An Introduction to Syntax
ROBERT D. VAN VALIN, JR.
Department of Linguistics, University at Buffalo, The State University of New York
p ub l i s he d b y t he p r e s s sy n d i c a t e o f t h e u n i v e r si t y o f c a m b r i d g e
The Pitt Building, Trumpington Street, Cambridge, United Kingdom
c amb r i dge uni ve r s i t y p r e ss
The Edinburgh Building, Cambridge c b2 2r u, UK
40 West 20th Street, New York, ny 10011–4211, USA
10 Stamford Road, Oakleigh, Melbourne 3166, Australia
Ruiz de Alarcón 13, 28014 Madrid, Spain
www.cup.cam.ac.uk
www.cup.org
© Robert D. Van Valin, Jr. 2001
This book is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without
the written permission of Cambridge University Press.
First published 2001
Printed in the United Kingdom at the University Press, Cambridge
Typeface 10/12pt Times
[gc]
A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data
Van Valin, Robert D., Jr.
An Introduction to Syntax / Robert D. Van Valin, Jr.
p. cm.
Includes bibliographical references and index.
i s b n 0 521 63199 8 (hdbk.) — i s b n 0 521 63566 7 (pbk.)
1. Grammar, Comparative and general—Syntax. I. Title.
P291.V357 2001
415—dc21
i s b n 0 521 63199 8 hardback
i s b n 0 521 63566 7 paperback
00–062129
Contents
List of figures
Preface
Acknowledgements
Abbreviations
page ix
xiii
xiv
xv
1
Syntax, lexical categories, and morphology
1.0 Introduction
1.1 Aspects of syntactic structure
1.2 Lexical categories
1.3 Morphology
Notes and suggested readings
Exercises
1
1
4
6
13
17
17
2
Grammatical relations
2.0 Introduction
2.1 Grammatical relations versus semantic roles
2.2 Properties of grammatical relations
2.3 Other systems of grammatical relations
2.4 Conclusion
Notes and suggested readings
Exercises
21
21
22
33
70
79
79
80
3
Dependency relations
3.0 Introduction
3.1 Syntactic dependencies
3.2 Dependency representations
3.3 Conclusion
Notes and suggested readings
Exercises
86
86
87
101
106
107
107
4
Constituent structure
4.0 Introduction
4.1 Constituents and their formal representation
110
110
111
vii
Contents
4.2 The universality of form classes
4.3 An alternative schema for phrase structure
4.4 The structure of complex sentences
4.5 Constituent structure and grammatical relations
Notes and suggested readings
Exercises
119
122
133
137
142
142
5
Grammar and lexicon
5.0 Introduction
5.1 Phrase structure rules
5.2 The lexicon and subcategorization
5.3 Relational-dependency rules and lexicon
5.4 Concluding remarks
Notes and suggested readings
Exercises
144
144
144
156
162
168
169
169
6
Theories of syntax
6.0 Introduction
6.1 Relational Grammar
6.2 Lexical-Functional Grammar
6.3 Government-Binding Theory
6.4 Role and Reference Grammar
6.5 Summary
6.6 Other syntactic theories
6.7 Conclusion
Notes and suggested readings
Exercises
172
172
173
182
193
205
218
221
224
225
225
References
Language index
Subject index
227
234
236
viii
List of figures
1.1
1.2
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
3.11
3.12
3.13
4.1
4.2
4.3
4.4
4.5
4.6
Language as a correlation between gestures and meaning
Language as a correlation between gestures and meaning
(revised)
Verb-specific semantic roles and thematic relations
Continuum from verb-specific semantic roles to grammatical
relations
Accusative versus ergative patterns
Structure of relative clauses
Structure of matrix-coding-as-subject construction
Structure of matrix-coding-as-object construction
Structure of control construction in (2.64a)
Structure of control construction in (2.64b)
Structure of conjunction-reduction construction in (2.70b)
Preliminary dependency representation
Enhanced representation
Yagua head-marking PP ‘in the canoe’
Swahili head-marking clause in (3.16)
Representation of Kalkatunga sentence in (3.2a′)
Representation of Croatian sentence in (3.1b)
Coordination of NPs and Vs
Active and passive clauses in English
Grammatical relations versus macroroles in Yidi¤ and English
Malagasy object complement
English infinitival complement
Dyirbal relative clause
Enhanced representations of English control construction and
Dyirbal relative clause
Aspects of a phrase-structure tree
Preliminary phrase-structure tree for (4.3a)
Phrase-structure trees for the two readings of (4.12a)
Tree diagram for (4.17b)
Constituent structure of (4.13b)
Constituent structure of (4.19a)
page 1
3
29
31
36
47
50
50
53
54
56
102
102
103
103
103
104
104
104
105
105
105
106
106
115
117
118
118
119
120
ix
List of figures
4.7
4.8
4.9
4.10
4.11
4.12
4.13
4.14
4.15
4.16
4.17
4.18
4.19
4.20
4.21
4.22
4.23
4.24
4.25
4.26
4.27
4.28
4.29
4.30
4.31
5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8
5.9
5.10
5.11
5.12
5.13
5.14
5.15
6.1
6.2
6.3
6.4
6.5
x
Revised preliminary phrase-structure tree for (4.3a)
Constituent structure of her proof of the theorem
General schema for phrase structure
Parallel phrase structure of N″ and V″
Tentative constituent structure representation for
this beautiful expensive red scarf
Iterated V-bars in a V″
Distinguishing argument PP from adjunct PP within V″
Structural contrast between argument and adjunct within N″
X-bar structure of P″ and Adj″
Final revision of the phrase-structure tree for (4.3a)
Coordinate Ns, N′s and N″s
General schema for coordination
Left-branching V′ and P′ in Lezgian
Left-branching clause structure in Hindi
Constituent structure of (4.35a)
Constituent structure of (4.37)
Constituent structure of English infinitival complement
Structure of left-branching control construction in Tindi in (4.38)
The constituent structure of English relative clauses
Constituent structure of the left-branching Quechua relative clause
in (4.39)
Basic clause configuration
Constituent structure of examples in (4.41)
Russian ditransitive sentence in (4.43)
Alternative structure for (4.43)
Possible constituent-structure trees for (4.47b) and (4.47c)
General X-bar schema for phrase structure
Correspondence between constituent structure and PS-rules
NP structures
English NP with adjunct PP
Correspondence between VP structure and PS-rules
English VP with adjunct PP
English object complement (embedded S′)
Result of first application of PS-rules in (5.13)
English VP containing NP with embedded PP
Three possible English V’s
Relational-dependency representations of English sentence
Relational-dependency representation of (5.35)
Relational-dependency representation of infinitival complement
Relational-dependency representation of that-clause complement
Revised representations for coordinate constructions
Relational network of the Russian sentences in (6.1)
Relational network for (6.1d)
Relational network for WH-question
Relational network for a passive construction
Relational network for dative shift
124
124
125
125
126
127
127
128
129
130
131
131
132
132
133
134
134
135
135
136
137
138
139
139
141
145
146
147
149
150
150
153
154
155
156
163
164
167
167
168
173
174
174
175
176
List of figures
6.6
6.7
6.8
6.9
6.10
6.11
6.12
6.13
6.14
6.15
6.16
6.17
6.18
6.19
6.20
6.21
6.22
6.23
6.24
6.25
6.26
6.27
6.28
6.29
6.30
6.31
6.32
6.33
6.34
6.35
6.36
6.37
6.38
6.39
6.40
6.41
6.42
6.43
6.44
6.45
6.46
Relational network for dative shift + passive
Organization of TG (Chomsky 1965)
TG derivation of (6.6)
An ill-formed relational network
Relational network for (6.8)
Relational network ruled out by 1 Advancement Exclusiveness Law
Relational networks for unergative and unaccusative verbs
C-structure and f-structure representations for ‘Juan sees a dog’
Grammatical functions in simple sentences
F-structure of Russian example in (6.13)
Enhanced c-structures for (6.10b) and (6.10c)
C-structure of (6.15)
F-structure of (6.15)
Impossible f-structures
C-structure and f-structure of English passive sentence
C-structure and f-structure for a WH-question in English
Organization of GB
General X-bar structure for clauses
Possible configurations for government
Case and +-role assignment in GB
D-structures of active and passive sentences
S-structure of Pat ate the snack
S-structure of passive sentence The snack was eaten by Pat
D-structure and s-structure of WH-question in English
Extended functional projections in the clause
S-structure of Pat sees Chris
Organization of the Minimalist Program
Universal oppositions underlying clause structure
Layered structure of the clause
Layered structure of clauses in English and Lakhota
The representation of operators in RRG
The organization of RRG
Representing focus structure in the layered structure of the clause
Constituent-, operator-, and focus-structure projections of the clause
The Actor-Undergoer Hierarchy
Combining syntactic templates from the syntactic inventory
Summary of the RRG linking system
Active and passive linkings from the logical structure in (6.24)
Linking between semantics and syntax in a WH-question
Historical development of the theories discussed in this chapter
HPSG representation of They sent us a letter
176
177
178
179
180
180
181
183
184
185
186
187
188
189
191
192
193
194
196
197
198
199
199
200
202
203
204
205
206
207
208
209
209
210
211
213
214
215
217
218
222
xi
Syntax, lexical categories, and morphology
CHAPTER 1
Syntax, lexical categories, and
morphology
1.0
Introduction
This book is an introduction to the basic concepts of syntax and syntactic
analysis. Syntax is a central component of human language. Language has often been
characterized as a systematic correlation between certain types of gestures and meaning, as represented simplistically in Figure 1.1. For spoken language, the gestures are
oral, and for signed language, they are manual.
Figure 1.1. Language as a correlation between gestures and meaning
It is not the case that every possible meaning that can be expressed is correlated with a
unique, unanalyzable gesture, be it oral or manual. Rather, each language has a stock of
meaning-bearing elements and different ways of combining them to express different
meanings, and these ways of combining them are themselves meaningful. The two
English sentences Chris gave the notebook to Dana and Dana gave the notebook to
Chris contain exactly the same meaning-bearing elements, i.e. words, but they have
different meanings because the words are combined differently in them. These different combinations fall into the realm of syntax; the two sentences differ not in terms of
the words in them but rather in terms of their syntax. Syntax can thus be given the
following characterization, taken from Matthews (1982:1):
The term ‘syntax’ is from the Ancient Greek s)ntaxis, a verbal noun which
literally means ‘arrangement’ or ‘setting out together’. Traditionally, it refers
to the branch of grammar dealing with the ways in which words, with or
without appropriate inflections, are arranged to show connections of meaning
within the sentence.
First and foremost, syntax deals with how sentences are constructed, and users of
human languages employ a striking variety of possible arrangements of the elements in
sentences. One of the most obvious yet important ways in which languages differ is the
order of the main elements in a sentence. In English, for example, the subject comes
before the verb and the direct object follows the verb. In Lakhota (a Siouan language
of North America), on the other hand, the subject and direct object both precede the
1
An introduction to syntax
verb, while in Toba Batak (an Austronesian language of Indonesia; Schachter 1984b),
they both follow the verb. This is illustrated in (1.1), in which the teacher, waXspekhiye
ki and guru i function as subjects, and a book, wówapi wN and buku function as direct
objects.
(1.1)
a. The teacher is reading a book.
b. WaXspekhiye ki wówapi wN yawá.
teacher
the book a read
c. Manjaha buku guru i.
read
book teacher the
English
Lakhota
Toba Batak
The Lakhota and Toba Batak sentences also mean ‘the teacher is reading the book’, and
in the Lakhota example the subject comes first followed by the direct object, whereas in
the Toba Batak example the subject comes last in the sentence, with the direct object
following the verb and preceding the subject. The basic word order in Toba Batak is
thus the opposite of that in Lakhota. There are also languages in which the order of
words is normally irrelevant to the interpretation of which element is subject and which
is object. This is the case in the following Russian sentences.
(1.2)
a. UAitel’nica Aitaet knigu.
teacher
read book
b. Knigu Aitaet uAitel’nica.
book read teacher
c. Bitaet uAitel’nica knigu.
read teacher
book
Russian
Again, all three of these sentences mean ‘the teacher is reading the book’, and in
these Russian examples the order of the words is not the key to their interpretation,
as it is in the sentences from the other three languages. Rather, it is the form of the
words that is crucial. The -a on the end of uAitel’nica ‘teacher’ signals that it is the
subject, and the -u on the end of knigu ‘book’ indicates that it is the direct object. If
the word for ‘teacher’ were the direct object in a sentence, then it would end in -u, as
in (1.3).
(1.3)
a. FenCAina videla uAitel’nicu.
woman saw teacher
b. UAitel’nicu videla EenCAina.
teacher
saw woman
‘The woman saw the teacher.’
Russian
These changes in the form of the words to indicate their function in the sentence are
what Matthews referred to as ‘inflections’, and the study of the formation of words
and how they may change their form is called morphology. These examples illustrate
the important relationship between syntax and morphology: something which may
be expressed syntactically in some languages may be expressed morphologically in
others. Which element is subject and which is object is signalled syntactically in the
examples from English, Lakhota and Toba Batak, while it is expressed morphologically in the Russian examples. Syntax and morphology make up what is traditionally
referred to as ‘grammar’; an alternative term for it is morphosyntax, which explicitly
recognizes the important relationship between syntax and morphology. Even though
2
Syntax, lexical categories, and morphology
this book is focussed on syntax, morphology will nevertheless be an important part of
the discussion.
Thus a more complex picture of the nature of language emerges than that given in
Figure 1.1; it is summarized in Figure 1.2.
Figure 1.2. Language as a correlation between gestures and meaning
(revised)
All of the examples looked at so far involve simple sentences, but one of the most
important syntactic properties of language is that simple sentences can be combined
in various ways to form complex sentences. In terms of Figure 1.2, one could say that
syntax makes possible the formulation of expressions with complex meanings out of
elements with simple meanings. One of the defining features of human language is its
unlimited nature; that is, the number of meaningful expressions that can be produced
by users of a human language is potentially infinite, and this expressive potential comes
from the combination of the basic meaningful elements with syntactic principles. Much
of the interest in language in psychology and cognitive science comes from what the
study of the cognitive mechanisms underlying language use and acquisition can reveal
about the human mind.
This book has three goals: first, to introduce the basic concepts of syntax; second,
to elucidate the principles and tools of syntactic analysis, which make it possible for
linguists to analyze the grammatical systems of human languages; and third, to give an
overview of the typological range of phenomena found in human languages which syntacticians seek to describe. The content of this book is presupposed by more advanced
courses in syntactic theory, and hence it is intended to prepare the reader for such
courses. The perspective of the book is primarily descriptive, and theoretical issues will
be raised only in chapter 6. To many people the term ‘grammar’ evokes bad memories
of prescriptive rules learned in school, e.g. ‘don’t split infinitives!’ Since the early
part of the twentieth century, linguistics has rejected the prescriptive tradition which
underlies school grammars and focusses instead on describing what users of human
language actually do, not on prescribing what they should do.
A central part of the description of what speakers do is characterizing the grammatical (or well-formed) sentences of a language and distinguishing them from
ungrammatical or (ill-formed) sentences. Grammatical sentences are those that are in
accord with the rules and principles of the syntax of a particular language, while ungrammatical sentences violate one or more syntactic rules or principles. For example,
(1.1a) is a grammatical sentence of English, while Teacher the book a reading is would
not be. Ungrammatical sentences are marked with an asterisk, hence *Teacher the book
a reading is. This sentence is ungrammatical because it violates some of the word order
rules for English, that is (i) basic word order in English clauses is subject–verb–object,
(ii) articles like the and a precede the noun they modify, and (iii) auxiliary verbs like
is precede the main verb, in this case reading. It is important to note that these are
English-specific syntactic rules; this word order is perfectly grammatical in Lakhota,
3
An introduction to syntax
as (1.1b) shows, and if the Lakhota words were arranged in the English order, e.g.
*Ki waXspekhiye yawá wN wówapi [the teacher reads a book], the result is thoroughly
ungrammatical. Well-formed sentences are those that are in accord with the syntactic
rules of the language; this does not entail that they always make sense semantically.
For example, the sentence the book is reading the teacher is nonsensical in terms of
its meaning, but it violates no syntactic rules or principles of English; indeed, it has
exactly the same syntactic structure as (1.1a). Hence it is grammatical (well-formed),
despite being semantically odd.
The organization of the book is as follows. In this chapter a number of distinctions
that are relevant to the discussion in the remainder of the book are introduced. First,
two aspects of syntactic structure are distinguished, one of which will be the main topic
of chapters 2 and 3, and the other will be the main topic of chapter 4. Second, the
traditional notion of parts of speech are reviewed, as these categories will be important
throughout the book. Finally, a brief introduction to some of the basic concepts of
morphology and morphological analysis is presented, with emphasis on those notions
that will be especially pertinent to the discussion in the succeeding chapters.
The next three chapters present basic syntactic phenomena from two different analytic perspectives and introduce the concepts and analytic tools used in each. Many of
the same grammatical phenomena will be analyzed from each perspective. In chapter 5
the basics of writing a grammar to describe syntactic phenomena will be presented; the
formulation of rules to express the generalizations arising from syntactic analysis and
the role of the lexicon in a grammar will be discussed. Different linguistic theories
make different sets of assumptions about the nature of syntactic structure and accordingly employ different analytic principles and tools. In chapter 6 the basic ideas of four
linguistic theories will be summarized, and their approaches to important grammatical
phenomena, including the formation of information questions (e.g. What did you see?)
and the passive voice (e.g. The bread was eaten by the mouse), will be compared and
contrasted. These two phenomena are especially revealing for a comparison of theories,
because the accounts given by the various theories highlight the conceptual and analytic
differences among them.
1.1
Aspects of syntactic structure
In the syntactic structure of sentences, two distinct yet interrelated aspects must
be distinguished. The first one has already been mentioned: the function of elements as
subject and direct object in a sentence. ‘Subject’ and ‘direct object’ have traditionally
been referred to as grammatical relations. Hence this kind of syntax will be referred
to as ‘relational structure’. It includes more than just grammatical relations like
subject and direct object; it also encompasses relationships like modifier–modified,
e.g. tall building or walk slowly (tall, slowly = modifier, building, walk = modified) and
possessor–possessed, e.g. Pat’s car (Pat’s = possessor, car = possessed). Relational
structure will be the primary focus of chapters 2 and 3.
The second aspect concerns the organization of the units which constitute sentences.
A sentence does not consist simply of a string of words; that is, in a sentence like The
teacher read a book in the library, it is not the case that each word is equally related
to the words adjacent to it in the string. There is no direct relationship between read
and a or between in and the; a is related to book, which it modifies, just as the is related
4
Syntax, lexical categories, and morphology
to library, which it modifies. A is related to read only through a book being the
direct object of read, and similarly, the is related to in only through the library being
the object of the preposition in. The words are organized into units which are then
organized into larger units. These units are called constituents, and the hierarchical
organization of the units in a sentence is called its constituent structure. This term
will be used to refer to this second aspect of syntactic structure. Consider the eight
words in the sentence The teacher read a book in the library. What units are these
words organized into? Intuitively, it seems clear that the article the or a goes with, or
forms a unit with, the noun following it. Is there any kind of evidence beyond a
native speaker’s intuitions that this is the case? Determining the constituent structure
of sentences is the major topic of chapter 4, but a brief preliminary look at the kind
of evidence needed follows.
If the article forms a unit with the noun that follows it, we would expect that in an
alternative form of the same sentence the two would have to be found together and
could not be split up. Thus in the passive version of this sentence, A book was read
by the teacher in the library, the unit a book serves as subject, and the unit the teacher
is the object of the preposition by. The constituent composed of a noun and an article
is called a noun phrase [NP]; as will be shown later, NPs can be very complex. The
preposition in and the NP following it also form a constituent in this sentence (in the
library); it is called a prepositional phrase [PP]. The fact that the PP is a constituent
can be seen by looking at another alternative form, In the library the teacher read
a book. Finally, the verb plus the NP following it form a unit as well, as shown by a
sentence like I expected to find someone reading the book, and reading the book was a
teacher. The constituent composed of a verb plus following NP is called a verb phrase
[VP]. As with NPs, VPs can be quite complex. In each of these alternative forms, a
combination of words from the original sentence which one might intuitively put
together in a single unit also occurs together as a unit, and this can be taken as evidence
that they are in fact constituents. Using square brackets to group the words in constituents together, the constituent structure of The teacher read a book in the library
may be represented as in (1.4). (‘S’ stands for ‘sentence’.)
(1.4)
[S [NP The [N teacher]] [VP [V read] [NP a [N book]] [PP [P in] [NP the [N library]] PP]
] S]
VP
Note the nesting of constituents within constituents in this sentence, e.g. the NP the
library is a constituent of the PP in the library which is a constituent of the VP read a
book in the library. In chapter 4 constituent structure will be explored in detail.
At the beginning of this section it was noted that the two aspects of syntactic
structure, relational structure and constituent structure, are ‘distinct yet interrelated’,
and it is possible now to see how this is the case. For example, a VP was described as
being composed of a verb and the following NP, but it could alternatively be characterized as involving the verb and its direct object. Similarly, a PP is composed of a
preposition and its object. NPs, on the other hand, involve modifiers, and accordingly
the relation between the and teacher could be described as one of modifier–modified.
Thus, these two aspects of syntactic structure are always present in a sentence, and
when one or the other is emphasized, the sentence is being described from one of the
two perspectives. It will be seen later that different grammatical phenomena seem to be
more easily analyzed from one perspective rather than the other.
5
An introduction to syntax
1.2
Lexical categories
In the discussion of the constituents of sentences, reference has been made
to nouns and noun phrases, verbs and verb phrases, and prepositions and prepositional
phrases. Nouns, verbs and prepositions are traditionally referred to as ‘parts of speech’
or ‘word classes’; in contemporary linguistics they are termed lexical categories. The
most important lexical categories are noun, verb, adjective, adverb and adposition,
which subsumes prepositions and postpositions. In traditional grammar, lexical
categories are given notional definitions, i.e. they are characterized in terms of their
semantic content. For example, noun is defined as ‘the name of a person, place or thing’,
verb is defined as an ‘action word’, and adjective is defined as ‘a word expressing
a property or attribute’. In modern linguistics, however, they are defined morphosyntactically in terms of their grammatical properties.
Nouns may be classified in a number of ways. There is a fundamental contrast
between nouns that refer uniquely to particular entities or individuals and those that do
not; the best example of the first kind of noun is a proper name, e.g. Sam, Elizabeth,
Paris or London, and nouns of this type are referred to as proper nouns. Nouns which
do not refer to unique individuals or entities are called common nouns, e.g. dog, table,
fish, car, pencil, water. One of the important differences between proper and common
nouns in a language like English is that common nouns normally take an article, while
proper nouns do not, e.g. The boy left versus *The Sam left (cf. *Boy left versus Sam
left). Common nouns may be divided into mass nouns and count nouns. Count nouns,
as the name implies, denote countable entities, e.g. seven chairs, six pencils, three dogs,
many cars. Mass nouns, on the other hand, are not readily countable in their primary
senses, e.g. *two waters, *four butters, *six snows. In order to make them countable, it
is necessary to add what is sometimes called a ‘measure word’, which delimits a specific
amount of the substance, e.g. two glasses/bottles/drops of water, four pats/sticks of butter,
six shovelfuls of snow. Measure words can be used with count nouns only when they are
plural, e.g. *six boxes of pencil versus six boxes of pencils, *two cups of peanut versus
three jars of peanuts. Pronouns are closely related to nouns, as they both function as
NPs. Pronouns are traditionally characterized as ‘substitutes’ for nouns or as ‘standing
for’ nouns, e.g. John went to the store, and he bought some milk, in which he substitutes
or stands for John in the second clause. This, however, is true only of third-person
pronouns like he, she, it, or they; it is not true of first-person pronouns like I or
second-person pronouns like you. First- and second-person pronouns refer to or index
the speaker and addressee in a speech event and do not replace or stand for a noun.
Verbs can likewise be categorized along a number of dimensions. One very important dimension which will be discussed in detail in chapters 2 and 3 is whether a verb
takes just a subject (an intransitive verb), or a subject and a direct object (a transitive
verb), or a subject, direct object and indirect object (a ditransitive verb). This will
be referred to as the ‘valence’ of the verb. Another dimension concerns the kind of
situation it represents. Some verbs represent static situations which do not involve
anyone actually doing anything, e.g. know as in Chris knows the answer, or see as in
Pat sees Dana over by the bookcase. Some symbolize actions, e.g. run as in Kim ran
around the track, or sing as in Leslie sang a beautiful aria. Others refer to a change
of state, e.g. freeze as in The water froze (the change in the state of the water is from
liquid to solid), or dry as in The clothes dried quickly (the change in the state of the
6
Syntax, lexical categories, and morphology
clothes is from wet to dry). Some represent complex situations involving an action plus
a change of state, e.g. break as in Larry broke the window with a rock (Larry does
something with a rock [action] which causes the window to break [change of state]).
This classification of verbs is quite complex and is more appropriately in the domain
of semantics rather than syntax. However, some syntactically relevant aspects of the
meaning of verbs will be investigated in chapter 2.
Some examples of adjectives in English include red, happy, tall, sick, interesting,
beautiful, and many others. Adjectives typically express properties of entities, e.g. a
red apple, a tall woman, a beautiful sunset. Some properties are inherent attributes of
an entity; for example, some apples are red because they are naturally so, whereas some
barns are red because they have been painted red, not because they are inherently red.
Hence color is an inherent property of apples but not of barns. Some languages signal
this distinction overtly. In Spanish, for example, the adjective feliz means ‘happy’, and
whether it is an inherent or permanent property of the person referred to is signaled by
the verb it is used with, i.e. Maria es feliz ‘Maria is happy (a happy person)’ versus
Maria está feliz ‘Maria is happy (now, at this moment but not necessarily always)’.
Spanish has two verbs meaning ‘be’, ser and estar, and one of the differences between
them is that ser plus adjective (es in this example) is used to signify inherent or
permanent attributes, while estar plus adjective (está in this example) serves to indicate
non-permanent, transitory attributes.
English adverbs typically, but not always, end in -ly, e.g. quickly, happily, beautifully, rapidly and carefully. Fast and friendly are exceptions; fast is an adverb without
-ly (it can also be an adjective), and friendly, despite the admonitions of road signs in
Texas to ‘drive friendly’, is an adjective, e.g. a friendly waiter. Adverbs modify verbs,
adjectives and even other adverbs, and they can be classified in terms of the nature of
this modification; manner adverbs, for example, indicate the manner in which something is done, e.g. The detective examined the crime scene carefully, or The ballerina
danced beautifully, while temporal adverbs, as the name implies, express when something happened, e.g. Kim talked to Chris yesterday, or Dana will see Pat tomorrow.
Yesterday and tomorrow do not end in -ly and have the same form when functioning as
an adverb that they have when functioning as a noun, e.g. Yesterday was a nice day,
Tomorrow will be very special. The most common adverbial modifiers of adjectives
and adverbs are words like very, extremely, rather, e.g. a very tall tree, the extremely
clever student, rather quickly. This class of adverbs is referred to as degree modifiers.
Prepositions are adpositions that occur before their object, while postpositions occur
after their object. English and Spanish have only prepositions, e.g. English in, on,
under, to, Spanish en, a, con, whereas Japanese and Korean have only postpositions.
German has both: in dem Haus ‘in the house’ (preposition in) versus dem Haus
gegenüber ‘over across from the house’ (postposition gegenüber).
There are a number of minor categories. The category of determiners includes
articles like a and the, and demonstratives like this and that. Determiners modify
nouns in relation to their referential properties. Articles indicate roughly whether the
speaker believes her interlocutor(s) can identify the referent of the NP or not; an
indefinite article like a(n) signals that the speaker does not assume the interlocutor(s)
can identify the referent of the NP, while a definite article like the indicates that the
speaker does assume that the interlocutor(s) can identify it. Demonstratives, on the
other hand, refer to entities in terms of their spatial proximity to the speaker; English
7
An introduction to syntax
this refers to an entity close to the speaker, while that refers to one farther away.
(Which book do you mean? This one here or that one over there? versus *This one over
there or that one here?) Many languages make a three-way distinction: close to the
speaker (English this, Spanish esta [FEM]), away from the speaker but not far (English
that, Spanish esa [FEM]), and farther away from the speaker (archaic English yon,
Spanish aquella [FEM]). These distinctions are also expressed by locative demonstratives, e.g. English here, German hier, Spanish aqui versus English there, German
da, Spanish ahí versus English yonder, German dort, Spanish allí. Quantifiers, as
the label implies, express quantity-related concepts. English quantifiers include every,
each, all, many, and few, as well as the numerals one, two, three, etc., e.g. every boy,
many books, the seven sisters. Classifiers serve to classify the nouns they modify in
terms of shape, material, function, social status and other properties. They are found in
many East and Southeast Asian and Mayan languages, among others. They are similar
in many respect to the measure words that occur with English mass nouns, but they
occur with all nouns regardless of the count–mass distinctions, e.g. Cantonese yat bei
séui [one cl water] ‘one cup of water versus yat ja séui ‘a jug of water’, versus yat
jbun séui ‘a bottle of water’ with a mass noun, nc ga dihnlóuh [this cl computer] ‘this
computer’ (classified as machine) versus nc bouh dihnlóuh ‘this computer’ (classified
as model) versus nc go dihnlóuh ‘this computer’ (classified as object) with a count
noun (Matthews and Yip 1994). Conjunctions, like and, but and or, serve to link the
elements in a conjoined expression. There are conjoined NPs, e.g. a boy and his dog,
conjoined verbs, e.g. Leslie danced and sang, and conjoined adjectives, e.g. Lisa is
tall and slender. All major lexical categories can be linked by conjunctions to form
conjoined expressions; this will be discussed in more detail in chapters 3 and 4.
Complementizers mark the dependent clause is a complex sentence, e.g. English that
as in Sally knows that Bill ate the last piece of pizza. The final category is particles,
which is a classification often given to elements which do not fall into any of the other
categories. Many particles have primarily discourse functions, e.g. English indeed,
German doch, Spanish entonces.
There is an important opposition that divides lexical categories into two general
classes, based on whether the membership of the class can readily be increased or not.
Languages can usually increase their stock of nouns, for example, by borrowing nouns
from other languages or creating new ones through compounding (e.g. black + board
yields blackboard) or other morphological means (e.g. rapid + -ly = rapidly), but they
do not normally create or borrow new adpositions, conjunctions or determiners. Lexical
categories such as noun and verb whose membership can be enlarged are termed open
class categories, whereas categories such as adposition, determiner or conjunction,
which have small, fixed membership, are called closed class categories.
The definitions of lexical categories given so far are primarily the notional ones from
traditional grammar. These definitions seem intuitively quite reasonable to speakers
of Indo-European languages, and they seem to correlate nicely with the syntactic
functions of the different parts of speech. Let us define three very general syntactic
functions: argument, modifier and predicate. In a sentence like the teacher read an
interesting book, the teacher and an interesting book are the arguments, read is the
predicate, and the, an and interesting are modifiers. Similarly, in Kim is tall, Kim is the
argument and is tall is the predicate. The term ‘argument’ here includes NPs and PPs
functioning as subject, direct object or indirect object. The notions of predicate and
8
Syntax, lexical categories, and morphology
argument will be discussed in more detail in the next chapter, but for now one can say
simply that in a sentence the predicate expresses the state of affairs that the referents of
the arguments are involved in. (The terms ‘predicate’ and ‘argument’ are also used in
semantics with a different meaning; they are being used here and elsewhere to refer to
syntactic notions, unless otherwise noted.) It is usual to distinguish 1-place, 2-place and
3-place predicates, depending on how many participants there are in the state of affairs
depicted by the predicate. Being sick is a state of affairs involving only one participant,
hence be sick is a 1-place predicate which takes one argument, e.g. Kim is sick. In the
teacher destroyed the note, there is an action of destroying involving a teacher and
a note. Destroying involves a destroyer and something destroyed; hence destroy is a
2-place predicate and takes two arguments. Finally, giving involves a giver, something
given and a recipient, and therefore give is a 3-place predicate and takes three arguments, e.g. The teacher gave an interesting book to Kim. Given these distinctions, it
seems intuitively clear that nouns would be arguments, verbs would be predicates and
adjectives would be modifiers, and this is in fact the case very often.
But not always. Nouns and adjectives can function as part of a predicate, as in
Dana is a phonologist and Chris was sick. Even though they are part of the predicate,
they are still formally distinct from verbs; they do not take tense suffixes like verbs
do, i.e. *Dana phonologists or *Chris sicked. The copula be, a kind of verb, carries
these verbal inflections. Contrast this with the situation in Lakhota, in which nouns and
adjective-like words do bear verbal inflections when functioning as predicates, in this
instance agreement in number with the subject.
(1.5)
1
a. WiAháCa ki hená lowT-pi.
man
the those sing-pl
‘Those men are singing.’1
a′. Lakhóta ki hená lowT-pi.
Sioux the those sing-pl
‘Those Siouxs (Indians) are singing.’
b. WiAháCa ki hená lakhóta-pi.
man
the those Sioux-pl
‘Those men are Siouxs (Indians).’
b′. Lakhóta ki hená wiAháCa-pi.
Sioux the those man-pl
‘Those Siouxs (Indians) are men.’
c. WiAháCa ki hená khúEa-pi.
man
the those sick-pl
‘Those men are sick.’
Lakhota
In most examples from languages other than English, there will be an interlinear gloss with a translation for
each meaningful element in the sentence directly under it as well as a free translation into English in the
third line. In the interlinear gloss, the translation will be lined up directly under the element being translated.
Complex words will be broken up into their meaningful parts (see section 1.3 below) separated by hyphens,
and the translation for each part will be joined to the translations for the other parts by hyphens and placed
below the whole word. Thus in (1.5a), for example, wiAháCa means ‘man’, ki means ‘the’ and hená means
‘those’; the last word, lowTpi, is broken up into two parts, lowT and pi, which are linked by a hyphen, and
each part is translated (lowT means ‘sing’ and pi means ‘plural subject’), with the translations linked by a
corresponding hyphen and placed below the Lakhota word. If an element requires a translation involving more
than one English word, the words will be joined by a ‘.’, e.g. ‘was.washed’ in (2.4a). Finally, grammatical
notions like tense and number are glossed using abbreviations which are listed at the beginning of the book.
9
An introduction to syntax
Nouns in Lakhota do not normally carry any indication of number; the only way to tell
that the NP containing wiAháCa ‘man’ is plural in (1.5a) is by means of the plural demonstrative hená ‘those’ (cf. hé ‘that’). In particular, the plural suffix -pi is impossible on
the noun wiAháCa in (1.5a); based on (1.5a), (1.5a′), one could conclude that it occurs
only on verbs. But this would be incorrect, as the sentences in (1.5b, c) show. Nouns
like wiAháCa ‘man’ do take -pi when they function as a predicate, rather than as an
argument. Hence nouns in Lakhota seem to function readily as predicates, something
their English counterparts do not do. Adjective-like words also function directly as
predicates, as (c) illustrates; there is no copular element analogous to English be in
either of these sentence types. Verbs and adjective-like words can also serve as arguments
in Lakhota, as in (1.6).
(1.6)
a. HokCíla ki hená Ahéya-pi.
boy
the those cry-pl
‘Those boys are crying.’
b. Bhéya ki hená hokCíla-pi.
cry
the those boy-pl
‘The ones crying are boys.’
c. KhúEa ki hená wiAháCa-pi.
sick the those man-pl
‘The sick ones are men.’
Lakhota
(cf. (1.5c))
The verb Ahéya ‘cry’ serves as the predicate in (1.6a) and the argument in (1.6b). Note
that in the English translation the verb cry cannot simply function as the subject; it
must, rather, occur in a complex expression the ones crying. In Lakhota, by contrast,
hokCíla ‘boy’ and Ahéya ‘cry’ simply exchange positions in the sentence without any
formal modification. The same is true of the noun wiAháCa ‘man’ and the adjectivelike word khúEa ‘sick’ in (1.6c). Thus, the expected correlations between noun and
argument, verb and predicate and adjective and modifier are not as strong in Lakhota as
they are in English.
An even more striking example of this lack of correlation between lexical class and
syntactic function can be seen in Nootka, a Wakashan language spoken on Vancouver
Island in British Columbia, Canada (Swadesh 1939).
(1.7)
10
a. Wa|a:k-ma qo:=as-=i.
go-3sgpres man-the
‘The man is going.’
a′. Qo:=as-ma wa|a:k-=i.
man-3sgpres go-the
‘The one going is a man.’
a″. Qo:=as-ma.
man-3sgpres
‘He is a man.’
b. =i:t-ma
qo:=as-=i.
large-3sgpres man-the
‘The man is large.’
b′. Qo:=as-ma =i:t-=i.
man-3sgpres big-the
‘The large one is a man.’
Nootka
Syntax, lexical categories, and morphology
c. CiqC9-ma
=o:kwi| qo:=as-=i.
speak-3sgpres to
man-the
‘He speaks to the man.’
c′. =i:t-ma
ciqC9-=i.
large-3sgpres speak-the
‘The one speaking is large.’
c″. =o:kwi|-ma qo:=as-=i.
to-3sgpres man-the
‘He is [in relation] to the man.’
d. Wa|a:k-ma =attiya.
go-3sgpres night
‘He is going at night.’
d′. =attiya-ma wa|a:k-=i.
night-3sgpres go-the
‘His going is at night.’
The basic pattern in these Nootka sentences is predicate-ma argument-=i, with -ma
signalling both tense and subject agreement in terms of person and number. The
sentences in (1.7a), (1.7a′) and (1.7b), (1.7b′) look like their Lakhota counterparts in
(1.5) and (1.6). The striking examples are in (1.7c) and (1.7d). (1.7c) contains a
preposition, =o:kwi| ‘to’, and in (1.7c″) it functions as the predicate, as indicated by its
occurring first in the sentence with the suffix -ma. In (1.7d) there is an adverb, =attiya
‘at night’, and in (1.7d′) it is the predicate. Thus in Nootka, the expected correlations
between lexical category and syntactic function appear to be even weaker than in
Lakhota.
This has important implications for the traditional view of lexical categories. This
view assumes that the semantics of words predict their category and hence their
function. The examples from Lakhota and Nootka call this seriously into question,
since functioning syntactically as predicate, argument or modifier does not follow
from the meaning of words as expected in many instances. Moreover, the first link
in the chain of inference, from meaning to category, does not even hold up from a
cross-linguistic perspective. The notional account assumes, at least implicitly, that the
major lexical categories are universal, but this turns out not to be true for all of them.
Every language has noun and verb as lexical categories, even Nootka. This reflects the
fundamental role of reference and predication in communication. One of the most
important functions of language is to allow speakers to depict states of affairs in the
world, and in order for them to do this, there must be linguistic devices which refer to
the participant(s) in a state of affairs and other devices which denote the action, event
or situation in a state of affairs. Lexical items specialized for the first task are nouns,
those specialized for the second are verbs. Even though in examples from Lakhota and
Nootka verbs function as arguments (and hence as referring expressions) and nouns
as predicates, it is nevertheless the case that the basic use of words like wiAháCa ‘man’
in Lakhota and qo:=as ‘man’ in Nootka is as an argument; similarly, the basic use of
words like Ahéya ‘cry’ in Lakhota and wa|a:k ‘go’ in Nootka is as a predicate.
What about the other major lexical categories? There are languages which lack
adpositions altogether; they express the semantic content of prepositions and postpositions by means of the kind of suffixes on nouns in the Russian examples in (1.2)
11
An introduction to syntax
and (1.3). The concepts expressed by these endings are called ‘case’, and the endings
are called ‘case markers’. Case will be discussed a great deal throughout this book.
Russian has both case suffixes and prepositions, but Dyirbal, an Australian Aboriginal
language (Dixon 1972), has only case suffixes and no adpositions at all. Hence the
lexical category ‘adposition’ is not universal. It also appears that adjective is not
universal. In Lakhota, for example, the words expressing properties like ‘red’, ‘tall’,
‘big’, etc., are formally verbs and have basically the same morphosyntactic properties
as verbs, as the examples in (1.6) showed in part. Hence there is no reason to posit a
category ‘adjective’ distinct from that of ‘verb’; the words corresponding to adjectives
in a language like English, e.g. khúEa ‘sick’ in (1.5) and (1.6), are really a subtype
of verb in Lakhota. In Dyirbal and Quechua, spoken in the northern Andes mountains
in South America (Wölck 1987), on the other hand, words of this type have the same
morphosyntactic properties as nouns, and therefore they should be analyzed as a
subtype of nouns. Finally, there has been much less research done on adverbs crosslinguistically than the other major categories, and therefore it is difficult to draw any
conclusions about their universality.
Thus, it appears that noun and verb are universal lexical categories, but adposition
and adjective are not. It is crucial to keep in mind that when it is claimed that adjective
is not a universally valid lexical category, it does not mean that there are languages
which lack words expressing properties like ‘red’, ‘big’, ‘happy’, etc. Rather, it means
that the words expressing these notions behave morphosyntactically like members of
one of the other classes (verb in Lakhota, noun in Dyirbal and Quechua).
In modern linguistics, the determination of the category of a word is not based on its
meaning but rather on its morphosyntactic behavior, i.e. the elements it cooccurs with
and the morphosyntactic environment(s) it occurs in. Meaning is not irrelevant to the
function of a word, but it does not reliably predict it either. The term which is used to
refer to classes based on their morphosyntactic properties is form class. Consider the
similarities and differences between common and proper nouns in English, which was
initially characterized semantically. They are both a type of noun, because they both
occur in the major morphosyntactic environments which nouns (and NPs) occur in, e.g.
as the subject or direct object of a verb, as the object of a preposition in a PP, and with
be as a predicate nominal (The girl gave a book to the teacher, Pat introduced Kim to
Dana; Max is my lawyer, My lawyer is Max). Other form classes cannot occur in these
positions, e.g. *The yellow put a clumsily on the receive. However, they differ in that
common nouns can be modified by determiners and adjectives, while proper nouns
cannot, e.g. a tall girl versus *a tall Dana. Furthermore, common nouns, if they are
count nouns, can take plural inflection, while proper nouns cannot, e.g. the tall girls
versus *Danas. Thus there are both syntactic and morphological differences between
common and proper nouns which can be used to distinguish them as belonging to two
distinct subclasses of the category noun.
English verbs can be differentiated from the other major classes by both morphological and syntactic criteria. Morphologically, only verbs take the suffixes -ing ‘progressive’, -ed ‘past tense’, or ‘past participle’, -s ‘third-person singular subject–present
tense’ and -en ‘past participle’. Syntactically, they occupy a unique position in a clause,
and they may be modified by adverbs but not by adjectives or demonstratives. There
are no consistent morphological properties that characterize English adjectives; there
are distinctive endings that some adjectives carry, e.g. -y as in slimy (related to the noun
12
Syntax, lexical categories, and morphology
slime) or tricky (related to the noun trick), and -ic as in toxic (related to the noun toxin)
or metric (related to the noun meter). Many adjectives take -er for their comparative
forms, e.g. taller, faster, and -est for their superlative forms, e.g. tallest, fastest.
However, many do not, e.g. *beautifuler, *beautifulest; these adjectives take more and
most to indicate their comparative (more beautiful) and superlative (most beautiful)
forms. English adjectives occupy a specific position within NPs, i.e. DEM-QNT-ADJN, as in the seven tall trees (*tall the seven trees, *the tall seven trees), and they may
function predicatively only in combination with the copula be, e.g. The tree is tall,
*The tree talls). Finally, English adverbs, as noted earlier, often (but not always) end
in -ly; they function only as modifiers (but never of nouns), e.g. the extremely quick
rabbit, the rabbit ran very quickly, *the quickly rabbit, and never as predicates, e.g.
*The rabbit is quickly.
This brief discussion of the morphosyntactic properties of the major English classes
has not been exhaustive, but it does illustrate how morphological and syntactic criteria
can be used to characterize the form classes in a language. Even though the criteria for
the classes are ultimately morphosyntactic, the labels for the classes reflect the traditional notional distinctions. That is, after having established the existence of a form
class based on the morphosyntactic properties of its members, the semantic properties
of the prototypical members of the class determine the name of the class. Hence if the
prototypical members of a class include elements that function as the name of a person,
place or thing, then the class will be given the label ‘noun’.
1.3
Morphology
Even though this book is about syntax and syntactic analysis, it is not possible
to get very far without some basic knowledge about morphology. It was already shown
in section 1.0 that some languages use morphology to express what other languages
express syntactically, and in the previous section it was noted that the inflectional properties of words are relevant to determining their category. In this section some basic
concepts of morphology and the basic techniques of morphological analysis will be
introduced; both will play a role in the syntactic analyses in later chapters. The discussion will be limited to those aspects of morphology which are relevant to syntactic
analysis; this is not intended to be a general introduction to this complex and important
part of linguistics.
Morphology is concerned with the structure of words, and morphological analysis is
the process by which linguists break complex words down into their component parts.
Consider the Lakhota word wahi, which means ‘I arrive’. Just looking at it by itself, it
is not possible to determine whether it is a simple or complex form. If it is compared
with another form, yahi ‘you [singular] arrive’, it can be seen immediately that there is
a common part to each of the words, -i, -hi or -ahi, and a different part in each, wah-,
wa- or w- and yah-, ya- or y-. It is clear that these are complex forms made up of more
than one component. There are two parts to each: one meaning ‘arrive’, which is -i,
-hi or -ahi, and the other meaning first-person singular (‘I’) or second-person singular
(‘you’) subject. It is not possible to tell from just these two forms, however, exactly
what the two components are, since there are at least three ways to divide up these
forms. Is the form for ‘arrive’ -i, hi or ahi? Is the form for first-person singular subject
wah-, wa- or w-? The answer to this question becomes somewhat clearer when the form
13
An introduction to syntax
Rhi ‘we [dual inclusive] arrive’ (i.e. ‘you [sg] and I arrive’) is examined. The common
parts to the three forms are -i or hi, and this would seem to eliminate the -ahi possibility.
But we are still left with two possibilities for ‘arrive’ and for ‘I’ (wah- or wa-), ‘you’
(yah- or ya-) and ‘we [dl incl]’ (Rh- or R-). In order to resolve the issue, it would useful
to look at some other verbs, as in (1.8).
(1.8)
a.
a′.
b.
b′.
c.
c′.
d.
d′.
e.
e′.
walowN
yalowN
waAhP
yaAhP
wa=u
ya=u
nawaEP
nayaEi
awaphe
ayaphe
‘I sing’
‘you [sg] sing’
‘I want’
‘you [sg] want’
‘I come’
‘you [sg] come’
‘I stand’
‘you [sg] stand’
‘I wait’
‘you [sg] wait’
Lakhota
In the first example in each pair the only common semantic element is first-person
singular subject, and the only form common to all of them is wa-. Similarly, in the
primed examples in the set the only common semantic element is second-person singular
subject, and the only form common to all of them is ya-. Therefore wa- must mean ‘I’
and ya- must mean ‘you [singular]’. If ‘I’ and ‘you’ are wa- and ya-, respectively, then
‘arrive’ must be -hi, not -i. Moreover, this means that ‘we [dual inclusive]’ must be
R-, not Rh-. Thus, the correct analysis of the first three forms is wa- ‘I’, ya- ‘you [sg]’,
R- ‘we [dl incl]’ and -hi ‘arrive’. It may further be concluded that lowN means ‘sing’,
AhP means ‘want’, =u means ‘come’, naEP means ‘stand’ and aphe means ‘wait’. Each
of these words wahi, yahi and Rhi, as well as those in (1.8), is composed of two
meaningful parts, which are called morphemes. Morphemes are the smallest meaningful units in language. Hi cannot be broken down into h + i, nor can wa- or ya- be broken
down into w- or y- + a; these smaller forms have no meaning. Hence, wa-, ya- and hi,
as well as R-, are all morphemes in Lakhota, as are lowN, AhP, =u, naEP and aphe.
The part of the complex form to which a morpheme is added is called the stem.
In these examples, the stems are hi, lowN, AhP, =u, naEP, and aphe. Morphemes like
wa- and ya- which occur before the stem are called prefixes. Morphemes which occur
after the stem are called suffixes; examples of suffixes can be found in the Russian
sentences in (1.2) and (1.3), i.e. the markers -a and -u, and in the Lakhota examples
in (1.5) and (1.6), i.e. -pi ‘plural’. In (1.8d), (1.8e) wa- occurs within the stem itself; in
these forms it is an infix. The general term which covers prefixes, infixes and suffixes
is affix.
This example has been very simple, but it illustrates the basic principle used in
breaking words down into their component morphemes: look for recurring forms that
correlate with consistent meanings. In all of the Lakhota words examined above, the
form wa- correlates with the meaning ‘first-person singular subject’, the form yacorrelates with the meaning ‘second-person singular subject’, and the form correlates
R- with the meaning ‘first dual inclusive subject’. Similarly, in comparing (1.8a) and
(1.8a′), the form lowN consistently correlates with the meaning ‘sing’.
The basic principle of morphological analysis stated above (‘look for recurring forms
that correlate with consistent meanings’), requires four very important qualifications.
14
Syntax, lexical categories, and morphology
The first is that a single meaning, e.g. plural, may be expressed by several different
forms. Consider the following simple example from English.
(1.9)
a. dog /d7g/
b. cat /kæt/
c. rose /roz/
a′. dogs /d7gz/
b′. cats /kæts/
c′. roses /roz8z/
The forms in the second column have a consistent semantic difference from the forms
in the first column, but there does not appear to be a consistent formal distinction
correlating with the semantic difference. Rather, there is /z/, /s/ and /8z/. However, if
one looked at a large number of English nouns, one would find that nouns ending in a
voiced sound take /z/ to indicate plural, that nouns ending in a voiceless, non-sibilant
sound take /s/, and that nouns ending in a sibilant take /8z/. These forms are clearly
related phonologically, and the form of the plural suffix is predictable from the phonological shape of the end of the word to which it is added. Hence these three forms
may be viewed as conditioned variants of a single morpheme; they are referred to as
its allomorphs. They are phonologically conditioned allomorphs, since the choice
of allomorph is determined by the phonological shape of the stem. There are other
allomorphs of the plural morpheme which are not phonologically conditioned. For
example, the plural of box is boxes, but the plural of ox is oxen, not *oxes. The plural
of ox is said to be morphologically conditioned, because it is not phonologically
predictable and is an idiosyncratic property of the word ox. Other examples of morphologically conditioned plurals in English include mice for mouse and teeth for tooth.
Words like mice and teeth illustrate the second qualification: meanings need not
be represented by segmentable parts of words. With words like dogs (/d7gz/) and cats
(/kæts/), it is easy to break them up into two parts, one meaning ‘dog’ or ‘cat’ and the
other meaning ‘plural’. But it is not obvious that mice (/mays/) and teeth (/ti+/) can be
broken down into two comparable parts, one meaning ‘mouse’ or ‘tooth’ and the other
meaning ‘plural’. Forms like these have posed profound problems for morphological
analysts and theorists alike. For the purposes of this book, it is enough to state that
such a form is morphologically conditioned but unsegmentable (cf. oxen, which is
morphologically conditioned but segmentable) and to state the meanings expressed by
the form.
When there is a group of phonologically conditioned allomorphs like those of the
English plural in (1.9), it is customary to select one of them as the basic allomorph
to represent the morpheme. Typically, the allomorph occurring in the widest range
of environments is taken as basic, although other factors may come into play. With
respect to the three allomorphs in (1.9), /z/ occurs in the greatest number of environments, and therefore it is a good candidate for the basic allomorph.
As noted in the initial discussion in this section, there is a choice as to how to divide
up a complex form. With respect to Lakhota wahi ‘I arrive’ and yahi ‘you [sg] arrive’,
there were three initial hypotheses: w- ‘I’, y- ‘you’ and -ahi ‘arrive’, versus wa- ‘I’, ya‘you’ and -hi ‘arrive’, versus wah- ‘I’, yah- ‘you’ and -i ‘arrive’. Given just these two
forms, there is no reason to choose one analysis over the others, but when the form Rhi
‘we [dl incl] arrive’ is considered, the possibilities are reduced to two, -hi or -i, etc.
Consideration of the data in (1.8) leads to the conclusion that the simplest analysis of
all three forms is wa- ‘I’, ya- ‘you’, R- ‘we [dl incl]’ and -hi ‘arrive’. Why is this the
simplest analysis? Because it avoids positing allomorphs for any of the morphemes.
15
An introduction to syntax
If the analysis of ‘arrive’ as being -i were maintained, then one would have to claim
that wa- ‘I’ has an allomorph wah- before the verb -i. While such an alternation is
not impossible, positing it nevertheless results in a more complex account than the
alternative analysis, which does not postulate any allomorphic variation for any of the
morphemes. Hence, all things being equal, the simplest analysis is to be preferred,
and one criterion for simplicity is positing the least amount of allomorphic variation
compatible with the facts.
Not only can a single meaning be expressed by multiple forms, but a single form
can express multiple meanings. This is the third qualification to the basic principle of
morphological analysis. For example, in Russian the -a suffix on uAitel’nica ‘teacher’
in (1.2) expresses three distinct concepts: nominative case, singular number, and
feminine gender. The -s suffix on verbs in English likewise expresses three concepts:
third-person subject, singular subject, present tense. Thus, even though there are simple
instances in which a single form consistently pairs with a single meaning, the kinds of
complexities involving allomorphic variation and multiple concepts in a single form
are very common.
The fourth qualification is that structural patterns in a language may require the
analyst to posit that a meaning is expressed by the absence of a form. Consider the
following paradigm for the verb hi ‘arrive’ in Lakhota.
(1.10)
a.
b.
c.
d.
e.
f.
g.
wahi
yahi
hi
Rhi
Rhipi
yahipi
hipi
‘I arrive’
‘you [sg] arrive’
‘he/she arrives’
‘we [dl incl] arrive’
‘we [pl] arrive’
‘you [pl] arrive’
‘they arrive’
Lakhota
It has already been established that wa- signals first-person singular subject, yasecond-person subject and R- first-person dual inclusive subject, and in (1.5) and (1.6)
it was shown that -pi indicates plural, which is the case here as well. What, then, signals
third person singular subject? The absence of -pi in (1.10a)–(1.10d) indicates that
the subject is non-plural, and it is the absence of wa-, ya- and R- that signals that the
forms in (1.10c) and (1.10g) are third person. Thus, third person is marked by the
absence of a prefix, and it is customary to represent this by ‘·’. Hence the form for
‘he/she arrives’ would be ·-hi. This meaningful absence of a phonological form is
called a zero morpheme. Zero morphemes are normally posited only within paradigms such as that in (1.10) in which one form is distinctive by virtue of the absence
of an affix.
Morphemes may be divided into two general classes: lexical morphemes, which
have substantive semantic content, e.g. English dog, rose, cat, or Lakhota hi ‘arrive’,
and grammatical morphemes, which lack substantive semantic content and express
grammatical notions like person, number, gender, tense or case, e.g. English -s on verbs,
Lakhota wa-, or Russian -a. The lexical versus grammatical opposition correlates with
the earlier distinction made between open and closed classes: typically, lexical morphemes are open-class items, while grammatical morphemes are closed-class items. This
opposition also relates to another important contrast, that between free morphemes and
bound morphemes. Free morphemes are elements that can stand alone as independent
16
Syntax, lexical categories, and morphology
words, e.g. dog, car, to, the and or, whereas bound morphemes cannot occur by themselves as independent words, e.g. -s, -ing, -ed in English. Bound morphemes are usually
grammatical morphemes, while free morphemes may be both lexical and grammatical,
as the above examples from English illustrate. There are languages in which lexical
morphemes can be considered bound morphemes, in that they cannot occur without
an accompanying grammatical morpheme. In Russian, for example, vide- means ‘see’,
which makes it a lexical morpheme, but it is not a free morpheme, since it cannot
occur as a complete word without the addition of a suffix indicating its tense and
subject agreement, e.g. -l-a ‘past-femsg’, yielding videla ‘saw’. These distinctions are
summarized in Table 1.1.
Table 1.1. Types of morphemes
Lexical
Grammatical
Free
Bound
dog, sing; Lakhota hokCíla, Ahéya
the, a; Lakhota ki, hená
Russian vide-s, -ing, -ed; Lakhota wa-, -pi
Since the focus of this book is on syntax, primary concern will be given to grammatical morphemes which express syntactic notions, e.g. subject, or syntactically relevant
notions, e.g. person agreement on a verb.
In the next chapter, the notion of subject and grammatical relations in general will
be examined.
Notes and suggested readings
For an excellent overview of systems of lexical categories across languages, see Schachter
(1985). Hopper and Thompson (1984), Langacker (1991) and Croft (1991) argue for the universality of noun and verb as lexical categories based on the fundamental role of reference and
predication in language. Jacobsen (1979) takes a detailed look at Nootkan languages and argues
that there is evidence in favor of postulating noun and verb as lexical categories in these
languages. Dixon (1977a) investigates the category ‘adjective’ cross-linguistically.
An excellent introduction to morphological analysis is Nida (1946); two more recent texts,
which include discussion of morphological theory, are Bauer (1988) and Spencer (1991).
Exercises
1. Pretend the italicized nonsense words in the following sentences are real words of English.
Identify the form class of each one, and state the morphosyntactic properties of each that lead
you to assign it to a particular category. [section 1.2]
(1) a. The dog wugged the ball.
b. The dog is wugging the ball.
c. The dog likes to wug the ball.
d. The dog gently wugged the ball.
e. *The wug kicked the ball.
f. *The dog chased the wug cat.
17
An introduction to syntax
(2)
(3)
(4)
(5)
(6)
(7)
a.
b.
c.
d.
e.
f.
a.
b.
c.
d.
a.
b.
c.
d.
e.
f.
a.
b.
c.
d.
e.
f.
a.
b.
c.
d.
e.
f.
a.
b.
c.
d.
e.
f.
g.
The tall blick sat by the river.
The blicks played in the park.
Mary sent a present to her favorite blick.
Sam is not a blick.
*Max blicked the cat.
*The blick animal ran away.
A nork person walked by the car.
Mary is very nork.
*Sam norks.
*The nork called me yesterday.
Li cat slept by the fire.
I bought li three interesting books.
Mary didn’t like li one.
I don’t care for li.
*Two li dogs barked at the cat.
*Sam lis every day.
Max walked blishly down the corridor.
Max walked down the corridor blishly.
Blishly, Max walked down the corridor.
Sam did so extremely blishly.
*Pat is blishly.
*The blishly woman looked unhappy.
Larry placed the book za the table.
Za the table Sam found his glasses.
*Za green book fell on the floor.
*I don’t like za.
*Sam zas every day.
*Sam found his gloves za.
Anna bought nace rare books.
I liked nace of them.
Nace left the party early.
I thought she bought too nace.
*Anna bought rare nace books.
*Sam naces every morning.
*The tall red nace fell off the shelf.
2. In the following verse from Lewis Carroll’s famous poem Jabberwocky, identify the form
class of each of the italicized words. State the morphosyntactic properties that lead you to
assign it to a particular category. Give two different analyses of the words in the last line.
[section 1.2]
’Twas brillig, and the slithy toves
Did gyre and gimble in the wabe;
All mimsy were the borogroves,
And the mome raths outgrabe.
3. Consider the following examples from English:
(1) The break is just above the knee.
(2) Please maple syrup your pancakes from the dispenser on the table. (sign in a cafe in
Adelaide, South Australia)
(3) The climb up the north face is very difficult and dangerous.
(4) Kinko’s, the new way to office. (advertisement)
18
Syntax, lexical categories, and morphology
(5) The referee booked the player for the foul.
(6) A good cry often makes you feel better.
What are the implications of examples like these for the discussion of English lexical categories
in section 1.2? Compare these with the examples from Nootka and Lakhota; is English as
flexible in terms of the functions of its lexical categories as these languages? [section 1.2]
4. In the following Italian sentences, what formal properties do the nouns, articles, adjectives,
adverbs and verbs have? Do they have any special morphological marking indicating their
category? Do they have any specific or special syntactic properties? With respect to the latter,
assume simple declarative utterances and ignore elliptical answers to questions, e.g. Q: How
does Fred run? A: Slowly, and vocative expressions, e.g. ‘Hey man!’. [section 1.2]
(1)
La bella ragazza parla rapidamente.
‘The pretty girl speaks rapidly.’
(2)
Il giovane ragazzo parla chiaramente.
‘The young boy speaks clearly.’
(3)
Le belle ragazze parlano chiaramente.
‘The pretty girls speak clearly.’
(4)
I giovani ragazzi parlano rapidamente.
‘The young boys speak rapidly.’
(5)
Parlano.
‘They speak.’
(6) *I parlano.
(7)
Parlano rapidamente.
‘They speak rapidly.’
(8) *Bella ragazza parla.
(9)
Una bella ragazza parla.
‘A pretty girl speaks.’
(10) *Bella parla.
(11) La macchina rapida arriva.
‘The fast car arrives.’
(12) Le macchine rapide arrivano.
‘The fast cars arrive.’
(13) Il ragazzo povero arriva.
‘The poor boy arrives.’
(14) Chiaramente parla.
‘He/she speaks clearly.’
(15) La bella arriva.
‘The pretty one arrives.’
(16) Il giovane parla.
‘The young one speaks.’
(17) *Bella.
(18) *I.
(19) *Chiaramente.
(20) Parla.
‘He/she speaks.’
(21) Arrivano.
‘They arrive.’
(22) *Ragazzo.
(23) *Macchina.
5. Break down the following words from Ayacucho Quechua (Parker 1969) into their constituent morphemes and state their meaning. [section 1.3]
(1) Rimani.
‘I speak.’
(2) Warmita rikun.
‘He/she sees the woman.’
(3) Runtuta mikurqani.
‘I ate an egg.’
(4) Runa rimarqa.
‘The man spoke.’
(5) Warmi rikun.
‘The woman sees him/her.’
(6) Runakuna rimarqaku.
‘The men spoke.’
(7) Runata rikuni.
‘I see the man.’
(8) Warmikuna rimanku.
‘The women speak.’
(9) Runtuta mikuni.
‘I eat an egg.’
(10) Runa hatun.
‘The man is big.’
(11) Warmikunata rikurqani.
‘I saw the women.’
(12) Runa daliwan.
‘The man hits me.’
(13) Rimarqani.
‘I spoke.’
(14) Runtu hatun.
‘The egg is big.’
19
An introduction to syntax
(15)
(16)
(17)
(18)
(19)
(20)
Warmi rikurqawa.
Rikurqa.
Riman.
Warmikuna rikurqawaku.
Warmi runtuta dalirqa.
Mikurqaku.
‘The woman saw me.’
‘He/she saw him/her.’
‘He/she speaks.’
‘The women saw me.’
‘The woman hit an egg.’
‘They ate.’
What would the following forms mean:
(21) Mikun.
(22) Warmi runata rikurqa.
(23) Runata dalirqani.
(24) Runa rikurqawa.
How
(25)
(26)
(27)
(28)
would you say the following in Ayacucho Quechua?
The men saw an egg.
The woman ate an egg.
The women hit [PAST] the man.
The men see the woman.
6. Break down the words in the following sentences from Hungarian (de Groot 1989) into their
constituent morphemes and state their meaning. If a morpheme has more than one allomorph,
give all of them. [section 1.3]
(1) János a könyvet Marinak adta.
‘John gave the book to Mary.’
(2) A könyv az asztal alatt van.
‘The book is under the table.’
(3) Mari Jánosnak adta az órát.
‘Mary gave the clock to John.’
(4) A gyerek látta Marit.
‘The child saw Mary.’
(5) Az óra az asztal fölött van.
‘The clock is above the table.’
(6) Mari látja Jánost.
‘Mary sees John.’
(7) A gyerek tette az órát az asztalhoz.
‘The child put the clock near the table.’
(8) Mari Jánosnak adja a könyvet.
‘Mary gives the book to John.’
(9) A level az asztal alatt van.
‘The letter is under the table.’
(10) János a könyvet az asztalra tette.
‘John put the book on the table.’
(11) Mari látta a gyereket.
‘Mary saw the child.’
(12) A könyv van az asztal mögött.
‘The book is behind the table.’
(13) Mari írja a levelet.
‘Mary writes the letter.’
(14) Mari a gyereknek adta az órát.
‘Mary gave the clock to the child.’
(15) János írta a levelet.
‘John wrote the letter.’
20