved out of that structure by a transformation (for the purpose of
relativization, question-formati an n the previous
ample, this constraint would forbid the movement of the embedded
occurrence of the hat out of the structure containing the conjoined
NPs and into the main sentence (Mary hates the hat) for the purpose
of relativization. Many such constraints apply only to specific lan-
guages; but it is hoped that some constraints will hold for all lan-
guages, and thus reveal basic facts about the workings of language.
Not all constraints that have been proposed are of the same nature
as the constraint on conjoined phrases; some are decidedly more
functional in nature, that is they have more to do with actual utter-
ances. For example, it is impossible for a speaker to empathize, or
side with, more than one individual in a sentence. Thus, in the unac-
ceptable sentence *Then, Mary’s husband hit his wife (in which his
refers to husband), the speaker first empathizes with Mary by using
the phrase Mary’s husband (rather than the name of Mary’s hus-
band). But the focus of empathy is contradicted by the later occur-
rence of the phrase his wife, used to refer to Mary. Again, no sen-
tence may contain more than one focus of empathy. Unlike the
constraint on conjoined phrases, this constraint does not disallow the
application of a transformation under certain conditions. It simply
rejects any English sentence that is not internally consistent in the
way empathy is indicated.
The study of constraints in syntax was initiated by Chomsky, Ross,
and others in the mid-1960s. Originally, they concentrated on formal
restrictions on transformations, such as the constraint on conjoined
phrases. In the 1970s, however, many linguists extended the notion of
a constraint to include such functional restrictions as the empathy
constraint. In fact, the study of constraints — with the purpose of
reducing the power of transformational-generative grammar to avoid
the generation of unacceptable sentences — has been the major issue
in syntax during the past fifteen years.
 Summary 127

Summary
Transformational rules apply to deep structures to add, delete, or
change elements. According to the original standard interpretive
theory of transformational-generative grammar and generative seman-
tics, transformational rules do not change the basic meaning of sen-
tences. After all relevant transformations have applied, a surface
structure is produced.
During the past fifteen years, the study of constraints has been a
major issue in syntax. This study of constraints grew because trans-
formational-generative grammars, as originally conceived, generated
many unacceptable sentences. In order to prevent such sentences
from being generated, it is necessary to impose certain constraints on
the power of transformations.

SOURCES FOR FURTHER READING

TRANSFORMATIONAL-GENERATIVE SYNTACTIC THEORY

Akmajian, Adrian, and Heny, Frank. An Introduction to the Princi-
ples of Transformational Syntax. Cambridge, Mass.: MIT Press,
19753
Bach, Emmon. Syntactic Theory. New York: Holt, Rinehart and
Winston, 1974.
Chomsky, Noam. Aspects of the Theory of Syntax. Cambridge,
Mass.: MIT Press, 1965.
. Syntactic Structures. London: Mouton, 1957.
Katz, Jerrold J., and Postal, Paul M. An Integrated Theory of
Linguistic Descriptions. Cambridge, Mass.: MIT Press, 1964.
Lakoff, George. Irregularity in Syntax. New York: Holt, Rinehart
and Winston, 1970.

ENGLISH TRANSFORMATIONS
Burt, Marina K. From Deep to Surface Structure: An Introduction to
Transformational Syntax. New York: Harper & Row, 197A.
Culicover, Peter W. Syntax. New York: Academic Press, 1976.
Jacobs, Roderick A., and Rosenbaum, Peter S. English Transforma-
tional Grammar. Waltham, Mass.: Blaisdell, 1968.
Whitman, Randal L. English and English Linguistics. New York:
Holt, Rinehart and Winston, 1975.

EXERCISES

1. Give a tree diagram for the deep structure of the sentence The op-
erator has been ringing the number. How many times must the FF
rule apply to yield the surface structure?
128 Syntax II

2. Which of the following structures and rules are not permissible?

Why?

ge ia
Avis CDi E

b. S> {Bt+(C)+D, corresponding to

A
ie
B D
c. Lex[V]

farmer

3. Why can the sentence The shooting of the hunters was terrible be
accounted for by a transformational-generative grammar but not
by a structural or traditional grammar?
4. Give a tree diagram for the deep structure of the sentence Hasn’t
Jefferson been painting the house? .
5. Give a tree diagram for the deep structure of the sentence A clown
who gets a laugh is happy. What transformations must apply to
yield the surface structure?
6. How many embedded sentences are permissible in an English sen-
tence?
7. Explain how the constraint on conjoined phrases prevents the sen-
tence *The cats which John likes and the dogs are sick.
 CHAPTER

SEMANTICS I

Semantics may be defined as the study of meaning. As such, it is a dis-

cipline that has fascinated philosophers since the beginning of West-
ern civilization, but it is also a field that has been sadly ignored by
linguists until very recently. For example, the structuralists’ rigid
view of science as absolute objectivity forced them to study language
without reference to meaning, which they believed to be inaccessible
to observation. When Chomsky began the tradition of transforma-
tional-generative grammar in 1957, he recognized the importance of
meaning; but he too was unable to give it a suitable place in linguistic
theory. Only within the past fifteen years has the study of meaning
been approached head-on in modern linguistics.
What might be expected from a theory of linguistic semantics?
Most basically, it should specify the meanings of words and sen-
tences, and explain how they interrelate. This will entail studying the
nature of semantic representations, the relationships among words
and larger constructions such as phrases and sentences, and the ways
in which semantics might be incorporated within grammar. Also,
some traditionally nonlinguistic notions must be considered in order
to understand the conditions under which sentences make sense. Al-
though the goals of linguistic semantics might seem rather ambitious,
the recent work done in this area has provided a vital starting point
for the study of meaning and a well-grounded optimism about its po-
tential.

129:
130 Semantics I

Words and Word Relationships

The concept of a word has long plagued linguists and philosophers;
for words are entities that everyone can identify intuitively but that
are rather slippery to work with. Given the existence of words, the
basic question is what words mean. For example, consider the word
book, v;hich has many varieties of meaning, as indicated by expres-
sions like book of matches, the officer threw the book at him, and
bookkeeper. Even if we restrict ourselves to the most basic sense of
the word book and perhaps define it as a number of bound pages,
usually with a cover, our intuitions as to what a book is will vary. For
instance, what is the minimum number of pages that constitute a
book? One? Two? Eleven? Seventy-nine? And what are the minimum
binding requirements? Glue? A paper clip? Clearly, words can only
be defined in terms of other words, and definitions of words are exact
only to the extent that everyone understands the same thing by them
and uses them in the same way.
Despite the problems in specifying the meaning of a word through a
definition, communication can take place because people do have
some basic idea of what words refer to. In this respect, words have a
denotative_ Ineaning, which is akin to a definition. Book refers to an
object tthat is ‘‘a number of bound pages, usually with a cover’’; this
is its denotative meaning. However, words also have a ute
-4meaning,-which includes varied aspects. For example, the notions 6
an object collected in libraries, a form of print media, and best-seller
all are closely related to the word book, but they have different con-
notations. The connotative meaning of a word may include shades of
feeling and judgment. Gal and woman, for example, have the same
denotative meanings, but their connotations are quite different. A full
account of the meaning of a word must include both its denotative
and connotative meanings.
Philosophers and linguists have also long be interested in certain
relationships among words. We are all familiar with such pairs of
words as sun and son, which have different meanings and are spelled
differently but are pronounced the same. In studying language usage,
pronunciation is more important than spelling, so both words would
be represented phonetically as [san]. This relationship is called
_mophony: two words with different meanings have the same bounds
The term may be extended to apply to words like bank,oa
degree, each of which has several meanings but one pronunciation
and spelling. All the meanings of bank, for example, are referred to by
the single phonetic form [bznk].
A more interesting relationship is. synonymy, in which the meanings
of words are supposedly alike but the pronunciations are different. It
 Dictionaries and Lexicons 131

has often been argued that no two words can have exactly the same
meaning. For example, such words as little and small might be in-
terchangeable in nearly all situations, but individual speakers may in-
sist that there are minor differences. Still, we refer to such pairs as
synonyms because they seem to mean pretty much the same thing.
Antonymy is the relationship that holds between words that are op-
posite in meaning. Although opposites can usually be recognized eas-
ily, the exact nature of antonymy is more subtle. A great many ant-
onyms are pairs of words that represent opposite degrees of some
semantic property. For example, such pairs as warm and cool, hot
and cold, and torrid and frigid each represent approximately the same
distance away from some neutral point along a continuum of climate,
but they lie in opposite directions. However, this property of op-
posites is not the only relationship subsumed by antonymy. Pairs like
buy and sell exhibit a reciprocal relationship, whereas pairs like white
and black exhibit a presence/absence (of color) relationship.
Synonymy and antonymy may be viewed as merely the most ex-
treme cases of relatedness and incompatibility. For example, many
groups of words like golf, putt, club, and caddy exhibit a close-knit
relationship; other groups of words — most obviously, color terms
like red, yellow, green, and blue — exhibit a mutual incompatibility.
Such relationships among words need to be fully characterized in a
study of semantics.
Other word relationships involve units that are smaller or larger
than single words. For example, many words take affixes like pre-,
un-, re-, -able, and -ize to augment their meanings; thus, energy
does not mean the same thing as energize. Also, some words can
enter into relationships with other words to form compounds or
idioms; for example, sidewalk and kick the bucket. Some of the most
common idioms involve the particles on, off, in, and out in construc-
tions with a variety of verbs, particularly in slang expressions like tie
one on, knock it off, sit in, and freak out. Tie one on has a specialized
meaning that goes beyond the meanings of its components.

Dictionaries and Lexicons

The meaning of a word is inherent in it in a subtle way that is difficult

to specify. However, the use of a word is a matter of record; it exists
in sentences that have been uttered. Dictionaries do not give the
meanings of words; they simply indicate the ways in which words are
used. Modern editors of dictionaries generally do not sit down, select
a word, and ask what it means. Instead, they glean sentences in
which that word occurs from the literature of the language, sort the
sentences into groups in which the word appears to be used the same
132 Semantics I

way, and select the best paraphrases to characterize each usage.

From sentences-of-record, some sort of intended meaning is inferred;
and this meaning is further described through paraphrases. Thus, a
good dictionary will give at least one example for each ‘“‘meaning’’ of
a word; the chosen example was actually used (among many others)
to establish that ‘‘meaning.”’
Consider, for example, the sample dictionary entries in Figure 11- _
In addition to examples of usage, this entry provides other helpful in-
formation, such as pronunciation, part of speech, linguistic history,
colloquial usage, and synonyms. From this entry, it is clear that dic-
tionaries provide not only semantic information but-also_ phonological,
syntactic, and sociolinguistic information (that is, social distinctions
among usages).
Much of this information is incorporated by the linguist in a lexicon,
which is a kind of specialized dictionary. The basic differences be-
tween dictionaries and lexicons are as follows. Pictionanes rovid
linguistic | history of the word,,
synonyms, and_e. te
in_ which the word is used. Furthermore, they ssometimes Aenea
among standard, colloquial, archaic, and otherrTusages of a word. Lex-
iconsare arranged according to semantic content or phonological con-
‘tent rather than spelling. “Lexicons may ‘also provide information
about the possible logical relations assumed by an item. For example,
the tah buy assumes four nouns related to it: a buyer, a seller, some
good that is bought, and the money used to obtain the good. Finally,
lexicons must list units other than words; namely, affixes and idioms,
some of which may be listed in dictionaries.
One basic question concerning the lexicon is how items are to be
“‘stored’’ in it. We must be cautious when using the metaphor of
physical storage in describing the lexicon; how lexical items are actu-
ally stored in the mind is a major issue in psycholinguistics and will be
discussed further in Chapter 13. But lexical items might be organized
in two major ways: according to their phonological content or accord-
ing to their semantic content.
As an example of the first method, consider the noun bank. It is
possible to list at least three basic definitions for this term, as follows:

Bank 1 — a place of storage for money or other valuables

Bank 2 — land that rises from the edge of a body of water
Bank 3 — a row of objects
The lexicon might list all three of these definitions under the form
benk. This scheme is an oversimplified approximation of the method
employed by traditional dictionaries, which would group the three
definitions under a single form spelled b-a-n-k. But lexical items may
also be grouped according to their semantic, rather than their phono-
 ¥ ° oF 7
Dictionaries and Lexicons 133

FIGURE 11-1
Sample Dictionary Entries

in-fallible \(‘)in, an+\ adj [ML infallibilis, fr. L in- tin- + LL

fallibilis fallible — more at FALLIBLE] 1 $ not fallible : incapa-
ble of error } UNERRING (~ marksman) (~ ear for pitch in
music) (~ memory) 2 $ not liable to mislead, deceive, or
disappoint $ SURE, CERTAIN, INDUBITABLE <~ remedy) <his
accentis an almost ~ index of his family background and
education —Richard Joseph) (~ scheme for making money)
3: apeanahis of error in defining doctrines touching faith or
morals
SyMN INERRABLE, INERRANT, UNERRING: INFALLIBLE describes
that which is exempt from possibility of error or mistake or
that which has been errorless (no mathematician is os terns
he may make mistakes —A.S.Eddington) <believed in an
infallible Bible —W.W.Sweet) INERRABLE and INERRANT are
erudite synonyms for INFALLIBLE sometimes used in its stead
to escape connotations arising from the discussion of papal
infallibility; the latter may imply that whatever is described has
not so far erred <the Church was ubiquitous, omniscient,
theoretically inerrant and omnicompetent —G.G.Coulton)
{at the moment we lack, in all English-speaking countries, the
inerrant literary sense which gave us the Prayer Book Collects,
often quite as beautiful in translation as in the original Latin —
W.L.Sperry) UNERRING may imply freedom from error coupled
with sureness, reliability, and exactness (an unerring marks-
man) (a man’s language is an unerring index of his nature —
Laurence Binyon) (the unerring scent of the hounds in pursuit
—George Meredith
in-fal-li-ble-ness n 3 INFALLIBILITY
in-fallibly \“+\ adj : without fail : suRELY, CERTAINLY (must
make the proper use of his opportunities or he will ~ perish —
—A.C.McGiffert)
in-fam-a-tory \sn'famo,toré, -tér-\ adj [ML infamatorius, fr.
L infamatus (past part. of infamare) + -orius -ory] archaic
: DEFAMATORY
in-fame \dn'fam\ vt -ED/-1ING/-s [ME enfamen, fr. MF enfamer,
infamer, fr. L infamare, fr. infamis] archaic DEFAME
in-faemize \‘in,fo,miz\ vt -ED/-ING/-s [infamous + -ize]
archaic ; to make infamous $ DEFAME
in-fa-mous \‘infomes\ adj [ME infamis, infamous, fr. L
infamis, fr. in- \in- + -famis (fr. fama fame) — more at FAME]
1: having a reputation of the worst kind : notorious as being
of vicious, contemptible, or criminal character $ DETESTABLE,
ABHORRENT (one of the most ~ spies and bullies of all time —
Time) (~ or traitor) <~ dog has got every vice
except hypocrisy —W.M.Thackeray) 2 3: causing or bringing
infamy ¢: deserving hatred or detestation (~ conduct) (~
vices) (~ treatment of prisoners) (men to whom totalitarian-
ism is ~ —Jerome Frank) (most ~ of quack nostrums — Time)
3: mayne a bad name as being associated with something
disgraceful or detestable (the street outside Newgate had not
obtained one ~ notoriety that has since attached to it —
Charles Dickens) 4: convicted of an offense judged infamous
<~ person) syMN see VICIOUS ‘

Source: By permission. From Webster’s Third New International Dictionary. © 1976

by G. & C. Merriam Co., Publishers of the Merriam-Webster Dictionaries.

logical, similarity. Bank 1 would then be grouped with such items as

safe and vault because all three terms have similar meanings.
Another problem in lexical representation is the multiplicity of
meanings of certain common words such as good and eat. Thus, in
expressions like good government, good apples, good time, and good
upbringing, the meaning of good varies with the meaning of the noun
that it modifies. Similarly, the process of eating varies, depending on
whether one is eating soup, peanuts, or steak or whether one is ‘‘eat-
ing his heart out,’’ or ‘‘eating his words.’’ This is one of the problems
that made the study of meaning seem futile to linguists for many
years. Hard-core philosophers even point out that the ‘‘meaning’’ of a
given chair is different at any point in time because the actual molecu-
lar structure of the chair is constantly being altered. Nevertheless,
134 Semantics I

most linguists feel that meaning must be described, and so they ab-
stract certain basic features inherent in the meanings of words. There-
fore, it may be true that a word like eat will vary in meaning accord-
ing to a wide variety of conditions, but there is a basic sense of eating
that is understood to involve taking matter into the mouth and swal-
lowing it. Hence, one can isolate such basic notions as matter, mouth,
and swallow as essential components of the meaning of the word eat.
In conclusion, most linguists would agree that lexical entries spec-
ify a phonological form for the word, a statement of the syntactic rela-
tions into which it may enter, and its meaning. Many other linguists
would insist that the meaning include information about the item’s
logical relations. For example, the verb buy presupposes four nouns
that are related to the verb: a buyer, a seller, money given to the
seller, and goods received by the buyer. Although these nouns may
not all be expressed in the surface structure of sentences, they are all
logically assumed by the meanings of buy. Finally, it is unwise to
view the lexicon as just a collection of words, because units that are
larger or smaller than words — namely, idioms and affixes — need to
be listed somehow in a lexicon.

The Semantics of Sentences

Even equipped with a complete lexicon of a language, we would still

confront questions of meaning that could not be resolved by the
meanings of words alone. The meanings of words are affected by their
relationships within sentences, and even by relationships among sen-
tences. Among the most important of these relationships are presup-
position, entailment, contradiction, context, anomaly, and ambiguity.
Presupposition refers to the conditions that must be met in order for
the intended meaning of a sentence to be regarded as acceptable. For
example, in the sentence John blamed Mary for doing that, the verb
blame involves the presupposition that the following clause is un-
desirable to John, no matter what doing that may refer to. An ex-
ample of another type of presupposition occurs in the sentence The
Bears lost to the Giants, in which it is presupposed that the Bears
played the Giants. If the Bears did not play the Giants, this sentence
would not be acceptable in the real world.
For many philosophers and linguists, entailment and contradiction
are essential concepts in getting at the meaning of sentences. For in-
stance, the sentence Helen has three nice grandchildren entails the
sentence Helen had children. Entailment means that the meaning of
one sentence is contained in that of another. In other words, the séc-
ond sentence follows logically from thesfirst. If the first sentence is
true, then the second sentence must be true. But now consider the
 The Semantics of Sentences 135

sentences Wilkins does not wish to sell his farm and Wilkins hopes to
sell his farm. The second sentence does not follow logically from the
first; in fact, it contradicts the first sentence. Thus, contradiction
means that if one sentence is true, the other must be false.
Time, place, and context in general also play a part in under-
standing the meaning of sentences. Thus, in the sentence Please take
this garbage can outside, the speaker must be making a request to at
least one other individual. We can also assume that the speaker and
the garbage can are both inside at the time the sentence is uttered
because of the word outside and because the verb take rather than
bring is used. Also, the garbage can must be near the speaker because
this rather than that is used. All of these conditions must be met in
order for the sentence to ‘‘make sense.”’
Sentences_that_violate_possible_semantic relations are said to be
anomalous;, one example is Chomsky’s famous sentence Colorless
green ideas sleep furiously. In this sentence, each successive pair of
words involves a semantic contradiction; that is, colorless green is a
semantic contradiction, as is green ideas, as are ideas sleep and sleep
furiously. There is a place for this type of sentence in poetry and liter-
ature, which indicates that humans often can understand and appreci-
ate such sentences. Thus, it is perfectly all right for Shakespeare to
write that ‘‘all the world’s a stage.’’ In an everyday literal com-
munication — for example, if two students are discussing theories of
cosmology — Shakespeare’s sentence would be anomalous. But in
the literary context, this sentence makes perfect sense; indeed, the
sentence has a greater impact because of the metaphor. Thus, such
cases of anomaly need to be accounted for within a complete theory
of semantics.
Ambiguity involves sentences with more than one normal interpre-
tation. The sentence They are looking for the bank, for example, may
be interpreted three different ways, depending on what bank refers to.
Such cases of ambiguity may be contrasted with the sentence The
shooting of the hunters was terrible, in which the ambiguity arises
because hunters may be either the subject or object of the shooting in
deep structure. Thus, there are two main types of ambiguity: ambigu-
ity caused by the semantic properties of words that are phonetically
identical, and ambiguity that involves the accidental merger of dif-
ferent syntactic structures.

Summary
Semantics may be regarded as the study of meaning. A complete
theory of semantics will specify the meanings and relationships of
words and sentences.
136 Semantics I

Words have both a denotative and a connotative meaning. Rela-

tionships among words include homophony, synonymy, and an-
tonymy; and words may further be analyzed through the techniques
of morphology. Every linguistic approach to semantics includes some
sort of dictionary or lexicon that will list all the affixes, words, and
idioms of a language. The lexicon will specify for each lexical item its
phonological representation, its syntactic properties, and its meaning.
A theory of semantics must also deal with relationships within and be-
tween sentences, and thus a complete theory will incorporate a treat-
ment of presupposition, entailment, contradiction, context, anomaly,
and ambiguity.

SOURCES FOR FURTHER READING

INTRODUCTORY SEMANTICS
Brown, Roger. Words and Things. New York: Macmillan/Free Press,
1968.
Dillon, George L. Introduction to Contemporary Linguistic Seman-
tics. Englewood Cliffs, New Jersey: Prentice-Hall, 1977.
Leech, Geoffrey. Semantics. Hammondsworth: Penguin, 1974.
Ullmann, Stephen. An Introduction to the Science of Meaning. New
York: Barnes and Noble, 1979.

SEMANTIC THEORY
Fodor, Janet Dean. Semantics. Hassocks, England: Harvester Press,
1977.
Kempson, Ruth M. Semantic Theory. Cambridge: Cambridge Univer-
sity Press, 1977.
Steinberg, Danny D., and Jakobovits, Leon A. Semantics: An Inter-
disciplinary Reader in Philosophy, Linguistics and Psychology.
Cambridge: Cambridge University Press, 1971.

EXERCISES

1. How many nouns are presupposed by the use of each of these

verbs in a real-world situation? (a) steal; (b) evaporate; (c) like
2. What facts relating to context are presupposed or implied in the
sentence Who else is appearing on ‘‘The Tonight Show’’ this eve-
ning?
3. If both homophony and ambiguity may be characterized as ‘‘one
form, more than one interpretation,’’ are they really one and the
same relationship?
 f Summary = 137

4. The following sentence indicates that there are other types of am-
biguity besides homophony and structural ambiguity: J dreamed
that I won the race and then I had a beer. In what way is this sen-
tence ambiguous?
 CHA
P-T ER

SEMANTICS I

A complete theory of linguistic semantics should indicate how the

relationships discussed in Chapter 11 fit into grammar as a whole.
Chomsky’s original exposition of transformational-generative gram-
mar said little about meaning, but since then there have been many at-
tempts to incorporate semantics into a transformational-generative
grammar, including interpretive semantics, generative semantics, case
grammar, and relational grammar. These theories suggest different
relationships between syntax and semantics, and we will briefly ex-
amine these differences. We will also examine two very different
approaches to syntax and semantics: the standard predicate calculus,
adapted from logic, which might be a powerful tool for transforma-
tional-generative grammar, and speech-act theory, which in certain
respects seems to challenge the transformational-generative ap-
proach.

Interpretive Semantics
In 1963, Katz and Fodor published ‘‘The Structure of a Semantic
Theory,’’ which marked the first attempt to incorporate a semantic
component within a transformational-generative grammar. Although
their paper was criticized in later years, its basic ideas provided the
impetus for a unified theory of grammar, which we have been calling
the standard interpretive theory. The ideas offered by Katz and Fodor
were originally called interpretive semantics.
138
 Interpretive Semantics 139

FIGURE 12-1
Katz and Fodor Entry for Bachelor
bachelor

noun

(human) (animal)

[who has the ete

first or lowest
ie aca, academic degree] (male)

[who has never [young knight [young fur seal

married] serving under without a mate
the standard of during the
another knight] breeding time]
Source: Jerry A. Fodor/Jerrold J. Katz, The Structure of Language; Reading in the
Philosophy of Language © 1964, p. 496. Reprinted by permission of Prentice-Hall,
Inc., Englewood Cliffs, New Jersey 07632.

THE SEMANTIC COMPONENT

Katz and Fodor proposed that a semantic component has two parts: a
dictionary and a set of projection rules. Dictionary entries provide
part-of-speech information and semantic information about the dif-
ferent aspects of meaning that a lexical item may have. For example,
the Katz and Fodor entry for bachelor would appear as in Figure
12-1. The features enclosed in parentheses are referred to as semantic
markers or, more generally, semantic features. They attempt to char-
acterize the noun bachelor in terms of basic semantic concepts. Se-
mantic markers are also used to define such relations as ambiguity
and anomaly, as well as other relations. The remaining idiosyncratic
aspects of the four meanings of bachelor are contained in distin-
guishers, which are enclosed in brackets.
Thus, according to interpretive semantics, the main components of
lexical items are a part-of-speech marker and semantic features. The
semantic features are binary features such as (+concrete), (living),
(+animate), and (t+human).
These features are not to be considered real English words but ab-
stract concepts given in a metalanguage, which is a precisely defined
language used in an exact discussion of another language. Languages
that are used to program computers, such as FORTRAN and
COBOL, are metalanguages. In theory, no language — including
English — may be used as a metalanguage to describe itself; this fact
accounts for much of the confusion surrounding binary features, since
features such as (human) look very much like English words indeed.
140 Semantics IT

The lexicon will also contain semantic redundancy rules, which rep-
resent the relationship between features. Thus, the rule (+hu-
man)—(+animate) indicates that anything that is (+human) must au-
tomatically be (+animate).
The projection rules are statements about how the different aspects
of the meaning of lexical items may be combined within successively
larger grammatical categories until a reading for the entire sentence ts
producéd> Thus, in the sentence The man hit the colorful ball, the
projection rules would ensure that the particular meaning of colorful
is consistent with the particular meaning of ball. In other words, the
projection rules would exclude the combination in which colorful
means ‘‘red, blue, and so on’’ and ball means ‘“‘a gala affair.’’ Then
the noun phrase the colorful ball is amalgamated with the other con-
stituents of the sentence. For instance, the colorful ball must refer to
a spherical object in order to be consistent with the meaning of the
verb hit. This procedure is continued until an interpretation for the
entire sentence is produced.
To illustrate how the projection rules interact with the insertion of
lexical items in deep structures, consider Figure 12-2, which is the
phrase-structure marker for the sentence The dog chewed the bone
before any lexical items have been inserted. Each insertion of a lex-
ical item into this structure replaces a symbol like Lex[N] with a
complex of features such as (+living) and (—human) associated with a
particular lexical term. The first occurrence of Lex[N] would accept
the lexical item marked as a noun and as (—human), and the second
occurrence of Lex[N] would accept a noun that is (—living). After all
the lexical items have been inserted, the projection rules would then
work upward, summing up the feature paths of these lexical items and
excluding any combinations that involve semantic contradictions. The
projection rules supposedly would eliminate combinations like The
bone chewed the dog in its nonmetaphorical reading.

FIGURE 12-2
Phrase-Structure Marker for The dog chewed the bone.

Det

Lex[Det] Lex[N] Past Lex[V] Lex[Det] Lex[N]

 Interprétive Semantics 141

EXTENDED STANDARD THEORY

Interpretive semantics defines deep” structure as the level at which all
lexical items have been insertedand no other transformations have
yet applied. In earlier versions of the theory, it was claimed that the
deep structure provides all the necessary syntactic information from
which an adequate semantic interpretation can be extracted. But later
versions of interpretive semantics, particularly that of Chomsky
(1971), acknowledged that certain properties of surface structure also
are relevant to semantic interpretation; two of these surface proper-
ties are ‘presupposition and focus.
~Focus, which refers to information not previously shared by
speaker and hearer, may be viewed as being largely determined by
the placement of intonation. Consider the following sentences:
1. Mary left the book on the table.
2. Mary left the book on the table.
The placement of intonation may be seen as contributing to a dif-
ference in meaning between these sentences. In the first sentence,
both speaker and hearer share the information of Mary’s leaving a
book, and the speaker asserts that the book was left on the table, and
not on the chair, bed, or somewhere else; this is the focus, the infor-
mation not previously shared by speaker and hearer. In the second
sentence, both speaker and hearer share the information that Mary
left something on the table, and the speaker asserts that it was the
book, and not the newspaper, pen, or whatever that was left on the
table. In other words, intonation is contributing to meaning; but in-
tonation is not assumed to occur in deep structure, as we saw in
Chapter 7. Thus, a property of the surface structure has contributed
to semantic interpretation.
Word order in surface structure also appears to influence the mean-
ing of sentences in a certain way:
1. It is simple to catch mice in this trap.
2. Mice are simple to catch in this trap.
3. This trap is simple to catch mice in.
Although such sentences ‘are usually considered paraphrases, they
differ with respect to topic; that is, whatever seems to be emphasized
as the principal concern of the sentence.
Examples like these led Chomsky to state that deep structure alone
could not supply the basis for a complete understanding of a sen-
tence’s meaning. Chomsky’s theories have at present changed even
more drastically than at the time he first suggested this. Most stan-
dard interpretivists have agreed with him, and this new view of syn-
tax and semantics is referred to as the extended standard theory.
142 Semantics II

Generative Semantics

Generative semantics developed in the mid-1960s as a reaction to the

Katz and Fodor theory of interpretive semantics. theory
In the of in-
terpretive s semantic com rovides an interpreta-

(ae oa Pa emateremarioe 12-2 afier allsepical tems

h ninse ansformations appl =
ever, in generative semantics, the_semantic component is generative
and the_syntactic_component.is—interpretive;—that_is,
the syntactic
component provides suitable structures for the meanings of sen-
tences.
Semantic representation in generative semantics, then, consists of a
set of generated semantic structures, often called propositions, along
with related information about presupposition and other semantic
properties of the sentence. Generative semanticists claim that there is
no real boundary between syntax and semantics. The ‘‘deepest’’
structures are actually semantic representations; like the syntactic
deep structures of interpretive semantics, these structures are labeled
trees. Transformations relate these deep semantic structures to sur-
face structures. Also, since generative semanticists claim that there is
no level of deep structure at which all lexical items are inserted and
before which any transformations apply, it is possible to have the lex-
ical-insertion rules interspersed among the transformations.

—<=RELATING SEMANTIC STRUCTURE AND SURFACE STRUCTURE

~*’ To clarify these concepts, consider a now-classic example from gen-
> erative semantics. It is claimed that the lexical item kill is derived
from an abstract semantic representation CAUSE BECOME NOT
' ALIVE, as in Figure 12-3. This structure gives the underlying repre-
¥~ sentation for the sentence John killed Fred. Keep in mind that it is not
..» a deep structure as in the standard interpretive theory, but a semantic
\representation containing metalinguistic terms such as CAUSE,
) ALIVE, and so on. Notice, also, two further peculiarities of this
structure. One is the small number of grammatical categories (S, V,
NP), which reflects the claim that only a few basic grammatical
categories exist in underlying structures. NOT, for example, is
treated as a verb. The other is that V appears first under each sen-
tence, which embodies the claim that English hasunderlying
an verb-_
_subject-object word order, Supposedly, this assumption will simplify
the statement of several transformations.
The process by which this semantic structure is turned into a sur-
face structure involves a transformation called predicate raising. This
transformation raises a verb from an embedded sentence into the next
 Pe, “
Generative Semantics 143

FIGURE 12-3
Semantic Representation of John killed Fred.

S
YAS
V NP
ALIVE Fred

“‘highest’’ sentence in the tree diagram, where it is joined to the verb

of that sentence. For example, predicate raising would lift ALIVE out
of the most deeply embedded sentence in Figure 12-3 and join it to
NOT, giving the following structure:

ek een
|
NOT ALIVE Fred
After the predicate-raising transformation applies several times —
that is, on successive cycles — the constituents CAUSE, BECOME,
NOT, and ALIVE are all grouped together as a single verb. At this
point, a lexical-insertion rule may replace this semantic represen-
tation by the actual lexical item kill, giving Kill John Fred. A later
transformation will move John to the front of the sentence, giving
John killed Fred (ignoring the matter cf tense). Thus, a lexical item
has been inserted after several applications of the predicate-raising
transformation, and the concept of deep structure is challenged as a
result.

COMPARING GENERATIVE AND INTERPRETIVE SEMANTICS

Some support for the analysis of generative semantics comes from the
existence of sentences such as John almost caused Fred to die and
John caused Fred to almost die, which represent different meanings
144 Semantics II

of the ambiguous sentence John almost killed Fred. In these sen-

tences, the adverb almost is seen to modify different parts of the
complex verb CAUSE BECOME NOT ALIVE, thus indicating the
source of the ambiguity. In other words, if we do not apply predicate
raising the final time, we get the structure

ea
ioe NP
CAUSE John S

Vv NP

BECOME NOT ALIVE Fred

in which BECOME NOT ALIVE may be replaced by the lexical item

die, and CAUSE may be replaced by the lexical item cause, giving
cause John die Fred. If almost modifies die, we get the sentence John
caused Fred to almost die, rather than John almost caused Fred to
die, in which almost modifies cause. '
This discussion involves some very complex theoretical issues, but
what it amounts to is that the interpretive semanticists view deep
structure as syntactically defined, whereas the generative semanticists
view it as semantically defined.
Both generative semanticists and interpretive semanticists claim
that their respective theories are not intended to be models of sen-
tence production or interpretation. Thus, in generative semantics one
does not ‘‘start’’ with a semantic representation, “‘then’’ give it a syn-
tactic structure, and ‘‘then’’ give it a pronunciation. For this reason
some interpretive semanticists claim that the debate between the two
theories is meaningless: that generative semantics is interpretive se-
mantics. Still, generative semanticists do challenge the centrality of
syntax and the concept of deep structure proposed by interpretive
semantics.

Case Grammar and Relational Grammar

Another theory of syntax and semantics that has enjoyed a good deal
of popularity is Charles Fillmore’s case grammar, first presented in
‘‘The Case for Case’’ in 1968. The form of a deep structure in case
grammar is radically different from that of interpretive semantics and
somewhat different from that of generative semantics. In case gram-
mar, a sentence consists of two constituents, modality and proposi-
‘tion. A typical tree diagram would look something like the following:
 #

Case Grammar and Relational Grammar 145

Modality Proposition
| o~
Verb Case A Case B Case C CaseD

In this analysis, udes such information as

tense, mood,il,negation, and aspect. A Proposition involves a verb and
one or more noun phrases that are in a tenseless set of case Tela-
tionships to the verb.
Fillmore identified several cases, such as Agent, Instrument, Ob-
ject, Experiencer, Benefactor, and so on; each is defined in terms of
its relationship to the verb. A specific verb can then be described in
terms of the cases that it takes. For example, the verb give takes an
agent (the person who performs the act of giving), an object (the thing
given), and a goal (the person who acts as the destination of the thing
given).
These cases should not be confused with the traditional cases, such
aS nominative, possessive, and objective. For example, the Object
case designates the NP that is being affected or talked about. Thus, in
the sentence The book lies on the table, the book is an NP in the Ob-
ject case, despite the fact that it is traditionally considered to be the
subject of the sentence and thus in the nominative case.
Each case is associated with a case marker, as well as noun phrase,
as follows:
Case A

Case Marker NP

In English, the case marker is often expressed as a preposition. The

case marker for an Instrument, for example, is the preposition with,
and that for a Goal is the preposition to. Not all cases have explicit
case markers; in English, the Object case has no marker, and the
Agent case is marked by the preposition by only in passive sentences.
Thus, a deep structure for the sentence John gave a book to Mary
might appear something like
S

Modality Proposition
a, ae
(declarative) Verb ac Object Goal

as
(past)

give
fe pf
NP:

byJohn
NP
IS

Qa book
NPi

to Mary
146 Semantics II

Transformations would then apply, in this instance moving the Agent

into a subject position and giving the verb an appropriate form rela-
tive to the specifications under Modality.

ADVANTAGES OF CASE GRAMMAR

In addition to giving a more explicitly semantic basis to deep struc-
ture, case grammar has the capacity to account for certain facets of
English grammar that are not easily handled within the standard in-
terpretive theory of transformational-generative grammar. For in-
stance, i i k_may occur in a variety of ically
different sentences, such as
—_——~.
0

1. John broke the glass with the hammer.

2. The hammer broke the glass.
3. The glass broke.

Case grammarians argue that the deep structures of these three sen-
tences are basically similar, as they could not be in the standard in-
terpretive theory. These deep structures are shown in Figure 12—4. In
each structure, break is specified as a verb that takes an Agent, an

not be realized. If there is an Agent, the transformational rules au-

tomatically move it into subject position and by is not realized. If
there is no Agent but there is an Instrument, then the Instrument is
automatically moved into subject position and with is not realized.
Finally, if there is neither an Agent nor an Instrument, then the Ob-
ject is moved into subject position. Thus, in case grammar, these sen-
tences can be generated from similar deep structures.

RELATIONAL GRAMMAR
Another approach to grammar has been developed by David Perlmut-
ter, Paul Postal, Edward Keenan, and others. It is called relational
grammar because it is concerned with grammatical relations like sub-
ject and direct object. In the standard interpretive theory, such no-
tions were defined in terms of syntactic structures; for example, the
subject was defined to be the highest NP under S in the tree diagram.
However, in relational grammar, subject, direct object, and other
relations are undefined notions that are basic to the description of lan-
guage. Since a sentence in relational grammar is defined as a verb and
one or more NPs bearing relation(s) to it, there is an immediate simi-
larity between relational grammar and case grammar, although gram-
matical relations and case relations are not to be confused. Neverthe-
less, some linguists even view case grammar as a variant of relational
grammar.
 /.
Case Grammar and Relational Grammar 147

FIGURE 12-4

1. John broke the glass with the hammer.

OE ATG VWI I
AT
Proposition

Verb Agent Instrument Object

NP NP
| eet pre
/*

break © John with the hammer @ the glass

2. The hammer broke the glass.

Sees Proposition
ad ~
Verb Agent Instrument Object

[se NP [Xe
| re HN
break © @ © the hammer © the glass

3. The glass broke.

Proposition
eaAgent
Verb
ek a ere SeObject
Instrument

[re /\ [Se
| La
break © @O @ © © the glass

Central to relational grammar is the concept of the relational hier-

archy:
Subject > Direct object > Indirect object > Nonterms.
The relational hierarchy arranges relations according to their impor-
tance to grammatical processes; for example, no transformation may
apply more freely for objects than for subjects. The symbol > may
thus be read as ‘‘outranks.”’ In this theory, subject, direct object, and
to-a
indirect objects_te
arerefered rms;_oth
relations er
(indicating
Jocatlon. meals, 20 Soon) arecalled nonterms.
The relational hierarchy is useful in the statement of transforma-
tions in relational grammar. Passive, for instance, may be defined as a
tule that_promotes the direct _object to subject. In the sentence The
148 Semantics I

ball was hit by John, the NP the ball has been promoted along the
relational hierarchy from direct object to subject. This aspect of the
passive transformation is assumed to hold for all languages, but
the grammar of each particular language must include other idiosyn-
cratic details. In English, these details include morphological change
of the verb, change in word order, and the insertion of by.
Relational grammar is a relatively new approach to grammar, and
its implications are still being considered. It is actually an outgrowth
of generative semantics, which at present has developed into several
subtheories.

Linguistics and Logic

Throughout the first decade of transformational-generative grammar,
semantics received little attention. However, linguists were able to
formulate a remarkably improved theory of grammar based on the
syntactic concepts of phrase-structure rules, deep structure, transfor-
mational rules, and surface structure. Meanwhile, philosophers and
logicians had long been concerned with semantics and had developed
logical systems to deal with semantics. They could not have hoped to
capture all the richness of the syntactic relations evidenced in the
world’s languages with their systems. As a result, it was only natural
that they join forces with the linguists.
The joint efforts of linguists and logicians during the past decade
have by no means solved the problems of syntax and semantics, and
attempts to merge logical systems and linguistic descriptions have
been extremely diverse. Nevertheless, some real benefits from these
efforts seem likely to emerge in the near future. In order to appreciate
the work now being done, a basic knowledge of a specific type of
logic is needed.
The logical systems that have been applied to linguistics are many,
but they typically are modifications of the standard predicate calculus,
which is a system for representing logical relations. The standard
predicate calculus is a universal system of syntax and semantics, and
thus it is valuable to the transformational-generative theory of gram-
mar, which holds that all human languages have essentially the same
underlying characteristics. Sentences (or statements, as they are
called in the standard predicate calculus) are of three types: atomic
statements, quantified statements, and combined statements.
Atomic statements, such as The dog chews the bone, are expressed
by meansof terms and predicatees. Te ms _are the things that are
talked about in statem riables, which ate
usually represented by | - 1 . In the analy-
sis of actual sentences, these variables are replaced by specifics like
 Linguistics and Logic 149

dog and bone. Predicates expr I f terms or the rela-

tions between terms; they are usually represented by capital letters
like F, G, and H. Each predicate in a statement has one or more
terms related to it. Thus, in the statement John laughs, laughs is a
one-term predicate; and the expression F(x), in which x is restricted
to John and F stands for the predicate laughs, captures the traditional
subject-predicate distinction. In the standard predicate calculus, pred-
icates can include common nouns when the common noun represents
a property predicated of some particular term. For example, in the
statement Jane is a girl_Jane is aterm and girlis a predicate, because
the word girl says something about Jane. For a statement like The
dog chews the bone, the representation in the standard predicate
calculus would be F(x,y) where x is restricted to the dog, y is re-
stricted to the bone, and F refers to chews. Here, the representation
~F@y) captures the subject-verb-object relation. A
Quantified netherTHEY
contain all pagible uence orera

able terms or to onl f thei jable terms. In

English, some words that are quantifiers are few, many, some, each,
and every. Here are two quantified statements, which represent the
sentences Everyone smiles and Someone smiles, respectively.

1. (Wx)F(x), F=smiles allx, x smiles

2. (Ax)F(x), F=smiles onnr eety thot x Smiles

Statement 1 may be read ‘‘For all x, x smiles.”” The symboLV repre-

sents the universal quantifier; it indicates that whatever x is, x smiles.
Statement 2 may be read ‘‘There is an x such that x smiles.” The
symbol_3_4 represents the existential quantifier; it_indicates only tha
there isafleast one xthat smiles.
The value of quantifiers is that they allow a certain type of ambigu-
ity to be resolved. Consider the sentence Someone smiles and some-
one laughs. This sentence is ambiguous: is there one person who both
smiles and laughs, or is there one person who smiles and another per-
son who laughs? In the standard predicate calculus, this sentence may
be given the two appropriate interpretations as follows, the first for
one person who both smiles and laughs and the second for one person
smiling and another laughing: .

(Ax)F@)G(), F=smiles and G = laughs

(Ax)(Ay)F(x)G(y), F =smiles and G =laughs

As in natural languages, the standard predicate calculus also allows

for combined statements. The symbols used to represent the possible
combinations are
150 Semantics II

~ negation; read ‘‘not’’

& conjunction; read ‘‘and’’
V disjunction; read ‘“‘either. ..or..
9 conditional; read ‘“‘if. . . then. .
9°

These operations allow the logical structure of sentences like the fol-
lowing to be described:
1. Tom doesn’t smile.
2. Tom smiles and laughs.
3. Tom smiles or laughs.
4. If Tom laughs, then he smiles.
Logical relations between sentences also can be described; for ex-
ample, sentence 2 contradicts sentence | and entails Tom smiles. In
general, if two or more true statements are combined, then the con-
junction must be true. Similarly, if a statement is true, then its nega-
tion is false, and vice versa.
With the minimal apparatus described here,- the standard predicate
calculus is able to generate an infinite number of sentences. Many
linguists have thus regarded the standard predicate calculus as a can-
didate for the underlying representation of sentences in natural lan-
guages, and they assume that transformations apply to such represen-
tations to..yield surface structures. But, of course, the standard
predicate calculus, as outlined here, cannot describe natural lan-
guages; it is useful for only a certain type of declarative sentences and
must be modified to describe questions, imperatives, and many other
types of sentences that characterize natural languages. Such modifica-
tion has been the prime objective of linguists working in logic during
the past ten years.

Speech Acts
Still another approach to meaning during the past decade has been the
theory of speech acts. This theory views semantics in the larger con-
text of human communication and establishes the speech act, rather
than the sentence, as the basic unit of human communication.
There are three types of speech acts: locutionary, illocutionary, and
perlocutionary acts. The Jocutionary act is the actual atta of a sen-
tence with a particular meaning; the illocutionary act is intent tha
the speaker has in uttering the sentence; and the perlocutionary act is
the result achieved in uttering the sentence. For instance, the actual
uttering of the sentence Close the door! along with its particular mean-
ing is a locutionary act. The request or order to close the door that
the speaker is making is an illocutionary act. The result of someone’s
actually closing the door is a perlocutionary act.
 Speech Acts 151

Linguists generally have been most concerned with describing illo-

cutionary acts, and Searle (1975) has isolated five categories of ‘Such
acts, as follows:

1. Representatives — acts that represent.a.particular state of affairs;

for example, statements, descriptions, assertions, and so on.
2. Directives — acts that have the intent_of arer to do
something; for example, “orders, requests, instructions, and so on.
3. Commissives— acts that commit the spea some future action,
for example,aS. threats, offers, and so on.
“my

2 ¢ speaker about
sara 1a RTT for example, thanks,apologies, welcomes,
and so on.
2 Declaratives — acts that bring abou nce that corre-
said; for example, declaring war, pro:
bamnae a couple husband sit wife, firing an employee, and so
on.

Each type of illocutionary act is associated with certain types of

deep structures. Thus, the representatives may take either of the
forms
I VERB (that)+S
I VERB NP, + NP, be + Predicate

The first of these forms is assumed to underlie sentences like I claim

that the Smiths have two children and I predict that Jones will be
President, whereas the second form underlies sentences like J diag-
nose this ailment as the common cold and I classify this insect as a
spider. Even such a simple sentence as J own a car might be assumed
to have an underlying structure approximately equivalent to ‘I say to
you that I own a car,’’ because language is used for communication
and thus involves a speaker — J or we — and a hearer — you (singu-
lar) or you (plural).
This analysis provides some insight into the various types of sen-
tences in natural languages: imperatives (‘‘I order you to Close the
door!’ or ‘*‘I request that you Close the door!’’), questions (“‘I ask
you How far is the nearest town?’’), and so on. The deepest underly-
ing clauses, such as J order you and I ask you, may be deleted in the
surface structure. The verbs in these deepest underlying clauses
(order, request, ask, and so on) contain the essence of the illocu-
tionary act.
Some linguists have suggested that the very basic distinction be-
tween competence and performance may be endangered by a theory
of speech acts. Because speech-act semantics regards the meaning of
sentences largely in terms of the conditions under which they may be
used, it seems that competence has become all of performance. In
152. Semantics II

other words, within such a theory, it becomes difficult to draw a line

between linguistic knowledge and use. This issue is one of the most
important of many that have arisen in the attempt to integrate seman-
tics and transformational-generative grammar.

Summary
Among the theories of syntax and semantics that have enjoyed a great
deal of popularity in recent years are interpretive semantics, genera-
tive semantics, case grammar, and relational grammar. The first
attempt to deal with semantics in the framework of transformational-
generative grammar was the Katz and Fodor theory, later called in-
terpretive semantics, which has been modified-in_several
ways since
1963. One important innovation was Chomsky’s declaration that sur-
face structure also may play a role in determining the meanings of
. sentences. The theory of interpretive semantics has been challenged
by generative semantics, which argues against: ality of syntax
and the level of syntacti¢ deep structure. Cas@ grammar’s most signif-
icant contribution has been a clear, explanatory statement of the logi-
cal case relationships that hold between verb and noun phrases,
whereas relational grammar has proposed that such grammatical rela-
tions as subject, direct object, and indirect object are of primary im-
portance in describing language.
Two other important approaches to semantics are the linguistic
adaptations of logical systems and the theory of speech acts. Logic
provides a universal system of syntax and semantics in the standard
predicate calculus, which can express logical sentence relationships
for an infinite number of sentences. The theory of speech acts views
semantics in the larger context of human communication.«In deter-
mining the different forms of deep structures for various types of
illocutionary acts, speech-act semantics assumes that all sentences
have an abstract underlying clause that contains references to the
speaker, the hearer, and the illocutionary force of the act.

SOURCES FOR FURTHER READING

SEMANTIC THEORY
Chafe, Wallace L. Meaning and the Structure of Language. Chicago:
University of Chicago Press, 1970.
Fillmore, Charles F. ‘‘The Case for Case.’? In Emmon Bach and
Robert Harms, eds., Universals of Linguistic Theory. New York:
Holt, Rinehart and Winston, 1968. (Case grammar)
Fodor, Janet Dean. Semantics. Hassocks, England: Harvester Press,
1977.
 Summary 153

Fodor, Jerry A., and Katz, Jerrold J. The Structure of Language.

Englewood Cliffs, New Jersey: Prentice-Hall, 1964.
Katz, Jerrold J. Semantic Theory. New York: Harper & Row, 1972.
Kempson, Ruth M. Semantic Theory. Cambridge: Cambridge Univer-
sity Press, 1977.
McCawley, James. “‘The Role of Semantics in a Grammar.’’ In
Emmon Bach and Robert Harms, eds., Universals of Linguistic
Theory. New York: Holt, Rinehart and Winston, 1968. (Generative
semantics)
Searle, John R. Expression and Meaning. London: Cambridge Uni-
versity Press, 1979.
Steinberg, Danny D., and Jakobovits, Leon A. Semantics: An Inter-
disciplinary Reader in Philosophy, Linguistics, and Psychology.
Cambridge: Cambridge University Press, 1971.

LOGIC
Keenan, Edward L. “‘Lijnguistics and Logic.’’ In Theo Vennemann
and Renate Bartsch, eds., Linguistics and Neighboring Disciplines.
Amsterdam: North-Holland, 1975.
Leblanc, Hughes, and Wisdom, William A. Deductive Logic. 2d ed.
Boston: Allyn and Bacon, 1976.

SPEECH ACTS
Austin, John L. How To Do Things with Words. Cambridge, Mass.:
Harvard University Press, 1962. :
Sadock, Jerrold M. Toward a Linguistic Theory of Speech Acts. New
York: Academic Press, 1974.
Searle, John. ‘‘Linguistics and the Philosophy of Language.’’ In
Theo Vennemann and Renate Bartsch, eds., Linguistics and
Neighboring Disciplines. Amsterdam: North-Holland, 1975.
. Speech Acts. London: Cambridge University Press, 1969.

EXERCISES

1. Write a semantic redundancy rule that shows the relationships be-

tween the features (+ woman) and (+ child-bearing).
2. Despite the fact that interpretive semantics, generative semantics,
case grammar, and relational grammar are distinct theories of syn-
tax and semantics, they are all similar in certain respects. List as
many of these similarities as you can.
3. Interpret the following sentence, which is expressed in the stan-
dard predicate calculus: (Wx) (Ay) F(y,x), where x refers to things,
y refers to God, and F(y,x) expresses that ‘‘y created x.”’
154 Semantics II

4. If p is a true statement and gq is a false statement, is the following

statement true or false? (pVq) & ~q. If necessary, think of p and q
as actual English sentences.
5. What would be the locutionary, illocutionary, and perlocutionary
acts involved for the sentence What time is it? in its literal in-
terpretation?
 II
LINGUISTICS:
A
BROA DER
VIEW
 ”
y

Ad ele

®
»
af soe i .
2 gee ether
=
hs
gine gen og* 7 72
a
7
: _
7
ety 3%
- <p Love
a
i %
- a
-
"

4»
:

:
—_

i
i
¥
!
et =

- ied .

7
4
Lad a
:

dy a
*
pa | i’ .
:@ Ow
_,

7
|
f a
x
#1
t
ai
aL i

‘
<
na

ae ‘
_
ni i }
ca .

r ——
=
=
-
a ae
’
’
7 aie
S 7
7
_
_ ¥ "s
i)
’ =v.
f
Pa
7 Poe _

*
-

é a2 ome a
>

7
 CHAPTER

PSYCHOLINGUISTICS

In Language and Mind, Noam Chomsky (1968) proposed that linguis-

tics might best be considered a branch of cognitive psychology. This
suggestion might seem rather startling at first; linguists and psycholo-
gists, after all, have worked together closely for only the past twenty
years. Nevertheless, linguistics is ultimately dependent on psychol-
ogy. Language is a learned ability, and only through a study of psy-
cholinguistcs can we ever hope to understand it.
As a preliminary oversimplification, we may say that psycholinguis-
tics is concerned with the acquisition, perception, and production of
language. Although psycholinguistic research has been increasing at a
rapid pace, many language processes are still not understood well, if
at all. If we are to have a thorough understanding of language, psy-
cholinguists must develop (1) a verifiable theory of the mental pro-
cesses and strategies involved in understanding speech; (2) a specifi-
cation of how these processes and strategies relate to the formal
theory of grammar; and (3) a theory of language acquisition — one
that will be consistent with what we know about the universal aspects
of language. In addition, like linguists, psycholinguists are searching
for a better understanding of what semantic representations are like.
The development of psycholinguistics is critical for linguistics; for
psycholinguists test the psychological reality of the theoretical con-
structs proposed by linguists. In the next two sections, the issue of
the psychological reality of the constructs of transformational-genera-

157.
a 158 Psycholinguistics

tive grammar is briefly explored. In the remaining sections of the

chapter, the primary subject matter of psycholinguistics is introduced
and some of the chief findings in the field are summarized.

The Psychological Reality of Transformations and Constituent

Structure
For the purposes of this section, a review of the standard interpretive
theory of transformational-generative grammar, as schematized in
Figure 13-1, will be helpful. The phrase-structure rules of the syntac-
tic component generate a set of structures that identify grammatical
relations and provide a syntax for the sentence. The lexicon, which
resembles a dictionary, provides specific formatives, or lexical items,
for these structures. This set of syntactic structures enriched with lex-
ical items defines the level of deep structure, to which the set of trans-
formational rules applies to yield the surface structure. The semantic
component acts on the deep structure to supply a meaning for the
sentence, and the phonological component acts on the surface struc-
ture to supply a phonetic representation.
It is tempting to assume that this model of a grammar actually

FIGURE 13-1
The Standard Interpretive Theory of
Transformational-Generative Grammar

(Phrase-
Structure SEMANTIC
Rules Deep Structure . COMPONENT
and ee rlral
Lexicon)

AS
<n
AZ
SSeS

o
O
M
P (Transfor- PHONOLOGICAL
O mational Surface Structure COMPONENT
N Rules) oe
eT (Sound)
E
N
ik
 Transformations and Constituent Structure 159 ¥

describes the human process of creating sentences, but is this the

case? And if it is, do we assume that the process of understanding, or
decoding, sentences is a transformational-generative grammar applied
in reverse? To answer these questions, we must return to the distinc-
tion between competence and performance introduced in Chapter 9.
The standard interpretive theory of transformational-generative
grammar is a model of competence: it attempts to describe and ex-
plain what an ideal speaker/hearer knows about his or her language.
As such, it is not intended to describe a process (how sentences are
produced or interpreted), and it does not entail any concept of di-
rectionality. In contrast, psycholinguistics is more concerned with
performance and the strategies or heuristics that humans employ to
utilize their competence. Still, many psycholinguists in the early 1960s
were inspired by the elegance and popularity of transformational-
generative grammar — to the extent that they regarded it as a pos-
sible model of how human language is produced and perceived.

TRANSFORMATIONS
In the early days of transformational-generative grammar, a great deal
of emphasis was placed on the notion of a kernel sentence — a sen-
tence that is simple, affirmative, active, and declarative. It was as-
sumed that such sentences are basic in the sense that other families of
sentences like questions, negatives, and passives are derived from
them by transformations. But later studies in transformational-genera-
tive grammar resulted in the standard theory, which claimed that neg-
ative, question, and passive sentences are triggered by underlying
markers in deep structure and that kernel sentences, therefore, have
no privileged status. Early psycholinguistic experiments, however,
employed the concept that kernel sentences are basic.
The results of the earliest psycholinguistic experiments’ on the
relationships between sentence types seemed to indicate that the
transformational operations involved in deriving negatives, questions,
and passives from kernel sentences corresponded to psychological
reality. To illustrate, consider the following examples of a kernel
sentence (sentence 1) and its negative (2), passive (3), and negative-
passive (4) transforms:

1. John hits the ball.

2. John does not hit the bail.
3. The ball is hit by John.
4. The ball is not hit by John.

14 summary of many of the experiments described in this section can be found in

Fodor, Bever, and Garrett (1974).
{ 160 Psycholinguistics

If one adopts the early model of transformational-generative gram-

mar, then relating sentences 2 and 3 involves two transformations: an
undoing or reversal of the negative transformation and an application
of the passive transformation. But relating sentences 2 and 4 involves
only the passive transformation. Thus, some experimenters reasoned,
if the transformations are real, it should be more difficult — and take
longer — to relate sentences 2 and 3 than to relate sentences 2 and 4.
And, indeed, early experiments found this to be the case. For ex-
ample, in the first experiment of this nature, subjects were presented
with a sentence like 2 and were asked to locate a transform like sen-
tence 3 from a list of sentences; search times were recorded and com-
pared. Experimenters found that subjects took longer to relate sen-
tences like 2 and 3 than sentences like 2 and 4. Therefore, the early
model of transformational-generative grammar appeared to be consis-
tent with the experimental results.
But if one adopts the standard theory, the conversion of sentence 2
to sentence 4 appears to involve three transformations: one must first
perform a reversal of the negative transformation to create a structure
to which the passive transformation can apply, namely, John hits the
ball. Then one applies the passive and negative transformations. Oth-
erwise, if one applied the passive transformation directly to sentence
2, one would generate the ungrammatical string * The ball does not be
hit by John. Relating sentences 2 and 4 thus involves more transfor-
mations than relating sentences 2 and 3, which seems to conflict with
the experimental results. But sentences 2 and 4, which differ by three
transformations, can be related more easily than sentences 2 and 3,
which differ by only two transformations, because they share an un-
derlying structural marker: Negative (Neg).
The results of these experiments do not necessarily confirm the
psychological reality of transformations; instead, they confirm the
psychological reality of particular sentence structures, such as nega-
tives, questions, and passives, as emphasized in the standard theory.
The early experiments were flawed, however, in that certain impor-
tant factors, such as the length of the sentences, were ignored. Fur-
thermore, only an extremely limited number of sentence types were
considered. No one has yet devised an experiment that can conclu-
sively demonstrate the psychological reality of transformational
operations, and optimism about achieving this goal has waned consid-
erably since the early 1960s.

CONSTITUENTS
The issue of the reality of constituent structure — the grammatical re-
lationships of words that fit together in a unit — deserves a closer look.
The most famous experiments designed to test the psychological real-
 Phonological and Semantic Representations

ity of constituent structure are the click experiments, in which sub-

jects hear a sentence with a click superimposed somewhere within it
and are asked to report the exact location of the click within the sen-
tence. In one of these experiments, a complex sentence was recorded
with clicks variably placed before, after, and at the major constituent
break. For example, in the sentence That Rex is a prude is acknowl-
edged by all, the main constituent break occurs after prude; a click
was produced sometimes before, sometimes after, and sometimes at
this main break. Errors in locating the clicks were greatest when the
clicks did not appear at the major constituent break, and the subjects
tended to hear the clicks as occurring at the major break. It has also
been shown that subjects have the most difficulty locating clicks ac-
curately when the click occurs just before the major clausal break.
The subjects’ tendency to hear clicks (better) at the major clausal
boundaries, thus resisting any interruption in the processing of the
sentence’s major units, supports the idea that sentences are perceived
in terms of their constituent structure. Although certain aspects of the
click experiments have been challenged, these experiments are gener-
ally regarded as providing support for the psychological reality of
constituents.
A simpler experiment showing the psychological reality of constitu-
ents consists of asking a linguistically naive speaker to group isolated
words of a sentence according to their relatedness. Thus, given the
sentence Ballplayers require quick instincts, and the array
INSTINCTS QUICK
REQUIRE
virtually everyone will group quick and instincts as a constituent,
rather than require and quick. Although such an experiment does not
explain how speakers use constituent structure in processing lan-
guage, it indicates that speakers conceptually organize language into
constituents.

The Psychological Reality of Phonological and

Semantic Representations
Psycholinguistic studies of grammar also have dealt with the psycho-
logical reality of phonological and semantic representations. In fact,
the psychological (that is, perceptual) reality of the phoneme was
assumed by many linguists even before the birth of psycholinguis-
tics and transformational-generative grammar. Sapir and others dem-
onstrated that when a person learns a foreign language, he or she
perceives that language in terms of the phonemic structure of the
A 162 Psycholinguistics

person’s own language. The modern transformational-generative

conception of the phoneme as a bundle of distinctive features, how-
ever, led to experiments that tested the psychological reality of the
features of which phonological representations are composed.

DISTINCTIVE FEATURES AND MARKEDNESS

In an experiment designed to explore the perception of distinctive
features, Greenberg and Jenkins (1964) asked subjects to judge the
similarity of paired comparisons involving the following English
segments: /p/, /t/, /k/, /b/, /d/ and /g/. For instance, a subject was asked
to assign a rating to the ‘‘distance’’ between the initial segments of
the pairs [pa] and [ba], [ga] and [pa], and so on. The data were then
correlated with the distinctive features that constitute the segments.
Greenberg and Jenkins found that sounds differing in only one dis-
tinctive feature were regularly judged to be more similar than sounds
differing in two distinctive features. The data also reflected the basic
facts of articulation: labials were perceived to be more similar to al-
veolars than alveolars were to velars, and alveolars were perceived to
be more similar to velars than labials were to velars. -The most impor-
tant feature in judgments of similarity is voicing: sounds that agree in
the feature [+ voice] were regularly judged to be more similar than
sounds agreeing in the feature [—voice]. This fact supports the
greater psychological weight of marked, as opposed to unmarked,
features.
The marked/unmarked distinction (described on pages 91-92) ap-
pears to have psychological significance. It can help to characterize
the semantic difference between the members of such word pairs as
long/short and horse/mare. In the first pair, the unmarked unit is long
because the normal way of measuring things is in terms of length, as
exemplified by such sentences as The bridge is one hundred yards
long. In the second pair, the unmarked unit is horse because the term
horse may refer to either a male or female animal, but the term mare
specifies the sex (female) of the animal. There is evidence from other
experiments that the presence of marked adjectives (rather than un-
marked adjectives) causes added difficulty in the understanding of a
sentence.

THE LEXICON AND MEMORY

One of the major experiments dealing with the psychological reality of
phonological and semantic representations was Brown and McNeill’s
(1966) investigation of the tip-of-the-tongue phenomenon; that is, the
unsettling state of mind that occurs when we have difficulty thinking
of a particular word that we are certain we know. In the experiment,
a list of definitions of uncommon words were read until one elicited
 Phonological and Semantic Representations

the tip-of-the-tongue state. When subjects acknowledged that they

were in such a state, their guesses at the correct word were noted.
The subjects’ guesses resembled the desired word either in sound
or meaning. Thus, if the desired word were sextant (a navigational in-
strument used in measuring angular distances, especially the altitude
of sun, moon, and stars at sea), some typical guesses would be astro-
labe and compass or two-syllable words such as secant and sextet.
The nature of these erroneous guesses provides some indication of
how words and definitions are stored in long-term memory.
The fact that the subjects produced words similar in meaning to the
desired word implies that the semantic representations of words inter-
sect to some degree and are stored together in memory. It would be
compatible with the standard interpretive theory of transformational-
generative grammar that what these words have in common are se-
mantic markers or features, as discussed in Chapter 12. Thus, in
searching for the correct word, one looks for words that share
markers like (navigation), (instrument), and (used in geometry). As-
sociated with this set of markers is some sort of phonological and or-
thographic tag that is partially specified; for example, se t. If the
desired word cannot be located from its semantic markers and its cor-
responding tag, another search is made for words with similar tags.
Actually, the tags are probably fully specified, but recognition is
weakest for their middle portions. .
Brown and McNeill’s experiment also revealed that subjects often
were able to guess the number of syllables in the desired word, the
location of the word’s primary stress, and the beginning and end let-
ters of the word. These data support the notion of phonological
storage in the brain.
Of course, the experiments discussed in this and the previous sec-
tions do not demonstrate how humans use language; but they do
provide some insight into the language process, and they permit a
comparison of contemporary linguistic models. In the case of
transformational-generative grammar, experiments indicate that some
of its units of representation do, in fact, have psychological signifi-
cance. However, there is no solid evidence that the process by which
we understand and produce sentences is indeed the set of rules pro-
posed by transformational-generative grammar.

OTHER APPROACHES
By no means has all psycholinguistic research been dedicated to as-
sessing the psychological reality of transformational-generative gram-
mar. One other line of research is the quantificational approach to
semantics, which originally was closely associated with behaviorism,
a theory of psychology. The quantificational approach attempts to
164. Psycholinguistics a

quantify and graph semantic relations as expressed by speakers. One

aspect of this approach is the concept of a semantic space: words are
plotted as points along basic axes, representing such dimensions as
power and affect, and the words that lie closest to each other in the
space are regarded as being most similar semantically. A common
criticism of this approach is that it is grounded in the emotional re-
sponses of speakers and does not really represent the content of
words.
Other approaches to semantics exist as well, and these have en-
joyed varying degrees of success. Still, no approach can claim to pro-
vide a complete understanding of semantic representation and the use
of meaning, and this area remains one of the most inaccessible to
psycholinguistic research.

Language Acquisition
Learning a first language is about as natural to humans as walking
upright. Humans have been able to devise impressive linguistic mod-
els that describe and attempt to explain their knowledge of language.
In light of these facts, it is rather disheartening to admit how little we
understand about the origin of language and the nature of language
acquisition. About the origin of language we know very little. Our
earliest records of language reveal no such thing as a primitive lan-
guage; instead, they indicate that language has always been as com-
plex as it is today. All we can say is that the evolution of language
was intimately related to the evolution of the human mind. With
regard to language acquisition, we have access to more data and can
perform psycholinguistic experiments with young children. As a re-
sult, it is possible to provide the following account of the typical
progression by which a child learns a language.

PHONOLOGICAL ACQUISITION
By the age of six months, the child has entered a babbling period, or
prelanguage state, in which almost any sound can conceivably be
produced. This babbling period allows the child to develop articula-
tory prowess. By the end of the first year, the rudiments of a phono-
logical system are evident. Normally, the first vowel produced with
regularity is a low, somewhat fronted vowel such as [a] or (al: and
the first consonant is generally a bilabial stop, such as [p] or [b]. The
acquisition of additional sounds continues with the development of a
nasal consonant in opposition to the oral consonant. This explains
why mama and papa are early occurrences in English as well as in
many other languages. However, these ‘‘words’’ are not originally
uttered with the meaning ‘‘mother’’ and ‘‘father.’’ They may simply
 ¢ A

Language Acquisition

indicate some physical or emotional need of the child and are only
associated with ‘‘mother’’ and ‘‘father’’ through the parents’ encour-
agement. Afterwards, a word like mama can be used to mean any-
thing from ‘‘There’s mama’’ to ‘‘Mama, I’m hungry.’’ Throughout the
entire period of phonological acquisition, contrasts are perceived that
cannot be produced. For example, the child may acknowledge the
difference between papa and baba, while only being able to produce
baba.

THE TWO-WORD STAGE

Naturally, the child will use his or her first words to describe the im-
mediate environment, including such objects as a ball, milk, and dog,
and will guess generally at the meanings of these words and how to
use them. For instance, dog may first be used to refer to all animals;
its meaning is narrowed later. Early in the second year of life, the
child will produce fifty or more of these one-word utterances; they
may be used to indicate roles, such as agent or affected object, as
well as the objects themselves. Also in the second year, the child will
begin to develop two-word sentences. These two-word sentences,
combined with a growing vocabulary, allow the child to express a rap-
idly increasing number of messages.
These two-word sentences have been analyzed in several ways.
Braine (1963) viewed such sentences as combinations of words from
two classes: a small class called pivot words and a large class of open
words, which previously may have been one-word sentences. In the
pivot class are such words as pretty, more, and my; and in the open
class, such words as milk, shoe, and Mommy. The pivot and open
words stand in a fixed relation to each other; that is, a pivot word
always occupies one position in the speech of a specific child, and a
word from the open class always occupies the other position. For ex-
ample, one child would always use the sequence pretty shoe; another
might say shoe pretty.
Other types of two-word sentences are also found in the child’s
grammar; for example, noun-plus-noun sequences like cup glass. In
this case, the child’s intent is something like ‘‘There is a cup and a
glass.’’ Lois Bloom (1970) has shown that in addition to conjunction,
the noun-plus-noun sequence can express attribution (party hat), pos-
session (Daddy hat), a subject-locative relation (Mommy home), and
a subject-object relation (Mommy ball, meaning, for example,
‘‘Mommy threw the ball’’). Of course, the child may also use two-
word sentences consisting of a noun plus a verb.
Thus, it is conceivable that the child’s grammar has a rule of the
form S>N+V+N, but only two of the constituents on the right-
hand side of the arrow may appear in a given sentence because of
~ 166 Psycholinguistics e

the child’s limited means of syntactic expression. Actually, Bloom

claimed that each of the N+N sentences has a different underlying
deep structure and that the different semantic relationships expressed
by them are reflected in different transformational derivations. Al-
though this analysis postulates semantic relations or roles, it is still
unclear whether these two-word sentences incorporate grammatical
relations such as a subject.

SYNTACTIC ACQUISITION
Following the two-word stage, children add morphological endings,
articles, prepositions, and so on to their speech. This acquisition of
syntax takes place by hypothesis testing. For example, a child may
hypothesize that the past tense ending for all English verbs is -ed,
thus producing such forms as * goed and *singed. In fact, the correct
forms may be produced before the appearance of * goed and * singed
through random imitation. The correct forms appear with regularity
only when the child learns that go and sing are irregular verbs that do
not take -ed.
Between the ages of two and three, the child develops a more
complete syntactic system and is often able to correct incomplete or
misconstructed sentences that he or she has just uttered. The devel-
opment of this syntactic system continues with the appearance of
complex sentences at about the age of three. Complex sentences
allow the combination of several meanings, or propositions, within a
sentence through syntactic devices such as relativization, coordina-
tion, and complementation.
As the child’s syntax develops, he or she also expresses increas-
ingly complex messages. The expression of the desire for a cookie
progresses from Cookie to I want a cookie to It would be nice to have
a cookie. In the first case, the child simply utters the word cookie,
expressing a basic need of hunger. In the second case, the child is
able to express his or her relation to the cookie and actually verbalize
the desire. In the third case, he or she can express a state that would
result from having a cookie, and can make a weaker, more polite
request that may be more influential. By the end of the fourth year,
the child has mastered many of the phonological rules of the adult
grammar. By the age of five or six, the child has more or less incorpo-
rated the adult grammar, although refinement certainly continues for
some time thereafter.

AN ALTERNATE VIEW: THE BEHAVIORIST APPROACH

To repeat, the key to language acquisition appears to be a system of
hypothesis testing, or a set of operating principles that are used as
 Language Acquisition 167 A

strategies in acquiring language structure. Some of these operating

principles are given in Table 13-1. Of course, operating principles
may be changed and modified as the child grows older.
This view of language acquisition seems preferable to the theory
that (1) children imitate the speech of their parents; (2) they are posi-
tively or negatively reinforced, depending on whether their sentences
are grammatical; and (3) this reinforcement leads children to general-
ize their verbal behavior accordingly. Such a view, usually associated
with the behaviorist school of psychology, enjoyed great popularity
until the early 1960s.
The view that imitation is central to language acquisition has been
challenged by several types of findings. Recent studies — see Clark
and Clark (1977) — show that children imitate only words or struc-
tures that they themselves have already used. Imitation, by itself,
does not seem to provide for the acquisition of more complex sen-
tence structures. Also, it is assumed that children must speak a lan-
guage in order to imitate their parents; thus, by the behaviorist
theory, children cannot learn a language without themselves speak-
ing. But there are recorded cases of speechless children who have
learned languages and of children who suddenly spoke in sentences
after years of silence. Another argument against imitation is based on
the fact that the articles a and the are among the most frequently used
forms in adult language; yet children often acquire prepositions like
on and in, the plural form -s, the present progressive form -ing, and
even some irregular past tense forms before they use articles regu-
larly. Finally, parents only occasionally provide reinforcement; they
do not always recognize their children’s grammatical errors. Even
when children are corrected, they may be uncertain of the nature of

TABLE 13-1
Operating Principles Used by Young Children

Semantic Coherence Surface Structure

1. Look for systematic modifica- 1. Pay attention to the ends of

tions in the forms of words. words.
2. Look for grammatical markers 2. Pay attention to the order of
that indicate underlying semantic words, prefixes, and suffixes.
distinctions clearly and make 3. Avoid interruption or rearrange-
semantic sense. ment of linguistic units.
3. Avoid exceptions.

Source: From Clark and Clark (1977), based on Slobin (1973).

* 168 Psycholinguistics

their errors and may learn the correct form for only one particular in-
stance.

AN ASSESSMENT
Transformational-generative grammar has not provided a complete,
or completely correct, theory of language acquisition. However, it
has circumvented these problems by stressing that the productivity of
language can be accounted for by the learning of rules that are applied
to particular instances. In other words, to learn a language is to learn
its rules.
But we should not be surprised to find that some rules have excep-
tions, which themselves must be learned as rules. Consider the sen-
tence John told Mary to leave the room, which has a deep structure
equivalent to John told Mary for Mary to leave the room. In such for-
to complement sentences, a rule operates to delete the second (subor-
dinated) occurrence of Mary and for when the closest noun in the
main clause is also Mary. But in the sentence John promised Mary to
leave the room, the deep structure is equivalent to John promised
Mary for John to leave the room. Here, the second occurrence of
John is deleted, even though the nearest noun in the main clause is
Mary. One way to describe the deletion of John is to invoke the
marked/unmarked distinction on a syntactic level and note that the
verb promise is marked as an exception to the rule of identical noun-
phrase deletion in for-to complements. Children learn to understand
sentences containing told (which is unmarked) earlier than sentences
containing promise.
It also has been found that sentences like John is eager to please
are understood earlier than sentences like John is easy to please. No-
tice that in the former sentence, the surface structure more accurately
reflects the deep structure, which is approximately John is eager for
John to please someone, than in the latter sentence, where the deep
structure is approximately For someone to please John is easy. Pre-
sumably, the child learns the former sentence earlier because it re-
flects the normal, expected position of the logical subject.
Fodor, Bever, and Garrett (1974) have pointed out that at two years
of age, children consider any noun that immediately precedes the
main verb to be the logical subject. By the age of three and a half,
they consider the first noun in the sentence to be the actor, regardless
of the noun’s surface relation to the verb. Such data are valuable
because they relate the linguistic development of the child to a system
of strategies, or perceptual heuristics. These strategies certainly
should be explored further in developing an adequate theory of lan-
guage acquisition.
 #

Perception and Production of Language 169

Perception and Production of Language

The study of speech perception is concerned with how a person
hears, recognizes, processes, and understands the linguistic message
spoken by another individual. The first step in this process occurs
when a person receives the acoustic signal; that is, a stream of
speech. Making use of acoustic cues, such as the characteristic hiss-
ing of [s], and other facts, such as intonation and rhythm, the hearer
attempts to construct a phonological representation. One view of
how this is accomplished is called the analysis-by-synthesis model.

THE ANALYSIS-BY-SYNTHESIS MODEL

According to the analysis-by-synthesis view, the hearer internally
generates, or synthesizes, a set of speech sounds until he or she finds
a match with what has been heard. In the process, the hearer con-
tinually adapts to the voice characteristics of the speaker and uses
acoustic cues and factors such as knowledge of permissible sequences
of segments in the given language to limit the number of sounds that
are generated. After a match is found, the hearer constructs a phono-
logical representation, which is stored in working memory.
Next, the phonological data must be organized into functional syn-
tactic constituents, and an underlying semantic, or propositional, rep-
resentation must be determined. How this is done is the subject of
much debate, although it is probable that hearers use some combina-
tion of syntactic and semantic strategies.
For example, when a person encounters the articles a, an, and the
or a quantifier such as some or many, he or she knows that a noun
phrase is being introduced for which the noun is forthcoming. Simi-
larly, relative pronouns such as which and that provide a signal that a
new clause is beginning. Sometimes these strategies can lead the
hearer afoul, as in odd sentences like The boat sailed down the river
sank. Here, the hearer associates the verb sailed with the noun boat
and groups these constituents into a main sentence, The boat sailed
down the river. However, upon encountering sank, the hearer must
go back and reanalyze the sentence, finally associating sank with boat
in the main sentence. The hearer then understands this sentence to
mean The boat that was sailed down the river sank.
Semantic strategies make use of the fact that people usually speak
about things that make sense and that being understood is the chief
goal of communication. Thus, given the following sequence of sen-
tences in discourse:
5. Tom wanted to deposit $500.00.
6. The bank was too far away from him to walk.
170 Psycholinguistics

the hearer can easily supply the proper sense of bank in sentence 6
because of the logical situation described in sentence 5. Similarly, the
hearer can assume that him in sentence6 refers to Tom.
After the hearer has understood the sentence, its essential meaning
is retained in long-term memory. The phonological representation and
the syntactic structure, however, are allowed to decay from working
memory.
Here is a greatly oversimplified account of how the analysis-by-
synthesis model is used for the syntactic construction of sentences.
An individual hears a sentence that he or she must interpret. The per-
son determines the structural description of the sentence and then
proceeds internally to generate structural descriptions until a match
with the one perceived is obtained. The random generating of sen-
tences with the purpose of obtaining a match (in order to understand
the sentence) would obviously take a long time in the real world;
thus, the speaker must supplement this process with a number of
strategies to reduce the number of searches to be made. Even granted
such strategies, the analysis-by-synthesis model requires a ridicu-
lously large amount of real time. Furthermore, if the subject of the
sentence is dog, for example, the speaker would have to generate
many rules of the form Lex[N]—boy, cat, and so on until coming
across the rule Lex[N]—dog. This difficulty might be overcome by
assuming that the speaker reverses rules, so that on hearing dog, he
or she could apply a perceptual rule of the form dog—Lex[N]. But a
transformational-generative grammar would then not be directly in-
volved in the perception of sentences.

THE SPEECH-PERCEPTION GRAMMAR

During the past decade, linguists such as George Lakoff and Henry
Thompson (1975) have renounced transformational-generative gram-
mar as an abstract model of language and insist that it is of little value
in understanding language use. Instead, they claim, linguists should
be concerned with grammars of sentence perception and sentence
production.
In order to illustrate how a sentence-perception grammar works,
Lakoff and Thompson discuss the example The aardvark was given a
bagel by Irving. In this sentence, the individual first hears the aard-
vark and interprets this noun as the subject. But on hearing was
given, the individual reinterprets the aardvark as the direct object.
Next, the individual perceives a bagel and is forced to change strat-
egy again in order to regard the aardvark as the indirect object and a
bagel as the direct object. Finally, Irving will be regarded as the sub-
ject. This method of analysis is thus a step-by-step, left-to-right sen-
tence-processing model that interprets certain.pieces of the sentence
 Perception and Production of Language 171

in succession. It is part of a more general model of language use

called cognitive grammar. :

SPEECH PRODUCTION
Less is known about the way language is produced than about the
way it is perceived. Clearly, speakers form sentences by beginning
with some sort of idea about what they wish to express, and the basic
components of this message are propositions. These propositions are
combined and organized in various ways (syntax). Then they are
supplied with actual words, which are encoded into a set of phono-
logical units. Finally, the articulatory apparatus must be activated to
produce the appropriate waveforms. Beyond these few vague re-
marks, it is difficult to say much more about sentence production with
confidence.
Nevertheless, one of the few accessible sources of data on speech
production is speech errors, and the most significant work in this area
has been done by Victoria Fromkin (1973). Some of the types of
speech errors that she has uncovered are the following:
1. The anticipation of a segment. For example, cup of coffee is re-
placed by cuff. . . coffee.
2. The preservation of a segment. Instead of gave the boy, gave the
goy is produced.
3. The reversing of segments. Rather than keep a tape, teep a cape is
produced.
4. The blending of two forms. Switched and changed becomes
swinged.
5. The elimination of a syllable or sequence of segments.
Tremendously is replaced by tremenly.
6. The misderivation of a form. Instead of an intervening node, an
intervenient node is produced.
7. The substituting of one word for another. Before the place opens
becomes before the place closes.

Such errors indicate that there are certain basic units in speech
production; namely, distinctive features, phonetic segments, sylla-
bles, words, and larger units (for example, phrases). Before producing
an utterance, the speaker organizes what he or she wishes to say ac-
cording to an articulatory program that makes use of these units.
Thus, after deciding on the meaning to be expressed, a syntactic
structure will be sketched that contains word slots indicating stress.
These slots are then filled in with the major content words (nouns,
verbs, adjectives, and adverbs). Afterwards, the function words (ar-
ticles, conjunctions, and prepositions) and affixes are supplied in
proper phonological form. Finally, the complete phonetic specifica-
172. Psycholinguistics

tion is provided one syllable at a time. If a mistake is made during any

stage of this program, one of the typical speech errors will be the
result.
Chomsky’s theory of transformational-generative syntax in the fif-
ties led to psycholinguistic experiments concerning that theory in the
sixties; Fromkin’s psycholinguistic considerations of the syllable (in
dealing with speech errors) in the sixties led to a greater emphasis on
the syllable in the linguistics of the seventies. This is an example of
the interdependence of psycholinguistics and linguistics — a rela-
tionship that will surely continue as we learn more about language in
the eighties.

Summary
During the past twenty years, psycholinguistics has earned a posi-
tion of prominence among the sciences related to linguistics. In deal-
ing with how language is acquired, understood, and produced, psy-
cholinguistics has faced head-on the most basic questions of language
study.
Although transformational-generative grammar has its limitations
and in its most direct interpretation appears to be an inadequate
model of language processing, some of its levels of description — par-
ticularly constituent structure and distinctive features — have psy-
chological significance and are useful in explaining certain aspects of
verbal behavior. The usefulness of the transformational-generative
model in explaining the data on language acquisition is still being
explored. But the transformational-generative model is superior to the
behavioristic model as a theory of both competence and language
learning.
Speech perception involves determining a phonological represen-
tation, a syntactic structure, and a semantic representation from the
acoustic data of speech. The analysis-by-synthesis model is a much-
discussed view of how this might be accomplished; speech-perception
grammar offers another approach. Speech production involves organ-
izing an intended message by means of an articulatory program. Thus
far, most of what is known in this area derives from studies of speech
errors.

SOURCES FOR FURTHER READING

INTRODUCTORY TEXTS
Clark, Herbert, and Clark, Eve V. Psychology and Language: An In-
troduction to Psycholinguistics. New York: Harcourt Brace Jovan-
ovich, 1977.
 Summary 173

Foss, Donald J., and David T. Hakes. Psycholinguistics: An Intro-

duction to the Psychology of Language. Englewood Cliffs, New
Jersey: Prentice-Hall, 1978.
Slobin, Dan Isaac. Psycholinguistics. Glenview, Illinois: Scott-Fores-
man, 1971.

PSYCHOLINGUISTIC THEORY
Brown, Roger W. Psycholinguistics. New York: Free Press, 1970.
Cairns, Helen S., and Cairns, Charles E. Psycholinguistics: A Cogni-
tive View of Grammar. New York: Holt, Rinehart and Winston,
1976.
Chomsky, Noam. Language and Mind. New York: Harcourt Brace
Jovanovich, 1968.
. ‘Review of Skinner’s Verbal Behavior.’’ Language, 1959.
Fodor, Jerry A.; Bever, Thomas G.; and Garrett, M. F. The Psychol-
ogy of Language: An Introduction to Psycholinguistics and
Generative Grammar. New York: McGraw-Hill, 1974.
Fromkin, Victoria A., ed. Speech Errors as Linguistic Evidence. The
Hague: Mouton, 1973.
Halle, Morris; Bresnan, Joan; and Miller, George A., eds. Linguistic
Theory and Psychological Reality. Cambridge, Mass.: MIT Press,
1978.
Jakobovits, Leon A., and Miron, M. S. Readings in the Psychology
of Language. Englewood Cliffs, New Jersey: Prentice-Hall, 1967.

LANGUAGE ACQUISITION
Bloom, Lois. Language Development: Form and Function in Emerg-
ing Grammars. Cambridge, Mass.: MIT Press, 1970.
Chomsky, Carol. The Acquisition of Syntax in Children 5 to 10. Cam-
bridge, Mass.: MIT Press, 1969.
Clark, Eve. ‘‘What’s In a Word? On a Child’s Acquisition of Seman-
tics in His First Language.’’ In Cognitive Development and the Ac-
quisition of Language, edited by T. E. Moore. New York:
Academic Press, 1973.
Dale, Philip S. Language Development: Structure and Function.
Hinsdale, Illinois: Dryden Press, 1972.
Donahoe, John W., and Wessells, Michael G. Learning, Language,
and Memory. New York: Harper and Row, 1980.
Jakobson, Roman. Child Language, Aphasia and Phonological
Universals. Trans. by A. R. Keller. The Hague: Mouton, 1968.
Lee, Victor, ed. Language Development. New York: John Wiley,
1979.
McNeill, David. The Acquisition of Language. New York: Harper
& Row, 1970.
174 Psycholinguistics

Menyuk, Paula. Language and Maturation. Cambridge, Mass.: MIT

Press, 1977.
Tough, Joan. The Development of Meaning: A Study of Children’s
Use of Language. New York: Halsted Press, 1977.

EXERCISES

le If transformational-generative grammar does not actually describe

the process by which people produce or understand sentences,
what is its value, if any, in linguistics?
. Center-branching sentences like The book the professor Fred
knows wrote revolutionized linguistics are difficult to understand.
Imagine a center-branching sentence ten times as long as the ex-
ample given. For all practical purposes, it would be impossible to
understand such a sentence. Is there any sense in which such a
sentence could be considered grammatical, that is, a proper En-
glish sentence?
. Which member of the following pairs is the marked member and
why? (a) the means of past-tense formation in walk/walked and
sing/sang; (b) wide and narrow
. According to the model of pivot and open classes, why would it be
unlikely to hear a child speak both of the two-word sentences on
table and book my?
. If knowing a language is knowing its rules and speakers of a lan-
guage find it virtually impossible to state the rules of that language,
how does a speaker of a language prove that he or she knows its
rules? Explain how you prove your knowledge of certain phono-
logical rules of English every time you utter a word. Explain how
you prove your knowledge of English syntax and semantics every
time you utter a sentence.
. Which of the following two sentences would you suspect is more
difficult for a child under the age of three to understand? Why?
a. John is told to respect a secret.
b. John is glad to respect a promise.
 CHAPTER

14,
Sociolinguistics

Sociolinguists concentrate on analyzing the diversity of language.

There are many ways of describing the diversity of language; we can
look at changes over geographical areas, at the various dialects of a
language, at special forms of language such as pidgin, and at slang; in
this chapter we will take a brief look at each of these. But sociolinguis-
tics is concerned not only with the description of such diversity but
also with the analysis of how linguistic differences are related to
sociological differences among individuals — for example, to dif-
ferences in socioeconomic class or ethnic group. As individuals’ so-
cial characteristics vary, so does their language. This variation within
languages, as we shall see, poses a challenge to transformational-
generative grammar.

Diversity in Language and Society

When the Pilgrims and Jamestown settlers arrived on this continent in
the early seventeenth century, they spoke as they had in the cities
and villages of England during Shakespeare’s time. Consequently,
there was no difference between American and British English at that
time. Although separated by the Atlantic Ocean, the colonists and the
British continued to speak similarly for some time, and interrelations
were numerous enough to bring educated Americans some knowledge
of cultivated British forms. Indeed, at the time of the American Revo-
175
176 Sociolinguistics

lution, and for some time after, many educated people in America had
pronunciations and vocabularies that were similar to those of edu-
cated people in England. Today, we recognize a great diversity of lan-
guage in America. Where does this diversity come from?
Although many of the early settlers of the United States were edu-
cated, others were not. The settlers were a diverse group — com-
posed of people seeking relief from religious persecution, people who
were criminals in England, and fortune hunters from all walks of life,
including some nobles who were anxious to expand their fortunes
with land granted by the king. These people, who used different
words, pronunciations, and syntactic patterns, came from different
parts of England: London, small towns, and rural areas. The language
differences that existed in England as a result of geographic separa-
tion are maintained there to a fairly high degree even today. When the
English settled in various parts of the eastern section of what was to
become the United States, they brought their diversity of language
with them; and as colonial settlements in America expanded, the
speech forms of the colonists’ descendants changed, increasing the
diversity.
The linguistic diversity that had existed in England was not caused
exclusively by geographic factors. Many of the language differences
among the English of the seventeenth and eighteenth centuries can be
traced to clearly defined distinctions among social classes; and again,
to some extent, this distinction in pronunciation still remains. Proba-
bly the best-known reference to this class difference in language is
George Bernard Shaw’s play Pygmalion, in which Professor Henry
Higgins insists that a flower girl can be turned into an aristocrat in a
short time — primarily as a result of her learning a new pronuncia-
tion. The important point is that all the original English settlers in the
United States did not speak alike; instead, they arrived speaking dif-
ferently, and they developed even further differences in their new
home.
One ideal of many settlers was the abolition of class differences. As
Thomas Jefferson stated in the Declaration of Independence, ‘‘All
men are created equal... .’’ Nevertheless, everyone recognizes that
there are social and-class differences among people. These class dif-
ferences are not as clearly defined in the United States as they are in
such countries as Great Britain, where some people are formally rec-
ognized or designated as members of the upper class by right of birth
or through purchase of a title. In America, the class system is rather
fluid and not formally structured; movement from one class to an-
other is accomplished more easily than in Great Britain.
Thus, when we speak of such groups as the working class or the
upper middle class in America, we are actually referring to a group of
 Linguistic Geography 177

individuals who share similar socioeconomic characteristics, such as

income, occupation, and education. Other factors — like age, sex,
ancestry, ethnic group, and geographical location — further help de-
fine an individual’s position in American society. How these socio-
logical factors interact with the observed diversity of language is what
sociolinguistics is all about.

Linguistic Geography
If six young people from different parts of the United States were
brought together and asked to name a sandwich with several kinds of
meat and cheese, lettuce, pickles, and other trimmings on a long roll,
they might give such names as hoagie, grinder, submarine, poor boy,
torpedo, or blimp. They would all be naming the same basic sand-
wich, but they would have different names for it, depending on where
they were from. For example, Philadelphians use the term hoagie;
New Englanders use grinder; and Southerners, especially in New
Orleans, use poor boy. These different names for the same item tell
little about the speakers, except to place them in particular locales.
Usually, no information about the speaker’s socioeconomic condition
can be determined from his or her name for an item, because the
names for most things are the same for all people of the same genera-
tion from a particular area. We might say that the names for many
things are geographically determined.
At the end of the nineteenth century, some linguists became inter-
ested.in preserving a record of distinctions in the names for things,
they compiled descriptions of such ‘‘dialects’’ in several coun-
tries. They also compiled linguistic atlases, one of the earliest of
which was Jules Gillieron’s and Edmond Edmont’s Atlas Linguistique
de la France, published between 1902 and 1920. Linguistic atlases
trace different word usages, such’ as the means for telling time
(quarter to ten, quarter of ten, quarter till ten, or nine forty-five), and
the pronunciation of particular words, such as the pronunciation of
greasy as [grisi] or [grizi] and of tomato as [tameto], [tamato], [tame-
to], or [tamato]). A linguistic geographer can draw lines called _iso-
hat mark boundaries withi ich one word or pr jation
_is typical. Isoglosses for many words and pronunciations sometimes
may be bundled together to mark dialect boundaries. These isoglosses
may then be correlated with geographical, political, or cultural bound-
aries, such as waterways, railroad lines, mountain ranges, city limits,
folk customs, and agricultural or seafaring practices.
An excellent example of the methodology of linguistic geography is
the Handbook of the Linguistic Geography of New England by Hans
Kurath and others, published in 1939. One of the maps in this collec-
178 Sociolinguistics

FIGURE 14-1
Distribution in New England of the Preconsonantal and Final r

6 @
‘dl

e
@e io)

©@~@
(@)
.
otra! ° @
@eoeccecee|:|:-@
|F

y e@
mit
/
|ees
@

®
e

5
e
e

e
A
[4
>

/ @ ® e
ieeg SOR, So
®@

at
One |, ieae ys Ren ae e
|oce®, ee ° * , if“A a °

°@
\@@
eye 2 5
a3
oe
" mah
Note: Anr preceding a vowel, as in road, borrow, far out, is pronounced in all parts of
New England. But before consonants and finally, as in hard, how far?, usage is re-
gional: in western New England and in New Brunswick the r is regularly pronounced, in
most of eastern New England it is dropped, while the Connecticut Valley is mixed and
unstable in practice.
Martha’s Vineyard, Marblehead, and Cape Ann, all secluded communities, appear as
r “‘islands’’ in eastern New England, where this r is still losing ground. On the other
hand the r is gaining ground in the Connecticut Valley.-
The largest circles indicate regular use of this r, the smallest ones sporadic use, and
the two intermediate sizes rather evenly divided usage.
Source: Hans Kurath, Handbook of the Linguistic Geography of New England (Provi-
dence: Brown University Press, 1939), chart 16. Used by permission of the American
Council of Learned Societies.

tion (see Figure 14-1) plotted the distribution in New England of

preconsonantal and final r in words like yard and hear. At that time,
the r sound was found to be present in most of western New England
([yard]) and absent in most of eastern New England ([ya:d]).. (Two
dots indicates added length.) In this case, the chief isogloss clearly
corresponded to the Connecticut River Valley. One place in eastern
New England where the presence of r predominated was the island of
 Dialect Differences 179

Martha’s Vineyard, which is isolated both geographically and cul-

turally from the rest of New England. As a rule, communities that are
geographically and culturall isolated tend to be conservative in their—
farcemeee tesa aici frase eet bea remot
munities.
Such analyses by linguistic geographers attempt to correlate linguis-
tic facts with nonlinguistic data, but they fail to explore fully the
complex web of social factors involved. Although fascinating to
peruse, the linguistic atlases provide only a limited form of socio-
linguistic description.

Dialect Differences

Dialect is one of those linguistic concepts that seem necessary to talk

about but are almost impossible to define. Dialect is a regional variety
of language that may differ from other varieties of the language in fea-
tures of its vocabulary, grammar, and pronunciation. Further, dialect
may be seen as a variety of language used by one occupational group
or one social class. How do dialects differ from languages? The tradi-
tional answer is that speakers of different dialects can communicate
with each other, whereas speakers of different languages cannot. But
New Yorkers and Cockney Englishmen might experience difficulty in
communication, despite the fact that they both speak English. Also,
the linguistic geographer’s isoglosses are of little use in defining dia-
lects, for these are only approximations that mask considerable
linguistic variation. Furthermore, if one speaks of a ‘‘New England
accent’’ or ‘‘dialect,’’ one must account for the variation between the
speech of Bostonians and that of other speakers from New England;
and if one speaks of a ‘‘Boston dialect,’’ one must account for the
variation in the speech of Boston’s various geographical and socio-
economic subgroups. In fact, every person speaks an idiolect, which
is to say that no two people speak exactly alike. Moreover, speakers
alter their speech through time and in different social contexts.
Nevertheless, in order to describe groups of speakers who share
particular linguistic characteristics at a particular time, it is necessary
to have a term; thus, we use the term dialect. In the United States,
Black English or the speech of the lower middle-class residents of
Brooklyn are treated as examples of dialects. These dialects are con-
sidered nonstandard with respect to Standard English, which is the
dialect of American English used by many leaders in government,
business, and education and heard over most of the national broad-
cast media. It is important to note that the term nonstandard does not
mean ‘‘inferior’’; nonstandard dialects are simply different from Stan-
180 Sociolinguistics

dard English. Each nonstandard dialect in each language is as logical,

legitimate, and linguistically sophisticated as the standard dialect but,
of course, not as socially acceptable.
Much discussion and controversy has been generated about Black
English, a variant of American English spoken by most blacks. Some
people advocate that Standard English be taught to blacks as a second
language on the grounds that this approach would permit blacks to
capitalize on opportunities in the worlds of business and politics
(where Standard English is the norm) without sacrificing their cultural
heritage. Others condemn this approach as racist and suggest that it is
American society, and not Black English, that needs changing.
Table 14-1 presents some differences between Standard English
and Black English. Some features of Black English also are found in
other dialects to some degree; for instance, the lack of distinction in
the pronunciation of tin and ten also occurs in the speech of most
speakers from the American South. Actually, no one possesses all the
linguistic features of Black English or Standard English, for language

TABLE 14-]
Some Differences between Standard English and Black English

Standard English Black English Relevant Difference

1. nice [nays] nice [na:s] No diphthongization of vowel

2. tin, ten tin, ten No contrast of /1/ and /e/ before
[tin], [ten] [tivn], [tin] nasals
3. death [dee] death [def] [eo] changes to [f] when in word-
final position
4. door {dor] door [dot] Final r deleted; preceding vowel
[0] is consequently raised
5. He lost a friend. He lost a friend. Simplification of certain consonant
{hi lost a frend) [hi los a frivn] clusters
6. Hit him in the Hit him upside _ Different choice of preposition
head! the head!
7. He doesn’t know He don’t know Loss of third-person-singular
anything. nothing. marker in do; indefinite (anything)
becomes negative indefinite (noth-
ing) in negative sentence
8. He walked home. He walk home. Loss of past-tense marker in
regular verbs
9. I’m going to I'ma take a Deletion of certain verbs
take a walk. walk.
10. You’re in a lot You in a lot of Deletion of verb to be in certain
of trouble. trouble. environments
 Lingua Francas, Pidgin's, and Creoles 18]

is characterized by much variation or switching from certain speech

forms to others.

Lingua Francas, Pidgins, and Creoles

When two or more people who do not normally speak the same lan-
guage adopt a commonly agreed on language for communication, this
language is called a lingua franca. The term is borrowed from the
name given to a language used along the Mediterranean coast in medi-
eval times for the purpose of trade. That language combined elements
of Greek, Arabic, and the Romance languages. Because the lingua
franca was commonly understood by many people speaking different
languages, it facilitated trade among those cultural groups. Today, the
term lingua franca has been extended to refer to any language that is
mutually agreed on to facilitate trade, diplomacy, or other social func-
tions.
A lingua franca may be a fully developed language that is com-
monly used by many speakers. For example, English has become the
lingua franca of trade around the world. Latin has been the lingua
franca for the Roman Catholic Church. For a long time, Yiddish has
served as a lingua franca for Jewish people throughout the world.
A lingua franca can also be a rudimentary language. Sometimes,
when speakers of two languages need to communicate, they resort to
a hybridized or watered-down language that has a small vocabulary
and a simple grammar. This type of language is called a pidgin.
Many vocabulary items and grammatical structures that have been
borrowed from a dominant or source language are combined with
some grammatical elements and vocabulary items borrowed from an
area or local language to make a pidgin. The pidgin is a lingua franca
for the two groups, neither of which speaks the pidgin as a native or
first language. A pidgin is considered a minimal language because of
its limited linguistic capacity.
Many people ascribe a negative connotation to pidgin languages
because they often developed as a result of colonization. For in-
stance, when the early settlers arrived on the shores of North
America and found themselves needing to communicate with Native
Americans, the two groups developed a pidgin language based on the
colonizers’ European languages. Some of the European languages
used as the basis of these pidgins were English, Dutch, French, Span-
ish, and Swedish. Because the settlers were convinced that Native
Americans were simple and childlike, they simplified their European
languages for the Native Americans. The Native Americans, not re-
alizing that the European language they heard had been simplified,
simply imitated what they heard as well as they could. They also
182 Sociolinguistics

added elements of their own language to fill in the gaps of vocabulary

and grammatical structure. Thus, the European colonizers provided
the source language for these pidgins and the Native Americans pro-
vided elements of the local languages. The resultant pidgin languages
were not native to either group and were used only when these groups
contacted each other for trade or diplomacy.
Not all pidgins are based on colonization. Many pidgins develop
when an individual travels in a foreign country where he or she does
not speak the language. If, for instance, an American travels in
France and speaks only very broken French, he or she might ask a
French person on the street, ‘““Comment aller Louvre?’’ (“‘How to go
Louvre?’’). The French person, recognizing the severe linguistic
handicap of the American speaker, might answer with broad gesticu-
lation, ‘‘Louvre allez autobus’’ (‘‘Louvre go bus’’). Both speakers
have reduced both the vocabulary and the syntactic structure of
French to a fundamental form. They are using a pidgin language for
communication. Both speakers have tacitly agreed to use this pidgin.
If the French person had answered in flawless French, no pidgin
would have been involved because it takes both-sides to make a
pidgin. When a pidgin develops, it is always a second language for
both speakers and not the native language of either.
Pidgins develop to serve a particular purpose, and the speakers
continue using their native languages in their own speech communi-
ties. Pidgin remains a second language and continues to exist only so
long as the need that originally brought it about continues to exist.
Some pidgin languages are still spoken today, some by a large number
of individuals. One important contemporary pidgin is called Tok Pisin
and is based on English and several languages native to New Guinea.
In North America, the Indians of the Pacific Northwest have devel-
oped a pidgin language called Chinook. The various tribes in the area
have their own native languages, but they use this pidgin when they
trade with one another.
Occasionally, a pidgin becomes the native or first language of a new
generation of speakers. At that point, the language is termed a creole.
Creole languages develop most easily when parents communicate
with each other only through a pidgin language, and the pidgin be-
comes the first language of the child. Reinforcement for a creole lan-
guage occurs when all the families in an area are forced to speak a
pidgin. For example, when Africans from many tribes that spoke dif-
ferent dialects were brought together on one plantation in the New
World, they learned a pidgin dominated by the language of the planta-
tion owner. Thus, in Haiti and southern Louisiana, French, the lan-
guage of the old plantation owners, is the dominant language of the
modern creoles.
 Z

The Correlation of Social Patterns and Linguistic Data 183

As a pidgin develops into a creole, it is expanded with more lexical

items and greater grammatical complexity. Often, a creole becomes a
fully developed language. This has happened to the Gullah dialects
spoken by islanders off the coast of Georgia and South Carolina and
to the English-based creole spoken in Jamaica, as well as the creole of
southern Louisiana and Haiti.

Slang
Regardless of social position, almost all people use slang from time to
time. Slang is another word that is difficult to define but expresses a
concept that is understood by almost everyone; probably the fun-
damental quality of a slang term is that it is not generally accepted.
Slang terms are usually ephemeral and quickly forgotten, such as
cat’s pajamas, zoot suit, and bebop; but some have become a part of
Standard English, such as phone, jazz, and TV. Slang terms vary from
generation to generation and from one geographical area to the next.
The slang spoken on college campuses in the 1950s differs from that
spoken on campuses today. For instance, beatnik, platters, and bop,
all of which were widely understood in the 1950s, are rarely used
today. Likewise, disco and wimp may well pass out of usage in the
next twenty years. Also, the slang used in Wisconsin may be quite
different from that used in Florida or California.
Slang words may come about by combining old words into a new
word, by assigning a new meaning to an old word, by introducing a
completely new word, or by abbreviating an old word and giving it a
new meaning. Uptight, laid back, and cop out are examples of com-
binations of old words used in new ways. Tough, gross, sucks, and
cool are words whose original meanings have been expanded. Goof,
boo boo, and blooper are completely new words that have been cre-
ated. The terms hyper, hype, and narc are examples of clipped forms
(abbreviations) bearing new meanings.

The Correlation of Social Patterns and Linguistic Data

One of the most significant discoveries of sociolinguistics is that lan-
guage varies regularly according to socioeconomic class, age, sex,
ethnic group, and other factors. Language also varies with the speech
context and with how much attention a person gives to his or her own
speech. Moreover, these variations can be measured and analyzed in
such a way that sound changes can be observed in progress and sig-
nificant patterns within a speech area can be isolated. These findings
are the result of innovative work by William Labov on Martha’s Vine-
184 Sociolinguistics

yard and New York City in the mid-1960s, and they have been dupli-
cated by linguists working in other areas of the country.
Generally, this work involves recording interviews with a diverse
group of informants from a particular speech community. It may be
necessary to exclude certain speakers; for example, those with a
speech impediment or those who have recently moved into the com-
munity. Even a small group of speakers who reflect different social
strata, sexes, age groups, ethnic groups, and so on should reveal the
basic sociolinguistic patterns in the community. Next, certain sounds
and syntactic patterns called linguistic variables (indicated in paren-
theses) are selected for closer study. For example, the variant pro-
nunciations of (r) constitute a phonological variable; likewise, the
presence or absence of the indirect object me in the sentence I’m
going to buy (me) some candy is an example of a syntactic variable.
Syntactic variables generally are more difficult to elicit in conversa-
tions than phonological variables.
In the interviews, the linguist tries to elicit various styles of speech,
ranging from formal to casual. Formal speech may be elicited by ask-
ing the informant to read a list of words that contain examples of
selected phonological variables. Casual speech is more difficult to ob-
tain because the speaker usually is aware of the artificiality of the in-
terview. Furthermore, since accurate phonetic transcriptions are es-
sential, the microphone being used to tape-record the interview must
be close enough to the speaker to eliminate extraneous noise — a
condition that emphasizes the artificial nature of the situation. But if
the interviewer asks questions about such topics as religion, supersti-
tion, likes and dislikes, hobbies, travel, and close calls with death, the
speaker may become emotionally involved and pay less attention to
his or her speech.
After the tapes of the interviews have been transcribed, the linguis-
tic variables are correlated with sociological characteristics. For ex-
ample, in his work on Martha’s Vineyard, Labov studied the linguis-
tic variables (al) and (au), which roughly correspond in our
transcription to [ay] and [aw]. Specifically, he was concerned with the
tendency of many of the speakers to raise the initial element of the
diphthongs, that is, [a]. Labov created an index for [a] that combined
the degree of raising of [a] and the frequency with which [a] was
raised. He found that those speakers who had a positive orientation
toward Martha’s Vineyard had the highest index of raising, those who
had neutral feelings about Martha’s Vineyard had a lower index of
raising, and those who had a negative orientation toward Martha’s
Vineyard had the lowest index of raising. In such a case where a
linguistic variable is correlated with a sociological variable to yield a
regular progression, we may say that stratification is exhibited.
 The Correlation of Social Patterns and Linguistic Data 185

Crane (1973) has presented an interesting case of the social stratifi-

cation of the pronunciation of (r) in the casual speech of whites in
Tuscaloosa, Alabama. By a combination of education, income, and
occupation, he grouped his informants into upper, middle, and lower
classes. He also subdivided each of these classes into three age
groups: over 62 years, 23 to 62 years, and under 23 years. This divi-
sion made it possible to view pronunciation changes that might be
taking place over the span of several generations.
In his analysis of the variable (r), summarized in Table 14-2, Crane
illustrated a change that is taking place. Among the oldest speakers,
the upper-class informants produce a full r-sound in words like fourth
and floor only 10 percent of the time. Among the middle-class and
lower-class speakers, the percentages were 41 percent and 85 percent,
respectively. The large variations between the upper-class and lower-
class speakers indicate that during the early decades of the twentieth
century, speech in which r is regularly dropped was probably consid-
ered prestigious in Tuscaloosa.
In the intermediate-age group, however, the percentages for full r-
pronunciation, from highest to lowest social class, were 14 percent,
72 percent, and 90 percent. This fact shows that by about the middle
of the century, the middle class had joined in adopting full
r-pronunciation as their speech norm, but the stubborn upper class
was still clinging to the old, prestigious form.
By looking at the figures for the youngest speakers, the latest stage
of the change can be observed. Full r-pronunciation was present 85
percent of the time in the speech of young members of the upper
class; the percentages for the middle class and the lower class were 88
percent and 94 percent, respectively. Thus, the young upper-class
speakers now participate strongly in full r-pronunciation. This study
indicates that should the change continue, in two generations all
speakers in Tuscaloosa will have full r-pronunciation as their norm.

TABLE 14-2
Percentage of Full R-Sound in the Casual
Speech of Whites in Tuscaloosa

Age Groups

Socioeconomic Class 62+ 23+ 22-

Upper class 10 14 85
Middle class 41 72 88
Lower class 85 90 94
186 Sociolinguistics

Language Variation and Grammatical Theory

In Aspects of the Theory of Syntax, Chomsky (1965) stated that a
transformational-generative grammar is concerned with the compe-
tence of an ideal speaker/hearer in a homogeneous speech community.
But Labov’s work in the 1960s demonstrated the basic heterogeneity
of language within a speech community. As linguists continued to un-
cover numerous cases of variation in language, it became necessary
to question how such variation could be accommodated within a
transformational-generative grammar.
One of the earliest studies dealing with this problem was Elliott,
Legum, and Thompson’s paper, ‘‘Syntactic Variation as Linguistic
Data’’ (1969). In one of their examples of syntactic variation, the au-
thors considered the deletion of pronouns used as subject NPs in
while clauses, as in the following sentences:

1. Sophia Loren was seen by the people while enjoying herself. (Sub-
ject NP of the while clause is deleted upon identity with the subject
of the preceding passive clause.)
2. The people saw Sophia Loren while enjoying themselves. (Subject
NP of the while clause is deleted upon identity with the subject of
the preceding active clause.)
3. Judy was seen by the people while enjoying themselves. (Subject
NP of the while clause is deleted upon identity with the object of
the preceding passive clause.)
4. The people saw Karen while enjoying herself. (Subject NP of the
while clause is deleted upon identity with the object of the preced-
ing active clause.)

They found that speakers of English vary in assessing whether such

sentences are grammatical; that is, their grammars differ concerning
the types of sentences in which the subject pronoun in a while clause
may be deleted. Nevertheless, a clear pattern emerged in the study.
Although speakers disagreed on the acceptability of particular sen-
tences, any speaker who accepted sentence 4 accepted all the other
sentences as well. Anyone who rejected sentence 4 but accepted sen-
tence 3 also accepted sentences | and 2. And anyone who rejected
sentences 3 and 4 but accepted sentence 2 also accepted sentence 1.
This pattern held true for all speakers, with exceptions recorded for
only a few sentences. Later studies showed that such ponepatterns,
anescalled
eae
implicational hierarchies, exist at all levels of grammar.
The relevance of hierarchies-to-penerative theory may be seen
clearly in the following sentences:

5. Tomorrow Bill expects that his mother will arrive from Dallas.
? 6. Tomorrow I realize that banks are closed.
 Language Variation and Grammatical Theory 187

?? 7. Tomorrow Sue mentioned that her favorite program will be

televised. .
* 8. Tomorrow I believe the claim that the governor will make a
speech.
* 9. Tomorrow John quit his job and will look for another one.
*10. Tomorrow we saw the man who will juggle for six straight
hours.
In each of these sentences a transformation called adverb in
has applied; this transformation moves the adverb tomorrow from the
end of the sentence to the beginning of the sentence. In the standard
theory of transformational-generative grammar, a rule must either
apply or not apply unconditionally, depending on whether its struc-
tural description is met. But as sentences 5-10 show, the application of
the adverb-preposing transformation has varying effects, depending
on the context about which it is preposed. The relevant contexts in
these sentences can be presented as follows:
5. Main clause containing the verb expect.
6. Main clause containing the verb realize (such verbs are called
enchoative factives).
7. Main clause containing the verb mention (such verbs are called
emotive factives).
8. Main clause containing the + the noun complement claim.
9. Main clause is a conjunct; tomorrow is a part of one of the con-
juncts.
10. Main clause contains a noun (man) that is relativized (with the
clause who will juggle . . .).
This set of hierarchically arranged contexts is part of what is called
the island hierarchy. Adverb preposing applies most freely when the
context is at the upper end of the hierarchy. Thus, such hierarchies
may be significant for the statement of grammatical rules.
Note that the judgments indicated for sentences 5-10 are those of
one of the authors; speakers’ judgments may vary for any particular
sentence. But speakers should not find any sentence less grammatical
than a sentence lower on the list. For example, regardless of a
speaker’s judgment of whether sentence 7 is grammatical, he or she
should not find it less grammatical than sentence 8, and so on.
After he noticed that rules do not always apply in all contexts in an
individual speaker’s grammar, Labov proposed the notion of a vari-
able rule. As an example of the variable rule that describes the dele-
tion of is in Black English, Labov (1972) suggested the following:

ike +Vb
2D) |} Cons )##—## oe
188 Sociolinguistics

Here, ## represent word boundaries and the angular brackets con-

tain variable constraints. These constraints are arranged vertically so
that the highest ones reflect those conditions under which the rule
most often applies. Thus, the first part of the rule following the slash
states that is is more likely to be deleted when following a pronoun
(+Pro) ending in a consenant (for example, it). The latter part of the
rule states that is is more likely to be deleted when followed by a
verb, particularly if the verb is a future form. If the following element
is not a verb, then the rule is more likely to apply when this element
is not a noun phrase. For example, consider the following sentences:
11. He’s running to the store = He running to the store.
12. She’s in the third aisle = She in the third aisle.
13. He’s president of the club = He president of the club.
The rule predicts that is will be deleted in sentences like sentence 11
more often than in those like sentences 12 and 13, and is will be
deleted in sentences like sentence 12 more often than in those like
sentence 13. The writing of variable rules continues to become more
and more sophisticated, often drawing from mathematical and statis-
tical models.

Summary
Sociolinguistics is concerned with the diversity of language as it re-
lates to various sociological factors. Some of the earliest attempts to
deal with linguistic diversity were the linguistic atlases, which indi-
cated the location and frequency of particular pronunciations and lex-
ical items within a geographical area. By drawing isoglosses that are
later correlated with geographical, political, and cultural boundaries,
linguistic geographers provide a sketchy, first approximation to a so-
ciolinguistic description. Linguists have also tried to capture the di-
versity of language by analyzing dialects. Although the concept of
dialect is somewhat questionable, it is frequently used to discuss the
speech of individuals who share many linguistic characteristics at a
particular time, such as American blacks and the lower middle-class
residents of Brooklyn. Linguistic diversity is increased by the exis-
tence of lingua francas, which sometimes take the form of pidgin lan-
guages. A pidgin may develop into a creole language.
A more complete approach to sociolinguistics correlates linguistic
variables with such sociological factors as socioeconomic class, age,
sex, ethnic group, and orientation toward the community. Although
the standard interpretive theory of transformational-generative gram-
mar deals with an ideal speaker/hearer in a homogeneous speech com-
munity, many sociolinguists have revised this aspect of the theory by
attempting to write grammars that describe the language variation in
 Summary 189

the speech community as a whole. The measurement and analysis of

linguistic variation has become increasingly sophisticated, incorpo-
rating variable rules and other mathematical and statistical models.

SOURCES FOR FURTHER READING

INTRODUCTORY SOCIOLINGUISTICS
Farb, Peter. Word Play: What Happens When People Talk. New
York: Knopf, 1974.
Fishman, Joshua A. Sociolinguistics: A Brief Introduction. Rowley,
Mass.: Newbury House, 1970.
Platt, John T., and Platt, Heidi K. The Social Significance of Speech:
An Introduction to and Workbook in Sociolinguistics. Amsterdam:
North Holland, 1975.
Trudgill, Peter. Sociolinguistics: An Introduction. Harmondsworth:
Penguin, 1974.

DIALECTOLOGY
Bailey, Richard, and Robinson, Jay L., compilers. Varieties of
Present-Day English. New York: Macmillan, 1973.
Dillard, J. L. Black English: Its History and Usage in the United
States. New York: Random House, 1972.
Kurath, Hans, et al. Handbook of the Linguistic Geography of New
England. Providence, Rhode Island: American Council of Learned
Societies, 1939.
Kurath, Hans, and McDavid, Raven I. The Pronunciation of English
in the Atlantic States. Ann Arbor: University of Michigan Press,
1961.
Malstrom, Jean. Language in Society. Rev. 2d ed. Rochelle Park,
New York: Hayden Book Company, 1973.
Reed, Carroll E. Dialects of American English. Rev. ed. Amherst:
University of Massachusetts Press, 1977.
Shuy, Roger W. Discovering American Dialects. Champaign, Illinois:
National Council of Teachers of English, 1967.
Williamson, Juanita V., and Virginia M. Burke. A Various Language:
Perspectives on American Dialects. New York: Holt, Rinehart and
Winston, 1971.
Wolfram, Walt, and Fassold, Ralph W. The Study of Social Dialects
in American English. Englewood Cliffs, New Jersey: Prentice-Hall,
1974.

PIDGINS AND CREOLES

Hymes, Dell, ed. Pidginization and Creolization of Languages. Cam-
bridge, England: Cambridge University Press, 1971.
190 Sociolinguistics

SOCIOLINGUISTIC METHODOLOGY
Labov, William L. The Social Stratification of English in New York
City. Washington: Center for Applied Linguistics, 1966.
. Sociolinguistic Patterns. Philadelphia: University of Pennsy]l-
vania Press, 1972.

LANGUAGE VARIATION
Elliott, D. S.; Legum, Stanley; and Thompson, Sandra Annear.
‘*Syntactic Variation as Linguistic Data.’’ In Papers from the Fifth
Regional Meeting of the Chicago Linguistic Society, edited by
Robert I. Binnick, et al. Chicago: University of Chicago, Depart-
ment of Linguistics, 1969.
Gregory, Michael, and Carroll, Susanne. Language and Situation:
Language Varieties and Their Social Contexts. London: Routledge
and Kegan Paul, 1978.

EXERCISES

1. Why did educated speakers of English on both sides of the Atlantic

have similar pronunciations and vocabularies at the time of the
American Revolution?
2. A linguistic geographer recorded the names for two hundred dif-
ferent items among 50 percent of the population within a particular
geographical area. He then drew a map on which he indicated the
frequency of occurrence of each of the various words he had re-
corded. His finished product revealed absolutely no variation, and
he was unable to draw any isoglosses. Would he be justified in
concluding that the inhabitants of the area all speak the same way?
If pronunciation and syntax also were studied and the data re-
vealed only random variation, would he be justified in concluding
that the inhabitants of the area use language in a random way?
3. Is it true that every ethnic group within a city speaks its own
dialect?
4. It is commonly known that certain speakers in New York City pro-
nounce words like bird and shirt as boid ([bAyd]) and shoit ({Sayt]).
Other speakers who possess the same socioeconomic character-
istics produce the latter forms much less often in their speech, and
almost never in formal speech. How would you account for this
variation?
5. Suppose another sociolinguistic study in Tuscaloosa ten years
from now demonstrated a drop in the percentage of full r-pronun-
ciation among the middle-age, middle-class group and, perhaps,
certain other groups. What conclusion might be justified?
 Summary 191

6. In a particular speech community, the vowel [2] becomes [e"]

most frequently before [m] and [n], but rarely before [n]. Show
how this fact could be expressed in a variable rule. You need not
decompose [z] and [e"] into their distinctive features.
-

&
 GVH tA PHT ER

ANTHROPOLOGICAL
LINGUISTICS

Anthropological linguistics is primarily concerned with how languages

‘fit
into-the-larger-eonteXxt
of sociocultural behavior. This is a broad is-
sue — one that necessarily encompasses sociology, psychology, neu-
rology, and other areas. It must take into account both the variations
and the universals in language. In this chapter, we will scan the di-
verse languages found in the world’s many cultures. They may be
classified according to their origin or according to aspects of their
structure such as phonology or syntax. We will also examine various
types of universals in language. But the departure point of anthropo-
logical linguistics is usually the relationship between language and
thought.

Language, Thought, and Culture

Are language and thought the same thing? Probably not. Almost all of
us have had the experience of not being able to express properly what
we want to say,,and people are capable of solving analytic problems

ments against the identity of language and thought. On the other

hand, the fact that many people talk to themselves and are often not
aware of their behavior argues for an extremely close relationship be-
tween language and thought. If language is not thought, it is more
likely a system for expressing thought.
192
 ‘Language, Thought, and Culture 193

Some scholars have concluded that. language determines thought.

This is a rather frightening prospect. Are we, as speakers of English
and representatives of American culture, bound to some particular
view of the world because our language contains progressive verb
forms and such words as heaven and hell?

THE WHORFIAN HYPOTHESIS

Debate about the relationship of language, thought, and culture be-
came especially heated in the first half of the twentieth century, when
anthropologists and linguists were busy studying the various lan-
guages of the world. In North America, for example, they discovered
that the language of the Greenlandic Eskimos has many word forms
to describe the various types of snow found in the vicinity of the Arc-
tic Circle; they have no word meaning ‘‘snow’’ to encompass the gen-
eral concept. The Hopi Indians were found to have no overt grammat-
ical distinction in their language between past and present, although
they have three tenses: the factual, the future, and the general. And
cultures even divide the color spectrum in different ways. Whereas
Americans recognize green and blue as distinct colors, some other
cultures have a single word to represent both green and blue. On the
other hand, Russian has two words for what Americans consider
blue. The differentiation of kinship terms from culture to culture also
has been well documented.

language a es determines its ion of the is

view 1s often called the Whorfian hypothesis in honor of one of these
Be
past
EN CSD Se Ene oe Wher nea
few decades, the popularity of this view has significantly dimin-
ished; for there seems to be no way to prove that different types of
languages reflect different thought processes. Also, the dominance of
transformational-generative grammar since the early 1960s has led to
a greater emphasis on universal grammar. Thus, many linguists have
expressed the belief that the base component of every language is the
same, and that syntactic differences between languages are due to dif-
ferent sets of transformations.
Nevertheless, in recent years some studies have again argued for a
close relationship between language and the way people perceive the
world. For example, Cooper and Ross (1975) have studied the phe-
nomenon of frozen word order in such phrases as here and there, now
and then, sooner or later, and so on. One would never say *there and
here, *then and now, or “later or sooner. Cooper and Ross’s most in-
teresting finding was that frozen word orders in English provide evi-
dence for at least three basic orientations, or outlooks, on the world:
194 Anthropological Linguistics

1. Up/down, as in peak and valley

2. Right/left, as in East-West game
3. Vertical/horizontal, as in height and width
There is further evidence for these orientations in the everyday ac-
tions of English-speaking people. For instance, when printing capital
letters, English-speaking people usually begin at the top and draw any
vertical lines first. Baseball players run to first base, rather than third
base, after hitting a ball — thus proceeding to the right. hort, the .
study suggested that a language may reflect a basic perception of the
world.
=

CULTURE AND LANGUAGE

Anthropological linguists point out that use of language is but one
form of human behavior and, therefore, a grammar must be part of a
culture. Consider, for example, the forms of the personal pronoun
meaning ‘‘you’’ in French: tu is the singular and the familiar form
for ‘‘you,”’ and vous is both the plural and the polite form. In English,
the one form you is used in all cases — singular, plural, familiar, or
polite. But to state only these facts about the grammars of French and
English is insufficient because there are cultural factors to be consid-
ered. Americans would not know which form of “‘you’’ to use in
addressing a French person unless they also knew something about
French culture. For example, a typical young French student would
not address an elderly French public official as tu. But the same
student might address a fellow student of similar age and status as tu
shortly after their first meeting. In English, information pertaining to
speaker and hearer is coded in other ways. For example, an elderly
public official who is an American male might first be addressed as
Sir. Similarly, the terms buddy, man, and pal would usually be used
in reference to American males. The phrase hey, buddy is used when
the addressee is unknown, but his buddy indicates that the person is
well known.
Grammar is thus only a part of culture. The particular grammatical
forms chosen by speakers in particular contexts are determined by
their culture. Only by considering culture can we fully understand
language behavior.

Language Families and Language Types

One of the earliest known attempts to account for the multiplicity of
languages is the biblical story of the Tower of Babel, in which God
afflicted each worker with a different language to prevent com-
munication. Today, linguists generally believe that each language fam-
 Language Families and Language Types 195

ily— or group of related languages— developed from an original

tribal language.

THE INDO-EUROPEAN FAMILY

One of the most studied families is the Indo-European language fam-
ily. This family probably developed from a language spoken some
5000 years ago by a small tribe located in what is today southern
Poland and Russia. At some point, groups of people from this tribe
migrated to the south. Their language evolved into the so-called Indo-
Iranian languages, which include modern Persian (Parsi) and Hindi.
Another group moved westward. Their descendants speak most of the
Western European languages, including English. A third group of
migrants moved north and ultimately developed the Slavic language
communities, including Russian and Polish. The assumption that En-
glish, Russian, Hindi, and other languages are historically related was
verified by the results of the comparative method (see Chapter 16).
Table 15—1 presents the names for the numerals from one to ten in
several Indo-European languages. At first glance, no two of these lan-
guages may seem related, but a more careful examination brings out
unmistakable similarities in elements of the number systems. The
number one, for instance, is similar in German and Greek and in Ger-
man and Danish. The number two is similar in Russian and Sanskrit
and in Spanish and Greek. The number three seems very closely
related in the seven languages in Table 15-1. With sophisticated tech-
niques of historical-comparative linguistics, the relatedness of virtually
every number on the list can be demonstrated.
Perhaps the major division of the Indo-European family is between
the centum and satem Tanguages, named after the way the first sound —
- ees

TABLE 15-1
Numbers One through Ten in Several Languages

English German Spanish Greek Russian Danish — Sanskrit

one eins uno heis odin en ekas

two Zwel dos duo dva to dva
three drei tres treis tri tre trayas
four vier cuatro téttares chetyre fire catvaras
five fiinf cinco pénte pyat fem panca
SIX sechs seis heks shest seks sat
seven sieben siete hepta sem syV sapta
eight acht ocho okto vosem otte asta
nine neun nueve ennéa devvat ni nava
ten zehn diez déka desyat ti dasa
196 Anthropological Linguistics

TABLE 15-2
Hundred in Several Centum and Satem Languages

Centum Languages Satem Languages

Latin: centum Avestan: satom

Gothic: hund Lithuanian: Simtas
Welsh: cant Old Church Slavonic: suto
Greek: hekaton Sanskrit: satam

in the word a Bae ntum languages,

the origin = d usually is realized as [k] or [h]. whereas in the
satem lan aT I éd as a more fronted consonant
—tike [sy or [8]. Table 15-2 illustrates this division. The table includes
languages of the Indo-European family that are no longer spoken —
Latin, Gothic, Avestan, Old Church Slavonic, and Sanskrit.

OTHER LANGUAGE FAMILIES

With the exception of Indo-European and a few of the other major
language families, the classification of languages by family is still
rather tentative. Approximately a hundred families have been pro-
posed for the languages of Mexico, Central America, and South
America, where knowledge of proper genealogy is very limited. On
the other hand, such common languages as Japanese and Korean are
not yet firmly associated with a particular language family, although a
very tentative association has been made with the Mongol and Turk-
ish languages.
Nevertheless, some of the more solidly based language families
may be noted (see Figure 15—1). The Kinna-Uerj ily has two chief
representatives, Finnish_and Hungarian; but this family also includes
Lappish and Estonian and many minor languages found in northern
Russia as far as eastern Siberia. The Altaic family, which is spoken
across much of central Asia, consists aoe the Turkic lan-
guages and the Mongol and Manchu languages. Afro-Asiatic is a fam-
ily that consists mainly of languages spoken in northern Africa and
parts of the Middle East. It includes the Semitic, Hamitic, Berber,
jtic, and Chad subgroups. The Niger-Congo family, whose best-
known representative is probably Swahili, consists of Many languages
spoken in central Africa. _Malayo-Polynesian_is the major language
family of the Pacific, cOnsisting of the Indonesian, Melanesian, and
Polynesian languages. Other Pacific language families are found in
Australia, Tasmania, and (perhaps) New Guinea. The Sino-Tibetan
family includes languages spoken in Burma, Tibet, and many parts-of
——————"
——
 daNDIA
I~ST
ay] jodiouug adonsuvT Saud]

-
2".
i

cad

oARTR
oad
urodoinq-opuy
_- .oIyeisy-o1py ueyaqi[-oulg oeITV ueIpIAeiq oI}PISW-ONsNy
| dLIS;-OuUuL

:391N0SWO1 [VIUWOISIFY :sousiNsuIT

uy ‘voNonposuT
puodseg ‘UONIPY
Aq ‘gq pou -uuewYysT] 1ysuUAdoD
@
‘Z96I ¢€/6IAq “OH JeYyouTY
pue ‘uojsuIAA
‘suy payutidoy
Aq uorsstwied
Jo “OH JIeYoUTY
PUP “UO}SUIM
198 Anthropological Linguistics

Thailand and China. Two other language families are the Dravidian
family, which is based in southern India and parts of central India and
Pakistan, and the Austro-Asiatic family, which includes Vietnamese
and the Munda languages
or eastern Ifidia. In North America, several
families have been identified. The most important of these are Atha-
baskan, Algonquian, Iroquian, Siouan, and Uto-Aztecan. a

STRUCTURAL TYPES
Another important way to classify languages is according to structural
type. Classification by phonological systems, for example, has long
been in vogue in modern linguistics. Languages have been grouped
according to the similarities of their phonemic systems or, more spe-
cifically, their vowel, consonant, and suprasegmental subsystems.
For example, some languages exhibit two degrees of tongue height for
their front unrounded vowels and back rounded vowels — Polish and
Hawaiian are examples — whereas others, like English and Italian,
exhibit three degrees of tongue height for such vowels. And some lan-
guages, like German, have a distinctive contrast between word-initial
voiced and voiceless stops; others, like Thai, have word-initial
voiced, voiceless, and voiceless aspirated stops that all contrast. To
speakers of English, a distinctive contrast between word-initial voice-
less unaspirated and voiceless aspirated stops might seem rather un-
usual because there are none in our language. On the other hand,
speakers of many languages would find the English contrast between
/e/ and /8/ as in thin and then most peculiar.
Supr: mentals are significant in classifying certain languages

pitch, alone can distinguish among meanings. In Mandarin Chinese,

for instance, the syllable ma may have at least four meanings, de-
pending on the tone, or pitch, associated with it. With a high-level
pitch, it means ‘‘mother’’ or ‘‘old woman’’; with a rising pitch, it
means ‘‘hemp’’; with a low pitch, it means “‘horse’’; and with a high,
falling pitch, it means ‘‘to curse’’ or ‘‘to scold.’’ Another_supraseg-
mental, length, does not often occur in distinctive contrasts, although
Tatin hada contrast between long and short vowels andItalian has 2
contrast between long and short consonants.
, typologies of languages according to phonological
systems have concentrated on the distinctive features. The proposed
feature [delayed release], which is present in the segments /¢/ and /j/ in
English, is not found, for example, in Hawaiian. Similarly, the Afri-
can language Zulu requires a feature like [glottalized] to characterize
glottalized consonants like /p?, t?, k?/, which contrast distinctively with
nonglottalized consonants like /p, t, k/.
One popular method of classification by language structure was de-
 Language Universals 199

veloped in the nineteenth century. It represents all languages as one

of three types: analytic, agglutinative, or inflectional. An analytic lan-
guage, like Chinese, is one in which each word is an isolated root. At
the opposite extreme are the agglutinative languages, like Turkish and
Eskimo, in which words may include different types of elements, or
affixes, although these are kept distinct from the roots; each affix has
one distinct grammatical meaning. (In agglutinative languages, a root
may not be changed to show grammatical function as the root changes
in foot/feet.) Somewhere between these two types are the inflectional
languages, like Latin and Arabic, in which affixes serve a rammatical
function and become part of the I
dependent.
Classification by syntax usually focuses on the types of word order
preferred by individual languages. For example, English is a subject-
verb-object language; Japanese is a subject-object-verb language; and
Tagalog is a verb-subject-object language.
Syntactic and morphological classification has been greatly ad-
vanced by Joseph Greenberg. One of his suggestions was a synthetic
index, or ranking, that takes into account a number of criteria for
each language. These criteria may include ratio of morphs (roughly
equivalent to morphemes) per word, an agglutination index, mor-
pheme classes per word, and an index that expresses grammatical
relationships.
The classification systems presented in this section clearly are more
geared to linguistics than anthropology. Anthropologists would be
more interested in the types of terms that a culture uses to describe
its environment, such as color terms and kinship terms, and they
would incorporate this information into a language classification.
Thus, we may know that, linguistically, German has a language struc-
ture more closely related to English than French. However, anthro-
pologically, German is similar to French in that both have a familiar
and polite form for ‘‘you’’ (Du and Sie, respectively, in German),
whereas English does not; this reflects one way in which the German
culture is more similar to the French culture than to the English cul-
ture. One of the ultimate purposes of anthropological linguistics is to
unite cultural and linguistic information.

Language Universals
The theory of transformational-generative grammar is a universal
theory of grammar that attempts to isolate facts and principles that
are true for all languages. There is evidence that seems to support the
claim that language is universal. For instance, adults sometimes
mimic young children’s speech to produce baby talk. The phonologi-
200. Anthropological Linguistics

cal, lexical, and syntactic simplification that takes place in this baby
talk is pretty much the same, regardless of what language the adult
speaks. Also, whenever adults who speak two languages must in-
teract for a commercial purpose or some similar reason, they develop
a pidgin, a simplified form of one of the two languages. The type of
simplification that takes place in pidgins is characteristicaliy universal.
The search for universal characteristics of language has been ap-
proached in several ways. The simplest definition of a language uni-
versal is some element, as ; e tha oundtnreverytan--
guage Of the world. An example of a language universal, then, is the
fact that all languages have vowels and consonants; another is that all
languages have some means of expressing negation and questions.
Txansformational-generati ria ve traditionally clas-
sified language universals as substantive or formal. universality
The
of vowels and consonants and of negative and interrogatory for
“qualify as substantive universals because they represent the~act
substantive data of languages. Formal _universals, on the other hand,
are more abstract and more difficult to discover> One proposed formal
uriversat
transformational
isthe cycle in phonology; certain con-
straints in syntax also may qualify as formal universals. A form
universal, then, is generally some theoretical construct_that can_be
applied to all languages and that provides some insight into how lan-.
guage works. ,
A realistic treatment of universals will not restrict them to charac-
teristics that are found in al] languages. For example, the distinctive
features are believed to be a universal set of approximately thirty
phonetic features. The fact that they are called wniversals does not
mean that every language has all the distinctive features; it means
only that the phonology of each language may be described in terms
of features drawn from this fixed set. And, of course, this set is not
truly fixed, because new research continues to yield modifications of
the set; however, in theory, there is an optimal set of distinctive fea-
tures for language. Which languapesillGhave Sie hereatineceracdiare
determined bya theory of Universal grammar. =~=~S<SO:C~*”
~—~Another type oflanguage universal.is-thé implicational universal, in
which certain facts observed in languages imply other facts. Thus, if a
language has a contrast of tongue heights for nasal vowels, as in
French, it must also have a contrast of tongue heights for nonnasal

marked distinction. ample, in the languages of the world, nasal

‘Vowels are more highly marked than nonnasal (or oral) vowels. The

feature. Similarly, if a language has a gender distinction in the plural,

it must have a gender distinction in the singular. The converse of this
 4 Summary 201

statement is not necessarily true, and this fact also may be captured
by marking conventions. That is, the existence of an unmarked fea-
ture does not imply that the marked feature occurs in the language.
Related to the study of implicational universals is the discovery of
implicational hier, ies, such as the accessibility hierarchy. This hier-
—whi a distinct similarity tothe Telationathierarchy of
relational grammar (see pages 146-148) is given as follows:
Subject>Direct object>Indirect object>
Oblique>Genitive>Object of comparison
Oblique refers to an NP whose function is similar to the drawer in Rod-
ney placed the gun in the drawer; genitive usually refers to a posses-
sive NP, such as the clown’s in the children laughed at the clown’s
face; and object of comparison refers to an NP like the book in The
movie was better than the book. The accessibility hierarchy has
proved useful in discussing, for example, the type of NPs that may be
relativized — that is, that may serve as the head noun for a relative
clause. In the sentence The woman who lives on 4th Street is a violin-
ist, the head NP the woman is a subject NP that is relativized with the
clause who lives on 4th Street. It is a universal fact that all languages
must be able to relativize subjects. But what is interesting about the
accessibility hierarchy is that it ranks NPs in descending order ac-
cording to which NP can be most easily relativized. Furthermore, an
implicational relationship holds among the terms of the hierarchy. For
example, if a language can relativize oblique NPs, it must also be able
to relativize NPs higher in the hierarchy — subjects, direct objects
and indirect objects. The accessibility hierarchy, discovered by Ed-
ward Keenan and Bernard Comrie (1977), has led to a recent
reawakening of interest in cross-linguistic data gathering and implica-
tional hierarchies.
Finally, language universals may be studied by providing statistical
data regarding the occurrence of certain features in the languages of
the world. A statistical universal may take the following form: feature,
or aspect, X is present Y percent of the time in the world’s languages.
Much of the important work on statistical universals was done by
Joseph Greenberg in the 1960s.

Summary

Anthropological linguistics is concerned with how language fits into

the larger context of sociocultural behavior and, more specifically,
with how grammar is a part of culture. There is no convincing evi-
dence that language and thought are the same thing, or that language
determines thought.
202 Anthropological Linguistics

Languages may be classified by family or type. The main lan-

guage families are Indo-European, Finno-Ugric, Altaic, Afro-Asiatic,
Niger-Congo, Malayo-Polynesian, and Sino-Tibetan. Linguists have
grouped languages by type primarily through the similarities of their
phonological, morphological, and syntactic systems. Anthropologists
have shown a greater interest in semantic systems, as manifested by
particular terms used in cultures.
All languages have certain properties in common, and the theory of
transformational-generative grammar hopes to account for these simi-
larities. Language universals may be classified as formal or substan-
tive; other approaches to universals involve implicational and statis-
tical facts.

SOURCES FOR FURTHER READING

LANGUAGE, THOUGHT, AND CULTURE

Baumm, Richard and Sherzer, Joel, eds. Exploration in the Eth-
nography of Speaking. New York: Cambridge University Press,
1974. :
Carroll, John B. Language and Thought. Englewood Cliffs, New Jer-
sey: Prentice-Hall, 1964.
. Language, Thought, and Reality: Selected Writings of Ben-
Jamin Lee Whorf. Cambridge, Mass.: MIT Press, 1956.
Hall, Robert A., Jr. Pidgin and Creole Languages. Ithaca, New
York: Cornell University Press, 1966.
Hymes, Dell H. Foundations of Sociolinguistics: An Ethnographic
Approach. Philadelphia: University of Pennsylvania Press, 1974.
Hymes, Dell, ed. Pidginization and Creolization of Languages. Cam-
bridge: Cambridge University Press, 1971.
Jesperson, Otto. Language: Its Nature, Development and Origin.
Allen and Unwin, 1922. Reprint New York: Norton, 1964.
Jessel, Levic. The Ethnic Process: An Evolutionary Concept of Lan-
guages and People. The Hague: Mouton, 1978.
Silverstein, Michael. ‘‘Linguistics and Anthropology.’’ In Linguistics
and Neighboring Disciplines, edited by Theo Vennemann and Re-
nate Bartsch. Amsterdam: North-Holland, 1975.

LANGUAGE TYPOLOGY AND LANGUAGE UNIVERSALS

Greenberg, Joseph H. Anthropological Linguistics: An Introduction.

New York: Random House, 1968.
. Universals of Language. Cambridge, Mass.: MIT Press,
1963.
Jakobson, Roman. Child Language, Aphasia, and Phonological Uni-
versals. The Hague: Mouton, 1968.
 Summary 203

Muller, S. H. The World’s Living Languages. New York: Ungar,

1974,

EXERCISES

i: List several terms by which a man can be addressed in American

society. Do the same for women.
23 The following is a list of words meaning ‘‘ten’’ in various lan-
guages. Identify the language as centum or satem. Hint: the second
consonant should serve as the basis for your decision. (a) taihun;
(b) desimt; (c) sak; (d) tasn
Identify each of the following as a formal, substantive, implica-
tional, or statistical universal: (a) Any language that permits three
consecutive consonants at the beginning of a word automatically
permits two consecutive consonants at the beginning of a word. (b)
For all languages, no phrase of a specified type contained within a
larger phrase of the same type may be moved out of the larger
phrase by a transformation. (c) Redundancy, defined in phonologi-
cal terms, is about 50 percent in all languages. (d) All languages
have nouns and verbs.
 COHsAUP
YTS K

16.
HISTORICAL-
COMPARATIVE
LINGUISTICS

Rather than looking at variations in languages from culture to culture

as anthropological linguistics does, the field of historical-comparative
linguistics concentrates on variations in languages through time.
Much of this work is of necessity based on written, not spoken, lan-
guage, and it often requires the ingenuity and patience of a detective,
or a fan of jigsaw puzzles. We review here what is known about the
processes of linguistic change — in phonology, morphology, syntax,
and semantics — and the methods for reconstructing languages.

Language Change
Language is alive. Perhaps it is best to attempt to understand this
concept by considering what a dead language might be. A dead lan-
guage would be one whose pronunciation, grammar, and vocabulary
were frozen into an unchanging state. For example, Latin frequently*
is considered a dead language. It is assumed that the proper pronun-
ciation of Latin was that uttered by Cicero and Caesar and that the
vocabulary and the grammar of Classical Latin are defined by two-
thousand-year-old texts. In other words, Classical Latin today is the
language that was used in the Forum in pre-Christian, Imperial Rome.
A case can be made for the claim that Latin is not quite dead. Con-
sider, for example, the translation of Alice in Wonderland into Latin
by Carruthers (see Carroll 1964), in which several words had to be
204
 Language Change 205

coined for things nonexistent in the Roman Empire, such as feles-

copium (‘‘telescope’’) and aureorum decoctio malorum (‘‘orange mar-
malade’’). One sentence especially must have pushed Carruthers’s in-
ventiveness to its limits:
Saporem vero quodammodo mixtum habebat: repraesentabat enim scribli-
tam cerasorum — et lacticinia — et cydonia — et carnem gallinaceam as-
sam — et cuppedia — et panem tostum butyro illitum. [pp. 6-7]

which translates back into English as:

It had in fact a sort of mixed flavor: it represented cherry tart — and cus-
tard — and pineapple — and roast poultry meat — and toffies — and
toasted bread smeared ae

Latin is not quite dead, and modern languages are very much alive.
This means that their pronunciations, vocabularies, and grammatical
structures are constantly changing. Such changes sometimes involve
the outright loss of some forms and the acquisition of others. New
words are always coming into a language as the need arises; for ex-
ample, astronaut and beatnik. The shape of other forms is inevitably
altered over time. The following Old English passage from Matt. 2:7
clarifies this point:
Herodes pa clypode on sunderspréce da tungolwitegan
Herod then called in private-speech the astronomers
and befran hi georde hwenne sé steorra him z2teowde
and asked them eagerly when the star (to) them appeared
[Rigg (1968), p. 15]
Only one word in this passage, and, survives today without change
(at least in spelling). Others, such as hwenne (‘‘when’’) and steorra
(‘‘star’’) are recognizable even though their appearance has changed
markedly. Several other words — clypode, tungolwitegan, and be-

TABLE 16-1
Comparisons of Words from Four Periods of English

Early Modern
Middle English English Modern English
Old English (Wycliffe Bible, (King James (New English
(c. 1000) cr 1385) Bible, 1611) Bible, 1961)

clypode clepide called called

tungolwitegan astromyenes wisemen astrologers
befran lernyde inquired ascertained

Source: Derived from Rigg (1968), p. 22.

206 Historical-Comparative Linguistics

fran — dropped out of use at some point in the history of English, as

indicated in Table 16-1.
But the greater visibility of changes in vocabulary should not ob-
scure the fact that grammar also changes. Also, it is useful to consider
all language change as grammatical change, because a grammar that
describes a modern language or dialect is simply one particular gram-
mar in a succession of grammars through time.

Phonological Change
Although it is natural that languages should develop new lexical items
to describe new concepts and situations, it seems somewhat unex-
pected that languages can develop new sounds; yet they often do.
Sometimes an entire sound segment (perhaps one already present in
the language) is introduced in words that contain a particular phonetic
environment, as the b in thimble (from Old English (bymle) was in-
troduced and the initial e in Spanish escuela (from Latin scola,
‘“‘school’’). This phenomenon — phonological change through the ad-
dition of a sound segment in words containing a particular phonetic
environment — is known as excrescence and its 0 ite is loss. The
example of the French word école (also from Latin scola) exhibits
both excrescence and loss. Like Spanish, French developed an e
where Latin had an initial s plus a consonant; but the initial s was
subsequently lost in French.
Phonological change might seem accessible because it is manifested
in physical sounds, but some of the most basic questions concerning
phonological change are still to be answered decisively. Specifically,
these questions are (1) What causes phonological change? (2) How
does such change take place? and (3) How may phonological change
best be described?

THE CAUSES CF PHONOLOGICAL CHANGE

Several theories have been proposed to explain phonological change.
The simplest of these is actually no theory; namely, that the effects of
sound change may be described but their causes cannot be explained.
This view has cropped up in the writings of virtually every generation
of linguists. To admit a lack of knowledge is often courageous, but
this view seems to be unnecessarily defeatist.
One reasonable explanation for phonological change is that nge
makes certain sound combinations easier to pronounce. In fact, a
pe ee a eR RS Nees EE
by which particular features of sounds come to resemble their neigh-
bops Forexample, OldBagllth ibecame waco? tieseen ane
tween voiced sounds and following stress; thus, Old English heofon
 Phonological Change 207

became today’s heaven. Another example is the change in early Old

English from [g] and [k] to [y] and [€], respectively, in the neigh-
borhood of front vowels; for example, Old English geard became
yard and Old English pic became pitch. Thus, the consonants became
more similar to the vowels in their place of articulation. This frontin
of consonants to the soft- ct is =alled
palatalization, and it is common in the world’s languages.
e Opposite of assimilation is dissimilation, a relatively infrequent
process whereby sounds become phonetically less like nearby
sounds. Thus, from Latin peregrinus one of the two rs dissimilated to
I to yield Italian pellegrino. Although assimilation and dissimilation
are opposites, they both ordinarily function to simplify pronunciation.
Unfortunately, there are shortcomings to an explanation of sound
change based only on simplification of pronunciation. First, the lan-
guages of the world are not maximally simplified in terms of their pro-
nunciations. Furthermore, there are a number of counterexamples to
the theory, like the High Germanic Consonant Shift in which [p], [t],
-and [k] developed into [p‘], [tS], and [k*]. The latter involve forming
the mouth as for [p], [t], and [k] and actually beginning to pronounce
these sounds, but quickly ending with the fricatives [f], [s], and [x] (a
pharyngeal fricative). These certainly seem more difficult to pro-
nounce.
Another theory of the cause of phonological change is phonological
space or symmetry. For example, recall the case of the Great Vowel
Shift discussed in Chapter 4. The vowel shift that took place in late
Middle English affected all vowels. Thus, when the highest front
vowel /i/ became the diphthong /ay/, there was a corresponding change
of the highest back vowel /u/ to the diphthong /aw/. In this case, front and
back vowels exhibited a symmetry in undergoing a similar shift. An-
other example illustrating this theory is the introduction of [n] in En-
glish. Before [n] developed, English had voiced and voiceless stops in
three positions: labial, alveolar, and velar. However, English had
only the labial and alveolar nasals, [m] and [n]. By developing the
velar nasal‘ [yn], English achieved a kind of symmetry between the
stops and the nasals.
Any account of the causes of phonological change also must con-
sider the social and cultural aspects of language. In most Southern
dialects of American English, r-less speech had been prestigious until
very recently. The fact that r is now pronounced more regularly in the
South may, to some degree, reflect increased relations with other
areas of the country and exposure to the pronunciation of r on the na-
tional broadcast media. Similarly, the development of the phoneme /z/
in English was assisted by the many French words containing [z] in
positions where it was not previously found in English; these words
208 Historical-Comparative Linguistics

were borrowed when French was the ruling, prestigious language in

England.
More sophisticated approaches to the cause of phonological change
also exist. They consider the relationship between rules in a grammar
and between underlying representations and surface phonetic repre-
sentations. For example, it has been suggested that rules may shift to
an order that reduces the number of surface exceptions to a rule.
Such hypotheses imply that phonological change may take place in
order to make languages easier to learn.

THE PROCESS OF PHONOLOGICAL CHANGE

The next question to be considered deals with the process by which
sounds change. Specifically, are sound changes gradual or abrupt?
For example, Latin frater (‘‘brother’’) developed to French freére.
One might assume that the change from the Latin vowel a to the
French vowel é took place gradually, perhaps progressing along a
phonetic continuum from [a] to [a] to [2] to [ze] to [et] to [el] to
[et] to [el]. However, there is a good deal of evidence that sound
change is not gradual. The most convincing examples involve met-
at which two segments are interchanged, as in the
decclocmentaron Old English hros to the modern horse. It is also
difficult to imagine how the f_a phonetic segment c take
place gradually (for example, from Latin scola to French école).
und change throughout a speech community
is gradual, because all speakers do not adopt the new phonetic form
simultaneously. For example, a new form may be adopted by a partic-
ular socioeconomic class and then spread gradually to other classes.
Also, a change may take place in a restricted phonetic environment
and then spread gradually to other phonetic environments within an
idiolect or dialect. The process by which-a sound change spreads
throu xicon and is manifested in an increasing number of
words is called lexical diffusion.
i
RT,

DESCRIBING PHONOLOGICAL CHANGE

To describe phonological changes, three major classifications may be
used, as shown in Table 16-2.
Under the structural classification, significant phonological change
is i fects the number or distribution of phonemes in a lan-
guage, not mere change in pronunciation. For example, a speaker 0

1 Although these approaches to phonological change are too complex to pursue here,
the inquisitive reader should consult Paul Kiparsky’s important article ‘‘Historical
Linguistics’’ for further discussion.
 Phonological Change 209

TABLE 16-2
Classification Systems for Sound Change

Structural Changes Changes in Articulation Rule Changes

1. Merger (partial or complete) 1. Assimilation 1. Addition
2. Split 2. Dissimilation 2. Loss
3. Loss (partial or complete) 3. Metathesis 3. Reordering
4. Excrescence 4. Addition (excrescence) 4. Simplification
5. Loss
Source: The structural classification is derived from Anttila (1972); the articulatory clas-
sification, from Bolinger (1968); and rule changes, from King (1969).

American English may change the articulation of /e/ and /3/ from an in-
terdental position to a position in which the tongue is pressed lightly
behind the teeth, but this change is rather insignificant because no
new phoneme or distribution of phonemes develops as a result of the
change. But now consider an example from Old English in which the
front rounded vowels @ and y changed to e and i in all cases. Since e
and i already existed in Old English, the number of phonemes in En-
glish was reduced by two —a significant phonological change. This
d_to as complete merger; if the change had taken
é€ merger wou
have been partial. The opposite of a merger is a split, whe one
honeme develops~into two ore phonemes. Under the str
turalist classification, the Mi nglish phoneme /n/ split into [n]
before velars and [n] in all other environments. When the condition-
ing velars stopped being pronounced, the phoneme /n/ developed.
cation deals with cha in terms of articula-
the place and manner of articulation, the
position of the velum, and the activity of the glottis. This approach
has been applied in-structuralist phonetics and in transformational-
generative phonology by using distinctive features to characterize the
change. The chief type of phonological change in this scheme is as-
similation. In transformational-generative phonology, the process of
assimilation is described through the use of distinctive features pre-
ceded by Greek letters, as introduced on page 80. Thus, to express
the fact that in a certain language a nasal consonant assumes the same
point of articulation as the following consonant, a rule like the follow-
ing may be used:
C
Cc a anterior :
+nasal| B coronal ee ge
B coronal
210 Historical-Comparative Linguistics

Other major changes in articulation are addition (excrescence) and

loss; less frequent changes are dissimilation and metathesis.
The third system of classifying sound changes has been used
by transformational-generative theorists; it describes phonological
means of rules. For instance, [hw] has beco in many
dialects of Englis ugh innovation, speakers ha d the rule
w]—[w] to their grammars so that a wo n_is now pro-
nounced [wen] instead of [hwen]. The Children of these spé ;
however, are exposed to w only, and they have no reason to posit an
underlying hw. In other words, they have lost a rule and their gram-
mars have been simplified. <ln_transformarional generative theory. 2
change in underlying representations is called restructuring.
the phonological component of a transformational-generative
grammar is understood to contain at least some ordered rules, then
rule reordering consists of reversing the order in which the rules
“apply. Consider an example from Standard Finnish inwhich long mid=-
vowels become diphthongs; for example, vee changes to vie. Another
rule deletes medial voiced continuants, which are a type of consonant
involving a constriction or occlusion of the vocal cavity over a period
of time. Thus, the consonant y is deleted in texe, giving tee. But al-
though new long vowels are produced by the second rule, these
vowels do not become diphthongs by the first rule. These facts may
be captured by the following ordering of rules in Standard Finnish: (1)
diphthongization, (2) loss of medial voiced continuants. In certain
nonstandard dialects of Finnish, however, a form like tee does in fact
change to tie. For these dialects, the proper ordering must be (1) loss
of medial voiced continuants, (2) diphthongization. These two dif-
ferent orderings may be summarized as follows:

Standard Finnish Nonstandard Finnish

Underlying representation: reve Underlying representation: teve

Diphthongization: does not apply Loss of medial voiced continu-
Loss of medial voiced continu- ants: tee
ants: fee Diphtheongization: tie
Phonetic representation: tee Phonetic representation: fie

Transformational-generative grammarians have proposed that the dif-

ference between the Standard and nonstandard dialects of Finnish can
be characterized by a difference in rule ordering in their respective
grammars. The ordering of rules in nonstandard Finnish is an example
of feeding drder, inWHCH GueTale creates representaiane towhich,a

deletes a continuant, thus producing a form to which the diphthongi-

 Morphological, Syntactic, and Semantic Change 211

zation rule may apply. Bleeding order exists when a rule removes rep-
resentations to which a subsequent rule might have applied. Accord-
ing to certain transformational-generative grammarians, feeding order
is one type of ordering that is preferred by grammars; that is, rules
may tend to shift into an order that allows their maximal use in a
grammar.
The final type of rule change, simpli i i eneral-
ization of the_sou j d or the context_within which the
SO anged. This type of change may be illustrated by the in-
prefix before adjectives in English (see Chapter 6). The nasal segment
of this prefix had been pronounced [m] before labials and [n] before
all other sounds, but recently it has come to be pronounced [n] before
velars in many words for many speakers. The change of /n/ to [n]
before velars thus permits a simplification, or generalization: the nasal
segment agrees with the following consonant in point of articulation.
To the transformational-generative theorist, the only significant
aspect of change is restructuring, as a change in underlying represen-
tation is evidence of the difference between the parent’s and the
child’s grammar through time.

Morphological, Syntactic, and Semantic Change

Examples of morphological and syntactic change are abundant in the
history of English. For instance, Old English used a considerable
amount of inflection to express relationships within sentences. A
noun generally had four, five, or six different forms that corresponded
in large part to a system of grammatical endings inherited from Ger-
manic. The Old English noun faeder (‘‘father’’), for example, had the
following forms:

Singular nominative, dative, accusative — feeder

Sin itive — feeder(es)
Plural nominative, accusative — federas
Plural genitive — federa,
Plural dative — federum
a = os
Instead of these five forms in Old English, English today has only two
phonetic forms, [fadar] and [fadarz], the latter being used to express
both possession and plurality. Old English also had inflected adjec-
tives that agreed in case, number, and gender with the nouns they
modified; modern English has lost inflectional endings on adjectives
other than the -er and -est of the comparative and superlative forms.
With respect to verbs, Old English had seven classes of strong verbs,
which require different vowels for different forms. Thus, one of the
212 Historical-Comparative Linguistics

classes of strong verbs, exemplified by findan (‘‘to find’’), had the fol-
lowing forms:
Infinitive form of present: findan
Past tense, first and third person singular: fand
Past tense, plural: fundon
Past participle: funden
Modern English in this case has only two forms, find and found, al-
though verbs like sing have retained three forms (sing/sang/sung).
The plural formation of nouns provides another example of mor-
phological change in English. Old English had several ways of form-
ing plurals, some of which still appear today in forms like foot/feet,
oxloxen, man/men, and sheep/sheep. It also had many plurals that
were formed by adding the suffix -as, as in stan/stanas (‘‘stone/
stones’’). The -as plural ending developed into the -s ending of mod-
ern English, giving the plural stones for stone. Eventually, a noun like
hand, which had the plural form handa in Old English, became mod-
eled after the pattern of stan/stanas, so that its modern plural is
hands. This process by which forms become modeled after existing

regularization
or simplification. Thus, in case just described, rela-
tively few English nouns today would have to be marked as excep-
tions to the general rule of plural formation.
Changes in word order also have developed in English. In Old En-
glish, subject-object-verb word order usually occurred in dependent
clauses; for examplé-perni 2 pele cempan weron (“that they-noble~
‘soldters—were’’). Verb-subject-object word order often occurred in
Old English when the sentence began with certain adverbs of time or
the negative particte-re;-for
example, pa weron on pam campdome
Cappadoniscé cempan—Cthen were in the military service Cap-
padocian soldiers’’)_In English today, subject-verb-object word order
isthenotm indeclatalive- Sentences--<xceptatlel fegative-aduerbs
(for example, Rarely do we see films).
———_—
OC

SIMPLIFICATION AND ELABORATION IN SYNTACTIC CHANGE

Traugott (1969), representing the transformational-generative view of
syntactic change, has described syntactic changes in terms _of
__simplification
orelaboration, For example iaOF Baslish, modals Uke
cunn- (can), mag- (may), mot- (must), scul- (shall), and will- (will) did
not occur in the same sentence with the perfect form habban (have).
Thus, a sentence like They should have sold the house would have
been ungrammatical in Old English. Furthermore, progressive verb
forms, but not perfects, could occur in the same sentence with mo-
dals, so that a sentence like We had been arguing all day would have
 Morphological, Syntactic, and Semantic Change 213

been ungrammatical in Old English. These facts are captured in the

following phrase-structure rule from the grammar of Old English:

en ded: es
Matin + ENE)
j 2 S

However, since all combinations are now possible in English, the

grammar of English today has the following phrase-structure rule:
Aux— Tense + (Modal) + (Perfect) + (Progressive)
eR
ee |

[he grammar of English has thus been generalized or simplified.

Traugott illustrated elaboration with the case of surface-subject de-
letion in relative clauses. In a sentence like J have a son who is four
years old, the surfacé-subject relative pronoun who may be deleted
only when followed by the verb to be; and if it is deleted, the verb to
be also must be deleted. The application of this rule yields the para-
phrase I have a son four years old. In Early Modern English, how-
ever, all surface-subject relative pronouns could be deleted, as the
following examples from Shakespeare show:
1. ‘‘This is the man should do the bloody deed.’’ (King John, act 4,
scene 2, line 69)
2. ‘‘There is a man haunts the forest.’ (As You Like It, act 3, scene
2, line 377)
3. ‘‘I have a brother is condemned to die.’’ (Measure for Measure, |
act 2, scene 2, line 34)

Tr ition of restricti on the application of surface-

subject deletion in modern English constitute tion of the
grammar.
aT nf,

ASPECTS OF SEMANTIC CHANGE

Traditionally, semantic change has been regarded as a change in the

designated contexts of lexical items. It includes the extension or
eee ee For example,
consider the word poot, ch originally designated a rather small
body of water; in modern times it has been used in other contexts, as
in such expressions as pool our resources and typing pool. On the
other hand, the word meat originally referred to anything edible, but it
was later restricted to refer to the edible part of an animal’s body.
Certain semantic changes also might be viewed as changes in se-
mantic features. For example, the word fee derives from Old English
feoh, which meant ‘‘livestock’’ and ‘‘cattle,’’ as well as ‘“‘property”’

2 Traugott uses the term progressive as a cover term for beo- (‘‘be’’), wes- (‘‘be’’), and
weord (‘‘be’’), which require the present-participle marker on the following verb.
214 Historical-Comparative Linguistics

or ‘‘money.’’ In losing the former meanings, the word fee lost the
semantic features (+animate) and (+living).
The historical development of fee also illustrates that semantic
change often mirrors cultural change; that is, it reflects the fact that
cattle and livestock were once used as a form of payment. Another
example of the influence of cultural change is the use of the -ess end-
ing to express the feminine form of a noun. Until very recently, the
use of this suffix was in vogue, as in authoress, heiress, and seam-
‘-stress. But the women’s movement fostered an opposing trend — the
obliteration of sex distinctions through the use of the word person fol-
lowing nouns. (The movement also introduced the term Ms. to oblit-
erate the distinction between married and unmarried women.) Thus,
a major Philadelphia newspaper recently carried an advertisement for
a wait person, rather than waiter or waitress.
Borrowing is an important factor in semantic change, and it too is_
orien theresult ofsoctat-amt cunural interaction. Themost obvious
result of borrowing is the exceptionally large number of words im-
ported into English from French; paté, garage, and rendezvous are
examples. In turn, modern American English has provided a rich
source of borrowings for other languages, which have looked to
American terminology for such words as supermarket and napalm.
Finally, the concept of borrowing may be expanded to include phono-
logical adoptions from the written language. Thus, acronyms com-
monly seen in print like NATO and CIA enter the spoken language as
[neto] and [si+ay +e].

Methods of Reconstruction
THE COMPARATIVE METHOD

The most powerful means of reconstructing languages is the compara-

fidence that certain languages are members of a particular family and

to reconstruct the original forms of the parent language. Through the
comparative method, the Indo-European family, of which English is a
member, was established. The reconstructed ‘‘original’’ language of
the Indo-European family is usually referred to as Proto-Indo-
European; proto- indicates that a language has been reconstructed
without written records for confirmation. Reconstucicd forms, called
rotoforms, are preced ;
In app e€ comparative a “ linguist searches through
the individual languages for epee that_share similar
meanings and some phonetic similarity. Basic sets of words, such as
those for the lower numerals, familial relationships, and parts of the
 Methods of Reconstruction PANS)

body, are the best choices, because these words are most likely to be
native words; borrowings and words that are phonetically, but not
semantically, similar must be excluded because they do not reflect the
phonological development of the language.
The comparative method is illustrated by the examples in Table
16-3. The languages in the table are Romance languages. They pro-
vide a useful test for the comparative method because we can usually
compare reconstructions of Proto-Romance with written records in
Latin, the actual parent language of the Romance languages.
A glance at the various forms for ‘‘nose,’’ ‘“‘head,’’ and ‘‘goat”’
reveals striking similarities. In the words for ‘‘nose,’’ for example,
the first segment in all languages is [n]. In the second segment, four
of the languages have [a] but French has [e]. The_major principle of
the comparative method is that if a certain correspondence is noted
among the cognates, it should hold for all other cognates: if it does
ot uld_be_a logical explanation for the exception. Notice
that the equivalence of [a] in Italian, Ladin, Spanish, and Rumanian
for ‘‘nose’’ also holds for ‘‘head’’ and ‘‘goat.’’ But French has [e] for
“‘nose’’ and [e] for “‘head’’ and ‘‘goat.’’ In explaining the French dis-
crepancy, we may rely on other information, such as our knowledge
of French, written records, and natural language processes. The fact
that Italian, Ladin, Spanish, and Rumanian have [s] as the third seg-
ment for ‘‘nose’’ and the fact that the French word for ‘‘nose’’ is
spelled nez imply that French originally had a consonant following the
vowel. Also, French has [e] in ‘‘head”’ and ‘‘goat’’ (corresponding to
[a] in the other languages), and it is common for vowels to become
tensed at the end of a word (as in [ne]). The fact that French has [e
for ‘thead”’ and [e:] for “goat’’ poses no problem: vowels often are
lengthened before voiced obstruents, that is, a stop, fricative, or affri-
cate, In conclusion, we account for the differences in the French
word for nose by proposing three changes: the change from Latin [a]

TABLE 16-3
Word Correspondences, in Several Romance Languages

English
Meaning Italian Ladin French Spanish Rumanian

““nose”’ [naso] [nas] na [naso] [nas]

“head” — [ko] [kéf] [3 [kabp] [
“goat” [képra] [kévfa] [SeQy] [kaya] [kAphe]
Source: Derived from Bloomfield. (1933).
Note: Ladin is a language spoken in certain areas of Switzerland and Northern Italy. Two
dots indicate added length.
216 Historical-Comparative Linguistics

_to French [e]; the loss of the consonant; and the change from [e] to [e]
at the end of a word. ~
Another correspondence noted in these examples is [k}-[k}[§}[k}
[k], with French again being the exception. However, this corre-
spondence is regular since it holds for ‘‘head’’ and ‘‘goat,’’ as well as
other cognates. Accordingly, we assume that the initial [k] of Latin
Sc to [S] in French, since it is common for velars to become
palatalized in the neighborhood of front_vowels.
more interesting correspondence occurs in the third segment of
the words for ‘‘head’’ and ‘‘goat,’’ where the following situation is
noted:

Italian Ladin’ French Spanish Rumanian

Third segment of ‘‘head’’ p f f b p

Third segment of “‘goat’’ p Vv Vv b p

Although three phonetic segments are noted among the cognates for
each word, the correspondence is regular, except for the [f]/[v] dis-
crepancy in Ladin and French. Again, it is common for voiceless
sounds to become voiced when occurring between voiced sounds.
This process also accounts for the appearance of [b] in Spanish, and it
gives us further reason to reconstruct *p as the Proto-Romance form.
The comparative method thus allows us to reconstruct *nas-,
*kap-, and *kapra- as the Proto-Romance forms. These reconstruc-
tions are confirmed by the forms found in the written records of
Latin: nasum, caput, and capram. The endings of Latin words are
more difficult to reconstruct because they largely have been lost in
the languages that derive from Latin.
The verified Latin reconstructions give us confidence in the results
of the comparative method; consequently, we can reconstruct a
Proto-Germanic language from languages like Old English, Gothic,
Old High German, Old Norse, and so forth. In turn, Proto-Indo-
European is reconstructéd from languages like Greek, Latin,
Sanskrit, and Germanic. For example, from the Romance languages
we can reconstruct the word for ‘‘father’’ as *pdter (the form found in
Latin), and from the Germanic languages we can reconstruct *fader.
Comparing these forms with Sanskrit pita and Greek pater, and con-
sidering other information as well, we can reconstruct Proto-Indo-
European *poter. Of course, the comparative method can be used to
establish familial relationships and parent languages for other lan-
guage families, too.
To summarize, the comparative method may be viewed ompris-
ing the following steps: (1) the selection of appropriate SELTEE
 Methods of Reconstruction 217

TABLE 16-4
Grimm’s Law (with Examples from Sanskrit and English)

Rule Sanskrit English

Indo-European voiceless stops become Germanic trayas three

voiceless fricatives.
Indo-European voiced stops become Germanic yuga yoke
voiceless stops.
Indo-European aspirated voiced stops become rudhira red
Germanic unaspirated voiced stops.
Note: Sanskrit is indicative of Indo-European and English is representative of Germanic.

the determination of the sound correspondences among the cognates,

(3) the reconstruction of protoforms for the cognates, and_(4) the
statement of the sound changes that relate the protoforms to the indi-
vidual language forms.
The most famous result of the comparative method iSSamus as
a statement of the correspondences among certain consonants 1
Indo-European and Germanic. A summary of Grimm’s law, with ex-
amples from Sanskrit and English, is given in Table 16-4.
Although the comparative method has been applied most exten-
sively to phonological reconstruction, it has also been applied suc-
cessfully to syntactic reconstruction. For example, by considering the
word order in various Indo-European languages, linguists hav n
that Proto- -Eur jects b verbs and genitives
before nouns. The extension of the principles of the comparative
“method
to semantic data is still in its preliminary stages.
Despite the impressive results of the comparative method, it has
limitations. Fi in of the actual pronunciations of
recons forms; t nstructio honetic
as phonemic. Second, reconstructions presuppose a parent language
that is fi of dialects. But
dialect-frandee Latin and Paccenmane certainly were not.

INTERNAL RECONSTRUCTION
The other chief method of reconstructing languages is_internal recon-
struction, which establishes previous stages of a particular language
by considering data usually from that language alone. It is a very lim-
ited means of reconstruction because it r: lies on mor hophonemic al-

divinity. When two morphemes are semantically related but show a

linguist assumes that the difference originally
phonetic difference, thet
e
218 Historical-Comparative Linguistics

When such alternations become obscured by subsequent changes,

reconstruction becomes extremely difficult; the s/r alternation, which
occurs as a relic in was/were is an example.
One of the most noteworthy cases of internal reconstruction oc-
curred in 1879 when the Swiss linguist Ferdinand de Saussure ob-
served that the large majority of Indo-European stems reconstructed
by the comparative method took the form consonant-vowel-con-
sonant. As a result, he suggested that all Indo-European words had
the consonant-vowel-consonant structure at one time, but that certain
initial consonants were later lost. Approximately twenty-five years
later, excavations in Turkey uncovered texts from the Hittite Empire;
Hittite was later shown to be an Indo-European language. Working
with Hittite in 1927, J. Kurylowicz discovered that certain symbols
transcribed as fh (later calledJaryngeals) corresponded to the conso-.
nants originally proposed by Saussure, thereby increasing the credi-
bility of internal reconstruction. Thus, Indo-European had to be re-
constructed again in the light of the Hittite data; for example, *esti
(‘‘he is’’?) became *?esti, where the glottal stop symbol is the
laryngeal that had been transcribed as h.
Another example of internal reconstruction is drawn from Sanskrit
and Greek. In both languages, the perfect tense is formed by duplicat-
ing the first consonant or consonants of the root of the verb and add-
ing a vowel. Consequently, for the perfect tense of the verb meaning
“to give,’’ we find Sanskrit da-ddu and Greek dé-do-ka. In other
cases, where the root of the verb begins with an aspirated stop, one
would expect to find aspirated stops in consecutive syllables of the
perfect-tense forms. Instead, an unaspirated stop is followed by an |
aspirated stop. Thus, there is reason to believe that at some earlier
stage of the languages, aspirated stops did appear in consecutive syl-
_ lables of the perfect-tense forms but that a later change dissimilated
the initial aspirated stop to an unaspirated stop. A benefit of internal ,
reconstruction is that it yields forms that may be used in comparative
studies. The reconstructed forms of Sanskrit and Greek are useful in
reconstructing Indo-European.

Summary
Language change occurs constantly, regularly, and at all levels of lan-
guage. Although there is no known simple cause of phonological
change, several explanations. have been proposed, for example, artic-
ulation simplification and phonological space or symmetry. Also, .
many instances of phonological change are instigated by sociological
factors. Conceivably, phonological change might also make languages
easier to learn.
 Summary Z19

The process by which sounds change is another debated issue. It is

clear that some changes are abrupt, but the spread of a sound change
throughout a community and throughout the lexicon is character-
istically gradual.
At least three systems of classifying sound change have been pro-
posed: by change in phonemic structure, by change in articulation,
and by rule change. Both the structuralists and the transformational-
generativists have utilized the articulatory classification to explain the
many sound changes involving assimilation.
Many examples of change seem to involve the process of analogy,
which results in simplification or regularization. Syntactic change,
from a transformational-generative point of view, consists of simplifi-
cation or elaboration of grammars. Semantic change may be de-
scribed as a widening or narrowing of the contexts of lexical items, or
it may be described in terms of semantic features. Borrowing is also a
major factor in semantic change.
The comparative method allows linguists to group languages into
families and to provide earlier forms for the parent languages of the
families. The method of internal reconstruction permits reconstruc-
tion of forms for a language based on evidence from that language
only.

SOURCES FOR FURTHER READING

INTRODUCTIONS
Anttila, Raimo. An Introduction to Historical and Comparative
Linguistics. New York: Macmillan, 1972.
Arlotto, Anthony. Introduction to Historical Linguistics. New York:
Houghton Mifflin, 1972.
Bynon, Theodora. Historical Linguistics. New York: Cambridge Uni-
versity Press, 1977.
Hoijer, Harry, ed. ‘‘Language History.’ In Leonard Bloomfield,
Language. New York: Holt, Rinehart and Winston, 1965.
King, Robert D. Historical Linguistics and Generative Grammar.
Englewood Cliffs, New Jersey: Prentice-Hall, 1969.
Lehmann, Winfred P. Historical Linguistics: An Introduction. 2d ed.
New York: Holt, Rinehart and Winston, 1972.

THEORY OF LANGUAGE CHANGE

Hoenigswald, Henry M. Language Change and Linguistic Recon-
struction. Chicago: University of Chicago Press, 1960.
Kiparsky, Paul. ‘‘Historical Linguistics.’’ In William Orr Dingwall,
ed. A Survey of Linguistic Science. College Park: University of
Maryland, 1971.
. “Linguistic Universals and Linguistic Change.’’ In Univer-
220 AHistorical-Comparative Linguistics

sals in Linguistic Theory, edited by Emon Bach and Robert P.

Harms. New York: Holt, Rinehart and Winston, 1968.
Li, Charles, ed. Mechanisms of Syntactic Change. Austin: University
of Texas Press, 1977.
Stockwell, Robert P. and Macaulay, Ronald S. K., eds. Linguistic
Change and Generative Theory. Bloomington: Indiana University
Press, 1972.
Traugott, Elizabeth Closs. A History of English Syntax. New York:
Holt, Rinehart and Winston, 1972.
Weinreich, Uriel; Labov, William; and Herzog, Marvin I. ‘‘Empirical
Foundation for a Theory of Language Change.’ In Winfred P.
Lehmann and Yakov Malklel, eds. Directions for Historical
Linguistics. Austin: University of Texas Press, 1968.

EXERCISES

1. What changes from Middle English to modern English are illus-

trated by the following sentence from Chaucer: ‘‘Thou wost thou
wel what thinge is speche’’? What types of change are generally
easiest to identify from such an isolated example?
2. Name the processes of language change taking place in the follow-
ing hypothetical examples: (a) [opfen]—[otfen]; (b) [vina]—
[vinda]; (c) [eglo]— [elgo]; (d) [sapan]—[seban]; (e) [rand]— [ran].
3. Tell whether each of the following examples is a case of a merger,
a split, or neither:
a. Ina language with the voiceless aspirated stops /p", t", k"/ and no
unaspirated stops, the stops lose their aspiration only when at
the end of the word.
b. In a language with the vowels /1, e, 2, a, 0, u/, the pronunciation
of /z/ and /a/ both become [a] in all cases.
c. Vowels preceding word-final nasals become nasalized and the
word-final nasals are subsequently lost.
4. For many speakers of Black English, [6] is changed to [f] when it
is the word-final sound.
a. Specify what type of change this is, according to each of the
three classification systems of Table 16-2.
b. Next, suppose that some future development causes [6] to
become [f] in all environments. Again, specify the type of
change according to the three systems.
5. In each of the following hypothetical cases, tell whether the
change is possible or highly unlikely: (a) [tob]—[top]; (b)
[ap"o]— [apo]; (c) [ukit]—[usit]; (d) [skot ksi]—[skot ski];
(e) [ep]—l[up]; (f) [obeg]—[oneg]; (g) [posi]—[fosi]; (h)
{lem]— [lemu].
 i Summary 221

6. In England in the 1960s, many young men started to refer to

women as birds. Explain this semantic development in terms of
change in context and change in semantic features.
7. Consider the following cognates:
English Dutch | German Swedish
men man man man
hend hant hant hand

Assuming that the correspondences indicated here hold throughout

the languages, (a) state the correspondences, (b) reconstruct pro-
toforms, and (c) state how the protoforms developed into the given
forms.
 CH ALP RE

NEUROLINGUISTICS

The ultimate goal of linguistics has always been to give a full account
of how language works, and this account must include an explanation
of how the human brain operates to generate language. It had long
been assumed that such a goal was far beyond science. But in the last
decade, advances have justified the name of yet another subfield of
linguistics, neurolinguistics. It is concerned principally with the inter-
section of linguistic theory and brain structure. A tracing of linguistic
rules through the circuitry of the brain is still the stuff of science fic-
tion, but it is no longer in the realm of the impossible. Today, neuro-
linguists concern themselves with a number of issues that relate to the
neurological basis of language: the brain’s anatomy, the species-
specificity of language, and the relationship between language and
consciousness.

Aphasia and Brain Mapping

A good deal is now known about the gross anatomy of the speech
centers in the brain, much of it deriving from a century of aphasia
studies. Aphasia is the name for a variety of speech losses and speech
problems that are related to dysfunctions of the brain itself. Specific
kinds of aphasia are related to damage to specific areas of the brain;
so the kinds of symptoms experienced can give valuable hints about
the speech functions of the damaged brain areas.
222
 Aphasia and Brain Mapping 223

TYPES OF APHASIA
One of the first aphasias to be identified was Broca’s aphasia, named
after the French physiologist Paul Broca, who, in 1861, published a
paper in which he correlated an area of the brain with certain speech
processes. Broca’s area, as indicated in Figure 17-1, lies directlyin
front of the part of the mot that controls the cles in-
xolwe in
speechproduction, Bocaclaimedthatdamagetothis
ae
results ina Certain type of aphasia, later called Broca’s aphasia, which
is. characterized by extreme difficulty in speech, v oor pro ia-
‘ion,andustatly-the 40S6ofthesmallwords,orfunction words,ef
language (for example, a, in, for, and so on).
In 1865, Broca further claimed that aphasia results only when dam-
age occurs on the left side of the brain. The brain is divided into two
halves (or hemispheres) that are anatomically almost entirely indepen-
dent of each other but are connected by a number of nerve bridges.
The two halves of the brain are nearly identical, but each half controls
the physical activity of the opposite side of the body. For example,
the left hand is principally controlled by the right side of the brain,
and vice versa. Broca’s discovery seemed to indicate that speech is
controlled by the left side of the brain; this has proven to be so in
about 97 percent of humans.
In 1874, another physiologist, Carl Wernicke, announced the dis-
covery of a second major speech area, now called Wernicke’s area
(see Figure 17-1). It has since been determined that Broca’s and Wer-
nicke’s areas are connected by a long track erves.
ea a rina Broca’s. Speech is
fluent, pronunciation is unimpaired, and speech sounds more or less
normal until its meaning is considered. Wernicke’s aphasia is chara¢-
terized by a lack of content; speech_c ins Many circum i
and empty words, such as thing, as exemplified by the following pas-
sage:
Sometimes it happens to me here; I hadn’t done it this way; I just
wouldn’t do anything about it; now I’m trying to, and all of a sudden I find
out that it’s not doing the way it should, and I can see, I’ve said it time
after time after time; and I’m going to continue to push it on these people
until they get this thing back in shape here, so at least I have this much of
it pushed together; then I can go into things like this and start working
with it. Because a lot of the things that you and I are talking about right
now are just doing a little piece of the thing.!

Other kinds of aphasia have been described and associated with

damaged areas of the brain. One interesting type is that associated

1Harry A. Whittaker, ‘‘Linguistic Competence: Evidence from Aphasia,”’ Glossa 4,

1970, 48.
224 Neurolinguistics

FIGURE 17-1
Side View of the Brain (Left Side)

motor cortex

connecting nerve tract

front back

Heschl’s
area Wernicke’s
area

with Heschl’s area (see Figure 17-1), located near Wernicke’s area.
Heschl’s area is principally associated with the nerve pathways of
hearing, and damage to it results in what is called word deafness, a
form of aphasia in which the tis and words spoken
to him or her. Such people-are not deaf. But the damage to Heschi’s
—area has apparently prevented sounds from entering the speech com-
plex at the right point or in the right way. Speech cannot be under-
stood, even though it may be ‘‘heard’’ as sound. People suffering
from word deafness may have no difficulty at all in understanding
written language.

BRAIN AREAS AND LANGUAGE : .

Quite a few other types of aphasia exist, and all illustrate the same
point: specific locations of the brain are involved with pee aspects
of
of speech-ortanguage—tt ts difficult to be more precise,
“About exactly what that involvement is. Broca’s aphasia te
some sort of interruption in the actual production of speech sounds,
making phonation exceptionally difficult. The process of converting
abstract phonological units into actual speech sounds thus seems to
involve, at some point, actual electrical flow or activity in Broca’s
area; but this rather vague assertion is the best we can do for now.
Similarly, Wernicke’s area appears to be significant in semantic
processes. Interestingly, damage in Wernicke’s area does not appear
to affect either phonation or syntactic construction, suggesting that
these processes have some real independence from semantic pro-
 Speech and Species ae

cesses in the brain. No aphasia characterized by deformed syntax

alone has been described to date. In such a case, the patient would
produce semantically appropriate and well-pronounced sentences
whose syntactic structures would be regularly garbled.
In summary, interpretations of aphasia in relation to linguistic
theories must be extremely guarded. Linguists are very cautious
about asserting that any of the areas of the brain are actually as-
sociated with language per se, although they are obviously associated
with speech. A strong possibility exists that language rules depend on
activity in deeper areas of the brain, and that the speech complex that
includes Broca’s and Wernicke’s areas is made up of relatively super-
ficial pathways for turning language into speech.

Speech and Species

Researchers have compared the human brain to the chimpanzee’s to
determine whether the chimpanzee is equipped for speech. Such an
endeavor is significant, for the chimpanzee is our closest relative,
sharing a good deal of our genetic heritage. Zoologists work with the
assumption that brain structure is genetically determined, and ling-
uists further assume that language depends directly on brain struc-
ture. Thus, if any animal might be expected to have a brain structure
compatible with language, the chimpanzee would be it. Moreover, on
a gross scale, the chimpanzee’s brain is similar enough to the human
brain for physiologists to have considerable trouble telling them apart.
In several respects, however, differences exist that relate directly to
the capacity for speech.
Chimpanzees have no perceptible Broca’s area; the area equivalent
to Wernicke’s area is small and symmetrical in both hemispheres; and
there are no nerve bridges to connect the two. In other words, the
chimpanzee apparently lacks entirely what in the human are known to
be the major speech areas of the brain. This is not to say that chim-
panzees are incapable of vocalizations; chimpanzees, like other ani-
mals, are capable of a wide range of communicative noises, each of
which may be associated with different types of messages. But these
vocalizations may be regarded as stereotyped and automatic; that is,
the calls are automatic responses to specific stimuli in the environ-
ment, such as food or threats from other animals. Moreover, the
vocalizations are very limited; the chimpanzee has a stock of about
thirty distinct calls, which it may deliver with different intensity de-
pending on its emotional state.
The brain areas involved in the chimpanzee’s vocalizations are part
of what is known collectively as the limbic system, an area_dee
within the brain that a reat significance to
226 Neurolinguistics

human communication. In humans, the limbic system is involved in

Pe RE EERE BGR RSS ESET LEST ea .
speech only under circumstances of pain or extreme mental duress.
The limbic language is aroused, for example, when a person steps on
a nail.
The inescapable conclusion, then, is that humans are uniquely
equipped for speech with a brain that contains anatomical specializa-
“tions found in no other species. The question of language, however,
“cannot beTesolved asoatily Avstated before, linguists are not con-
vinced that language organization is necessarily closely identified with
the neural organization of speech. Thus, a lack of speech centers does
not necessarily define a lack of language ability. As a result, porpoises
and chimpanzees may indeed have the kind of mentality and neural
organization that is presumably required for language, even though
they lack humans’ unique capacity for speech. Because the neuro-
linguist has no idea of what the anatomical correlates of language or-
ganization might be, the matter must rest there for the moment.

Language and Consciousness

About 1940, it was discovered that epileptic seizures could be con-
trolled through a surgical technique in which the major neurological
bridges between the two halves of the brain are cut. Apparently,

FIGURE 17-2
Optic Connections

left field right field

YR optic nerve crossing

left hemisphere right hemisphere

 Language and Consciousness 227
- .

FIGURE 17-3
Optic Connections after Severing the Optic Crossing between the
Two Sides of the Brain

left field oe i
right i
field

major epileptic seizures involve an uncontrolled reverberation back

and forth between the two sides of the brain; when the bridges are
cut, the reverberation is forestalled. Subjects whose hemispheric con-
nections are severed are apparently able to lead normal lives, with no
diminution of intellectual capacities. Experiments with these split-
brain subjects have provided some fascinating insights into the rela-
tionship between speech and consciousness.
To understand these experiments, we need to know some basic
facts about vision. In normal persons, the retina of each eye is basi-
cally divided into left and right halves, with each half of the retina
connected, via the optic nerve, to one side of the brain, as indicated
in Figure 17-2. The left visual field of both eyes is connected to the
right hemisphere (see the solid lines in Figure 17-2), and the right
visual field to the left hemisphere (dashed lines). Notice also that
the optic nerve crosses to carry the right-field vision in the right eye
to the left side of the brain and the left-field vision in the left eye
to the right side of the brain. When the bridges between the two sides
of the brain are cut, the optic crossing is cut as well; the result, as
shown in Figure 17-3, is that the left field is perceived solely with the
right eye through the right visual area and the right field is seen only
with the left eye through the left visual area.
228 Neurolinguistics

FIGURE 17-4
Perception of Hatband with Optic Crossing Severed

hat Cee band

Because both fields can be seen, vision is apparently normal in

split-brain subjects. However, when split-brain subjects were shown
two-word compounds such as hatband with hat in the left visual field
and band in the right, as shown in Figure 17-4, they reported that
they saw the word band only. Because the brain had been split so that
the two halves were independent of each other, hat could not be
transmitted to the left side, where speech is produced, and so hat
could not be talked about.
A second experiment was quite similar. In this case, the right visual
field contained the picture of an innocuous object, an apple, but the
left field had a picture of a naked woman. The subject was a woman
who was known to react with blushes and nervousness when shown
off-color material, and in this case she also blushed and became ner-
vous. When asked why the apple was making her uncomfortable, she
replied that she did not know. This experiment demonstrated that she
was actually seeing with both eyes, but that the material being carried
to the right side of her brain was still not getting into the language
area even though it caused her to respond with blushes.
Both of these split-brain experi that the
relationship between consciousness and unconscious is actually a
matter of access to the speech areas of the brain. In other words, if a
person can talk about something, he or she is conscious of it; but if
 Summary pipes)

for some reason the information is withheld from the speech area, he
or she is unconscious of it.

Summary $
The field of neurolinguistics is still in its infancy. Much of the work in
progress today is still too uncertain or inconclusive to report here.
However, evidence has accumulated in at least three areas that war-
rant discussion: the brain’s anatomy as it relates to speech, the spe-
cies-specificity of language, and the relationship of consciousness to
speech.
Evidence from studies of aphasic individuals, supplemented by re-
lated research, demonstrates that specific areas of the brain are in-
volved in the production of speech and that these areas are located on
the left side of the brain in most people. Anatomical comparisons of
chimpanzee and human brains demonstrate that chimpanzees lack the
specific areas of the human brain that are associated with speech.
Finally, evidence from people whose brains have been surgically
divided suggests that consciousness of something is intimately
sociated with the ability to speak about it.

SOURCES FOR FURTHER READING

Gazzaniga, Michael S. The Bisected Brain. New York: Appleton-Cen-

tury-Crofts, 1970.
Lenneberg, Eric. Biological Foundations of Language. New York:
John Wiley, 1967.
Penfield, Wilder, and Roberts, Lamar. Speech and Brain Mecha-
nisms. Princeton: Princeton University Press, 1959.
Whittaker, Harry A. ‘‘Linguistic Competence: Evidence from Apha-
sia,’ Glossa 4, 1970, 46-54.

QUESTIONS FOR DISCUSSION

1. Why are neurolinguists interested in aphasia and epilepsy?

2. How does a chimpanzee’s brain differ from a human’s? What con-
sequences does this have for the linguistic capacity of each?
3. What part of the brain is involved in speech production for most
people?
 CHAPTER

18) “
ACOUSTIC
PHONETICS

Since the nineteenth century, linguists have been concerned with pho-
netics; however, the exact study of acoustic phonetics — the study of
the physical properties of speech — did not begin to progress until
the invention of the speech spectrograph in the late 1940s. With this
instrument, phoneticians could precisely analyze the waveforms and
frequencies of the sounds that make up human speech. In order to ap-
preciate the physical nature of speech, we must discuss some basic
physics as it applies to sound. We can then discuss the characteris-
tics of vowels and consonants and some applications of acoustic
phonetics.

The Physical Basis of Speech Sounds

Sound is a form of energy; specifically, it is energy reflected in vibra-
tions intheair, although
travels
sound through
water
and other gases
and liquids as well. There are many kinds of sounds, the simplest of
which is a pure tone, or the sound given off by a tuning fork. If you
strike the tines of a tuning fork — for example, a fork of concert pitch
A — it begins to vibrate exactly 440 times per second, or at 440
cycles per second. Cycles per second_are now designated as Hertz
(abbreviated Hz). Thus, the tone A involves a_simple vibration 0
Hz, which is its frequency — the-rate_of vibration. The vibrations of
the tuning-fork tines set up equivalent vibrations in the air around the
230
 The Physical Basis of Speech Sounds 231

FIGURE 18-1
Pure Tone

ee fork
i
ea
al ae

CN eosin

tuning fork, as indicated in Figure 18-1. If 440 Hz is the note A, then

880 Hz is the A that is one octave above the original note, and so on.
Vibrations occur with different amplitudes as well as aie OLA The,
amplitude of the wave is a measurem ¢ of the sound.

sound wave that are directly related to our psychological responses to

OMEN cent aerelatharia: bigypeicentioh OLuiigntampttadede-re*
Tated
tuning fork obeys the laws of the universe; it will not vibrate for-
ever. But it will not slow down; the frequency of the fork’s vibration
will remain constant while the amplitude decreases to zero as it runs
out of energy. One way to represent frequency and amplitude is to
use a frequency-response diagram, in which the amplitude is repre-
sented by the darkness of the line, as indicated in Figure 18-2.
Almost all sounds are more complex than that of a tuning fork
because they are made up of many frequencies of differing strengths.
Waves add together, resulting in a composite wave that is usually dif-
ficult to evaluate. However, waves may be analyzed by a spec-
trograph, a machine that sorts out the components of a complex
wave. The spectrograph makes use_of another phenomenon called
sympathetic vibration. ng fork is vibrating, other A tuning

FIGURE 18-2
Frequency-Response Diagram

loud A note quieter A note

440

frequency
(Hz)

time in seconds
232 Acoustic Phonetics

FIGURE 18-3
Sympathetic Vibration of Tuning Forks

forks in the vicinity will begin to vibrate ‘‘sympathetically,’’ being set

in motion by the sound waves emanating from the first. However,
tuning forks other than A will not vibrate (see Figure 18-3). More-
over, the amplitude of the sympathetically vibrating tuning fork will
depend directly on the amplitude of the waves that set it vibrating.
A sound spectrograph is little more than a bank of many tuning
forks, each of which is different from the next. A complex wave con-
sisting of four wave components will cause four of the spectrograph
tuning forks to vibrate sympathetically, making it possible to deter-

FIGURE 18-4
Spectrograph

ee
complex
wave
 Vowels 229

FIGURE 18-5
Spectrogram

a ie)(=)

ro)S
psy

frequency
(Hz)

time

mine the frequency and amplitude of each component wave (see Fig-
ure 18-4). The spectrograph prints this information on a frequency-
response chart called a sound spectrogram (see Figure 18-5).

Vowels

Speech sounds are complex; in fact, they are among the most com-
plex of naturally occurring sounds. They are made up of sound waves
ranging from 20 Hz to 20,000 Hz — the entire audible sound range.
However, most of the range above 5000 Hz is not very important to
speech.
A vowel, such as [i], is made up of several characteristic frequen-
cies, as indicated in Figure 18—6. The lowest of these characteristic
frequencies (V in Figure 18-6) is the voice bar; it represents the fun-
damental frequency of vocal-cord vibration in voicing. All vowels, of
course, are voiced. The fundamental frequency will depend on the in-
dividual’s voice, and there are great differences among male, fe-
male, and children’s voices. When [i] is produced, a significant band of

FIGURE 18-6
Sound Spectrogram of {i]

2300
N
F2
x
>
oO

3 270 FI
i=y
8
= 50 Vv

time
234 Acoustic Phonetics

FIGURE 18-7
Spectrogram of the Major Vowels

2500

1500

1000

750

500

250

wave energy is observed at about 270 Hz; this is called the first for-
mant, or Fl. No significant wave energy is found at higher frequen-
cies until about 2300 Hz, which for [i]is the second formant, or F2 4
formant may be thought f-sound_caused
y the particular shape of the vocal tract durin
Other highe -$3,F4-¥5, SO On, exist as well; but the first
two formants carry enough information to distinguish all vowels.
Each vowel, then, is associated with a characteristic Fl and F2, as
shown in Figure 18-7. For example, for [e], Fl is around 520 Hz and
F2 around 1870 Hz.
How are formants made in the mouth? The mouth is a reasonating
chamber that works in essentially the same way that a pipe organ
does. The air in a pipe will |vibrate at certain frequencies depending
on three aspects of its construction: (1) its length, (2)_its_cross-sec-
—~tional
area, and (3) whether it is open Kaciacedice the_end. A pipe
organ consists of a*bank of pipes of different sizes, each pipe resonat-
ing to one specific frequency; each note on the keyboard is connected
to a shutter that allows air to get into the different pipes to set their
air into vibration. By moving the tongue and lips, different-sized
shapes can be created in. the mouth, each of which has different res-
onating characteristics (see Figure 18-8). As the vocal cords vibrate,
a column of vibrating air ascends into the mouth, where it resonates,
—
T against F2, as in
in Figure y,
: one finds the vowel
triangle familiar from the vowel chart.discussed in Chapter 5. In some
 “UOISSIULIOd YIM Posy) ‘SS2ig O8vdIYD Jo AjIs
-IDATU) OY, AQ 7961 © IYWSUAdOD *16-96 “dd ‘(Zg6] ‘SSoid OBVIIYD Jo AjISIOATUA) OY, :O3VIIYD) SoyaUuoY Isnooy fo sjuawaly ‘pasojapey] Jo} /294n0g
000°€ 000° 000°T 000°€ 000°7 000°T
p,oym

pey
~ 000°€ 000°7 000°T 000°€ 000°% 000'I
ee tod ate it ee
YW Uy
58
peoy
000°¢ 000°7 000°T 000°€ 000°Z 000°T
|; | pomey | } PH

000'€
000‘%
000°T
000°€ 000°2 000‘T
—————— j

peoy
DAJIadg J2MOA O]
PalVJAY JO
SuOI]ISOg JVIOA JIVAT
8-81
AUNOIA
236 Acoustic Phonetics

FIGURE 18-9
Fl versus F2

F2
2500 1100 850
i u 250

e fe) 500 Fl

a 750

way, F1 corresponds to tongue height and F2 corresponds to tongue

frontedness, although it is really a little more complicated than that.

Consonants

The acoustic picture of consonants is slightly more complex than that

of vowels. Stop consonants generally have a sharp burst of energy at
all frequencies, and there is no apparent difference in the appearance
of the bursts of any two stops (see Figure 18-10). There is an interval
of as much as 1/20 of a second between the burst and the beginning of
the vowel formants of the vowel that follows it. These formants, how-
ever, have not yet stabilized at their characteristic levels; they are in
transition toward them. It turns out that the transitions can be ex-
tended back to the burst, thus locating a particular point on the burst
(see Figure 18-11). This point is characteristic of particular articula-
tory positions. For example, [t] and [d] are characterized by a burst
locus at about 1800 Hz, and the locus for [p] and [b] is at about 700

FIGURE 18-10
Stop Consonant Burst

3000 +

frequency
(Hz)
200 L

time
 Some Applications 2a

FIGURE 18-11
Finding a Consonant Locus

3000
<)
mo)
e
Ss oe
——
S tS,
3 Cr
cha ee

250

| sligoaectons ass
ail z2 3 4 a)
time in seconds

Hz; the locus for [k] and [g] varies, depending on whether they are
followed by a front or back vowel.
The differentiation of voiced and voiceless consonants is not — at
least in English — the presence or absence of a voice bar during the
consonant burst, as it is in French and Russian. In English, there is
no voice bar simultaneous with the burst. Voicin j ter the
burst; it will be considerably delayed (up to 1/20 of a second) after a
voiceless stop, but much less delayed a to about 1/50 of a second)
after a voiced stop. The discovery that the voiced consonants of En-
glish are not really voiced during the initial stages of their production
has led some linguists to suggest that they not be labeled voiced. In-
stead, linguists suggest that the terms tense (for [p,t,k]) and /ax (for
[b,d,g]) be used.
The fricatives are characterized by diffuse energy scattered in sev-
eral spectral regions. Most of the component frequencies for the con-
sonant [s] are above 6600 Hz, whereas the concentration of energy for
[8] is between 3000 Hz and 4500 Hz. The energy of [f] and [0] is gen-
erally weak; they are distinguished by transitional formants in the
same manner as the stop consonants. Unlike the stops, however, the
voiced fricatives are accompanied by a voice bar.

Some Applications
Knowledge of the acoustic characteristics of individual speech sounds
has made it possible for scientists to construct talking machines. A
computer can be hooked up to a bank of vibrating units, and informa-
tion fed into it concerning those frequencies and intensities that it
should simulate. The result is recognizable speech. One of the earliest
efforts of this kind resulted in a recording of a rather inhuman but
238 Acoustic Phonetics

completely recognizable rendition of the song ‘‘Daisy, Daisy’’ on a

record called Music for Mathematics produced by Bell Labs. Later,
this song was featured in the movie 200]: A Space Odyssey and sung
by Hal, the computer, as he (it) died.
Applications are widespread and, for the most part, obvious. Doors
might be computerized to recognize and open to certain voice com-
mands; voice-operated typewriters and computer consoles could re-
place secretaries and some programming technicians; “‘Home,
James’’ might be directed at the chauffeurless computer car, and so
on. Why do these things not yet exist? The problem is that at the
acoustic level, no two voices are exactly alike; indeed, no individual
person says the same sound exactly the same way twice. The
frequencies given in Figures 18—6 through 18-11 are the average fig-
ures for adult male voices. Variations are enormous, and about the
only constancies are the relationships between the Hz values for par-
ticular sounds for a given individual.
Simple machines have been constructed to respond to numerals
given vocally; but they respond only to a few individuals, and even
then not all the time — for instance, not if an operator becomes care-
less with his or her pronunciation. That these problems will some day
be solved is almost certain. The machines will have to be extremely
sophisticated, however, and will likely have a preprogram in which
the machine tunes into the acoustic values and range of variation for
each speaker who wishes to use it.

Summary
The study of acoustic phonetics was accelerated in the late 1940s by
the invention of the spectrograph, which allowed analysis of the
waveforms and frequencies that make up human speech. Vowels and
consonants have different acoustic characteristics. Vowels may be —

pared sounds. Consonants, too, may be distinguished from each other

rough analysis of their wave frequencies, and the articulatory dis-

SOURCES FOR FURTHER READING

Fant, Gunnar. Acoustic Theory of Speech Production. s’Gravenhage:

Mouton, 1960.
Fry, D. B., ed. Acoustic Phonetics: A Course of Basic Readings.
Cambridge: Cambridge University Press, 1976.
Ladefoged, Peter. Elements of Acoustic Phonetics. Chicago: Univer-
sity of Chicago Press, 1962.
 Summary 239

Lehiste, Ilse. Suprasegmentals. Cambridge, Mass.: MIT Press, 1970.

Lieberman, Philip. Speech Physiology and Acoustic Phonetics. New
York: Macmillan, 1977.

QUESTIONS FOR DISCUSSION

1. What do frequency and amplitude correspond to in our perception

of language?
2. Explain how a spectrograph works.
3. Explain the significance of formants in spectrograms. What do the
first and second formants roughly correspond to in speech produc-
tion?
4. Briefly describe the acoustic differences between vowels and con-
sonants.
 ANSWERS
TO
EXERCISES

Chapter 5 Phonetics
iI In beginning to pronounce the word ski, the tongue is closer to the front of
the mouth than it is in beginning to pronounce the word school; more no-
ticeably, there is no rounding of the lips for ski as there is for school.
These initial differences are determined by the following vowel in both
words. In ski, the vowel is the fronted [i], and in school, the vowel is the
backed and rounded [u]. The effect of the vowel on the pronunciation of
[s] is noticed even across the [k] boundary, which both words share. The
nature of these boundaries is thus called into question. In fact, studies in
acoustic phonetics show that a clear break between the individual phonetic
segments of a word simply does not exist. Nevertheless, each segment has
some distinctive acoustic characteristics that we perceive, and the division
of words into phonetic segments makes speech easier to discuss.
. No. The letters of the English alphabet do not always represent only one
sound. It is the requirement of a phonetic alphabet that every symbol rep-
resent only one sound.
. (a) [50]; (b) [baksaz]; (c) [twink!j; (d) [ersal]; (e) [fyu]; (f) [jo]; (g) [sizon];
(h) [ciz]; (i) [strip]; (j) [Sud]; (k) [keetostrafik] ~
Ja: oyll
b. [a:]
. Languages like English have midvowels that are neither [high] nor [low].
In English, the sounds [e], [e], [A], and [o] lie midway between the two
extremes. However, sounds cannot be both [+high] and [+low] for the
same reason that something cannot be both hot and cold: it is a logical and
physical impossibility.
(a) [+continuant]; (b) [+anterior]; (c) [+sonorant]; (d) [—consonantal]:
(e) [—coronal]
. The suprasegmentals cannot be segmented and cannot be described easily
in terms of articulatory apparatus. Instead, they are more closely related
to the acoustic nature of speech.
. The vowel represented by o in convict is [a]; that in convict is [9].
. (a) resign; (b) agonizing; (c) petinia; (d) present; (e) polysyllabic

240
 Chapter 7 Phonology II 241

Chapter 6 Phonology
1; Phonetics is concerned with describing and representing speech sounds.
Phonology deals with how sounds are organized and how they function.
Phonology is an abstraction from the basic data of phonetics.
. (a) yes; (b) no; (c) no; (d) no; (e) yes; (f) yes
. Underlying representations express the relatedness of forms containing
alternations, such as those represented in the forms sane/sanity or
sign/signature. Underlying representations allow one set of rules to gener-
ate a large number of related forms.
. No, because [m] and [m], and [n] and [n] do not contrast in minimal pairs.
They exist in complementary distribution in English; that is, syllabic [m]
and [n] cannot appear in mother and nor.
. (a) Other silent letters include b (bomb/bombard) and n (hymn/hymanal,
solemn/solemnity). Many silent letters, however, have no pronounced al-
ternant; for example, knife and gnu. (b) In most cases the “‘silent’’ letter, if
pronounced, would violate restrictions on permitted sequences in English.
There simply are no [-mn] pronounced words in English, nor are there any
[gn-] pronunciations. One hesitates to say why there should be no such
words — perhaps such pronunciations should be considered difficult for
our delicate American mouths — but at least we may be sure that there
are none. If the underlying representation should, for some reason, contain
such a sequence, then one or the other of the sounds must be eliminated
by some sort of rule.
6. A is rewritten as B or C when followed by D or DE at the end of a word.
ik a. C—>[+voice] / V Vv
bal.G
+voice |-9 7 $

Chapter 7 Phonology II
I. (a), (d), and (e).
je The meaning of sign is related to the meaning of signify, signal, and other
words in which the /g/ surfaces in phonetic form. We can account for these
regular and common alternations by assuming an underlying /g/ in sign and
stating the rule that /g/ is deleted before syllable-final nasals. The word
gnat does not meet the requirements of this rule; but more importantly,
gnat does not enter into alternations with other words in which the /g/
surfaces.
. a. [—-high], [—low], [+back], [+syllabic]
b. [+nasal]; [anterior] and [coronal]
. Two consecutive stops are not permitted to occur word initially.
. On the first cycle, apply the main stress rule to the expression
[vp L, young], fy grand], [ychildhy lyhp; A means adjective. On the second
cycle apply the compound stress rule and on the third cycle apply the
nuclear stress rule. The derivation proceeds as follows:
242 Answers to Exercises

[vel young ]a[n[, grand ]4[y child ]yInInp

1 1 1 Cycle |
1 2 Cycle 2
2 1 3 Cycle 3

. a. By the main stress rule.

b. No. Jn is a preposition and the rule carries the constraint that it works
only with nouns, adjectives and verbs.
a. /f/ is unmarked; /v/ is marked. /v/ incorporates the additional feature of
voicing.
b. /f/ is marked; /v/ is unmarked. It is more natural for two-word final ob-
struents (stops or fricatives) to agree in the feature of voicing. Thus, if
the word-final obstruent is voiced, it is more natural for /v/ rather /f/, to
occur before it.
c. [s] is unmarked; @ is marked. [s] is one of three regular plural endings
in English (along with [z] and [az]). Forming the plural by not changing
the word is irregular, occurring in only a few cases. By the notion of
markedness as addition, one would suspect that [s] is the marked form
of plural formation, since it incorporates the addition of a feature that is
not present in the plural form of fish. And it certainly seems odd that
the marked form can be @. Nevertheless, most of the world’s languages
have some marker that indicates plurality; the lack of such a marker (as
in the irregular form fish) would thus be considered marked.

Chapter 8 Morphology
— . {The}, {old}, {-er}, {gentle}, {man}, {vote}, {-d}, {wise}, and {-ly}.
. The word lawyer is composed of two morphemes, {law} and {-yer}. The
morpheme {law} has a clear meaning and the morpheme {-er} is a suffix
indicating an agent. The problem is what to do with y. It is not a separate
morpheme because it carries no meaning by itself. Even if we say that -y is
an adjective marker in English, y clearly does not fulfill this function in
lawyer. Therefore, y is best analyzed as part of a morpheme {-yer}, which
occurs as an allomorph of {-er} in the word Jawyer. Support for this deci-
sion is exhibited by other words such as savior and collier.
. No. These two morphemes represent entirely different meanings.
. Free morphemes: The, child, skip, rope, and, play, game, and joy. Bound
morphemes: -ren, -ed, -ed, -s, and -ly. It can be argued that -ful can be ei-
ther a free or bound morpheme, since it appears independently as full in
full house. However, it is commonly used as the suffix -fu/ in modern En-
glish. Derivational morphemes: -ful, -ly. Inflectional morphemes: -ren, -ed,
-ed, and -s.
of a. word-formation
b. adjustment (Standard transformational-generative phonology treated
this as a phonological rule.)
c. adjustment
d. word-formation
6. Both of the words proof and prove would be marked with identical fea-
 Chapter 9 Syntax I 243

tures in the lexicon, with the exception that proof would be marked
[+noun] and prove would be marked [+verb].

Chapter 9 Syntax I
1. The element wh is not intentionally produced; it represents a pause while
the speaker is formulating his or her thoughts. The elements st- st- are
stuttering errors; the speaker certainly did not intend to produce st- three
times. These elements, then, are a part of performance and need not be ac-.
counted for by a competence model.
red 24, (55 G7, B14
else 4 wy Gare Oil 12513
ne oe Pec eer, epoca few fenCae
syntactic
. syntactic
eagcpogp
. semantic
4. The first phrase-structure rule may be reformulated as:

S—NP+ Aux + VP+ (Adv)

Adv must be placed in parentheses because not all sentences contain ad-
verbs.

Chapter 10 Syntax II

>
Lex[Det] Lex[N] Lex[{V] Det N
|
Lex[Det] Lex[N]

the operator Pres have -en be -ing_ ring the number

The FF rule must apply three times: once to Pres + have, once to -en + be,
and once to -ing + ring.
2. a. Permissible.
b. The rule says ‘‘choose between A and B”’ but the tree diagram shows
that both A and B have been chosen.
c. Lex[V] requires a verb; farmer is a noun.
3. The sentence is ambiguous: the hunters is the subject NP under one in-
terpretation and the object NP under another. Transformational-generative
grammar accounts for this ambiguity by assuming two distinct deep struc-
tures for the sentence — an option not originally available to the struc-
turalists or the traditionalists.
244 Answers to Exercises

Ques Neg Aux VP

7 ao
N Tense Perf Prog Vv NP
teenie
Det N
|
Lex[N] Lex[V] Lex[Det] mes

Jefferson Pres have -en_ be -ing He Ay house

Lex[Det] Lex[N] NP Aux VP Tense Adj

| Ea aS ee
Det N Tense V NP : Lex[Adj]
| | eo
Lex[Det] Lex[N] Det N
| |
Lex[V] Lex[Det] Lex[N]
|
a clown a clown — Pres get a laugh Pres be happy

The relative transformation moves the second identical occurrence of a

clown into the main sentence, leaving only one occurrence of the clown
and inserting who into the embedded sentence. The constituents Pres + get
and Pres+be must undergo FF. Finally, LF must apply to get + Pres to
give gets, and to be + Pres to give is.
6. In theory, an infinite number of embeddings are permissible in English
(competence). In actuality, the number of embedded sentences that one
can produce is limited (performance).
7. In the deep structure of this sentence, the cats and the dogs constitutes
conjoined NPs. Thus, moving the NP the cats out of the conjoined struc-
ture for the purpose of relativization is forbidden.

Chapter 11 Semantics I
1. (a) 3 nouns (x steals y from z); (b) 1 noun (x evaporates); (c) 2 nouns (x
likes y).
2. This sentence presupposes or implies the following facts: The speaker is
asking someone a question in reaction to a statement just made usually by
that person. This statement supplied the name of at least one guest who
would be appearing on ‘‘The Tonight Show.’’ Moreover, the speaker pre-
supposes that this particular show will actually be aired that evening; that
is, the evening of the day on which the sentence is uttered. Also, the show
 Chapter 13 Psycholinguistics 245

must be aired during the time or later than the time that the sentence is
uttered.
. In one sense, homophony might be viewed as a specialized case of ambi-
guity, as in sentences like John walked by the bank. This sentence is am-
biguous because bank (and thus, the sentence) may have any of three
different meanings. However, there is a basic distinction between
homophony of words and the structural ambiguity of sentences. The in-
terpretations in homophony are in terms of word definitions, perhaps ex-
pressed by semantic features or markers; whereas in structural ambiguity,
interpretations involve different syntactic relations.
. This sentence may be considered ambiguous because the clause then I had
a beer may be either part of the dream or a separate event in reality.

Chapter 12 Semantics II
. (+child-bearing)— (+woman)
. Interpretive semantics, generative semantics, case grammar, and relational
grammar are all theories of transformational-generative grammar. They all
employ rules to relate observed surface structures to underlying struc-
tures. All these theories have the same basic goal: to relate sound and
meaning. They all incorporate a lexicon, and they all specify semantic rep-
resentations in terms of a metalanguage — a specialized language used to
describe another language — such as semantic markers or formal logical
representations. In fact, all four theories hope to explain essentially the
same semantic facts, but their approaches and capabilities vary widely.
Ww. “For all things, there is a God such that God created all things.”’

. True. (pVq) is true because if either of the statements is true, as is p, the

combined statement is true. ~g is true because the negation of a false
statement is true. Thus, two true statements are combined by & to yield
another true statement.
. The locutionary act is the actual utterance of the sentence What time is it?
with its particular meaning. The illocutionary act is the request of the
speaker for information. The perlocutionary act is the result that someone
tells the speaker the time.

Chapter 13 Psycholinguistics
1. Transformational-generative grammar is probably the best current model
for describing what a speaker knows about his or her language. For exam-
ple, a speaker of English knows that the phrase the shooting of the hunters
can mean either ‘‘the hunters’ shooting’’ or ‘‘someone’s shooting of the
hunters.’’ The transformational-generative grammarian posits different
deep structures for these two interpretations to reflect the speaker’s knowl-
edge of this aspect of English. Furthermore, if certain features of
transformational-generative grammar are shown to have psychological cor-
relates, these may provide useful information in the construction of an
adequate model. Finally, transformational-generative grammar may ulti-
mately shed some light on other psycholinguistic issues like language ac-
quisition, and it may provide insights for other areas of study.
246 Answers to Exercises

2. Such a sentence would not violate any of the rules of sentence formation
in English. Moreover, the rules used to generate such a sentence are
known by all speakers of English. If we accept Chomsky’s concept of
competence, then such a sentence would be grammatical but subject to
severe performance restrictions. However, it is safe to say that such a sen-
tence would never be spoken in English; and in a real sense, our sentence
would be ungrammatical or, to use a more proper term, unacceptable.
3. In (a), sing/sang is the marked member because the past tense of sing is
not formed in the normal, regular manner; that is, by adding the -ed suffix.
In (b), narrow is the marked member because a standard dimension of
measurement is width, as in The field is 180 yards wide.
4. In the first sentence, the pivot word occupies the first position. In the sec-
ond sentence, the pivot word occupies the second position. According to
the model of pivot and open classes, the positions of pivot and open words
are fixed relative to each other for any given child.
5. The rules of a language are simply another way of describing a speaker’s
knowledge of a language. Thus, the speaker proves that he or she knows
the rules of his or her language by learning the language as a child and by
speaking grammatical sentences in everyday life. A speaker of English
would never utter the ‘‘word”’ *fpgmlni because it disobeys the rules of En-
glish pertaining to the possible structuring of sound sequences. Actually,
this “‘word”’ is probably impossible in all the languages of the world. How-
ever, note that the German word Pferd (‘‘horse’’) is not a possible word in
English because the initial pf-sequence disobeys the constraints on possi-
ble segment sequences in English. Also, if a speaker of English pro-
nounces a word like consumable, he or she “‘knows’”’ the phonological rule
of English that in this case assigns the principal word-accent to the second
‘syllable. Furthermore, the speaker can generalize these rules to new in-
stances. A speaker of English who utters a sentence proves knowledge of
English syntax by producing proper word order. Thus, the speaker utters
sentences like The mailman is coming today rather than Today is coming
the mailman. The word order in the latter sentence would be possible in
German. Finally, a speaker of English proves knowledge of English se-
mantics by the fact that he or she utters sentences that are understood by
other speakers of English. The listener acknowledges the objects, con-
cepts, and actions that the speaker is referring to.
6. One would suspect that (a) is more difficult for a child under the age of
three to understand. In (a) John is the underlying object, whereas in (b)
John is the underlying subject. Children under the age of three would con-
sider John to be the logical subject in both sentences. The fact that (a) con-
tains told and (b) contains promise is irrelevant, for promise is a noun in
(b).

Chapter 14 Sociolinguistics
1. The educated speakers were in most cases members of the upper-social
strata who spoke a prestigious dialect that was somewhat resistant to
change. This dialect approximates the standard dialect of British English
that is, to some degree, maintained today in similar form.
 Chapter 15 Anthropological Linguistics 247

2. The linguistic geographer would not be justified in making either conclu-

sion because he did not consider the socioeconomic characteristics of his
informants.
3. No. Although shared ethnicity is one factor that may lead to shared
linguistic features, and although a given ethnic group in a given situation
may display greater linguistic homogeneity than other ethnic groups, eth-
nicity alone does not determine dialect. Other factors like age, sex, in-
come, education, and occupation may cause great linguistic diversity
among speakers of the same ethnic group.
4. Since the shoit and boid pronunciation has been ridiculed for quite some
time, it has become a stigmatized feature associated with the lower
classes. Speakers who shun this pronunciation are concerned about being
marked as lower-class, linguistically unsophisticated speakers. Although
there is some possibility that certain speakers are unaware of the stigma
attached to their pronunciation, it is more likely that they simply do not
wish to change — whether out of pride, apathy, or some other reason.
5. The most general conclusion would be that some factor has caused a shift
back toward r-less speech among this group. Perhaps a progressive com-
mercial role for the South and a Southern president might have caused a
re-emergence of Southern pride and a shift back toward what had long
been considered a Southern speech trait (r-less speech). Remember that
this is a hypothetical example that does not purport to predict such a
change. In fact, full r-pronunciation will probably continue to be the
norm — for at least some time, and at least in Tuscaloosa.

+nasal
fle) = +anterior

Chapter 15 Anthropological Linguistics

1. A man may be addressed as sir (young boys to adult men; one adult to
another, usually when the person addressed is unknown); mister (usually
by a younger person or child to an adult); buddy, pal, and others (usually
among males); man (used primarily by blacks and the hip young when re-
ferring to each other); my dear fellow (used on occasion by the upper class
or in a formal or stuffy situation). Obviously, many other terms exist.
A woman may be addressed as miss (although frowned on by women’s
liberation, this is still the most common term used to refer to an unknown
woman); ma’am (used by children or conservative speakers); madam
(used in more formal situations).
When the more formal terms are used (sir, madam, and so on), they are
preceded by equally formal expressions like pardon me whereas the infor-
mal terms (like man) would be preceded by an informal term, usually hey.
2. (a) centum (Gothic); (b) satem (Lithuanian); (c) centum (Tocharian); (d)
satem (Armenian);
3. (a) implicational; (b) formal; (c) statistical; (d) substantive
248 Answers to Exercises

Chapter 16 Historical-Comparative Linguistics

1. This example illustrates the loss of the lexical items wost and thou and the
loss of verb-subject word order (wost thou) in declarative sentences. An-
other syntactic curiosity is the phrase what thinge is speche, which would
appear today as what sort of thing speech is. From the orthography, one
might guess that the medieval vowel in speche has changed in pronuncia-
tion. Similarly, one might guess that there was a (reduced) final vowel in
the words thinge and speche; but note the modern English words breathe,
time, and so on. Although morphological, syntactic, lexical, and possibly
semantic changes are usually the easiest to identify, one would need addi-
tional data to realize that the loss of thou is part of a larger change in the
pronoun system and that wost contains a second-person -st ending that
was subsequently lost.
in). (a) dissimilation of point of articulation; (b) excrescence; (c) metathesis; (d)
assimilation of voicing; (e) loss
. (a) neither; (b) merger; (c) split
. a. partial merger; change in place in articulation; rule addition
b. complete merger; change in place of articulation; rule simplification
. (a) possible; (b) possible; (c) possible; (d) possible; (e) highly unlikely; (f)
highly unlikely; (g) possible; (h) highly unlikely
. The use of the term birds represented a widening or extension of context;
the term could also be described as acquiring the semantic features (+hu-
man) and (+female).
a. M-m-m-m; @-a-a-a; n-n-n-n; h-h-h-h; d-t-t-d
b. *man; *hand
c. The vowel [a] became fronted to [gz] in English only.
The consonant [d] became voiceless in Dutch and German, i.e., [t].
 GLOSSARY

Italic terms within definitions are defined elsewhere in the glossary.

Accessibility hierarchy is an ordering of relationships between types of noun

phrases; it is useful in studying language universals.
Acoustic signal is a stream of speech.
Adjustment rule changes the shape of certain morphemes in the immediate
environment of other specific morphemes.
Affricates are a group of sounds produced by combining a stop with a frica-
tive.
Agglutinative language is one in which words may include different types of
elements or affixes, and these are kept distinct from roots. Contrast with
Analytic language and Inflectional language.
Allomorph is any of the variant forms of a morpheme. For example, the En-
glish plural morpheme is realized through its various allomorphs, /-s/, /-z/,
and /az/, as in cats, dogs, and houses.
Allophones are the actual speech sounds that make up phonemes. Compare
with Phonemes.
Alphabetic writing is that which uses a symbol or letter to represent a particu-
lar spoken sound. Contrast with Syllabic writing and Logographic writing.
Alveo-Palatal fricatives are sounds produced when the center of the tongue is
arched toward the hard palate.
Ambiguity refers to a surface structure that represents different meanings,
which are expressed by different deep structures. Ambiguity may also refer
to the different possible meanings of certain words, such as bank.
Amplitude is a measurement of the energy of a sound; it is related to the
perception of loudness.
Analogy is the process by which forms become modeled after existing pat-
terns in a language.
Analysis-by-synthesis-model is a theory of speech perception in which the
hearer internally generates or synthesizes a set of speech sounds until he or
she finds a match with what has been heard.
Analytic language is one in which each word is an isolated root. Contrast
with Agglutinative language and Inflectional language.

249
250 Glossary

Anglo-Saxon. See Old English.

Anomaly is the irregularity or abnormality resulting when possible semantic
relations are violated; for example, burning cold, colorless green.
Anterior is a phonological feature designating those sounds formed in the
front of the mouth.
Antonymy is the relationship between two words that are opposite in mean-
ing, such as sweet and sour.
Aphasia is the name for a variety of speech losses and speech problems that
are related to dysfunctions of the brain.
Aspiration is the explosion of air accompanying the production of some
sounds.
Assimilation is a process of phonological change in which sounds are altered
to resemble surrounding sounds with respect to particular features. Con-
trast with Dissimilation.
Atomic statements in standard predicate calculus are sentences that have no
quantifiers, negations, conditions, and so on.

Babbling period occurs in a child by about the age of six months. This pre-
language state is a time when almost any sound can conceivably be pro-
duced, and it allows a child to develop its articulatory prowess.
Back is a phonological feature that designates those vowel sounds produced
when the tongue is further back than a neutral position.
Base component of the syntactic component of a grammar consists of
phrase-structure rules and a lexicon.
Bilabials are sounds produced when the lips are pressed together.
Bleeding order in transformational-generative grammar exists when a rule
removes representations to which a subsequent rule might have applied.
Borrowing is a semantic change in which a lexical item from one language is
incorporated into another.
Bound morpheme is a morpheme that cannot exist alone. It must be attached
to a free morpheme.
Braces in linguistic notation are used for writing linguistic rules; they indicate
and. They are also used in structuralist morphology to enclose morphemes.
Broca’s aphasia is a speech malfunction characterized by extreme difficulty
in speech, very poor pronunciation, and usually the loss of the small words
or function words of language such as a, in, for, and so on.
Broca’s area of the brain was named after the French physiologist Paul
Broca, who first identified this area directly in front of the part of the motor
cortex that controls those muscles involved in speech production.
Breath control is the act of directing the muscles principally involved in air
expulsion.

Case grammar is a theory of syntax and semantics in which several cases

such as agent, instrument, and so on are identified. These cases are defined
in terms of their relationship to the verb.
Celtic language is an Indo-European language spoken by the Celts, who were
early settlers in the islands that are known today as Britain.
Centum languages form a major division of the Indo-European language
 Glossary Poe

family. In this group of languages, the original k-sound is realized as [k] or

[h]. Contrast with Satem languages..
Cognates are words that share similar meanings and some phonetic similarity.
Cognitive grammar is a model of language that is concerned with sentence
perception and sentence production.
Combined statement in standard predicate calculus is a sentence that con-
tains negation, condition, conjunction, or disjunction.
Commissives are illocutionary speech acts that commit the speaker to some
future action, such as promises, threats, offers, and so on.
Communication is a process in which information is transmitted from a
source to a goal.
Comparative method is a process of language. reconstruction through the
analysis of cognates.
Competence is a speaker’s knowledge of his or her language; that is, his
knowledge of its sound structure, its words, and its ihe rules. Con-
trast with Performance.
Complementary distribution means that the allophones of a particular
phoneme occur in different phonetic environments.
Connotation is the suggestive or implied meaning of a word. Contrast with
Denotation.
Consonantal is a phonological feature that designates those sounds produced
by the constriction or occlusion of the oral cavity.
Consonants are sounds produced when the oral cavity is constricted or oc-
cluded.
Constituent structure is the syntactic relationships of linguistic units in sen-
tences.
Content words are nouns, verbs, adjectives, and adverbs. Contrast with
Function words.
Contextual meaning is that which is dependent on the utterances surrounding
it.
Continuant is a phonological feature that designates sounds produced by a
constriction or occlusion of the oral cavity for a period of time, such as [p]
and [f].
Contradiction means that if one sentence is true, the other must be false.
Coronal is a phonological feature that designates those sounds made with the
front half or blade of the tongue.
Creole is a pidgin language that is learned as a first language by a child who
heard his parents or associates, speaking a pidgin language.
Cyclic rule is the name given by the transformational-generative grammarians
to any rule that can reapply to successive cycles where each cycle is a
block of ordered rules.
Cuneiform writing was the wedge-shaped script of ancient Mesopotamia. It
contained characters for syllables.

Declaratives are illocutionary speech acts that bring about some performance
that corresponds to what is being said. Declaring war, firing an employee,
and pronouncing a couple husband and wife are examples.
Deep structure is an abstract level representing the basis for the meaning of a
252 Glossary

sentence. It consists of a structure generated by phrase-structure rules and

contains lexical items from the lexicon, but no transformations have yet
applied to it.
Denotation is the clearly defined meaning of a word. Contrast with Connota-
tion.
Derivational morpheme is a morpheme that is added to a root (word) to form
a new word that differs, usually, in its part-of-speech classification.
Descriptive linguistics is the sub-discipline of linguistics that is concerned
with providing an accurate description of the phonology, syntax, and
semantics of languages.
Diachronic linguistics is historical linguistics; it deals with the relationship
between different states of language through time.
Diacritic marks are symbols used to achieve greater accuracy in the transcrip-
tion of vowel variants. Arrows, dots, and lines are commonly used for this
purpose.
Dialect is a regional variety of language that may differ from other varieties of
the language by features of vocabulary, grammar, and pronunciation.
Further, dialect may be seen as a variety of language used by one occupa-
tional group or one social class.
Diphthongal vowels are those that are made up of two sounds in sequence such
as [ay] in ride. :
Directives are illocutionary speech acts that have the intent of getting the
hearer to do something, such as orders, requests, instructions, and so on.
Dissimilation is the process in which sounds become phonetically less like
nearby sounds. Contrast with Assimilation.
Distinctive feature is a binary attribute of individual sounds.
Distinguisher is the idiosyncratic aspect of the semantic concept of a word.

Early Modern English is the name given to the English that was spoken from
about the end of the fifteenth century through approximately the middle of
the eighteenth century.
Embedded sentence is a sentence that is embedded within another. The most
obvious example is a sentence containing a relative clause, such as The dog
that you saw belongs to me.
Entailment occurs when the meaning of one sentence is contained in that of
another.
Excrescence is phonological change through the addition of a sound segment
in words containing a particular phonetic environment.
Expressives are illocutionary speech acts that express an attitude of the
speaker about some state of affairs, such as thanks, apologies, welcomes,
and so on.
Extended standard theory states that deep structure alone cannot supply the
basis for the meaning of a sentence.

Feeding order in transformational-generative grammar occurs when one rule

creates representations to which a subsequent rule may apply.
Focus refers to information not previously shared by the speaker and hearer.
 Glossary 253

Formal universal is a language universal that is a theoretical construct that

can be applied to all languages.
Formant is a characteristic band of sound that is caused by the particular
shape of the vocal tract during speech production.
Free morpheme is a morpheme that can exist as a word by itself.
Frequency in acoustic phonetics is the rate of vibration of a sound.
Fricatives are sounds produced by forcing air through a narrow opening in the
oral cavity, such as [f] and [oe].
Function words are articles, conjunctions, and prepositions. Contrast with
Content words.

Generative semantics holds that semantics, or meaning, is primary, and that

syntactic and phonological rules operate on semantic structures to provide
a shape and a sound for sentences.
Gesture is unspoken communication through body movement or manipulation
of the environment.
Glides are the [w] and [y] sounds in English. They always appear either be-
fore or after a vowel and the sound moves or ‘‘glides’’ to or from the ar-
ticulatory position for that vowel.
Glottal stop is produced by closing the air flow at the glottis.
Glottis is the area between the upper part of the larynx and the vocal cords.
Grammar is a system of rules of a language. There are phonological, syntac-
tic, and semantic components to grammar.
Great Vowel Shift is the term used to refer to the major changes in the pro-
nunciation of vowels that took place early in the Middle English period.
Vowels generally moved to the front of the vowel chart and up on the
vowel chart; high vowels became diphthongs.
Grimm’s law is a regular correspondence of sounds in the Indo-European and
Germanic languages noted by Jakob Grimm in 1822.

Hertz (Hz) is a unit of frequency that is the equivalent of one cycle per sec-
ond. It is used by phoneticians in the measurement of sound.
Heuristics are strategies that people employ to simplify their processing of
sentence structures.
Hieroglyphs are pictures that represent sounds.
High is a phonological feature that designates those vowel sounds produced
when the blade of the tongue is above the neutral position.
Historical linguistics. See Diachronic linguistics.
Homophony occurs when two words with different meanings and spellings
have the same sound, such as sail and sale. The definition may be ex-
panded to include one word with more than one meaning, such as bank.

Idiolect is the language pattern of any given individual.

Illocutionary speech act is the intent that a speaker has in uttering a sentence.
Immediate-constituent analysis was a system used by structural linguistics to
indicate words’ relationships to each other in syntactic structures.
Implicational universal is a language universal that is inferred from observed
254 Glossary

facts. For example, a contrast of tongue heights for nasal vowels in a lan-
guage implies a contrast of tongue heights for non-nasal vowels.
Indo-European language family is a large group of related languages: Ger-
manic, Celtic, Italic, Baltic, Slavic, Greek, Armenian, Hittite, Tocharian,
Iranian, and Indic.
Inflection is the process of changing words to indicate case, tense, person,
number, mood, or gender.
Inflectional language is one in which affixes serve a grammatical function and
become part of the word on which they are necessarily dependent. Contrast
with Agglutinative language and Analytical language.
Inflectional morpheme is normally a suffix in English, and its function is to
indicate certain grammatical properties associated with nouns and verbs,
such as gender, number, case, and tense.
Innate characteristics are those with which a person or animal is born. These
attributes are genetically determined.
Innovation in transformational-generative grammar is a process of sound
change.
Interdental fricatives are the sounds produced when the tongue is pressed
against the upper teeth or between the teeth and air is forced between
them, e.g. [0] or [6].
Internal reconstruction establishes previous stages of a particular language
through a consideration of data usually from that language alone.
Interpretive semantics is a theory that attempts to incorporate a semantic
component within transformational-generative grammar by providing
semantic interpretations to syntactic deep structures.
Intonation is a suprasegmental feature that in English designates a rising or
falling pitch over a group of words.
Isogloss is an imaginary geographical line that shows the boundaries within
which one word or pronunciation is typical.

Kana syllabaries of Japanese is an example of syllabic writing.

Kinesics is the study of positioning and moving the body during a conversa-
tion.

Labiodental fricatives are sounds produced when the lower lip is pressed
lightly against the upper teeth and air is forced between them, e.g. [f] or
[v].
Language is a form of communication that is nonstereotyped and nonfinite. It
is learned and unlimited in scope.
Language family is a group of related languages.
Language reconstruction is the reconstruction of linguistic forms in languages
that no longer exist through the comparison of forms in several existing
languages and through internal reconstruction.
Language universal is some element, aspect, or principle that is found in
virtually every language of the world.
Larynx (the voice box) forms the upper part of the trachea and contains the
vocal cords.
Lexical diffusion is a process of phonological change in which a sound change
 Glossary 255

spreads through the /exicon and is manifested in an increasing number of

words.
Lexicon is a list or dictionary of words, sounds, and so on, that provides
additional linguistic information as well.
Lingua franca is a language that is mutually agreed upon by two or more
speakers whose native language is not the same, as a means of communica-
tion to facilitate trade, diplomacy, or other social functions.
Linguistics is the study of language, especially as it is spoken.
Linguistic atlas is a volume of maps which show particular word use or pro-
nunciation within a given geographical area.
Linguistic variable is the term that refers to any sound, word, or syntactic
structure that varies in pronunciation, use of word, or syntactic pattern to
convey the same message.
Liquid sounds are formed when the tip of the tongue is flexed in connection
with the alveolar ridge.
Locutionary speech act is the actual uttering of a sentence with a particular
meaning.
Logographic writing is that which uses a symbol to represent a word. Con-
trast with Alphabetic writing and Syllabic writing.
Loss is phonological change in which a sound is deleted in a particular pho-
netic environment.
Low is a phonological feature that designates those vowel sounds produced
when the blade of the tongue is above the neutral position.

Markedness is a theoretical device that characterizes what is normal, simple,

expected, and plausible in a language and what is not.
Merger is a process of phonological change in which two or more phonemes
become one.
Metaianguage is a precisely defined language used in an exact discussion of
another language.
Metathesis is a process of phonological change in which two segments are
interchanged.
Middle English is the term given to the language spoken in five major dialect
forms in Britain from the twelfth century through the fifteenth century.
Miminal pair is any two words that contain the same number of segments,
differ in meaning, and exhibit only one phonetic difference.
Modal refers to the mood of a verb.
Modal auxiliary is a verb used with other verbs to express mood or tense; for
example, can, could, may, might, must, ought, shall, should, will, and
would.
Modality in case grammar includes such information as tense, mood, nega-
tion, and aspect.
Modern English is the term used for English that has been spoken from about
the middle of the eighteenth century until today.
Modistic grammar. See Speculative grammar.
Monophthongal vowel is a vowel that has a single sound, such as the [1] in
kit.
Morpheme is the smallest meaning-bearing unit of language.
256 Glossary

Morphology is a study of the words of a language and the inflections used to

form new words.
Morphophoneme was a term used by structural linguists to represent
phonemic alternations in related words such as foot and feet.

Narrowing is a semantic change in which the meaning of a lexical item is

restricted.
Nasal feature designates those consonantal sounds that are produced when
the nasal cavity is open.
Nasal sounds are produced when the uvula is open and air is allowed to flow
through the nose.
Natural phonology is a theory of phonology that is heavily based on natural
segments, classes, and rules.
Neo-grammarian school of linguists developed in the nineteenth century.
These grammarians viewed laws of sound as a natural science and believed
that sound change laws operated without exception.
Neurolinguistics is the study of the physiological basis of language and brain
functions underlying language acquisition and use of language.
Nonterms in relational grammar indicate relations such as location, means,
and so on.

Obstruent is a stop, a fricative, or an affricate, i.e., any sound made by ob-

structing the flow of air in the vocal cavity.
Old English is the English that was spoken from roughly the beginning of the
eighth century to the middle of the twelfth century. Sometimes this is called
Anglo-Saxon.
Onomatopoeia is the formation of words whose sounds suggest their meanings.
Open words are a class of lexical items in a young child’s vocabulary. This is
a fairly large class of words that may have been one-word sentences. Con-
trast with pivot words.
Oral cavity is the mouth.

Palatalization is the fronting of consonants to the soft palate.

Palatal sounds are those produced when air is forced through the oral cavity
while the tongue is pressed against the hard palate, e.g. [€] or [j].
Paralanguage is the modulation of the voice in communication.
Paraphrase is the communication of the same thought with any two different
syntactic structures which convey the same basic meaning.
Parentheses in linguistic notation are used to indicate optionality.
Performance is an actual utterance or the act of speech itself. Contrast with
Competence.
Perlocutionary speech act is the result the speaker hopes to achieve in utter-
ing a sentence.
Pharynx connects the nasal cavity to the larynx and then becomes continuous
with the esophagus.
Philology is the study of speech primarily through the analysis of written
texts.
Phoneme is an abstract phonological unit representing a class of real sounds,
termed the allophones of the phoneme.
 Glossary 257

Phonemic alternation is the changing vowel sound in such related words as

vain and vanity.
Phonetic alphabet is a series of symbols that describe distinct sounds.
Phonetic environment is the sounds that surround a phoneme.
Phonetics is the study of speech sounds.
Phonological derivation is an operation that begins with an underlying
representation and through the application of a set of specific rules yields
the actual sound the speaker produces.
Phonological rules are operational statements that may add, delete, or change
some phonological entity, resulting in a new linguistic entity.
Phonology is the study of how speech sounds are organized and how they
function in a language.
Phrase-structure grammar is a series of rules that establishes the basic struc-
tures of sentences.
Pictographs are an attempt to represent a concept through a drawing or pic-
ture.
Pidgin is a simplified form of a language which is marked by a small vocabu-
lary and an uncomplicated grammar.
Pivot words are a class of lexical items in a young child’s vocabulary. Com-
mon examples are more, my, and pretty. These words are added to the
words in the open class to form two-word sentences.
Plosives. See Stops.
Predicate in standard predicate calculus expresses the properties of terms or
the relations between terms. ;
Predicate raising is a transformation that raises a verb from an embedded
sentence into the next highest sentence in the tree diagram, where it is
joined to the verb of that sentence.
Presupposition refers to the conditions that must be met in order for the in-
tended meaning of a sentence to be regarded as acceptable.
Projection rules are statements about how the different aspects of the mean-
ing of lexical items may be combined within successively larger grammati-
cal categories until a reading for the entire sentence is produced.
Proposition in case grammar is a constituent that includes a verb and one or
more noun phrases that are in a tenseless set of case relationships to the
verb.
Protoforms are reconstructed forms of a parent language.
Proxemics is a field of study concerned with the space maintained between
two speakers in conversation.
Psycholinguistics is the study of the acquisition, perception, and production
of language.
Pure tone is the sound given by a tuning fork.

Quantified statements in standard predicate calculus are sentences that con-

tain quantifiers and define whether the statement holds true in all cases or
particular cases.

Recursive rule is a rule that can apply over and again and permit sentences of
infinite length to be generated in a language.
Redundancy refers to the prediction of certain linguistic information from
258 Glossary

other information. It is useful in communication for providing more infor-

mation than is absolutely necessary in a noise-free environment.
Reed-Kellogg sentence diagram places the subject and predicate of a sen-
tence on a horizontal line with all modifying words and phrases arranged to
indicate their function and relationship to each other.
Regularity. See Productivity.
Relational grammar is an approach to grammar that is concerned with
grammatical relations like subject and direct object, which are seen as
undefined notions that are basic to the description of language.
Representatives are illocutionary speech acts that represent a particular state
of affairs, such as statements, descriptions, assertions, and so on.
Restructuring in transformational-generative theory is a change in underlying
representations.
Round is a phonological feature that designates those vowel sounds produced
when the lips are rounded.
Rule ordering refers to which linguistic operations must be performed on a
linguistic entity in a prescribed sequence. A change in the order in which
rules are applied may lead to quite different results.
Rule reordering occurs when the order in which rules apply is reversed.

Satem languages are a major division of the Indo-European language family.

In this group of languages, the original k-sound is realized as [s] or [s].
Contrast with Centum languages.
Semantic features indicate semantic properties of words, e.g. (animate),
(concrete), and so on.
Semantic redundancy rules represent the relationship between semantic fea-
tures.
Semantics is concerned with the meaning of words and sentences.
Semantic space is the concept, explored by the quantificational approach,
that words lying nearest each other on a graph, as determined by speakers’
judgments, may be regarded as similar in meaning.
Semivowels. See Glides.
Sentence is a group of words that expresses a complete thought, according to
prescriptive grammarians. But a sentence may be defined as whatever a
native speaker knows to be a sentence, according to transformational-
generative grammarians. According to this group’s definition, no definition
is necessary since sentences can be recognized intuitively.
Sentence patterns are acceptable word configurations in a language.
Sign language uses gestures to communicate concepts.
Simplification in transformational-generative grammar is a process of syntac-
tic change in which rules are dropped and language is generalized or
simplified.
Slang consists of vocabulary, often short-lived, that is spoken by a sub-group
within a society. It is often used for the purpose of keeping others ‘‘out’’ of
conversations and is frequently spontaneous and often racy.
Sociolinguistics is the study of the interaction of socioeconomic characteris-
tics and language.
Soft palate. See Velum.
 Glossary 239

Sonorant is a phonological feature that designates those sounds that have

- resonance.
Sound is energy, as reflected in vibrations in gases and liquids.
Spectrograph is an instrument used by phoneticians to analyze precisely
the waveforms and frequencies of human speech.
Speculative grammar is a study that sought a theoretical basis for Latin
grammar by attempting to define modes by which speakers perceive and
signify things. Also known as modistic grammar.
Speech is the expression of thoughts, or meanings, through sounds.
Speech act is a basic unit of communication that includes prelocutionary,
illocutionary, and locutionary aspects.
Split is a process of phonological change in which one phoneme develops into
two or more phonemes.
Standard American English is an idealization of the language spoken in this
country. As such, it is a speech not characterized by regional differences in
pronunciation, word use, or syntax.
Standard interpretive theory of transformational-generative grammar sees
syntax as central to competence. Sentences are formed within the syntactic
component of competence. Phonology and semantics are considered in-
terpretive in the sense that their rules operate on the product of the syntac-
tic component to clothe it with sound and to make sense out of it.
Standard predicate calculus is a system for representing logical relations in
the study of syntax and semantics.
Stopplosives. See Stops.
Stops, plosives, or stopplosives are produced by entirely closing off the flow
of air at some point in the mouth, e.g. [p], [t], and [k].
Stratification is exhibited when a linguistic variable is correlated with a
sociological variable to yield a regular progression.
Stress is a suprasegmental feature that designates pitch, length, or loudness.
‘ The basic unit containing stress is the syllable.
Strong verbs are those that change their form in conjugation, e.g. sing, sang,
sung.
Strident is a phonological feature that designates consonantal sounds that
have a high-frequency noisiness.
Structural linguisics emphasizes the investigation of concrete linguistic data
through isolating, classifying, analyzing, and segmenting an observed lan-
guage.
Substantive universals are language universals observed in actual data from
the world’s languages. Contrast with formal universals.
Suprasegmental refers to aspects of pronunciation that go beyond the produc-
tion of individual (segmental) sounds. They must be dealt with in conjunc-
tion with segmental sounds.
Surface structure is an actual speech utterance derived by a series of trans-
formations from a deep structure. Contrast with Deep structure.
Syllabic is a phonological feature that designates the role a sound plays in the
structure of a syllable.
Syllabic writing is that which uses a symbol to represent a syllable. Contrast
with Alphabetic writing and Logographic writing.
260 Glossary

Syllable consists of a vowel or a vowel accompanied by one or more conson-

ants.
Sympathetic vibration is the spontaneous vibration of vibrators (tuning forks,
piano strings, and so on) caused by sound waves of that frequency emanat-
ing from another vibrator of that frequency.
Synchronic linguistics deals with the analysis of a language at a particular
time.
Synonymy is the relationship between words with similar meanings but differ-
ent pronunciations, such as big and large.
Syntactic features are aspects of a word that govern its use in structures.
The part of speech of a word is one of the most important of these syntactic
features.
Syntax concerns the combining of message-bearing units (words) into the sen-
tences of a language.

Tense is a phonological feature that designates those vowel sounds that are
produced when the muscles of vocalization are tensed.
Terms in relational grammar are undefined notions such as subject, direct
object, and so on.
Tone languages employ pitch to communicate meaning.
Trachea is the duct through which air passes to and from the lungs.
Transformation is an operation that adds, deletes, or changes an element in
one syntactic structure to produce another syntactic structure.
Transformational component of a grammar is a set of rules that turns a deep
structure into a surface structure.
Transformational cycle refers to the reapplication of transformations to pro-
gressively larger constituents.
Transformational-generative grammar is a universal theory of language de-
veloped by Noam Chomsky and his students in the mid-fifties.
Tree diagram is the transformational-generative means of graphically repre-
senting the syntactic structures of sentences.

Unaspirated sounds are consonant sounds not accompanied by a puff of air.

Underlying representation is an abstract unit that is related to an actual
phonetic segment by means of phonological rules.
Universal grammar is the concept that languages are essentially the same in
nature and differ only in surface characteristics.
Unreleased stops occur when the lips are not opened to allow an explosion of
air at the end of the production of stops.
Uvula is the movable flap that controls the passage of air through the nasal
cavity.

Velar sounds are those produced when the tongue is pressed against the
velum.
Velum is the soft palate, the flexible part of the mouth that joins the hard
palate and the uvula.
Vocal cords are the folds in the larynx that help to produce sounds when they
are tensed or pulled together while air from the lungs is passed over them.
 Glossary 261

Vocal tract is all the speech mechanisms located from the pharynx through
the mouth and nose. These include the glottis, the epiglottis, the uvula, the
velum, the hard palate, the aleveolar ridge, the teeth, the lips, the tongue,
and the nasal cavity.
Voice bar is the lowest of the characteristic frequencies of vocal cord vibra-
tion on a sound spectrum in the voicing of an utterance.
Voice box. See Larynx.
Voiced sounds are those produced when the vocal cords are closed and air
from the lungs is passed over them, causing them to vibrate.
Voiceless sounds are those produced when the vocal cords are open.
Voice quality is the specific timbre or condition of a speaker’s vocal produc-
tion.
Vowels are sounds that are produced when ae oral cavity is relatively open
to air.

Wernicke’s aphasia is marked by a lack of content in speech. The speech of a

person suffering from this aphasia contains many circumlocutions and
empty words such as thing.
Wernicke’s area was discovered by the physiologist Carl Wernicke in 1874 as
a second major speech area in the brain next to and closely connected by a
long series of nerves to Broca’s area.
Weak verbs are those that form their tenses in a regular manner; that is in
English they add -ed to the stem to form the past tense and the past ne
ple, for example, walk, walked, walked.
Widening is a type of semantic change in which the meaning of a lexical item
is extended.
Word-formation rule is a morphological rule that consists of a base — a set of
words that are members of a grammatical category, such as noun, verb,
adjective, or adverb — and an operation that specifies the effect of the rule
on the base.
 BIBLIOGRAPHY

Abercrombie, David. Elements of General Phonetics. Edinburgh: Edinburgh

University Press, 1967.
Akmajian, Adrian and Frank Heny. An Introduction to the Principles of
Transformational Syntax. Cambridge, Mass.: MIT Press, 1975.
Akmajian, Adrian, Richard A. Demers, and Robert M. Harnish. Linguistics:
An Introduction to Language and Communication. Cambridge, Mass.:
MIT Press, 1979.
Anttila, Raimo. An Introduction to Historical and Comparative Linguistics.
New York: Macmillan, 1972.
Arlotto, Anthony. Introduction to Historical Linguistics. New York: Hough-
ton, Mifflin, 1972.
Aronoff, Mark. Word-Formation in Generative Grammar. Cambridge, Mass.:
MIT Press, 1976.
Austin, John L. How To Do Things with Words. Cambridge, Mass.: Harvard
University Press, 1962.
Bach, Emmon. Syntactic Theory. New York: Holt, 1974.
Bailey, Richard W., and Jay L. Robinson, comp. Varieties of Present-Day
English. New York: Macmillan, 1973.
Baugh, Albert C. A History of the English Language. 2nd ed. New York:
Appleton-Century-Crofts, 1957.
Baumm, Richard and Joel Sherzer, eds. Exploration in the Ethnography of
Speaking. New York: Cambridge University Press, 1974.
Bellugi-Klima, Ursula. ‘“‘Linguistic Mechanisms Underlying Child Speech,”
in E. M. Zale, ed. Proceedings of the Conference on Language and Lan-
guage Behavior. New York: Appleton-Century-Crofts, 1968.
Belvalkar, S. K. An Account of the Different Existing Systems of Sanskrit
Grammar. Poona: S. K. Belvalkar, 1915.
Bever, Thomas. ‘‘The Cognitive Basis for Linguistic Structures,’’ in John R.
Hayes, ed. Cognition and the Development of Language. New York:
Wiley, 1970.
. “A Survey of Some Recent Work in Psycholinguistics,’’ in W. J.
Plath, ed. Specification and Utilization of a Transformational Grammar:

262
 Bibliography 263

Scientific Report Number Three. Yorktown Heights, N.Y.: Thomas J.

Watson Research Center, International Business Machines Corp.,
1968.
Birdwhistell, Ray L. Kinesics and Context: Essays on Body Motion
Communication. Philadelphia: University of Pennsylvania Press, 1970.
Bloom, Lois. Language Development: Form and Function in Emerging
Grammars. Cambridge, Mass.: MIT Press, 1970.
Bloomfield, Leonard. Language. New York: Holt, 1933.
Bolinger, Dwight. Aspects of Language. New York: Harcourt, 1968.
. “On Defining the Morpheme.”’ Word, 1948.
Bourne, Geoffrey H. The Ape People. New York: Putnam, 1971.
Braine, Martin D. S. ‘‘The Ontogeny of English Phrase Structure: The First
Phase.”’ Language, 39, 1-13, 1963.
Brown, Roger W. Words and Things. New York: Macmillan/Free Press,
1968.
Brown, Roger W. and David McNeill. “‘The Tip of the Tongue
Phenomenon.” Journal of Verbal Learning and Verbal Behavior 5,
325-327, 1966.
Bruck, Anthony, Robert A. Fox, and Michael W. LaGaly. Papers from the
Parasession on Natural Phonology. Chicago: Chicago Linguistic Society,
1974.
Bursill-Hall, G. L. ‘“‘Medieval Grammatical Theories.’’ Canadian Journal of
Linguistics, 1963.
Burt, Marina K. From Deep to Surface Structure: An Introduction to Trans-
formational Syntax. New York: Harper & Row, 1971.
Bynon, Theodora. Historical Linguistics. New York: Cambridge University
Press, 1977.
Cadet, Felix. Port Royal Education. London: Swan, Sonnenschein, 1898.
Cairns, Helen S. and Charles E. Cairns. Psycholinguistics: A Cognitive View
of Grammar. New York: Holt, Rinehart and Winston, 1976.
Carroll, John B. Language and Thought. Englewood Cliffs, N.J.: Prentice-
Hall, 1964.
, ed. Language, Thought, and Reality: Selected Writings of Benjamin
Lee Whorf. Cambridge, Mass.: MIT Press, 1956.
Carroll, Ludovici (Lewis Carroll). Alicia in Terra Mirabili. Translated by
Clive Harcourt Carruthers. New York: St. Martin’s Press, 1964.
Catford, John C. Fundamental Problems in Phonetics. Bloomington: Indiana
University Press, 1977.
Chafe, Wallace L. Meaning and the Structure of Language. Chicago: Univer-
sity of Chicago Press, 1970.
Cherry, Colin. On Human Communication. 3rd ed. Cambridge, Mass.: MIT
Press, 1957.
Chomsky, Carol. The Acquisition of Syntax in Children from 5 to 10. Cam-
bridge, Mass.: MIT Press, 1970.
Chomsky, Noam. Aspects of the Theory of Syntax. Cambridge, Mass.: MIT
Press, 1965.
. Cartesian Linguistics. New York: Harper & Row, 1966.
——. ‘‘Deep Structure, Surface Structures, and Semantic Interpretation,”
264 Bibliography

in Danny D. Steinberg and Leon A. Jakobovits, eds. Semantics. Cam-

bridge: Cambridge University Press, 1971.
. Language and Mind. New York: Harcourt, 1968.
‘*Review of Skinner’s ‘Verbal Behavior,’ ’’ Language, 35, 26-58, 1959.
. Syntactic Structures. The Hague: Mouton, 1957.
Chomsky, Noam and Morris Halle. The Sound Pattern of English. New
York: Harper & Row, 1968.
Clark, Eve. ‘‘What’s in a Word? On a Child’s Acquisition of Semantics in His
First Language.’ In T. E. Moore. Cognitive Development and the Acquisi-
tion of Language. New York: Academic Press, 1973.
Clark, Herbert and Eve V. Clark. Psychology and Language: An Introduc-
tion to Psycholinguistics. New York: Harcourt, Brace, Jovanovich,
1977.
Clifton, C., I. Kurez, and J. J. Jenkins. ‘‘Grammatical Relations as Determi-
nants of Sentence Similarity.’ Journal of Verbal Learning and Verbal
Behavior 4, 112-117, 1965.
Clifton, C. and P. Odom. ‘‘Similarity Conditions among Certain English Sen-
tence Constructions.’’ Psychological Monographs 80 (whole no. 613), 1966.
Cooper, William and John Ross. ‘‘Word Order,’’ in R. Grossman, J. San, and
T. Vance, eds. Papers from the Parassession on CE RONGE: Chicago:
Chicago Linguistic Society, 1975.
Cousteau, Jacques-Yves. The Whale: Mighty Monarch of the Sea. New
York: Doubleday, 1972.
Cowan, George M. and Florence H. Cowan. ‘‘Mazateco Locational and Di-
rectional Morphemes.”’ Aboriginal Linguistics 1, 1-9, 1947.
Crane, L. Ben. “The Social Stratification of English among White Speakers
in Tuscaloosa, Alabama.’’ Ph.D. dissertation, University of Mas-
sachusetts, 1973.
Culicover, Peter W. Syntax. New York: Academic Press, 1976.
Dale, Philip S. Language Development: Structure and Function. Hinsdale,
Ill.: Dryden Press, 1972.
Dillard, J. L. Black English: Its History and Usage in the United States. New
York, Random House, 1972.
Dillon, George L. Introduction to Contemporary Linguistic Semantics. En-
glewood Cliffs, N.J.: Prentice-Hall, 1977.
Diringer, David. Writing. New York: Praeger, 1962.
Donahoe, John W. and Michael G. Wessells. Learning, Language, and
Memory. New York: Harper and Row, 1980.
Elliott, Dale, Stanley Legum, and Sandra A. Thompson. ‘‘Syntactic Variation
as Linguistic Data,’ in Robert I. Binnick et al., eds. Papers from the Fifth
Regional Meeting of the Chicago Linguistics Society. Chicago: University
of Chicago Department of Linguistics, 1969.
Ervin-Tripp, Susan M. ‘‘On Sociolinguistic Rules: Alternations and Co-
Occurrence,’ in J. Gumperz and D. Hymes, eds. Directions in
Sociolinguistics. New York: Holt, 1972.
Faddegon, B. Studies on Panini’s Grammar. Amsterdam: North-Holland,
1936.
Fant, Gunnar. Acoustic Theory of Speech Production. s’Gravenhage:
Mouton, 1960.
 Bibliography 265

Fant, Louie J., Jr. Say It with WONG: Silver Spring, Md.: National Associa-
tion for the Deaf, 1964.
Farb, Peter. Word Play: What Happens When People Talk. New York:
Knopf, 1974.
Fillmore, Charles. *‘The Case for Case,’’ in Emmon Bach and Robert Harms,
eds. Universals of Linguistic Theory. New York: Holt, 1968.
. “Types of Lexical Information,’’ in Danny D. Steinberg and Leon A.
Jakobovits, eds. Semantics. Cambridge: Cambridge University Press, 1971.
Fishman, Joshua, ed. Advances in the Creation and Revision of Writing
Systems. The Hague: Mouton, 1977.
. Sociolinguistics: A Brief Introduction. Rawley, Mass.: Newbury
House, 1970.
Fodor, Janet Dean. Semantics. Hassocks, England: Harvester Press, 1977.
Fodor, Jerry and Thomas Bever. ‘‘The Psychological Reality of Linguistic
Segments.’’ Journal of Verbal Learning and Verbal Behavior 4, 414-420,
1965.
Fodor, Jerry, Thomas Bever, and Merrill Garrett. The Psychology of
Language. New York: McGraw-Hill, 1974.
Fodor, Jerry and Jerrold J. Katz, eds. The Structure of Language: Readings
in the Philosophy of Language. Englewood Cliffs, N.J.: Prentice-Hall,
1964.
Foss, Donald J. and David T. Hakes. Psycholinguistics: An Introduction to
the Psychology of Language. Englewood Cliffs, N.J.: Prentice-Hall, 1978.
Fromkin, Victoria. ‘“‘Speculations on Performance Models.”’ Journal of
Linguistics 4, 47-68, 1968.
, ed. Speech Errors as Linguistic Evidence. The Hague: Mouton, 1973.
Fromkin, Victoria and Robert Rodman. An Introduction to Language. New
York: Holt, 1974.
Fry, D. B., ed. Acoustic Phonetics: A Course of Basic Readings. Cambridge:
Cambridge University Press, 1976.
Gardner, R. Allen and Beatrice T. Gardner. ‘‘Teaching Sign Language to a
Chimpanzee.”’ Science, 165, 664-672, August 1969.
Gazzaniga, Michael S. The Bisected Brain. New York: Appleton-Century-
Crofts, 1970.
Gelb, Ignace J. The Study of Writing. Rev. ed. Chicago: University of
Chicago Press, 1963.
Gillieron, Jules L. and Edmond Edmont. Atlas Linguistique de la France.
Paris: H. Champion, 1902-1910.
Gleason, Henry A. An Introduction to Descriptive Linguistics. Rev. ed. New
York: Holt, Rinehart and Winston. 1961.
Gordon, James Daniel. The English Language: An Historical Introduction.
New York, Thomas Y. Crowell, 1972.
Greenberg, Joseph H. Anthropological Linguistics: An Introduction. New
York: Random House, 1968.
. Universals of Language. Cambridge, Mass.: MIT Press, 1963.
Greenberg, Joseph H. and James J. Jenkins. ‘Studies in the Psychological
Correlates of the Sound System of American English.” Word 20, 157-177,
1964. Also in Jakobovits and Miron.
Gregory, Michael and Susanne Carroll. Language and Situation: Language
266 Bibliography

Varieties and Their Social Contexts. London: Routledge and Kegan Paul,
1978.
Hall, Edward T. The Hidden Dimension. Garden City, N.Y.: Doubleday,
1956.
. The Silent Language. Garden City, N.Y.: Doubleday, 1959.
Hall, Robert A., Jr. Pidgin and Creole Languages. Ithaca, N.Y.: Cornell
University Press, 1966.
Halle, Morris. ‘‘Phonology and Generative Grammar.’’ Word, 1962.
Halle, Morris, Joan Bresnan, and George A. Miller, eds. Linguistic Theory
and Psychological Reality. Cambridge, Mass.: MIT Press, 1978.
Harms, Robert. Introduction to Phonological Theory. Englewood Cliffs,
N.J.: Prentice-Hall, 1968.
Harris, Zellig. Structural Linguistics. Chicago: The University of Chicago
Press, 1951.
Hayes, Harold. ‘“‘The Pursuit of Reason.’’ New York Times Magazine, June
12, 1977.
Hockett, Charles F. A Course in Modern Linguistics. New York: Macmillan,
1958.
Hoenigswald, Henry M. Language Change and Linguistic Reconstruction.
Chicago: University of Chicago Press, 1960.
Hooper, Joan B. An Introduction to Natural Generative Phonology. New
York: Academic Press, 1976.
Hybels, Saundra and Richard L. Weaver, II. Speech Communication. New
York: Van Nostrand, 1974.
Hyman, Larry M. Phonology: Theory and Analysis. New York: Hoit, 1975.
Hymes, Dell H. Foundations of Sociolinguistics: An Ethnographic Approach.
Philadelphia: University of Pennsylvania Press, 1974.
, ed. Pidginization and Creolization of Languages. Cambridge: Cam-
bridge University Press, 1971.
, ed. Studies in the History of Linguistics: Traditions and Paradigms.
Bloomington: Indiana University Press, 1974.
International Phonetic Association. Principles of the International Phonetic
Alphabet. Rev. ed. London: International Phonetic Association, 1949.
Ivi¢, Milka. Trends in Linguistics. The Hague: Mouton, 1965.
Jacobs, Roderick A. and Peter S. Rosenbaum. English Transformational
Grammar. Waltham, Mass.: Blaisdell, 1968.
Jakobovits, Leon A. and Murray S. Miron, eds. Readings in the Psychology
of Language. Englewood Cliffs, N.J.: Prentice-Hall, 1967.
Jakobson, Roman. Child Language, Aphasia, and Phonological Universals.
Translated by A. R. Keller. The Hague: Mouton, 1968.
Jakobson, Roman, Gunnar Fant and Morris Halle. Preliminaries to Speech
Analysis. Cambridge, Mass.: MIT Press, 1963.
Jankowsky, Kurt R. The Neogrammarians: A Re-evaluation of Their Place in
the Development of Linguistic Science. The Hague: Mouton, 1972.
Jespersen, Otto. Essentials of English Grammar. Reprint. Montgomery,
Alabama: University of Alabama Press, 1964.
. Growth and Structure of the English Language. 8th ed. Leipzig:
B. G. Teubner, 1935.
 -
Bibliography 267

. Language: Its Nature, Development and Origin. n.p.: Allen and

Unwin, 1922. Reprinted New York: Norton, 1964.
Jessel, Levic. The Ethnic Process: An Evolutionary Concept of Languages
and People. The Hague: Mouton, 1978.
Johnson, Samuel. A Dictionary of the English Language. 1755. Reprint. New
York: AMS Press, 1967.
Katz, Jerrold J. “‘Mentalism in Linguistics.’’ Language 40, 124-137, 1964.
Also in Jakobovits and Miron.
. Semantic Theory. New York: Harper & Row, 1972.
Katz, Jerrold J. and Jerry Fodor. ‘‘The Structure of a Semantic Theory.”
Language 39, 170-210, 1963. Also in Fodor and Katz.
Katz, Jerrold, J. and Paul M. Postal. An Integrated Theory of Linguistic
Descriptions. Cambridge, Mass.: MIT Press, 1964.
Keenan, Edward L. ‘‘Linguistic and Logic,’’ in Theo Vennemann and Re-
nate Bartsch, eds. Linguistics and Neighboring Disciplines. Amsterdam:
North-Holland, 1975.
. “On Semantically Based Grammar.”’ Linguistic Inquiry 3, 413-461,
1972.
Keenan, Edward and Bernard Comrie. ‘‘Noun Phrase Accessibility and Uni-
versal Grammar.”’ Linguistics Inquiry 8, 63-99, 1977.
Kempson, Ruth M. Semantic Theory. Cambridge: Cambridge University
Press, 1977.
King, Robert D. Historical Linguistics and Generative Grammar. Englewood
Cliffs, N.J.: Prentice-Hall, 1969. :
Kiparsky, Paul. ‘‘Linguistic Universals and Linguistic Change,” in Emmon
Bach and Robert T. Harms, eds. Universals of Linguistic Theory. New
York: Holt, 1968.
Kurath, Hans and Raven I. McDavid. The Pronunciation of English in the
Atlantic States. Ann Arbor: University of Michigan Press, 1961.
Kurath, Hans, et al. Handbook of the Linguistic Geography of New England.
Providence, R.I.: American Council of Learned Societies, 1939.
Labov, William. ‘‘The Social Motivation of a Sound Change.’’ Word 19,
273-309.
. The Social Stratification of English in New York City. Washington,
D.C.: Center for Applied Linguistics, 1966.
. Sociolinguistic Patterns. Philadelphia: University of Pennsylvania
Press, 1972.
Ladefoged, Peter. A Course in Phonetics. New York: Harcourt, Brace,
Jovanovich, 1975.
. Elements of Acoustic Phonetics. Chicago: University of Chicago
Press, 1962.
. Preliminaries to Linguistic Phonetics. Chicago: University of
Chicago Press, 1971.
Lakoff, George. ‘‘Fuzzy Grammar and the Performance/Competence Ter-
minology Game,’ in Claudia Corum, T. Cedric-Stark Smith, and Ann
Weiser, eds. Papers From the Ninth Regional Meeting. Chicago: Chicago
Linguistic Society, 1973.
_Irregulartity in Syntax. New York: Holt, Rinehart and Winston, 1970.
268 Bibliography

. “On Generative Semantics,’’ in Danny D. Steinberg and Leon A.

Jakobovits, eds. Semantics. Cambridge: Cambridge University Press, 1971.
Lakoff, George and Henry Thompson. ‘‘Introducing Cognitive Grammar,”’ in
Cathy Cogen et al., eds. Proceedings of the First Annual Meeting of the
Berkeley Linguistics Society. Berkeley, Calif.: Berkeley Linguistics Soci-
hye NETS);
Langacker, Ronald. Language and Its Structure. New York: Harcourt, 1967.
Leblanc, Hughes and William A. Wisdom. Deductive Logic. 2nd ed. Boston:
Allyn and Bacon, 1976.
Lee, Victor, ed. Language Development. New York: John Wiley, 1979.
Leech, Geoffrey. Semantics. Harmondsworth: Penguin, 1974.
Lehiste, Ilse. Suprasegmentals. Cambridge, Mass.: MIT Press, 1970.
Lehmann, Winfred P. Historical Linguistics: An Introduction. 2nd ed. New
York: Holt, 1972.
Lenneberg, Eric. Biological Foundations of Language. New York: John
Wiley, 1967.
Li, Charles, ed. Mechanisms of Syntactic Change. Austin: University of
Texas Press, 1977.
Lieberman, Philip. Speech Physiology and Acoustic Phonetics. New York:
Macmillan, 1977.
Lilly, John C. The Mind of the Dolphin: A Non-Human Intelligence. Garden
City, N.Y.: Doubleday, 1967.
Lindauer, Martin. Communication Among Social Bees. Cambridge: Cam-
bridge University Press, 1971.
Linden, Eugene. Apes, Men and Language. New York: Saturday Review
Press, 1974.
Lorenz, Konrad. King Solomon's Ring. New York: Thomas Y. Crowell,
19524
Lowth, Robert. A Short Introduction to English Grammar. 1762. Reprint.
Menston, Eng.: Scolar Press, 1967.
Lyons, John. Noam Chomsky. New York: Viking Press, 1970.
McCawley, James. *“‘The Role of Semantics in a Grammar,’’ in Emmon Bach
and Robert Harms, eds. Universals of Linguistic Theory. New York: Holt,
1968.
McNeill, David. The Acquisition of Language. New York: Harper & Row,
1970.
Makkai, Valerie B., ed. Phonological Theory: Evolution and Current
Practice. New York: Holt, Rinehart and Winston, 1972.
Malmstrom, Jean. Language in Society. Rey. 2nd ed. Rochelle Park, N.Y.:
Hayden Book, 1973.
Mamyuk, Paula. Language and Maturation. Cambridge, Mass.: MIT Press,
1977.
Matthews, Peter H. Morphology: An Introduction to the Theory of Word
Structure. Cambridge: Cambridge University Press, 1976.
Miller, George A. ‘‘Some Psychological Studies of Grammar.’’ American
Psychologist 17, 748-762, 1962. Also in Jakobovits and Miron.
Miller, George A. and K. McKean. ‘‘A Chronometric Study of Some Rela-
tions between Sentences.’’ Quarterly Journal of Experimental Psychology
16, 297-308, 1964.
 Bibliography 269

Miller, Roy A. The Japanese Language. Chicago: University of Chicago

Press, 1967.
Morgan, Jerry. ‘‘On the Treatment of Presupposition in Transformational
Grammar,”’ in Robert I. Binnick et al., ed. Papers from the Fifth Regional
Meeting of the Chicago Linguistics Society. Chicago: University of
Chicago Department of Linguistics, 1969.
Muller, S. H. The World’s Living Languages. New York: Ungar, 1974.
Nida, Eugene A. Morphology: The Descriptive Analysis of Words. 2nd. ed.
Ann Arbor: University of Michigan Press, 1956.
Padley, G. A. Grammatical Theory in Western Europe, 1500-1700: The Latin
Tradition. New York: Cambridge University Press, 1976.
Parret, Herman, ed. History of Linguistic Thought and Contemporary
Linguistics. New York: Walter De Gruyter, 1976.
Pedersen, Holger. Linguistic Science in the Nineteenth Century. Cambridge,
Mass.: Harvard University Press, 1931.
Penfield, Wilder and Lamar Roberts. Speech and Brain Mechanisms. Prince-
ton: Princeton University Press, 1959.
Pike, Kenneth. Phonetics. London: Oxford University Press, 1943.
Platt, John T. and Heidi K. Platt. The Social Significance of Speech: An
Introduction to and Workbook in Sociolinguistics. Amsterdam: North Hol-
land, 1975.
Postal, Paul M. *“‘On the Surface Verb ‘Remind,’ ”’ Linguistic Inquiry 1,
37-120, 1970.
Premack, David. Intelligence in Ape and Man. Hillsdale, N.J.: L. Erlbaum
Associates, 1976.
. ‘‘Language in the Chimpanzee?’’ Science 172, 808-822, May, 1971.
Premack, David and Ann J. Premack. ‘‘Teaching Language to an Ape.”
Scientific American, October, 1972, 92-99.
Pyles, Thomas. The Origins and Development of the English Language. 2nd
ed., New York: Harcourt, 1971.
Reed, Carroll E. Dialects of American English. Rev. ed. Amherst: University
of Massachusetts Press, 1977.
Rigg, A. G., ed. The English Language: A Historical Reader. New York:
Appleton-Century-Crofts, 1968.
Robertson, Stuart and Frederic G. Cassidy. The Development of Modern
English 2nd. ed. Englewood Cliffs, N.J.: Prentice-Hall, 1954.
Robins, Robert H. Ancient and Medieval Grammatical Theory in Europe.
London: Bell, 1951.
. A Short History of Linguistics. Bloomington: Indiana University
Press, 1967.
Rumbaugh, Duene M., ed. Language Learning by a Chimpanzee: The Lana
Project. New York: Academic Press, 1977.
Sadock, Jerrold M. Toward a Linguistic Theory of Speech Acts. New York
Academic Press, 1974.
Salus, Peter H. On Language: Plato to Von Humboldt. New York: Holt,
Rinehart and Winston, 1969.
Sapir, Edward. ‘‘The Psychological Reality of Phonemes,’’ in Valerie B.
’

Makkai, ed. Phonological Theory, Evolution and Current Practice. New

York: Holt, 1972.
270 ~=Bibliography

Saussure, Ferdinand de. A Course in General Linguistics. New York:

McGraw-Hill, 1966.
Schane, Sanford. Generative Phonology. Englewood Cliffs, N.J.: Prentice-
Hall, 1973.
Searle, John R. Expression and Meaning. London: Cambridge University
Press, 1979.
. ‘Linguistics and the Philosophy of Language,’ in Theo Vennemann
and Renate Bartsch, eds. Linguistics and Neighboring Disciplines. Am-
sterdam: North-Holland, 1975.
. Speech Acts. London: Cambridge University Press, 1969.
Sebeok, Thomas A., ed. How Animals Communicate. Bloomington: Indiana
University Press, 1977.
Sebeok, Thomas A. and Alexandra Ramsay eds., Approaches to Animal
Communication. The Hague: Mouton, 1969.
Shuy, Roger W. Discovering American Dialects. Champaign, IIl.: National
Council of Teachers of English, 1967.
Silverstein, Michael. “‘Linguistics and Anthropology,’ in Theo Vennemann
and Renate Bartsch, eds. Linguistics and Neighboring Disciplines. Am-
sterdam: North-Holland, 1975.
Singh, Sadand and Kala S. Singh. Phonetics: Principles and Practices. Balti-
more: University Park Press, 1976.
Slobin, Dan I. Psycholinguistics. Glenview, Ill.: Scott, Foresman, 1971.
Smalley, William A. Manual of Articulatory Phonetics. Rev. ed. n.p.: Com-
mittee on Missionary Personnel, Division of Foreign Missions, NCCC,
1968.
Stanley, Richard. ‘“Redundancy Rules in Phonology.’ Language, 1967.
Steinberg, Danny D. and Leon A. Jakobovits, eds. Semantics: An Interdisci-
plinary Reader in Philosophy, Linguistics and Psychology. Cambridge:
Cambridge University Press, 1971.
Thorpe, William H. Learning and Instinct in Animals. London: Methuen,
1956.
Tomkins, William. Indian Sign Dictionary. New York: Dover, 1969.
Tough, Joan. The Development of Meaning: A Study of Children’s Use of
Language. New York: Halsted Press, 1977.
Traugott, Elizabeth C. A History of English Syntax. New York: Holt,
Rinehart and Winston, 1972.
. “Toward a Grammar of Syntactic Change.’’ Lingua 23, 1-27, 1969.
Trubetzkoy, N.S. Principles of Phonology. Berkeley: University of California
Press, 1969.
Trudgill, Peter. Sociolinguistics: An Introduction. Harmondsworth: Penguin,
1974.
Ullmann, Stephen. An Introduction to the Science of Meaning. New York:
Barnes and Noble, 1979.
Von Frisch, Karl. Bees: Their Vision, Chemical Senses, and Language.
Ithaca, N.Y.: Cornell University Press, 1950.
. The Dance Language and Orientation of Bees. Translated by Leigh
E. Chadwick. Cambridge, Mass.: Harvard University Press, 1967.
Von Humboldt, Wilhelm. Linguistic Variability and Intellectual Development.
 Bibliography pa]
4

Translated by George C. Buck and Frithsof A. Raven. Coral Gabels, FI.:

University of Miami Press, 1971.
Vorlat, Emma. Progress in English Grammar, 1585-1735, Luxemburg, n.p.,
1963.
Walsh, J. M. and A. K. Walsh. Plain English Handbook. Cincinnati:
McCormick- Mathers, 1972.
Wang, William S-Y. “‘The Chinese Language.’’ Scientific American 228, 2,
50-63, February 1973.
Wardhaugh, Ronald. Introduction to Linguistics. 2nd ed. New York:
McGraw-Hill, 1977.
Webster, Noah. Webster's New International Dictionary of the English
Language. 2nd ed. Springfield, Mass.: G. & C. Merriam, 1953.
Weinreich, Uriel. *‘Explorations in Semantic Theory,’’ in Thomas A. Sebeok,
ed. Current Trends in Linguistics. Vol. 3. The Hague: Mouton, 1966.
Weinreich, Uriel, William Labov, and Marvin I. Herzog. ‘‘Empirical Founda-
tions for a Theory of Language Change,’ in Winfred P. Lehmann and
Yakov Malkiel, eds. Directions for Historical Linguistics. Austin: Univer-
sity of Texas Press, 1968.
Whitman, Randal L. English and English Linguistics. New York: Holt, Rine-
hart and Winston, 1975.
Whittaker, Harry A. ‘“‘Linguistic Competence: Evidence from Aphasia.”’
Glossa 4, 46-54, 1970.
Williams, Joseph M. Origins of the English Language: A Social and Linguis-
tic History. New York: Free Press, 1975.
Williamson, Juanita V. and Virginia M. Burke. A Various Language: Per-
spectives on American Dialects. New York: Holt, Rinehart and Winston,
1971.
Wilson, Edward O. Sociobiology: The Abridged Edition. Cambridge, Mass.:
Belknap Press of Harvard University, 1980,
Wolfram, Walt and Ralph W. Fassold. The Study of Social Dialects in Ameri-
can English. Englewood Cliffs, N.J.: Prentice-Hall, 1974.
Wundt, Wilhelm. The Language of Gestures. The Hague: Mouton, 1973.
 a
on = _
i

S ‘ - oe Dy
—, ) gate 2» leet! hun dell -) eer hd
ote oe
CUT vse finite eeswet
maiticoe 413.) / Eee OF artogiay) Apien4 ianaons gchored ah
— a et a ‘te

rf emia Bh ; 4 Actes 7 -) aaew

A ha Garr yshe
4 Ge: i Sty hig? i. i fé. ; i a juz ne ona
: ‘tor BUS gal
{ Ui ies ke AML gL OF Mast MeN oe iN stadiuibs ie Ww
vl Mewes Soi ae
‘ : \ y= Passa | eh res esy ' ee) atte" teat a
ah 4 eam a aSee fed
: Sh ene 3) eet ORO Le CPi
ee, A mitts @ weal? ~anavase Fi AG! 6 ON " Ace a fie baie.) a
C mm * : “4 =
iA : ae
eh i» i ettgelt sat, 0 fv gas} ae {kt LeLeute :
RIO (waruiyeslt oe call baie ab Oat. 7 : pyc StSTNAS deena: ¢
‘ ° a
om Wagers a ia ae SRA aS means Psat PSone al
710.) niveau. nbs aR Ty A vl, Aner) wih bgt. Pee i
Lenniat | ar. as PA
Al Hatt Fee ! Se 4 ahean oe Meh Be A Bhat Cae >

7
.
ave a z 7
- ame ¢)
We av po) z
yeti
AVAIL TID. Says eo sieie = ere ES
rere a eae
a i»? 2 é 3 La

Sia H ¥ +" ‘ im9 he,

ven+ a4) a! A ‘ay fa wih! y . x

; Phan
oye
i TOFoytnid,Yn
wan
BE
Sr

“~ cay
wscealg Dap, 3 avis 4 » ited M yes ‘ie

ORTEGA LOTT, I ane: "Baty

0 ek. Meu ails, cy cm ai ‘i eth

nk? TARE jo loosania
te MA
nprh, wh Lily
git oyaN aywy 3 et read as
Lay Meet sO Be ae
tl OWING cues of gua D ie Mb
Hh ee Dre ory Kru) P y a, ex -”

e; Cas Yyer tia parry La : =.

mn et : fate} Tn ia% cael

_ | a :
F vile, & Sed 7) -< Penn ey ae tae ;

Ld Peay, Fi Sa : :
a / i tary, aN Sie qh eae
cen a He AY ae = , tel iipae é jh A Dy ‘7 ipod

oy NT : rat ; a ma.
J antaal! Venera shia!
a * | rppilpeat
a; ’ Pon ‘al oe
“oTilann, we the
ae)
Ror :
 INDEX

Accessibility hierarchy, 201 Analysis-by-synthesis model, 169-170

Acoustic phonetics, 57, 230 Analytic languages, 199
Acts, locutionary, 150 Anglo-Saxon, 39-44, 205-207, 211-213
Adjustment rules, 100 Animal communication, 3-12
Adverb preposing, 187 Anomaly, 135
Affricates, 59, 63 Anthropological linguistics, 192-203
African languages, 23-24 Antonymy, 131
influences on American English, 50-51 Ape communication, 9, 11
Afro-Asiatic languages, 196 Aphasia, 222-225
Agents, 145 Aristotle, 29
Agglutinative languages, 199 Articulatory classification, 209-210
Allomorphs, 98—100 Articulatory phonetics, 57-70
Allophones, 73, 83 Aspects of the Theory of Syntax (Noam
Alphabetic writing system, 19-20 Chomsky), 186
Altaic languages, 196 Aspirated stops, 61
Alveolar ridge, 63 Aspiration, 61, 73
Alveolars, 59, 61 Assimilation, 206
Ambiguity, 119-120, 135-136, 149 Assimilatory rules, 93
American English, 175. See also Dialects Atomic statements, 148
and Slang Austro-Asiatic languages, 198
African influence on, 50—S1 Auxiliary element, 110
British English influence on, 52
Dutch influence on, 50 Babbling period, 164-165
French influence on, 50 Bacon, Roger, 31
German influence on, 50-51 Base component (morphology), 100
Indians’ (of North America) influence Base component (syntax), 109
on, 50 Bee communication, 5, 7-8
Italian influence on, 50-51 Behaviorist approach to language acquisi-
Jewish influence on, 50-51 tion, 166-167
Scandinavian influence on, 50-51 Bever, Thomas, 168
Spanish influence on, 50 Bilabials, 59, 61
American sign language, 25 Black English, 180-181, 187-188
American standard English, 64, 179-180 Bleeding order, 211
Amplitude, 231 Bloom, Lois, 165-166
Analogy, 212 Bloomfield, Leonard, 34

213
274 Index

Body language, 22 Context, 135

Bopp, Franz, 32 Contradiction, 135
Borrowing, 214 Co-occurence restrictions, 118
Bound morphemes, 97-98 Cratylus, 29
Brain, speech areas of, 224-225 Creoles, 182
Brain mapping, 222-225 Culture, 193-194
Broca’s aphasia, 223 Cyclic rules, 86
Broca’s area, 223

Deaf, communication among the, 23, 25

Case grammar, 144-146, 152
Declaratives, 151
Case marker, 145
Decoding, 4, 159
Caxton, William, 47
Deep structure, 35, 109, 118-119, 141
Celtic language, 39-42
Denotative meaning, 130
Centum languages, 195-196
Derivational morphemes, 98
Cetacean communication, 5
Descartes, 31
Chaffinche communication, 8
Descriptive linguistics, 34, 104-106
Chaucer, Geoffrey, 47
Determiner, 111
Chimpanzee communication, 9, 11
Diachronic linguistics, 33
Chinese language, 17-18, 23
Diacritics, 64-65
Chinook language, 182
Diagrams
Chomsky, Noam, 31, 35, 75-83, 102, 108,
Reed-Kellogg, 33, 103-104
126, 129, 138, 141, 152, 157, 172, 186
tree, 111 :
Clark, Eve V., 167
Dialects, 179-181, 183-185
Clark, Herbert, 167
Dictionaries, 131-134, 139
Cognates, 214
A Dictionary of the English Language
Cognitive grammar, 170-171
(1755) (Samuel Johnson), 49
Combined statements, 149
Dionysius Thrax, 29
Commissives, 150
Diphthong, 65
Common nouns, 30
Directives, 151
Communication, definition of, 3, 12
Dissimilation, 207
Comparative method, 32, 214-218
Distinctive features, 66-68, 70, 162
Competence, 107, 115, 151, 159
Distinguishers, 139
model of, 108-109
Dolphin communication, 5
standard interpretive theory of, 108, 127,
Donatus, 30
138, 158
Double negatives, 49
Complementary distribution, 73
Dravidian languages, 198
Complete merger, 208
Drum communication, 24
Components
Dryden, John, 48-49
morphological, 100
Dutch language, influence on American
phonological, 16
English, 50
semantic, 16, 158
syntactic, 16, 108-109, 158
transformational, 109, 117 Early modern English, 47-48, 213
Compound stress rule, 87 Elaboration, 212-213
Conjoined phrases, 126 Embedded sentences, 122-123
Connotative meaning, 130 Encoding, 3
Consonants, 59-64, 236-237 English language
defined, 70 Anglo-Saxon, 39-44, 205-207, 211-213
voiced, 61 Early Modern English, 47-48, 213
voiceless, 61 history of, 35-53
Constituent structure, 158-161 influence of African languages on, 49
Constituents, 160-161 influence of French language on, 44-45,
Constraints, 125-126 49
 Index 21D

influence of Indian (Asian) language on, Generative semantics, 108, 127, 142-144
49 Geographical influence on language diver-
influence of Mexican language on, 49 sity, 176-179
influence of South American languages Geography, linguistic, 176-179
on, 49 Germanic tribes, 40-41
Middle English, 44-47 German language, influence on American
Modern English, 48-52, 211-213 English, 50-51
nonstandard, 179 Gesture, 22, 26
Old English. See Anglo-Saxon Glides, 59, 64
redundancy statements, 88-89, 94 Glottal stops, 62
spelling, 20 Goal, 145
Standard, 64, 179-180 Gorilla communication, 9, 11
Entailment, 135 Grammar, 16. See also Transformational-
Epilepsy, 226-227 generative and structural linguistics.
Excrescence, 206 cognitive, 170-171
Existential quantifier, 149 definition of, 26
Expressives, 151 history of, 28
Extended standard theory, 141 prescriptive, 103-104
Grammatical theory of language variation,
Face reading, 22-23 186-188
Features Great vowel shift, 46, 82, 207
distinctive, 66-68, 70, 162 Greek alphabet, 20
phonological, 77 Grimm brothers, 32
semantic, 139 Grimm’s law, 217
Feeding order, 210-211 Gullah dialect, 183
Fillmore, Charles, 144-145
Finnish language, 210 Halle, Morris, 36, 75, 79, 81
Finno-Ugric languages, 196 Harris, Zellig, 35
Flip-flop rule, 120 Henry VIII, King of England, 47
Focus, 141 Hertz (Hz), 230-238
Fodor, J. A., 36, 138, 142, 168 Heschl’s area, 224
Formal universal, 200 Heuristics, 107
Formant, 234 perceptual, 168
Formatives, 158 Hiragana, 18, 21
Free morphemes, 97-98 syllabary, 19
Free variation, 73 Historical-comparative linguistics, 32-33
French language Homophony, 130
influence on American English, 50 Human communication, 10
influence on English, 44-45, 49
Frequency (acoustics), 230-231 Idiolects, 179
Frequency (markedness), 91 Illocutionary acts, 150-151
Fricatives, 59, 62-63, 237 Immediate constituents, 105
alveolar, 62 Imperative transformations, 121-122
interdental, 62 Implicational hierarchies, 186-188
labiodental, 62 Implicational universals, 200
palatal, 62-63 Indian (Asian) linguists, 29
Fromkin, Victoria, 171 Indians of North America
Frozen word order, 193 influence on American English, 50
sign language, 24-25
Gardner, Beatrice T., 9 Indo-European languages, 32, 195
Gardner, R. Allen, 9 Indo-Iranian languages, 195
Garrett, M. F., 168 Inflectional languages, 199
General American English, 64, 179-180 Inflectional morphemes, 97-98
276 Index

Innate communication, 8 Larynx, 58

Innovation, 210 Latin language, 16, 39, 42, 181, 204-205,
Interdentals, 59 207, 216
Internal reconstruction, 217-218 grammar of, 30-31
Interpretive semantics, 108, 127, 138-144 influence on Anglo-Saxon, 41-42
Intonation, 69 Learned communication, 8
Island hierarchy, 187 Lexical diffusion, 208
Isoglosses, 177 Lexical-formation rule, 121
Italian language, influence on American Lexical insertions, 114
English, 50-51 Lexical items, 140
Lexicon, 90, 109, 132, 158, 162-163
Jakobson, Roman, 35 Limbic system, 225-226
Japanese language, 16, 18 Lingua francas, 181
writing of, 21 Linguistic atlases, 177
Jespersen, Otto, 33 Linguistic geography, 176-179
Jewish language, influence on American Linguistic Geography of New England
English, 50-52 (Hans Kurath et al.), 177-179
Johnson, Samuel, 49 Linguistic variables, 184
Jones, William, 32 Linguistics
anthropological, 192-203
Kana, 18 definition of, 28
Kanji, 21 descriptive, 34
Katakana, 18, 21 diachronic, 33
Katz, J. J., 36, 138, 142 historical, 32-33
Keenan, Edward, 146 structural, 34
Kernel sentences, 159 synchronic, 33
Kinesics, 22 Lip reading, 22-23
Koko, 9, 11 Liquids, 59, 63
Kurath, Hans, 177-178 Local language, 181
Locutionary acts, 150
Labiodentals, 59 Logic, 148-150
Labov, William, 183, 187 Logographic writing, 17
Lakoff, George, 108, 170 London dialect, 47
Langland, William, 47 Lorenz, Konrad, 6
Language, 3 Loss, 206
definition of, 10 Lowth, Bishop, 49
Language acquisition, 92, 164-168
Language and consciousness, 226-229 McCawley, James, 108
Language and culture, 194 McNeill, David, 162-163
Language and Mind (Noam Chomsky), Main stress rule, 87
157 Malayo-Polynesian languages, 196
Language and thought, 28, 192-194 Markedness, 91-92, 94, 162-163
Language change, 92, 204-206, 218-219 frequency, 91
Language diversity, 175—177 neutrality, 91
geographic influence, 176-179 productivity, 91
social class influence, 183-185 regularity, 91
Language families, 194-198 Mazatecan language, 23
Language reconstruction, 32, 214-218 Meaning, 119, 127
Language typology, 92 connotative, 130
Language universals, 31, 199-202 denotative, 130
Language variation, grammatical theory Meaning (semantics), 131-134
of, 186-188 Meaning (speech acts), 150-152
Languages, types of, 198-199 Medium, 4
 Index 277

Memory, 162-163 Oral cavity, 58

Merger Ordering of transformations, 82, 124
complete, 208
partial, 208 Palatal fricatives, 62-63
Metalanguages, 139 Palatals, 59
Metathesis, 208 Panini, 28
Middle English, 44-47 Paralanguage, 22
Minimal pair, 73 Paraphrases, 119, 132
Mixed writing system, 21 Partial merger, 208
Modality, 145 Perceptual reality of phonemes, 161-162
Modern English, 48-52, 211-213 Performance, 106, 115, 151
Modistic grammar, 30-31 Perlmutter, David, 46
Monophthongs, 65 Perlocutionary acts, 150
Morphemes, 90 Pharynx, 58
definition of, 96, 100 Philology, 28
types of, 97-98 Phonemes, 73-74, 161-162
Morphological change, 211-212 Phonemic alternation, 74
Morphology Phonemics, structural, 72
definition of, 96 Phonetic alphabet, 32, 59
rules of, 100 Phonetics
structural approach, 96-97 acoustic, 57, 230
transformational-generative approach, articulatory, 57-70
99-100 definition of, 57, 59
Morphophonemes, 74 Phonological acquisition, 164-165
Mouth, 58 Phonological change, 206-211
Phonological component, 16
Phonological derivation, 77
Narrowing, 213
Phonological feature, 77
Nasal cavity, 58
Phonological rules
Nasals, 59, 63
assimilatory, 93
Natural phonology, 93-94
compound stress, 87
Natural rules, 92
cyclic, 86
Negations, 121-122
definition of, 76-77
Neo-grammarian school of linguistics, 32
main, 87
Neurolinguistics, 222-229
natural, 92
definition of, 222
ordering of, 82, 124
Neutrality, 91
reordering, 210
Niger-Congo languages, 196
trisyllabic laxing, 81, 91
Nonstandard English, 179
velar softening, 81-82
Nonterms, 147
Phonological space, 207
Normans, 44-45
Phonology
Normative grammar. See Prescriptive
definition of, 70
grammar
structural and transformational-
Noun phrase, 110
generative compared, 75
Nouns, 30
Phrase-structure grammar, 109-113
Phrase-structure marker, 113
Objects (case grammar), 135 Phrase-structure rules, 109, 115, 158
Old English, 39, 44, 205-207, 211-213. Pictographs, 17
See also Anglo-Saxon Pidgins, 181-182
Onomatopoeia, 15 Pivot words, 165
Open word class, 165 Plato, 29
Operations (morphology), 100 Plosives. See Stops
Optic crossing, 227 Port Royal grammarians, 31
278 Index

Postal, Paul, 108, 146 Semantic change, 213-214

Predicate raising, 142-143 Semantic component, 16, 158
Predicates, 149 Semantic features, 114-115, 139
Premack, Ann J., 9 Semantic markers. See Semantic features
Premack, David, 9 Semantic redundancy rules, 140
Prescriptive grammar, 33, 103 Semantic space, 164
Presupposition, 134 Semantic sirategies, 169-170
Priscian, 30 Semantic structure, 142
Productivity, 91 Semantics, 16, 131-134
Projection rules, 140 definition of, 129, 136
Proper nouns, 30 generative, 108, 127, 142-144
Propositions, 142, 145 interpretive, 143-144
Protoforms, 214 quantificational approach, 163-164
Proto-Indo-European, 214 Semivowels, 59, 64
Proxemics, 22 Sentence pattern, 105
Psycholinguistic experiments, 159-161 Sentence-perception grammar, 170-171
Psycholinguistics, 157—174 Sentences (statements), 148
Psychological reality, 157-159, 160-162 Sentences
Pure tone, 230 definition of, 102
embedded, 122-123
Qualifiers, 145 kernel, 159
Quantificational approach, 163-164 semantics of, 134
Quantified statements, 149 Shakespeare, William,48
Questions, 121-122 Short Introduction to English Grammar
(1762) (Bishop Lowth), 49
Rask, Rasmus, 32 Sign language, 25
Recursive rules, 123 Simplification
Redundancy, 88-91, 94 phonological change, 211
Redundancy rules, 140 syntactic change, 212
Redundancy statements, 89-91 Sino-Tibetan languages, 196
Reed-Kellogg sentence diagrams, 33, Skylark communication, 9
103-104 Slang, 183
Regularity, 91 Social class
Relational grammar, 146-148 distinctions, 176
Relational hierarchy, 147-148 influence on language, 183-185
Relationship of sound and meaning, 15 Sociolinguistics, 175-190
Relative transformation, 123 Soft palate,58
Representatives, 151 Sound, 230
Restructuring, 210 Sound change, 208-211
Romance languages, 215 The Sound Pattern of English (Noam
Romanji, 21 Chomsky and Morris Halle), 75, 79, 83
Ross, John, 108, 126 Source language, 181
Rule ordering, 82 Spanish language, influence on American
Rule reordering, 210 English, 50
Spectrogram, 233
Sanskrit, 28 Spectrograph, 230-232
Satem languages, 195-196 Speculative grammar, 30-31
Saussure, Ferdinand de, 34, 218 Speech, 14, 26
Scandinavian languages Speech acts, 150-152
influence on Anglo-Saxon, 42 Speech errors, 171-172
influence on American English, 50-51 Speech perception, analysis-by-synthesis
Schane, Sanford, 93 model, 169-170
Searle, John, 151 Speech perception grammar, 170-171
 Index ie]

Speech production, 171-172 Symmetry, 207

Speech sounds, production of, 57-59 © Sympathetic vibration, 231-232
Speed writing, 21 Synchronic linguistics, 33
Spelling Synonymy, 130-131
English, 20 Syntactic acquisition, 166
reforms, 20 Syntactic change, 211-213
Split (phonological change), 209 Syntactic component, 108-109, 158
Split-brain experiments, 227-228 Syntactic features, 99, 114
Stampe, David, 93 Syntactic Structures (Noam Chomsky),
Standard English, 64, 179-180 102
Standard interpretive theory, 108, 127, 138, Syntax, 16
158 constraints on, 125-126
Standard predicate calculus, 148-150, 152 definition of, 102
Statements structural approach, 103-106
atomic, 148 transformational-generative approach,
combined, 149 106-109
quantified, 149
Statistical universals, 201
Stoics, 30 Tagalog, 16
Stops, 59-62 Tense, 112
Stratification, 184-185 Terms, 147-148
Stress, 69 Thompson, Henry, 170
Transformational cycle, 86 Thorpe, William H., 8
Stress contours, 93 Thrax, Dionysius, 29
Stress rules Tip-of-the-tongue phenomenon, 162-163
assimilatory, 93 Tok Pisin, 182
compound, 87 Tone languages, 23, 198-199
cyclic, 86 Topic, 141
definition of, 76-77 Trachea, 58
main, 87 Trachne grammatike (Dionysius Thrax),
natural, 92 29
ordering of, 82, 124 Traditional grammar, 33, 103
reordering, 210 Transformational component, 109, 117
trisyllabic laxing, 81, 91 Transformational cycle, 86
velar softening, 81-82 Transformational-generative grammar,
Strong verbs, 44, 211 34-36, 106-128
Structural change, 120 Transformations
Structural classification, 208 conjoined phrases, 126
Structural description, 120 constraints on, 125-126
Structural grammar, 104-106 definition of, 117
Structural linguistics, 34 embedded sentences, 122-123
Structural phonemics, 72 flip-flop rule, 120
Structuralists, 103 imperative, 122
Structure. See Syntax lexical formation, 121
Substantive universals, 200 ordering, 124
Suprasegmentals, 68, 198 recursive, 123
Surface structure, 109, 118-120, 152 relative, 122
Surface structure and semantic structure, Transformations and constituent structure,
142 159-160
Swift, Jonathan, 48-49 Tree diagram, 111
Syllabary, 19 Trisyllabic laxing rule, 81, 91
Syllable, 69 Trubetzkoy, Nikolaj Sergeevic, 35
Syllabic writing system, 18 Tuning fork, 231-233
280 Index

Turkish alphabet, 20 Vowel chart, 64-65

Typologies of languages, 198-199 Vowels, 59, 70, 233-236

Washoe, 9
Unaspirated stops, 62 Wesecneuas
Underlying representations, 75, 78-81 Webster-Neah 49
Universal grammar, 32 ie Aas j
. : Wernicke’s aphasia, 223-224
Universal quantifier, 149 Eee
: Wernicke’s area, 223
Universals, 31, 199-202
Uvula, 58 Whales, 5
: Whistling, 23
Whorf, Benjamin Lee, 193
Varro, 30 Whorfian hypothesis, 193
Velar, 59 Widening, 213
Velar softening rule, 81-82 Word classes, 29-30
Velum, 58 Word deafness, 224
Vennemann, Theo, 93 Word endings, 16-17, 30
Verb phrase, 110 Word-formation rule, 100
Verbs Word order, 16. See also Syntax
strong, 44, 211 Word relationships, 130
weak, 44 Words, 130
Vocal cords, 58, 61 open class, 165
Vocal tract, 14 pivot, 165
Voice bar, 233 Writing, 17
Voice box, 58 Wycliffe, John, 47
Voiced consonants, 61
Voiceless consonants, 61 Yerkes Primate Center, 9
Voicing, 61 Yiddish language
Von Frisch, Karl, 5, 7 influence on American English, 50-51
von Humboldt, Wilhelm, 32-33 lingua franca, 181
 Pa:
a

, a)
© ’ Ls .
- 3 ; § 4
i F
9 um

ard <5 ly
: 5 @ > - I
_

¥ . ©
re.

* . < 7
- ? - : F
_

i
o”
iKE
_—
=
_
yf
af

i) is.
sues
1]
- ~
= ge eea Le
ww
as .
ae
: 7 = >
y
ad
_ _
—— hd ree _ i
- 7 itd
Little, Brown and Company.
Boston

ISBN 0-316-160156