HÉCTOR CAMPOS BÙI HUỲNH THỦY THƯƠNG

ENGLISH SYNTAX
AND
UNIVERSAL GRAMMAR
An introduction to Syntax Analysis

VNU-HCM PRESS - 2018

i
ii
 CONTENTS

FOREWORD…………………………………………………………………………………………………………………………… .vii
PROLOGUE ...................................................................................................................................viii
CHAPTER 1: LINGUISTICS AS A SCIENCE OF THE MIND .......................................................... 1
1.1 What is linguistics? ............................................................................................................. 1
1.2 What are the different areas of Linguistics and where is Syntax in the big picture? ........ 4
1.2.1 Theoretical Linguistics or Core Linguistics ................................................................. 4
1.2.2 Applied Linguistics...................................................................................................... 5
1.3 Where is Linguistics in relation to other sciences? ............................................................ 6
1.4 An example of the scientific method in Linguistics ............................................................ 6
1.5 What do we know about syntax before studying this book? .....................................................8
1.6 The logical problem of language acquisition.................................................................... 12
1.7. Language Acquisition and Universal Grammar ............................................................... 14
CHAPTER 2: OUR MENTAL LEXICON ................................................................................... 18
2.1 The structure of the Catalog ............................................................................................ 18
2.1.1 Categorization .......................................................................................................... 18
2.1.2 C-selection ................................................................................................................ 21
2.1.3 S-selection ................................................................................................................ 23
2.1.4 Semantic information .............................................................................................. 26
2.1.5 Phonological information......................................................................................... 27
2.1.6 Special properties..................................................................................................... 27
2.1.7 Lexical entries for other categories.......................................................................... 29
2.1.8 Summary of the Catalog........................................................................................... 31
2.2 The Extended Catalog....................................................................................................... 31
2.2.1 Free morphemes ...................................................................................................... 32
2.2.2 Bound Morphemes and Derivational Morphology .................................................. 32
2.2.3. Bound roots.............................................................................................................. 38
2.2.4 Compounds. ............................................................................................................. 40
2.3 Summary .......................................................................................................................... 42
CHAPTER 3: THE PRINCIPLES AND PARAMETERS MODEL ................................................... 44
3.1 The modular system of the Principles and Parameters model ........................................ 44
3.2 Introduction to X'-syntax: The basic architecture of syntactic structures ....................... 45
3.3 From the Lexicon and X'-syntax to D-Structure and S-Structure ..................................... 48
iii
 3.4 Constituents ..................................................................................................................... 49
3.5 Summary .......................................................................................................................... 51
CHAPTER 4: THE VERB PHRASE ......................................................................................... 54
4.1 Complements of the verb phrase..................................................................................... 54
4.1.1 Intransitive verbs...................................................................................................... 54
4.1.2 Transitive verbs ........................................................................................................ 54
4.1.3 Prepositional verbs .................................................................................................. 55
4.1.4 Verbs with NP-PP complements .............................................................................. 55
4.1.5 Ditransitive verbs ..................................................................................................... 56
4.2. Modifiers in the verb phrase ........................................................................................... 63
4.2.1 Postverbal modifiers. ............................................................................................... 63
4.2.2 Preverbal modifiers. ................................................................................................. 66
4.2.3 An important note on NPs that function as adverbs. ................................................ 68
4.3 Phrasal versus Prepositional verbs.................................................................................... 68
CHAPTER 5: THE STRUCTURE OF THE NOUN PHRASE ......................................................... 74
5.1 The structure of noun phrases ......................................................................................... 74
5.2 Determiners as specifiers of NP. ...................................................................................... 74
5.3 NP modifiers ..................................................................................................................... 75
5.4 Complements of NP: PP-complements ............................................................................ 77
5.5 NP complements: non-PP complements. ........................................................................ 78
5.5.1 Prenominal complements ........................................................................................ 78
5.5.2 Postnominal Of-complements ................................................................................. 80
5.6 PP-modifiers ..................................................................................................................... 83
5.7 Explaining some ambiguous noun phrases ...................................................................... 84
5.8 A note on quantifiers........................................................................................................ 86
5.9 Partitives........................................................................................................................... 87
5.10 Non-deverbal nouns with complements........................................................................ 88
5.11 Summary ........................................................................................................................ 90
APPENDIX: REVISITING INDIRECT OBJECTS .............................................................................. 92
CHAPTER 6: THE STRUCTURE OF AP, PP AND ADVP ............................................................ 97
6.1 The structure of AP........................................................................................................... 97
6.2 The structure of PP ......................................................................................................... 101
6.3 The structure of ADVP .................................................................................................... 103
6.4 Summary ........................................................................................................................ 106
CHAPTER 7: MAKING SENTENCES : THE STRUCTURE OF IP .................................................107
7.1 Subject and Predicate..................................................................................................... 107
7.2 Diagramming sentences ................................................................................................. 109
7.3 Putting the verb and inflection together ....................................................................... 112

iv
 7.4 Incorporating Modal Verbs into IP. ................................................................................ 115
7.4.1 Differences between modal and main verbs ......................................................... 115
7.4.2 A syntactic representation for modal verbs........................................................... 117
7.4.3 A note on semi-modal verbs .................................................................................. 118
7.5 Auxiliary Verbs in English ............................................................................................... 120
7.6 Summary ........................................................................................................................ 126
CHAPTER 8: FROM NP TO DP ............................................................................................127
8.1 Some basic properties of X'-syntax model ..................................................................... 127
8.2 Revisiting our structure for NP ....................................................................................... 129
8.3 Revising our structure for NP ......................................................................................... 131
8.4 Why is our DP structure better than our old NP structure? .......................................... 131
8.4.1 Possessive constructions........................................................................................ 131
8.4.2 Deverbal nouns and their subjects ........................................................................ 135
8.4.3 Postnominal determiners ...................................................................................... 138
8.5 What elements can appear under D? ............................................................................ 141
8.6 QP and classifiers in Vietnamese.................................................................................... 143
8.7 Summary ........................................................................................................................ 153
APPENDIX: SOME PARALLELISMS BETWEEN DP AND IP: NOMINATIVE AND GENITIVE CASE .. 154
CHAPTER 9: THE STRUCTURE OF CP...................................................................................159
9.1 Types of sentences ......................................................................................................... 159
9.2 Types of C ....................................................................................................................... 160
9.3 CP structures: Embedded clauses .................................................................................. 160
9.4 CP Structures: Declarative main clauses ........................................................................ 162
9.5 Interrogative main clauses: Simple questions................................................................ 162
9.6 Simple yes/no questions in English. ............................................................................... 164
9.6.1 Simple questions with modals ............................................................................... 164
9.6.2 Simple questions with auxiliaries ........................................................................... 165
9.6.3 Simple questions with main verbs ......................................................................... 166
9.7 Negative sentences ........................................................................................................ 169
9.7.1 Negative sentences with modal verbs ................................................................... 169
9.7.2 Negative sentences with auxiliary verbs ................................................................ 171
9.7.3 Negative sentences with main verbs. .................................................................... 172
9.7.4 Exploring the syntax of NOT................................................................................... 176
9.8 Sentence adverbs ........................................................................................................... 184
9.9 Summary ........................................................................................................................ 186
CHAPTER 10: WH-MOVEMENT: WH-INTERROGATIVES .....................................................188
10.1 Wh-questions ............................................................................................................... 188
10.2 Wh-questions in main clauses...................................................................................... 189

v
 10.3 Wh-questions in embedded clauses ............................................................................ 195
10.4 Long distance wh-questions ......................................................................................... 199
10.5 What about Vietnamese and other languages that don't move wh-words? ..................... 203
10.6 What about subject wh-questions? ............................................................................. 205
10.7 What about whether? .................................................................................................. 207
10.8 Summary ...................................................................................................................... 209
CHAPTER 11: WH-MOVEMENT: RELATIVE CLAUSES ...........................................................210
11.1 Restrictive and Non-restrictive relative clauses ........................................................... 210
11.2 Restrictive relative clauses. .......................................................................................... 211
11.2.1 Restrictive relative clauses with Who or Which .................................................. 213
11.2.2 Restrictive relative clauses without a pronounced relative pronoun .................. 217
11.3 Non-restrictive relative clauses .................................................................................... 222
11.4 Summary ...................................................................................................................... 227
APPENDIX: RELATIVE CLAUSES WITHOUT A SILENT ANTECEDENT ........................................ 228
CHAPTER 12: NON-TENSED CLAUSES ................................................................................231
12.1 Infinitival clauses without an overt subject: subject-controlled PRO .......................... 231
12.2 Infinitival clauses with overt subjects .......................................................................... 234
12.2.1 Verbs that take an infinitival clause with a For complementizer ........................ 234
12.2.2 Verbs that take an infinitival clause with an overt subject and without For ....... 236
12.3 Infinitival clauses without an overt subject: Object-controlled PRO ........................... 241
12.4 Arbitrary PRO ............................................................................................................... 245
12.5 A brief look at Small Clauses ........................................................................................ 245
12.6 Summary ........................................................................................................................ 249
CHAPTER 13: DP-MOVEMENT ..........................................................................................252
13.1 Copulative verbs ........................................................................................................... 252
13.2 The verb to seem.......................................................................................................... 254
13.3 Passive Constructions ................................................................................................... 262
13.4 There-constructions ..................................................................................................... 264
13.5 Where is the subject of a sentence merged?............................................................... 269
13.6 Extraposed sentences?................................................................................................. 272
13.7 Tough-constructions .................................................................................................... 276
CHALLENGE ............................................................................................................................. 279
13.8 Summary ...................................................................................................................... 280
FINAL PRACTICE ......................................................................................................................... 282
EPILOGUE................................................................................................................................... 285
ONE HUNDRED IMPORTANT TERMS TO REMEMBER ................................................................ 291
RECOMMENDED FURTHER READINGS ...................................................................................... 300
INDEX ......................................................................................................................................... 302

vi
 FOREWORD

“I have read the book with my true enjoyment and happiness.” It is the case of the textbook
entitled “English Syntax and Universal Grammar –An introduction to Syntactic Analysis”,
which has been carefully crafted and prepared by Professors Héctor Campos and his co-
authors. As a dedicated visiting scholar at the Faculty of English Linguistics and Literature
(EF), the University of Social Sciences and Humanities (Vietnam National University –Ho
Chi Minh City, abbreviated as USSH, VNU-HCM) since 2015, Professor Héctor Campos
has greatly inspired the EF students and professors with his amazing ideas and proposed
principles for analyzing English syntactic structures. Thanks to his unstopped devotion to
research and teaching, the book is finally introduced to and welcomed by the EF students
and teaching staff for its “universal” values that are brought about with many
comprehensible examples, well-synthesized fundamental principles and paradigms, as well
as abundant practical activities and exercises related to syntactic analysis.

The EF staff are very much indebted to the professor and his team’s serious academic
dedication and endless contribution to our faculty's scholarship. We strongly believe that
“English Syntax and Universal Grammar –An introduction to Syntactic Analysis”, both the
students’ book and answer key book, will be of benefit to teachers and students alike. While
the novice readers will enjoy its simplicity and comprehensibility, the more experienced
ones will update their knowledge with another perspective in syntactic analysis as well as
with research-based projects and topics for further investigation and exploration in their
classes.

Beside our sincere appreciation to the authors and their team for producing this work, we
would also like to thank Georgetown University (USA) and the USSH (VNU-HCM,
Vietnam) for their financial support, making this publication part of the EF’s endeavor in
updating, supplementing and enriching its learning resources ever since its development in
1957. It is hereby our great honor to introduce this textbook to all readers.

Le Hoang Dung, PhD

Dean of Faculty of English Linguistics and Literature

vii
 PROLOGUE

There may be two reasons why this book is in your hands: you are required to take a syntax
course as part of your course requirements or you love English so much that you want to
understand how it works and why speakers say what they say the way they say it.
This textbook has been written to provide you with the basic tools to start analyzing English
as well as many other languages. The syntactic model introduced here is what is known as
the Principles and Parameters model, also known as the theory of Government and
Binding. This model has also been applied to study first, second and third language
acquisition, computational models of language, language processing in the brain/mind,
psycholinguistics, etc. After going through this book, you will not only be able to analyze
many structures of English and other languages, but also be ableto follow research in those
fields. This model will take you through the research in syntax done until the early 90's. To
understand current research, you will need to continue your study of syntax. Some basic
literature is given at the end of this book.
This books consists of thirteen chapters. It assumes that you have no previous knowledge of
linguistics, Universal Grammar or syntax. The book contains several PRACTICE sections
where you are asked to apply the concepts that have been introduced in the lesson. Some
practice section contain a CHALLENGE section. These are more advanced problems that
will require deeper analysis. Some chapters also contain a RESEARCH PROJECT section.
This section requires research and usually asks you to apply the concepts of the chapter to
Vietnamese or other languages. Instructors are advised to assign these research projects as
group projects.
This book was written during Héctor Campos's sabbatical term the spring of 2016
sponsored by Georgetown University. In the summers of 2015 and 2016, he had the
privilege of teaching English Morphosyntax for the Faculty of English Linguistics and
Literature at the University of Social Sciences and Humanities – Vietnam National
University – Ho Chi Minh city. The notes he made for those bright students and the fruitful
discussions he had with his coauthors grew into this textbook.
We would like to thank the Dean of the Faculty of English Linguistics and Literature, Mr.
Lê Hoàng Dũng; the Chair of the Department of English Linguistics, Mr. Lê Nguyễn Minh
Thọ; its former Chair, Ms. Võ Thị Nữ Anh; Ms. Cao Thị Quỳnh Loan, Ms. Nguyễn Bích
Hạnh, Ms. Nguyễn Thị Kim Loan and Mr. Nguyễn Xuân Triều, for their support with this
project. Héctor Campos owes special thanks to Ms. Trần Thị Lê Dung, Ms. Phó Phương
Dung and Ms. Trần Thị Tươi, who helped him become a visiting lecturer at their renowned
University. Our gratitude to the members of the faculty and students who attended Héctor
Campos's 2015 – 2017 lectures, as well as to those who attended our book presentation in
viii
August 2017, where different parts of the book were presented. We are indebted to the
faculty and brilliant students who attended the 2017 Saigon Syntax Club meetings where
this material was tested and revised. They definitely played an active role in shaping this
material.
Last but not least, special thanks to our students both in Saigon as well as in
Washington D.C. who went through this book and helped us improve the original
version. We would also like to thank Jonathan Horneman his help. The errors in the text
remain only ours.

The authors

ix
 English Syntax & Universal Grammar 1

CHAPTER 1
LINGUISTICS AS A SCIENCE OF THE MIND

1.1. What is linguistics?

We will start with a basic question before we delve deeper into the subject of syntax. What
is linguistics?

Although many answers are possible, we will adopt the answer in (1):

(1) What is Linguistics?

Linguistics is the scientific study of language.

In our definition in (1) there are two words that require further discussion: scientific and
language.

What do we mean when we say that linguistics is the "scientific" study of language?
What do we mean by "scientific"? We mean that we apply the scientific method in our
research. What is the scientific method about? The scientific method is the way we proceed
in our study. As in all sciences, we follow the following steps in our linguistic research:

(2) The scientific method:

(a) collect data

(b) analyze the data
(c) make a hypothesis
(d) provide evidence for our hypothesis
(e) make predictions
(f) verify the predictions
(g) revise or reject the hypothesis if necessary
(h) test the hypothesis again

As we can see, linguistics, and in particular theoretical linguistics, is not any different from
the way that physical scientists proceed with their research.
2 Héctor Campos & Bùi Huỳnh Thủy Thương

Now that we understand "scientific" in our definition in (1), let us tackle a more difficult
question. What do we mean when we say the scientific study of "language"? To understand
what we mean by "language" in (1), we need to differentiate two kinds of language: the
External Language (E-Language) and the Internal Language (I-Language). Many
languages have two words for for these two concepts. So Spanish, for instance, translates E-
language as lengua, while I-language is lenguaje. In his classic work On Language (1836),
Wilhem von Humboldt distinguished between energeia (energy) and ergon (work).
According to Humboldt, the creative ability we have in language is energeia (Tätigkeit,
Erzeugung) while what we produce thanks to this ability is ergon (Werk, Erzeugtes).
Ferdinand de Saussure in his Cours de linguistique générale (1916) defines langue as the
internal system while parole is the external system. In early transformational grammar this
difference was captured by the difference between competence (equivalent to I-Language)
and performance (equivalent to E-Language). The diagram in (3) shows the two different
types of languages:

(3)

EXTERNAL LANGUAGE
(E-Language)
INTERNAL LANGUAGE
(I-Language)

"competence" "performance"

The External Languageor E-Language is what we can hear, what we produce, what we
can record, transcribe, etc. It has a physical reality. The Internal Language or I-
Language, on the other hand, is the system that we have in our minds which allows us to
produce that external language. I-Language is our "knowledge of language" that allows
us to create our E-Language. What are some of the differences between I-Language and
E-Language?

(a) E-Language is external, I-Language is internal. Thus while our E-Language has
a physical reality, our I-Language is internal to the mind.

(b) E-Language is infinite, I-Language is finite. Notice that there is no such a thing
as "the longest sentence in the world". This is because we can always create longer
sentences. If my friend Joe has created a sentence with 100 words, I can always create a
new longer sentence with 101 words (or more!), if I have a good enough memory to
remember what I was saying! Thus we say that E-Language is (potentially) infinite. I-
Language, on the other hand, is finite, or a child would never be able to learn a language.
 English Syntax & Universal Grammar 3

We see that our I-Language consists of a finite set of rules that children are able to
learn, which allows them to produce an infinite number of sentences as their E-
Language.

(c) I-Language is individual. Each human being, regardless of intelligence or gender,

will have his/her own system of I-language.

(d) I-Language is unconscious. We do not know why we know what we know (we
will see examples of our linguistic knowledge in section 3 of this chapter). We just know
what to do when we talk.

(e) I-Language in intensional, E-Language is extensional. Let us take an example to

understand the difference between "extensional" and "intensional". Imagine that my friend
Tuoi speaks a strange mathematical language. Her sentences are {2, 4, 6, 8, 10, ..}, that is,
her sentences consist of only even numbers. The numbers she says and which I can collect
for my study constitute the "extension". The "intension", on the other hand, is the formula,
the system that makes my friend produce that particular extension of even numbers. We
could say that the "intension" or formula in her mind is "(2 x Y), where Y is the set of
natural numbers bigger than 0". Thus E-Language is what we produce (extension), while I-
Language is the system that allows us to produce that E-Language (intension).

Since we have distinguished two kinds of Language, which one is the one that we linguists
want to study: I-Language or E-Language?

Our task as theoretical linguists is to create a model for the I-Language. Imagine I have
an old fashioned mechanical clock. It tells time accurately. I want to know how it can tell
time. I can open the clock, take it apart and I can understand how it works. Now imagine
that I have a calculator. The calculator can add accurately all the time (provided I enter the
numbers correctly). I want to know how the calculator can do that. I open the calculator but
I can find no numbers, no addition or rule for addition inside. All I can find is wires, chips,
etc. Similarly with language. When we talk we use verbs, noun, verb phrases, sentences. I
want to know how this works, so I (illegally) open someone's head and all I can find is a
brain. Will I find noun phrases and sentences in the brain? No, because language, as
addition in my calculator, is a "program", it does not have any physical reality. So how can
we create a model for I-Language if I-Language is in the mind and has no physical reality?
This is where E-Language comes into play. We study E-Language in order to create a
model for I-Language. In other words, E-Language is our "window" to our I-Language.
The Chart in (4) represents the way that we would create a model for I-Language:
4 Héctor Campos & Bùi Huỳnh Thủy Thương

(4)

EXTERNAL LANGUAGE
(E-Language)
INTERNAL LANGUAGE
(I-Language)

"competence" "performance"

MODEL

As we can see in (4), we linguists want to create a model for I-Language. To create this
model we use evidence from E-Language. To test this model that we create, we use
evidence from E-Language. Thus I-Language is a "construct". It has no physical reality.
But it starts from and ends in our E-Language.

1.2 What are the different areas of Linguistics and where is Syntax in the big
picture?

We will divide Linguistics into two major fields: Theoretical Linguistics and Applied
Linguistics. We will look at these in turn.

1.2.1 Theoretical Linguistics or Core Linguistics. Theoretical Linguistics has six major
fields. We will divide them into the study of sound, form and meaning. These are the fields
and their subject of study:

(5) Areas of THEORETICAL LINGUISTICS

STUDY OF SOUND
(a) Phonetics: studies sounds, the way they are produced and perceived
(b) Phonology: studies sounds, processes that affect sounds, interactions among
sounds

STUDY OF FORM
(c) Morphology: studies words, the way words are formed
(d) Syntax: studies the way we combine words to form phrases and sentences
 English Syntax & Universal Grammar 5

STUDY OF MEANING

(e) Semantics: studies meaning devoid of context

(f) Pragmatics: studies meaning in context.

We consider these fields as the "core fields" (hence also the name of "core Linguistics")
because they will be the foundation for other areas of Linguistics. In this view, all the other
fields are "Applied Linguistics" because we will apply the basic notions from our core fields
to study those other fields.

1.2.2 Applied Linguistics. (6) represents some of the major fields of what we are calling
"Applied Linguistics" here.

(6) Areas of APPLIED LINGUISTICS

a. L1, L2, L3 acquisition: studies how first, second and third languages are
acquired. Sometimes also called psycholinguistics.
b. Neurolinguistics: studies the connection between brain and language
c. Computational linguistics: the brain is (like) a powerful computer. We can
test and see whether we can imitate the way the mind works by using
computers
d. Historical linguistics: studies the way languages evolve through time
e. Dialectology: studies the ways that a particular language varies across
particular geographical region.
f. Sociolinguistics: studies the relation between language and society, variables
that affect language variation, etc.
g. Forensic linguistics: language and the law

From (5) and (6), we can see that the study of syntax is part of Theoretical Linguistics.
But syntax can also be the subject of study in "Applied Linguistics". We may study how we
learn to use the present tense, for instance, in our first, second or third language (we would
be in the field of Language Acquisition). We may study what areas of the brain are affected
when we use different tenses, when we use regular and irregular forms (now we would be
in the field of Neuro-linguistics). We may simulate a program to see how a computer can
learn verbal paradigms of regular and irregular verbs (we would be in the field of
computational linguistics), etc. To do solid work in "Applied Linguistics" we must have a
solid knowledge of the theoretical field we are trying to investigate. Of course the applied
fields will also have their own methods and principles which we need to master in order to
become solid researchers in any of the fields in (6). So, it is actually harder to be an applied
linguist than a theoretical linguist since an applied linguist would actually need to master (at
least) two fields to be a solid linguist.
6 Héctor Campos & Bùi Huỳnh Thủy Thương

1.3 Where is Linguistics in relation to other sciences?

Since our goal is to create a model of our knowledge of language in the mind, linguistics is
part of psychology, which is the science that studies behavior and mind. Since the mind is a
human characteristic, psychology is part of biology, which is the study of living things,
including humans. And biology is part of physics, which is the mother of all sciences and
concerns itself with the study of all matter. The chart in (7) summarizes our conception of
Linguistics within the sciences:

(7)

Linguistics

Psychology

Biology

Physics

Since Linguistics is a part of physics and biology, it is no wonder that linguistics follows the
scientific method as its research methodology.

1.4 An example of the scientific method in Linguistics

Imagine you are presented with the following data from Spanish and are asked to find the
rule that allows us to form the present tense for singular pronouns:

(8) CANTAR -to sing SALTAR -to jump

Canto -I sing Salto -I jump
Cantas -You sing Saltas -You jump
Canta -He/she/it sings Salta -He/she/it jumps

Looking at the data in (8) we observe that both cantar 'to sing' and saltar 'to jump' take the
same endings for the same person. These endings are: -o, -as, -a. Thus we are ready to
make a hypothesis: -o, -as, -a are the endings for the present tense of singular verbs in
 English Syntax & Universal Grammar 7

Spanish. This sequence of endings is what we call a verbal paradigm. Thus we have a
hypothesis for the verbal paradigm for the present tense of singular verbs in Spanish! How
can we test our hypothesis? We can go and get more data from our speakers (or from our
sources). Imagine we get the new data (9):

(9) AMAR -to love COMER -to eat

Amo -I love Como -I eat
Amas -You love Comes -You eat
Ama -He/she/it loves Come -He/she/it eats

Does our hypothesis for this verbal paradigm work? It works very well for AMAR 'to love',
which takes the same endings as CANTAR 'to sing' and SALTAR 'to jump' in (8). But our
hypothesis fails with COMER, although it does work for the first person singular (I) when I
say 'I eat'.

Given the data with COMER 'to eat', we are forced to modify our hypothesis. But
immediately we see a way out!! CANTAR 'to sing', SALTAR 'to jump' and AMAR 'to love'
all end in –AR, while COMER 'to eat' ends in –ER. So we may hypothesize that there are
two classes of verbs: those that end in –AR, which take the endings –o, -as, -a and those
that end in –ER, which take the endings –o, -es, -e.

To test our revised hypothesis, we go and collect more data and now we get the following
new data:

(10) BEBER -to drink PERDER -to lose

Bebo -I drink Pierdo -I lose
Bebes -You drink Pierdes -You lose
Bebe -He/she/it drinks Pierde -He/she/it loses

We see that our hypothesis works. Both verbs show the paradigm we predicted, namely, -o,
-es, -e since they are verbs that end in –ER. But now we have a new dilemma. While all the
previous verbs keep the original stem (cant-, salt-, am-, com-, beb-) in front of the endings
we discovered, perd- changes to pierd-. So now we need to find out why perder undergoes
that stem change. Notice that, given our data so far, we cannot conclude that verbs that
include an -e- in the stem undergo dipthongization because dipthongization does not happen
with BEBER 'to drink'.

Trying to discover what is happening with PERDER in (10), we collect more data and we
get:
8 Héctor Campos & Bùi Huỳnh Thủy Thương

(11) ESCRIBIR -to write SENTIR -to feel

Escribo -I write Siento -I feel
Escribes You write Sientes -you feel
Escribe -He/she/it writes Siente -He/she/it feels

These new data complicate our hypothesis further, because we can now see that verbs
ending in –IR take the same paradigm as those ending in –ER. And we found another verb
that undergoes a stem change, sent- becomes sient-, the same way that perd- becomes
pierd-.

These few examples should give you a feeling for how the scientific method works. Starting
with some basic data, we come up with a hypothesis. We test this hypothesis against some
further data and we see that our hypothesis does not explain all the cases. We revise the
hypothesis, check against further data again and keep on finding further problems.

They say that a scientist (hence linguists are included) is like a person who is walking with
his/her arms full of oranges (the scientist wants to collect oranges). Happy with the oranges
in his/her arms, he/she sees some oranges on the floor. Since he/she has defined his/her task
as collecting oranges, he/she bends down to pick the oranges on the floor, picks them up,
but then some oranges fall from his/her arms to the floor. And so on. The point is that there
will always be oranges on the floor. A better theory is the one that leaves fewer oranges on
the floor. Oranges represent facts in our story. Thus in science, we will always have facts
in search of an explanation. The better theory will be the one that is able to explain more
facts. When two theories explain the same facts, we will pick the theory that is simpler, the
one that has less machinery to account for the data. This is what we call Occam's razor.
Occam's razor is as important in Linguistics as it is in the natural sciences.

1.5 What do we know about syntax before studying this book?

Modern Theoretical Linguistics starts from what is known as Plato's paradox:

(12) Plato's Paradox

How come we know so much with so little evidence?

In this section we will show that you know a lot about syntax even before you start
studying it!

(a) We recognize grammaticality. Consider the examples in (13) and (14):

 English Syntax & Universal Grammar 9

(13) a. John eats rice.

b. *John rice eats.

(14) John-wa gohan-o tabemasu. (Japanese)

John-subject rice-object eats

We say that English is an SVO language. This means that the unmarked word order in
English is Subject (S) > Verb (V) > Object (O). In (13a), we follow this order, so the
sentence is grammatical. In (13b), we have used the order SOV, so the sentence is
ungrammatical (we represent ungrammatical sentences by using an asterisk "*" in front of
the sentence). However, in Japanese, this order is grammatical, as we can see in (14). All
speakers of English and Japanese recognize (13a) and (14) as grammatical. Plato's paradox
asks us: How did we learn this? Who taught us that this is the rule? Who taught us the
notions of subject, verb, object? So here we have a challenge for the theory of Universal
Grammar that we will be studying in this course: How can we account for the contradictory
data in (13) and (14)? How can our system explain English and Japanese which exhibit
different grammatical word orders? We will start exploring this issue in Chapter 3.

(b) We recognize non-sensical sentences as different from ungrammatical sentences.

Consider the example in (15), a famous Chomsky sentence used to prove the autonomy of
syntax in his 1957 book Syntactic Structures:

(15) Colorless green ideas sleep furiously.

We know that (15) is grammatically correct (as opposed to 14b), but (15) does not
make sense!

This is similar to what we observe on (16):

(16)

(16) "looks" like a normal object, the same way that (15) "sounds" like a normal sentence.
But we cannot make sense out of (15) and (16).

How can we recognize non-sensical objects? We cannot claim that something is non-
sensical because we had never heard it or seen it before! We listen to new ideas and see new
objects all the time. And yet we know when something is non-sensical as opposed to
ungrammatical.
10 Héctor Campos & Bùi Huỳnh Thủy Thương

(c) We recognize ambiguity. Consider the example in (17):

(17) the old Vietnamese teacher

a. the teacher is old and from Vietnam
b. the teacher is old and teaches Vietnamese
c. the teacher teaches old Vietnamese
d. the teacher used to teach Vietnamese
e. the person is old, from Vietnam and used to teach

The simple noun phrase in (17) is 5 times ambiguous! If we think about it carefully, we will
get the five possible readings listed in (17). How did we learn that this structure was
ambiguous? Compare the noun phrase in (17) with the one in (18):

(18) the old Vietnamese man

The structure in (18) is not ambiguous, although we have only changed the final noun.
Plato's paradox asks us how we learned linguistic ambiguity! This is similar to what
happens in our minds when we observe visual ambiguity. Look at the pictures in (19):

(19)

What do you see in the picture on the left? You may see two black faces looking at each
other or a white vase. It is a bit harder to see the ambiguity in the picture on the right: it can
either be an old woman (ugly, with a big nose, dressed in rags) or a young woman
(beautiful, with a tiny nose, big eyelashes, very elegant). Similar to what we saw with
language we may apply Plato's paradox here: Who taught us that these pictures were
ambiguous? The answer is simple, and we would give the same answer for the linguistic
ambiguity in (17). Nobody taught us. We just know the answer! Our internal linguistic
and visual systems give us the answer.
 English Syntax & Universal Grammar 11

It is interesting to see that we talk about "learning to talk" but we would never talk about
"learning to see". Yet both processes are similar. Given the appropriate environment, we
automatically see, the same way that given the appropriate environment, we can talk.

This is the reason that linguists view language as an "organ", no different from vision.

(d) We can perceive invisible elements. Consider the examples in (19):

(19) a. Tuoi appeared to Huong [ ] to know the answer.

b. Tuoi appealed to Huong [ ] to know the answer.

In the examples in (19) we have changed the verb appeared to appealed, that is, we have
changed one sound. Yet the way we interpret the invisible element marked as "[ ]" in (19) is
different. In (19a) Tuoi may know the answer while in (19b), Huong may know the answer.
It is obvious from the examples in (19) that there is no element [ ] in front of the verb to
know and yet our minds interpret a noun phrase there. Plato's paradox comes up again: How
did we learn to understand these invisible elements?

The same can also be observed in our visual system. Consider the picture in (20). How
many triangles can you see in (20)?

(20)

Everybody can clearly "see" two triangles. But the triangle with white borders is actually an
illusion. It is not there, the same way that Tuoi and Huong are not in front of the infinitive in
(19). Our visual system makes us see two triangles in (20), the same way that our linguistic
system makes us understand that Tuoi or Huong may know the answer in (19). How did we
learn that? Well, it is clear that we did not learn that anywhere. We just know the answer!

(e) Finally let us look at some of the limits of our linguistic system. Consider the
sentences in (21):
12 Héctor Campos & Bùi Huỳnh Thủy Thương

(21) a. Quani thinks that hei is smart.

b. *Hei thinks that Quani is smart.

In (21a) he may refer to Quan although it can also refer to someone else mentioned in
discourse. We say that Quan and he are coreferent. We indicate coreference by using the
same subindex "i" on both Quan and he. We say that Quan is the pronoun's antecedent. In
(21b), on the other hand, he cannot refer to Quan. We may think that this is due to the fact
that the antecedent must precede the pronoun, but this suspicion is quickly proved wrong
when we consider sentences like (22), where the pronoun or possessive precedes the
antecedent and yet both the pronoun he and the possessive his can take Quan as their
antecedent:

(22) a. When hei arrived, Quani cooked dinner.

b. Hisi mother thinks that Quani is smart.

How did we learn that (21a) is possible but (21b) is not? Yet more curious, let us try the
sentences in (21) in Vietnamese:

(23) a. Quâni nghĩ [mình]i rất thông minh.

b. *[Mình]i nghĩ Quâni rất thông minh.

We observe exactly the same behavior as English in (21), only in (21a) can the pronoun
mình 'he/self' take Quân as its antecedent. In (23b) this relation is blocked, just as it is
blocked in English (21b). How did English and Vietnamese speakers decide to have the
same constraint on coreference as in (21)? If we check this pair of sentences in many
languages we will see that it is also impossible. Why do many, if not all, languages have the
same constraint? Where does this constraint come from? Who taught us that we cannot get
coreference in (21b), but it is possible to get it in (21a) and in (22)? The answer is that
nobody taught us. We just know the answer!!

As we can see from the examples in this section, we know more than we are aware of and
we cannot really explain how we acquired this knowledge. We just know that certain
combinations or relations are possible while others are not. This is similar to what happens
with our visual system: nobody taught us to see, we did not "learn" to see, we just see!

Perhaps we can find a clue to solve Plato's paradox when we look at how children acquire
a language. So we turn to language acquisition in our next section.

1.6 The logical problem of language acquisition

As we saw in the previous section, we "know" far more than we have been taught.
Language acquisition also presents us with Plato's paradox, as we will see in this section.
 English Syntax & Universal Grammar 13

There are some interesting facts about language acquisition which should make us think of
Plato's paradox once again.

(a) Children get incorrect and incomplete data, yet they are able to ignore or filter out
those facts and still come up with the correct linguistic system. When children hear a
language they hear incomplete sentences, slips of the tongue, etc. and yet, they are able
to figure out the correct system for a particular language. How is this possible? Imagine
that we are trying to learn to play chess by watching two people play but the players
have agreed that from time to time they will make a wrong move. These wrong moves
would considerably complicate our task of learning to play chess and, in fact, we may
not be able to distinguish the correct from the incorrect moves. So we would never learn
to play chess correctly! Imagine a similar situation with language, where the possible
combinations are much more complex than chess moves. Unless there is an internal
system guiding the child, language acquisition would be impossible.

(b) Children typically get simplified data from parents and siblings, yet they are able to
come up with the correct grammatical system. How is this possible? Again, an internal
system seems to be pushing the development.

(c) Children, no matter the language they are acquiring, go through similar
developmental stages in their language acquisition. How is it possible that a child
acquiring Vietnamese should go through similar stages as a child learning English?
Again, it must be the case that an internal system is guiding the child in the process of
acquiring a language.

(d) What role does correction play when children learn a language? It plays minimal to
no role at all! We have all probably seen parents or siblings correcting babies trying to
speak and yet the baby keeps on repeating the same mistake until one day the baby
magically produces the correct form. Again, there seems to be an internal system that
allows children to produce the correct form or structure at a particular point in
their development.1

(e) The strongest argument for an internal system comes from what is known as the
poverty of stimulus argument. Imagine I present you with the first numbers of
sequence of numbers that I have in my mind: {1, 2, 3, ...} and I ask you to tell me what
you think the next number in the sequence will be. You will probably automatically say
that "4" is the next number in the sequence. This is because you probably thought that
the intension or formula I had in my mind was X+1, where X= {0, 1, 2, 3, ...} (if you
1
You can watch a fun video about the role of correction when learning a language in
[Link]
14 Héctor Campos & Bùi Huỳnh Thủy Thương

did not think of the formula, your mind did it automatically for you!). You will probably
be very surprised and disappointed to find out that the next number I had in my mind
was actually "10" (!) because the formula I had in my mind was [(X-3)*(X-2)*(X-1)] +
X = Y where x > 0. When X has the value 1, 2 or 3, the result will always be the same as
X (try doing the math!). But when I take X=4, the result will actually be "10". So the
sequence I had in my mind was actually {0, 1, 2, 3, 10, 29, 66, ...}.

Let us apply this example to language acquisition now. Why did you think that "4" was the
next number in the sequence? Because you had only encountered the first 4 elements of the
sequence {0, 1, 2, 3, ..}. Had you encountered the next number in the sequence, namely
"10", then you would have known that I actually had a more complex intension than the one
you thought of. So this problem applies directly to language acquisition. The child hears
very limited evidence (and as we saw in (a,b) above, and many times the evidence is
defective), and yet the child is able to come up with the correct system. How is this
possible? Again, language acquisition would be impossible if the child were to do this
without the help of an internal system.

1.7. Language Acquisition and Universal Grammar

We saw in the previous section that Language Acquisition seems to be guided by an

internal system. This internal mechanism that facilitates language acquisition in children is
what Chomsky has labeled the Language Acquisition Device (LAD). One important
component of this LAD is UNIVERSAL GRAMMAR.

Universal Grammar is the set of Principles and Parameters that makes language
acquisition possible. We may think of a Principle as a light switch which we may turn on
(+) or off (-):

(24)
+
-
Let us take the principle known as the Pro-drop Principle (also called pro-drop
Parameteror Silent Subject parameter). This principle accounts for languages that have
optional subjects versus languages that don't. Compare English and Spanish in (25):

(25) a. I visit Vietnam in the summer.

b. (Yo) visito Vietnam en el verano.
(I) visit Vietnam in the summer
 English Syntax & Universal Grammar 15

As we can see in (25), the subject is optional in Spanish (we only use it when we want to
emphasize the subject). In English, on the other hand, the subject is obligatory. Thus the
Silent Subject Principle has a (+) value in Spanish (since subjects can silent) while it has a
(-) value in English. The child will not be able to fix this parameter until the child is
exposed to the language. Thus Universal Grammar provides the child with a Principle and
an open Parameter; exposure to the language will allow the child to fix the parameter as
(+) or (-), depending on the language.

(26)

+
- -
Silent Subject Principle Silent Subject Principle
English Spanish

Consider now Vietnamese and compare it with English:

(27) a. - Khiêm đến chưa? b. - Has Khiem come yet?

- Dạ, đến rồi. - Yes, he has.

As we can see in (27a), Vietnamese allows us to drop the subject but, different from
Spanish in (25b), in Vietnamese this is possible only when the context provides us with
enough information so as to be able to drop the subject. Notice that in English this is
impossible, as shown in (27b).

Vietnamese can also drop objects if the context is clear:

(28) a. - Hà gặp Quân chưa? b. - Has Ha met Quan yet?

- Dạ, gặp rồi. - Yeah, she met him already.

As we can see in (28), both the subject and the object can be dropped in Vietnamese
provided the context allows us to identify them. In (28b) we see that this is impossible in
English. Thus if we have a principle like "drop subject/object if context is clear", English
would have a negative value (-) for this principle while Vietnamese would have a positive
value (+). The Chart in (29) shows the values for the principle "drop subject/object in
context" for the three languages discussed in this section:
16 Héctor Campos & Bùi Huỳnh Thủy Thương

(29)

- -
Drop Subject/ Drop Subject/ Drop Subject/
Object in Object in Object in
context context context

English Vietnamese Spanish

In this course we will explore the principles of Universal Grammar that make English
possible. Although we will concentrate on English, we will discuss other languages to see
how we would account for languages that differ from English. The challenge for Universal
Grammar is to propose principles that will be narrow enough so as not to overgenerate and
produce structures that do not exist. On the other hand, our system must be wide or flexible
enough to allow us to explain the rich variety that is found in different languages.

To conclude this introductory chapter, we must insist that what we are proposing here is just
a model. There is a famous quote attributed to the biologist Thomas H. Huxley (1825-
1895):

There is no sadder sight in the world than to see a beautiful theory killed by
a brutal fact.

Fortunately (for theoreticians), facts do not kill theories, but better theories kill defective
theories. In science there is no truth. Every proposal must be subject to further scrutiny.
The model we propose in this book is not the "truth" about English syntax. It is a working
model of a very limited part of our syntactic knowledge of the English language.

With the due precaution in place, we are ready to start exploring the Syntax of English and
Universal Grammar.

Enjoy the adventure!

 English Syntax & Universal Grammar 17

PRACTICE

I. Answer the following questions briefly:

1. What do we mean when we say that linguistics is the "scientific" study of language?
2. Why do we need to distinguish I-Language from E-Language?
3. What are some of the differences between I-Language and E-Language?
4. What do we mean when we say that I-Language is "intensional"?
5. Why do we say that (theoretical) linguistics is a branch of psychology?
6. Why do we divide linguistics into "Core linguistics" and "Applied linguistics"?
7. Mention some of the fields in each of these areas.
8. Give original examples of Plato's paradox as it applies to syntax.
9. Why did the chapter compare visual ambiguity with syntactic ambiguity?
10. What is the function of the LAD? Is the LAD the same as Universal Grammar?

II. Watch Caroline Heycock's video "Prescriptive and Descriptive rules" (in
[Link]
BJFiOUfmMR4ew9TFwO), then answer the following questions:
1. What are some of the differences between a "prescriptive rule" and a "descriptive rule".
Give two examples of each.
2. Discuss the use of the word "law" in the following phrases:
a. laws of a country
b. laws of nature/physics
How do they relate to the prescriptive and descriptive rules you defined in Question 1
above?
3. How can we gather "evidence" for our descriptive rules?
4. Give original examples for the following descriptive rules discussed in the video:
a. coordinate structure constraint
b. that-trace effect
c. tough constructions
5. Consider the following sentences in Vietnamese:
a. Bài này rất khó hiểu.
b. Quân ham ăn.
c. Anh ấy rất khó ngủ.
d. Anh ấy rất dễ làm người khác buồn.
Translate the sentences above into English. Are these adjectives best analyzed as "easy" or
"eager" tough-constructions? What are some differences between English and Vietnamese?
18 Héctor Campos, Bùi Huỳnh Thủy Thương

CHAPTER 2
OUR MENTAL LEXICON

In Chapter 1 we saw that the goal of linguistics is to create a model of I-Language, or

our internal language system, that will allow us to generate our E-Language, or
external language. In this chapter we will create a model for our MENTAL LEXICON,
which is the repository of the basic elements that will allow us to build words and then
phrases and sentences.

What information must our Lexicon contain so that we are able to create syntactic
structures? What do we "know" when we say we know a verb, a noun, etc.? This is the
question we will try to answer in this chapter.

The LEXICON is where our knowledge of words, their meaning, and the basic
information to build our structures comes from. We will assume that our Lexicon has
two parts: a CATALOG and an EXTENDED CATALOG. We will discuss these
modules in turn.

2.1 The structure of the Catalog

Let us take two verbs as an example: to visit and to put. When we say that "we know"
these verbs, what exactly do we know? As you will see, we know more than you may be
aware of!

2.1.1 Categorization. When we say that we know the verbs to visit and to put, we
certainly know their syntactic category. We know that they are verbs (V). So, for
every element in the lexicon we must know their category. What are the possible
categories that we may find in English? In this book we will work with 11 basic
categories. Some may be familiar, others unfamiliar:

(1) The basic categories in our system

verbs (V), nouns (N), adjectives (A), prepositions (P), adverbs (ADV),
auxiliary verbs (AUX), determiners (D), quantifiers (Q), negation
(NEG), complementizers (C), inflection (I).
 English Syntax & Universal Grammar 19

You are probably familiar with most of these syntactic categories and may be
wondering why there are some categories missing. The category determiner (D) will
include both definite and indefinite articles (the, a/an), demonstrative adjectives
(this, these, that, those) and possessive adjectives (my, your, his, her, its, our, your,
their). The category noun will include all kinds of nouns including proper nouns (John,
Tuoi, Saigon, etc.) as well as common nouns (student, city, syntax, etc.). It will also
include pronouns (he, him, she, her, etc.). For the moment, inflection will include
modal verbs (can – could, will – would, shall – should, may – might, must). As we will
see in Chapter 7, it will also include the morphemes that indicate tense (present, past)
and agreement with the subject. Quantifiers (Q) are elements that show quantity, so Q
will include quantifiers like all, any, no, some, many, much, etc. Negation (NEG) will
include elements like –n't (which we can find in shouldn't) and not (as in should not).
Complementizers (C) will be elements that will help us connect two sentences and
these will be words like that or if. The system will also include conjunctions (CONJ)
like and, or, but, however, etc., but these conjunctions are the topic of a more advanced
syntax course, so we will not discuss them here.

We will use this category information to form syntactic trees. Syntactic trees represent the
way that these categories combine with each other to form more complex elements. A very
important point to learn before you start doing syntax is that TREES REPRESENT
CATEGORIES NOT FUNCTIONS. Take for instance the sentence I will go to Hanoi
next week. The noun phrase next week functions as an adverb since it modifies the verb to
go, but when we represent its structure, we will say that it is a noun phrase rather than an
adverb phrase. As we will see in the next chapters, functions are derived from the position
in the tree that we assign to our elements in their syntactic structure.

We will say that the Catalog contains a series of lexical entries or lexical grids. These
lexical entries will contain all the information that represents our knowledge of the
elements that we will use to form phrases and sentences. Lexical entries will contain
all the phonetic, phonological, morphological, syntactic and semantic information
that we need to combine these elements.

So let's go back to our original verbs to visit and to put and represent their lexical grids
to include their categories:

(2) Building a lexical grid (Part 1)

visit put
categorization→ V V
20 Héctor Campos, Bùi Huỳnh Thủy Thương

These categories will then combine with other categories to form larger units or
constituents called phrases. We will study the structure of the different possible
phrases in the next chapters. For example, a noun may combine with a determiner and
an adjective to form a noun phrase (NP) [NP the smart girl], an adjective may combine
with an adverb to form an adjective phrase (AP) [AP very smart], a preposition may
combine with a noun phrase to form a prepositional phrase (PP) [ PP near our house], a
verb may combine with a noun phrase to form a verb phrase (VP) [ VP do the homework],
and so on. We will study how to distinguish constituents in the next chapter.

PRACTICE 1

I. For the following sentences, write the category on top of each word as in the
example:

Ex: N V N
│ │ │
Ha loves syntax

a. She went to Dalat last week.

b. My father said that he would come to visit me on Saturday.
c. Our teacher has been working at this university for many years.
d. Khiem should not come to class late.
e. Our English teacher does not teach every day.

II. Label all the words in these sentences like exercise I and then identify all the
NP-, VP-, AP-, PP-phrases:

a. My boyfriend usually wakes up at noon.

b. Our best friend from Dalat went to Japan last week.
c. My parents will go to my house in the countryside for the summer.
d. Ha and Quan made a delicious coconut cake in a rice cooker in my friend's room.

RESEARCH PROJECT

Here's a Vietnamese sentence. Write the categories for each word:

Người Việt Nam có thể chưa từng ăn bánh bao, chưa từng ăn mì khô nhưng chắc
chắn ai cũng đã từng ăn phở.
What words are problematic for our categorization system?
 English Syntax & Universal Grammar 21

2.1.2 C-selection. The next piece of information that we will need for our lexical entry
is the c-selection. C-selection stands for c(ategory)-selection. This is also known as
subcategorization. C-selection refers to the category or categories the element in
question takes as complements. Thus the verb to visit requires a noun phrase as a
complement while the verb to put requires two complements: a noun phrase (NP) and a
prepositional phrase (PP). We indicate this information as an underscore "_" followed
by the selected category in the lexical grid:

(3) Building a lexical grid (Part 2)

visit put
categorization→ V V
c-selection → _ NP _ NP PP

This will allow us to distinguish different types of verbs depending on the number of
complements they take. A transitive verb is a verb that takes one NP complement
while a ditransitive verb is a verb that takes two NP complements. On the other hand,
an intransitive verb is one that takes no complements. An example for each of these
types of verbs is given in (4):

(4)
watch send sneeze
categorization→ V V V
c-selection → _ NP _ NP NP -
(transitive) (ditransitive) (intransitive)

We have verbs that may take different kinds of complements, as shown in (5):

(5) put talk behave

categorization→ V V V
c-selection → _ NP PP _PP _ ADVP

PRACTICE 2

I. Write lexical grids for the verbs in italics. Include their categorization as well as
their c-selection:
a. Tuoi wrote a long report.
b. The professor gave her students an exam.
c. Thuong brought her baby to class.
22 Héctor Campos, Bùi Huỳnh Thủy Thương

d. I need to withdraw money from the bank.

II. Consider the following sentences:
a. Ha made her friends a chocolate cake.
b. Ha made a chocolate cake for her friends.
Would you propose one or two lexical grids for this verb? Justify your answer.

It is important to remember that c-selection only includes the categories that are
necessary to complete the meaning of the verb. Take for instance the verbs read and
put in (6):

(6) a. Tuoi read the chapter at the library.

b. Quan put the icecream in the refrigerator in the morning.

In (6a), the verb read c-selects an NP, thus the PP at the library is extra information.
We say that this PP which serves as extra information is a modifierto the verb. Thus a
complement is necessary to complete the information of a verb while a modifier
adds extra information. In (6b) the PP in the refrigerator is c-selected by the verb put
but the PP in the morning is not. Thus we say that in (6b), the PP in the refrigerator is a
complement, while the PP in the morning is a modifier.

PRACTICE 3

Write the c-selection for the verbs in italics. Decide if the phrases in brackets are
‘complements’ or ‘modifiers’.
a. She withdrew money [PP from the bank] [NP last night].
b. She went [PP to the library] [PP in the evening].
c. She arrived [PP in the evening].
d. She behaved [ADVP very well] [PP at her grandparents' house].
e. She danced [ADVP very well] [PP at the performance] [NP last night].

CHALLENGE
Decide if the underlined constituents are COMPLEMENTS or MODIFIERS.
a. The woman wasted the following morning.
b. The woman died the following morning.
c. The professor blew up with his impolite answer.
d. The professor put up with his impolite answer.
e. They finally decided on the boat.
 English Syntax & Universal Grammar 23

f. They finally decided on the boat on the plane.

g. She put the letter on the shelf.
h. She read the letter on the shelf.

2.1.3 S-selection. When we say that we "know" the verbs visit and put we not only
know how many phrases are necessary to complete the full meaning of these verbs but
we also know what role these phrases play in relation to each other. This information is
encoded in the semantic selection or s-selection of the lexical entry. This information is
also called argument structure. The argument structure of a verb will be encoded in
particular roles which are called theta-roles ("theta" stands for "thematic", usually
abbreviated as "θ-roles"). In (7) we can see a brief definition of the basic theta-roles
we will use in this book:

(7) AGENT (AG): the doer of an action

EXPERIENCER (EXP): the perceiver of a state or an event
PATIENT (PAT): the argument that is perceived, experienced, or that
undergoes the action of the verb or a change of state; typically a direct
object
GOAL: the end point of movement
RECIPIENT (REC): the argument that receives something, typical with verbs
of transfer; indirect objects are typically recipients
LOCATION (LOC): the place where an action/event/state occurs
INSTRUMENT (INSTR): the tool with which an action is performed
SOURCE: the starting point of a movement
THEME (TH): a kind of agent which also receives the result of the action,
typical of verbs of movement. It may also be an argument that
receives a description
BENEFICIARY (BEN): the argument for whose benefit an action is performed
PROPOSITION (PROP): the theta-role assigned to sentences

The sentences in (8) exemplify these theta-roles:

(8) a. Quan (AG) sent flowers (PAT) to Ha (REC).

b. Ha (EXP) loves Quan (PAT).
c. The students (TH) went to Malaysia (GOAL).
d. The teacher (AG) put his laptop (PAT) on the table (LOC).
e. The butcher (AG) cut the meat (PAT) with a knife (INSTR).
f. Khiem (AG) withdrew money (PAT) from the bank (SOURCE).
24 Héctor Campos, Bùi Huỳnh Thủy Thương

g. The student (AG) carried the books (PAT) for her teacher (BEN)
h. This exercise (THEME) is hard.
i. I know [that Co Chi loves chocolate] (PROP)

Let's incorporate this information into the lexical grids of the verbs visit and put:

(9) Building a lexical grid (Part 3)

visit put
categorization→ V V
c-selection → _ NP _ NP PP
s-selection → AG,PAT AG,PAT,LOC

S-selection will also include semantic restrictions (also called selectional

restrictions). Even though both visit and put take a patient as a complement, they take
very different types of patients. We must visit "visitable" places: while we can visit a
church, we cannot visit sugar or a book. Constructions that violate these semantic
restrictions on s-selection are semantically odd or semantically deviant and we
typically indicate them with the pound sign "#". On the other hand, when we violate
syntactic conditions, sentences are judged ungrammatical, which we indicate with an
asterisk "*". Examples of both kinds of violations are given in (10).

(10) a. #John visited sincerity.

b. #Colorless green ideas sleep furiously.
c. *John visited.
d. *John slept a dream.
e. *John visited the house the church.
f. *Visited the house.

The sentences in (10a,b) are syntactically well-formed, however they are semantically
deviant. How can we detect that? They are syntactically well-formed because we
respect the c-selection of both verbs: visit, a transitive verb, takes an NP-complement in
(10a), sincerity, but this NP is not the appropriate kind of NP that can function as a
patient for visit. Hence the semantic deviance. Sleep, in (10b), on the other hand is an
intransitive verb. But the agent subject we have chosen for it is not of the appropriate
kind. Hence the semantic deviance. In (10a,b) we have respected the syntactic
conditions of having a complement and a subject, but we have violated the semantic
restrictions by assigning to these verbs a complement (10a) or a subject (10b) of the
 English Syntax & Universal Grammar 25

wrong semantic kind. When we include arguments that are not compatible with the
theta-roles the verb assigns we get semantic deviance.

In (10c,d,e,f) we have syntactic violations. Syntactic violations lead to

ungrammaticality, which we indicate with an asterisk "*" preceding the sentence. The
verb visit is transitive, but it is missing a complement in (10c). Hence the
ungrammaticality of (10c). In (10d), on the other hand, the verb sleep is intransitive, but
we have included an extra complement. Thus the ungrammaticality of (10d). In (10e)
we have the verb visit followed by two complements, but the verb visit requires only
one complement. Thus the second complement will not be recognized by the verb. This
also leads to syntactic ungrammaticality. Finally in (10f), we are missing the subject,
which will leave the theta-role of AGENT unassigned.

To guarantee syntactic well-formedness or grammaticality, we will assume three

universal principles. Consider the first two: the Projection Principle in (11) and the
Theta-Criterion in (12):

(11) The Projection Principle

The c-selection of a verb must be respected at every syntactic level.

(12) The Theta-Criterion

Each argument must receive one and only one theta-role and each theta-role
must be assigned to one and only one argument. (In brief: ARG ↔ Θ-role)

Let's look at the verb to visit. To visit c-selects NP as a complement, while its s-
selection tells us that it needs an AGENT and a PATIENT. The Projection Principle
guarantees that we will assign a complement to visit so we can assign the PATIENT
theta-role, but it does not guarantee that we will have a subject to assign the theta-role
of AGENT. Thus a third principle becomes necessary to ensure that all the theta-roles
will be assigned in the syntax. This principle will guarantee that we have a subject so
that all the arguments of the verb can be assigned, not just the complements. This
principle is known as the Extended Projection Principle, formulated in (13):

(13) Extended Projection Principle

Every sentence must have a subject.
26 Héctor Campos, Bùi Huỳnh Thủy Thương

As we will see in the next chapter, with these three principles we guarantee that every
verb will be introduced or merged in a construction not only with the correct number of
complements but also with the correct number of arguments to satisfy both its c- as
well as its s-selection.

PRACTICE 4

I. Identify the theta-role that the underlined noun phrases have. When a sentence
is ambiguous, indicate what alternative theta-role may have been assigned. Don't
be afraid to propose a new theta-role if none of the theta-roles that we studied
above fits.
a. Tuoi sent an email to Quan.
b. Tuoi sent Quan an email.
c. Quan received an email from Tuoi.
d. Thuong thinks that her husband made a cake for her.
e. Khiem rubbed the pizza with a garlic clove.
f. Quan cut the pizza with a fork.
g. It rains a lot in Saigon in November.
h. The professor left with her teaching assistant.
i. I played a tune on my iPod.
j. I listened to a tune on my iPod.

II. Decide if the following sentences are in violation of the Projection Principle, the
Theta-Criterion, the Extended Projection Principle or a combination of these.
Write "#" or "*" in the space provided. Justify your answer.
a. ___ John repeated the mountain.
b. ___ John opened.
c. ___ Susan went to Hanoi to Hue.
d. ___ Sincerity scared the cat.
e. ___ The cat scared sincerity.
f. ___ Studied math in highschool.

2.1.4 Semantic information. So far we have argued for three types of information that
need to be included with verbs: category, c-selection, s-selection + semantic
restrictions. Consider the lexical entries for send and give below:
 English Syntax & Universal Grammar 27

(14) send give

categorization→ V V
c-selection → _ NP NP _ NP NP
s-selection → AG, PAT, REC AG, PAT, REC

As you can see in (14) the verbs send and give have exactly the same specification or
information in their lexical grids regarding categorization, c-selection and s-selection.
While the information we have so far helps us handle the syntax of these verbs pretty
effectively, we are still lacking the full meaning that would help us distinguish these
two verbs. We will represent such missing semantic information as 'SEND' and 'GIVE'.
How to represent this meaning is a matter of research for LEXICAL SEMANTICS.
Lexical semantics will need to posit a system of features that will help us distinguish
one form from another.

(15) Building a lexical grid (Part 4)

send give
categorization→ V V
c-selection → _ NP NP _ NP NP
s-selection → AG, PAT, REC AG, PAT, REC
meaning → 'SEND' 'GIVE'

2.1.5 Phonological information. While we have a lot of syntactic and semantic

information in the lexical entries in (15), we have no information regarding how to
pronounce these words. Thus we need to include phonological information that will
help us decide how to pronounce these words. We will represent this information by
means of a phonological transcription, but phonologists will break these representations
into features that you will study or have already studied in a phonetics/phonology
course. Since this is a course in syntax and not in phonetics/phonology, we will not
complicate our lexical entries with further phonological or semantic information:

(16) Building a lexical grid (Part 5)

visit put
categorization→ V V
c-selection → _NP _NP PP
s-selection → AG, PAT AG, PAT, LOC
meaning → 'VISIT' 'PUT'
phonology → /ˈvɪzɪt / /pʊt/
28 Héctor Campos, Bùi Huỳnh Thủy Thương

2.1.6 Special properties. Finally the lexical entry must also include information about
any special (idiosyncratic) property or irregularity that these verbs may have. For
instance, if we have the verb write, we will have to include the fact that the past tense
(wrote) and the past participle (written) are irregular. For the noun foot, we need to
include information regarding the plural which is the irregular form feet as opposed to
the regular form *foots. These phenomena regarding plural and past tense formation are
examples of morphological irregularity. We must also include information about
syntactic irregularity. We will discuss some of these cases in future chapters, but let us
look at one case for illustration here. Consider the examples in (17):

(17) a. I gave the books to Mary.

b. I cooked lunch for/*to my mother.

As we can see in (17a), the verb give can use the word TO in front of the indirect object
or recipient. But the verb cook requires FOR instead of TO in front of the indirect object
or recipient. TO is the more common or unmarked form for indirect objects. Thus we
will need to specify in our Lexicon that the verb cook takes FOR in front of its indirect
object. This is a special property of cook. We do not need to specify that the verb give
takes TO: our system will assume that TO is the unmarked case and it will always
assign TO to an indirect object unless there is specific information in the lexical entry
that it must do otherwise. We will discuss Indirect Objects in more detail in Chapter 4.

(18) shows three lexical entries. The last two entries show irregularities that need to be
marked in the lexical entry as they do not follow the regular rules of the language.
Structures likes (18) represent our "knowledge" of verbs. Thus we see that when we say
we "know" the verb to cook we just do not know its meaning, we know all the
phonological, morphological, syntactic and semantic properties that make it possible for
this verb to combine with other elements to form phrases and then sentences.

(18) Building a lexical grid (Part 6)

visit write cook

categorization→ V V V
c-selection → _ NP _ NP NP _ NP NP
s-selection → AG,PAT AG,PAT,REC AG,PAT, REC
meaning → 'VISIT' 'WRITE' 'COOK'
phonology → /ˈvɪzɪt / /raɪt/ /kʊk/
special props. → - past: wrote indirect object: FOR
pp: written
 English Syntax & Universal Grammar 29

PRACTICE 5

Write the LEXICAL ENTRY for the underlined verbs:

a. She put the flowers in the vase.
b. Ha brought flowers from Dalat.
c. Ha made a cake for her boyfriend.
d. The teacher explained the lesson to the students.
e. Khiem swam 2 miles yesterday.

2.1.7 Lexical entries for other categories. In the previous sections we saw how to
capture our knowledge of verbs. In this section we will discuss other categories. Let us
start with nouns.

Consider the nouns girl, boys and syntax. What do we know when we say that we
"know" these nouns? We know their syntactic category: they are nouns. This
knowledge will allow us to combine them with adjectives and determiners to form noun
phrases. We also know their gender: we know that girl is feminine, boy is masculine
and syntax is neuter. This is why when we replace these words by a pronoun we get the
pronouns, he, she and it, respectively. We know their number: we know that girl is
singular and boys is plural. We also know whether they are countable or uncountable,
thus boy and girl are countable nouns while syntax is uncountable or a mass noun. This
will allow us to choose many as opposed to much when we want to include this
quantifier in our structure. We know whether they are human or non-human, which
allows us to be able to use boy and girl as agents, but not as instruments. We also need
the property animate which will help us differentiate humans/animals from objects.
Some verbs are sensitive to this feature. For example, the verb think applies only to
humans, but the verb eat applies to animates, including human. Of course we know their
meaning and phonological information. As with verb, the lexical entry or lexical grid
will need to include any special properties the noun may have. We show the lexical
grids or lexical entries for these three nouns in (19) below:

(19) girl boys syntax foot

Categorization N N N N
gender [feminine] [masculine] [neuter] [neuter]
± Count [+count] [+count] [-count] [+count]
number [singular] [plural] --- [singular]
animate [+animate] [+animate] [-animate] [-animate]
± Human [+human] [+human] [-human] [-human]
30 Héctor Campos, Bùi Huỳnh Thủy Thương

meaning 'GIRL' 'BOY' 'SYNTAX' 'FOOT'

phonological /gɜrl/ /bɔɪz/ /ˈsɪntæks/ /fʊt/
information
special inform. pl.: feet

Some nouns may take complements: linguistics student, author of the book (the
underlined elements indicate the complements). This information will have to be
specified in the lexical entry as a c-selection property, similar to the way we did for
verbs in section 1.2.

PRACTICE 6

I. Using the features in (19), write a lexical grid for the following nouns:
a. professor b. problems c. sincerity
II. Determiners and quantifiers are typically sensitive to features found in nouns.
What features are the following elements sensitive to:
a. many b. some c. one d. every e. all

Adjectives, like verbs, are considered logical predicates and they need other elements
(arguments) to complete their meaning. They assign a property to their arguments. Take
smart, for instance: it requires a noun phrase so it can assign the property of "being
smart". When smart means "intelligent" it needs a human NP, unless we extend the
property of having intelligence to machines, as in smart phones, smart refrigerator, etc.
The lexical entry for smart will also need to specify its full meaning and phonological
information.

Some adjectives may also take complements. Thus the adjective proud needs two
arguments: the person who is proud and something/someone to be proud of (Tuoi is
proud of her nephews).

Prepositions are also arguments. They typically take complements: under the table, on
the table. Notice that we need a second argument to assign the property of being under
or on the table. Prepositions have meaning. So this meaning will have to be specified in
the lexical entry as well.

Affixes will also have to be specified in the Lexicon. For instance, the –ed affix we need
to form the past tense in English will have to be specified as combining with a verb and
having the property of indicating past tense. As we will see in the next section, there
 English Syntax & Universal Grammar 31

will be three affixes –s in English: one to indicate third person singular present
(IIISG,PRES), another to indicate plural (PL) and a third one to indicate possession
(POSS). This latter one is typically written as 's when we write.

(20) -s -s -s

IIISG PL POSS

PRES

V- N- NP-

As we can see in (20), the first affix –s combines with verbs, the second one combines
with nouns and the third one with noun phrases.

2.1.8 Summary of the Catalog. We have seen that the first component of the
Lexicon, the CATALOG, contains a list of lexical entries that include all the
phonological, morphological, syntactic and semantic information necessary for us to
combine these elements in the syntax. The second component of the Lexicon, the
EXTENDED CATALOG will contain the words formed from the basic elements
found in the Catalog. This is where GENERATIVE MORPHOLOGY takes place.
Generative morphology expands the catalog into the extended catalog by creating
all the words that we will need in our syntactic derivations.

(21) The Structure of our MENTAL LEXICON

Catalog GENERATIVE MORPHOLOGY Extended Catalog

In the next section we will discuss the basic operations that take place going from the
Catalog to the Extended Catalog.

2.2 The Extended Catalog: A brief introduction to Morphology

As we saw in the Chapter 1, morphology is the study of words. The minimal unit in
morphology is the morpheme. Our "knowledge" of morphology includes knowing the
form (i.e. the sound) and the meaning of the morphemes that belong to a particular
language. We claimed in the previous section that all the elements necessary to create
words, phrases and sentences are in the Catalog. These basic elements are morphemes.
32 Héctor Campos, Bùi Huỳnh Thủy Thương

Thus all morphemes of a language are listed in the Catalog accompanied by their
phonetic, phonological, morphological, syntactic and semantic information.

Morphemes can be bound or free. Free morphemescan stand on their own. On the
other hand, bound morphemes must attach to roots or bases to form words. In the
following sections we will discuss how Generative Morphology makes use of those
roots or stems and bound morphemes to create new words. We discuss free morphemes
in section 2.1 and bound morphemes in section 2.2. Section 2.3 will discuss how
compounds are formed.

2.2.1 Free morphemes. Free morphemes can belong to open or closed classes of
words. The open class of words includes content or lexical words; these typically
include verbs, nouns, adjectives and adverbs. The closed class of words includes
function or grammatical elements. These function or grammatical words include
prepositions, determiners, pronouns, auxiliary verbs, modals, etc.
a. Content or {verbs, nouns, adjectives, adverbs}
(22) lexical words
(open class)
FREE MORPHEMES

b. Function or {prepositions, determiners, pronouns,

grammatical auxiliaries/modal verbs}
elements
(closed class)

Thus the words eat (V), boy (N), Ha (N), nice (A), freely (ADV), under (P), the (D), she
(pronoun, N), may (modal, I), have (AUX), etc. are all free morphemes in English.

2.2.2 Bound Morphemes and Derivational Morphology. Bound morphemes

cannot stand alone. Derivational morphology studies the way that these bound
morphemes interact with other morphemes to form words.

Bound morphemes are affixes. Affixes must attach to an appropriate root to be

licensed. If an affix precedes the root, we say it is a prefix (ex., mis-understand, un-
happy, etc.). If the affix follows the root, we say it is a suffix (dog(g)-y, teach-er, work-
ed, work-ing).
 English Syntax & Universal Grammar 33

Bound affixes can be derivational or inflectional. We will consider these in turn.

Section [Link] will discuss inflectional affixes. Section [Link] will discuss
derivational affixes.

[Link] Inflectional affixes. English has eight inflectional morphemes. These

morphemes are always bound affixes, that is, they must always attach to another
element to be licensed. Inflectional morphemes in English may attach to verbs, nouns
and adjectives, as shown in (23). Examples for each kind are given on the far right of
each morpheme:

(23) a. Verbal affixes: -s: third person singular, present indicative (works, is, has, puts)
-ed: past tense (worked, wrote, put)
-en: past participle (worked, written)
-ing: present participle or gerund (working, writing)
b. Noun affixes: -s : plural (cars)
-'s: possessor (John's)
c. Adjectival affixes:-er: comparative (smarter)
-est: superlative (smartest)

All of these affixes are attached to a root in the Extended Catalog, with three
exceptions that will happen in the syntax. The three exceptions are –s (present), -ed
(past) and –'s (possessor). The Chart in (24) shows how and where the categories that
combine with these affixes are formed:

(24) Inflectional Affixes in English

CATALOG EXTENDED CATALOG SYNTAX

a. past participle
work, -en work + en → worked
write, -en write + en → written
b. gerund
work, -ing write + ing → writing
c. plural
car, -s car + -s → cars
foot, -s foot + -s → feet
wife, -s wife + -s → wives
shrimp, -s shrimp + -s → shrimp
d. comparative
smart, -er smart + -er → smarter
good, -er good + -er → better
34 Héctor Campos, Bùi Huỳnh Thủy Thương

e. superlative
smart, -est smart + -est → smartest
good, -est good + -est → best
f. third person singular
present tense
work, -s → → → → → → → → work + -s → works
g. past tense
work, -ed → → → → → → → → work + -ed → worked
write, -ed → → → → → → → → write + -ed → wrote
h. possessive
John, -'s → → → → → → → → John + 's → John's
the, car, -'s → → → → → → → → the + car + -'s → the car's

As you can see in (24), our Catalog includes all the necessary elements to form words
and phrases. Past participles (24a), gerunds (24b), plurals (24c), comparatives (24d)
and superlatives (24e) are formed in the Extended Catalog. Thus we can use these
words directly in our syntax since these words have already been formed prior to our
syntactic operations.

Morphological structures do not care if the final forms we get are regular or
irregular. All forms are treated as regular in their morphological structures. Consider
the plural nouns cars and feet. We will label inflectional suffixes as "Infl" in
morphological structures:

(25)
a. cars b. feet

As we can see in (25), the morphological structure just contains the original root and
the inflectional affix. When phonology spells out (i.e. pronounces) these words, it
will spell out cars and feet, respectively. We thus see that the irregularity observed in
plural forms is not represented in the morphological structure. We observe a
similar system with adjectives, as shown in (26):
 English Syntax & Universal Grammar 35

(26)
a. smarter b. better c. more intelligent

As with nouns, the morphological structure of regular and irregular forms will be
the same. Our phonological system will spell-out good + -er as better and it will spell
out intelligent + -er as more intelligent.

How does the morphology know that feet, better and more intelligent are exceptions to
the rule of plural and comparative formation? As we claimed in section 1.6 of this
chapter, the respective lexical entries for these words include that information. So when
the rule of plural and comparative formation applies, it will know not to apply the
generative rule of adding –s and –er to these forms, but rather, it will take these
irregular forms from their respective lexical entries.

As we can see in the chart in (26), present tense, past tense and possessive
morphemes are not attached until the syntax. Although this may seem obscure at this
point, this will become clearer in the next chapters as we develop our syntactic model in
the next chapters.

We mentioned at the beginning of this section that bound affixes can be either
inflectional or derivational. In the next section we will discuss derivational affixes.

PRACTICE 7

Diagram the morphological structure for the following words:

a. computers
b. deer (plural)
c. whiter
d. brightest
e. worse
f. written
36 Héctor Campos, Bùi Huỳnh Thủy Thương

[Link] Derivational Affixes. We saw in the previous section that inflectional

affixes add grammatical information regarding tense (present, past), number (plural),
degree (comparative, superlative), etc. When we add an inflectional affix, we do not
create a new word, we simply add grammatical information. On the other hand,
when we add a derivational affix to a root, we create a new word with a different
meaning. As we saw in the previous section, inflectional affixes are always suffixes in
English, in other words, they always follow the root. Derivational affixes may be
suffixes or prefixes. The verb misunderstand and the adjective unhappy contain
prefixes. Their morphological structures are shown in (27):

(27) Morphological structures with prefixes

a. misunderstand b. unhappy

As we can see in (27), prefixes typically do not affect the original category of the
root or base. Adding the prefix mis- to the verb understand, we still get a verb:
misunderstand. Adding the prefix un- to the adjective root happy, we still get an
adjective as a result.

Suffixes, on the other hand, may (but do not necessarily) change the original
category of the root. Consider the morphological structure of the words doggy and
teacher:

(28)
a. doggy b. teacher

As we can see in (28), the diminutive suffix –y does not change the original category of
dog; doggy is still a noun.1 On the other hand, as we can see in (28b), the suffix –er

1
The extra -g- that appears in the orthography is just an artifact of our spelling system to avoid
pronouncing -gy as [ʤɪ] , as in the ending in morphology.
 English Syntax & Universal Grammar 37

changes the original category of the verb teach to the noun teacher. Thus we see that a
suffix may, but does not always, change the original category of a root.

It is possible to find more than one suffix at a time. Consider the word friendliness:

(29) friendliness

As we can see in (29), the original noun category of the noun friend was changed to an
adjective by the suffix -ly, which in turn was changed back to a noun by the suffix -ness,
giving us the noun friendliness.
Nothing prevents us from having a prefix and one or more suffixes, as shown in (30):

(30) unhappiness

We must be careful when we combine different morphemes. In (30) we formed the

adjective unhappy first and then added the suffix -ness rather than the other way around.
This is because the prefix un- typically attaches to adjectives and not to nouns (ex.:
uncommon, unequal, unfair, etc.). Another case where caution must be exercised is
deciding the category of words that belong to different categories. Take, for instance,
the derived word helpfulness. Should we consider help a noun (as in your help) or a
verb (as in Help me!)? To decide, we need to look at many examples that have the suffix
-ful and see if there is a generalization regarding what elements -ful typically attaches
to. If we do look at many examples, we will see that -ful typically attaches to nouns.
Thus helpfulness will have the structure in (31):
38 Héctor Campos, Bùi Huỳnh Thủy Thương

(31) helpfulness

2.2.3 Bound roots. We have seen that derivational affixes can be prefixes or suffixes.
There is yet a third class of derivational bound affixes that are neither prefixes nor
affixes. We will call these forms bound roots. They are problematic for morphological
analysis because, contrary to what we observed for other morphemes, these roots do not
have a meaning on their own; they get their full meaning only after attaching to other
morphemes. Some examples are -ceive (conceive, receive, deceive, etc.), -mit (admit,
commit, transmit, etc.), -fer (transfer, defer, refer, etc.), etc. We will represent these
bound roots as in (32):
(32)
a. conceive b. transmit

Derived words can serve as input for inflectional morphology. Consider the
morphological structures for the plural noun teachers, the comparative adjective
unfriendlier and the plural noun transmissions below:

(33)
a. teachers b. unfriendlier
 English Syntax & Universal Grammar 39

c. Transmissions

We can see in (33) that the inflectional affixes always attach after all derivational
affixes.

Summarizing this lengthy section, we have seen that bound affixes can be inflectional
or derivational. While inflectional affixes in English are always suffixes,
derivational affixes can be either prefixes or suffixes. This is summarized in (34):

(34)

PRACTICE 8

Diagram the morphological structure of the following words:

a. piglets f. untie k. tastier
b. reddish g. unusual l. arrivals
c. minibuses h. unuseful m. most unhappy
d. redness i. beautification n. mistreatment
e. painters j. organizers o. unsystematically
40 Héctor Campos, Bùi Huỳnh Thủy Thương

The chart in (35) summarizes our discussion in sections 2.1 and 2.2. As an exercise, try
to find examples of each type of morpheme in that chart.

(35)

2.2.4 Compounds. English has a rich system to create compound words. Two or more
words may be joined to form a compound. When two words belong to the same
grammatical category, the result is always the same category: boyfriend, red-hot, etc.
(36) a. boyfriend b. red-hot c. sleepwalk

We can also find compounds formed by different categories, as shown in (37):

(37) a. sky-high b. hand-wash c. blackboard

In English, typically the rightmost element is the head of the compound. In (37a) we
are talking about "high", in (37b) we refer to "wash", in (37c) we are talking about a
 English Syntax & Universal Grammar 41

"board". When the head is inside the compound, we say that the compound is
endocentric. Thus the compounds in (36-37) are all endocentric. If the head is not in
the components of the compound, we say that the compound is exocentric. Pickpocket
and skinhead are examples of an exocentric compound:

(38) a. pickpocket b. skinhead

A pickpocket is not a kind of pocket nor is a skinhead a kind of head (cf. blackboard).
The two compound words in (38) are thus exocentric.

Nouns are easier to compound than other categories. Notice the structure of the complex
compound nouns in (39):

(39) a. train station b. train station coffee

c. train station coffee house d. train station coffee house owner

PRACTICE 9

Diagram the following compounds:

a. bathroom
42 Héctor Campos, Bùi Huỳnh Thủy Thương

b. bathroom towel
c. bathroom towel collection
d. bathroom towel collection set
e. Turkish cotton towel collection sets
f. printer ink
g. laser printer ink cartridges
h. dry-cleaner
i. overcook
j. sing-along

2.3 Summary

In this chapter we looked at the structure of our mental Lexicon. We proposed that the
Lexicon has two components: a Catalog and an Extended Catalog. The Catalog
contains a list of all the morphemes necessary to create words in a language. Generative
Morphology creates these words and stores them in the Extended Catalog making
them available for the syntax, which is responsible for putting words together to
create phrases and sentences.

The morphemes in the Catalog will contain all the phonetic, phonological,
morphological, syntactic and semantic information we will need in the syntax. All this
information is included in the lexical entry or lexical grid. Some of this information
necessary for the syntax is the categorization, c-selection, s-selection, and the
semantic restrictions. The lexical entries also contain information about the meaning
and the phonetics/phonology for these morphemes as well as any special properties or
exceptions that these roots or morphemes may undergo. Three principles are necessary
to guarantee that a verb respects its c- and its s-selection in the syntax:

(40) The Projection Principle

The c-selection of a verb must be respected at every syntactic level.

(41) Extended Projection Principle

Every sentence must have a subject.
 English Syntax & Universal Grammar 43

(42) The Theta-Criterion

Each argument must receive one and only one theta-role and each theta-role
must be assigned to one and only one argument.

Generative Morphology may form words by attaching inflectional morphemes to

roots or to derived words. Inflectional morphemes add grammatical or functional
information; in English, inflectional morphemes are always suffixes, that is, they follow
the word. Inflectional affixes do not create new words. Generative morphology also
makes use of derivational morphemes adding prefixes and suffixes to roots or stems to
create new words.

Generative morphology also creates compounds by combining words and creating new
words. Compounds are endocentric if the head of the compound is inside the
compound itself. Compounds are exocentric if the head is not found in the compound.

The Lexicon is one of the six modules we will find in our syntactic model. The next
chapter will discuss the interaction of the Lexicon with other modules in order to derive
phrases and eventually, sentences.
 English Syntax & Universal Grammar 43

CHAPTER 3

THE PRINCIPLES AND

PARAMETERS MODEL AND X'-SYNTAX

Remember that our objective in theoretical linguistics is to create a model for I-

language. The basic relation we are trying to capture in our syntactic model is how a
particular sequence of words in the correct order can have meaning. To explain this
we will propose a model known as the Principles and Parameters model. This model
was introduced by Noam Chomsky in 1981 in his book Lectures on Government and
Binding and has been revised since then to what now we call the Minimalist Program.

3.1 The modular system of the Principles and Parameters model

The Principles and Paramenters model has 6 modules: the Lexicon, X'-syntax, D-
Structure, S-Structure, Phonetic/Phonological Form (PF) and Logical Form (LF). The
system is shown in (1):

(1) X’-Syntax Lexicon

D-Structure

S-Structure

PF LF

Phonetic/Phonological Form (PF) interfaces with our sound-creating system (our

Sensory-Motor system) while Logical Form (LF) interfaces with our system of meaning
44 Héctor Campos & Bùi Huỳnh Thủy Thương

(our Conceptual-Intentional system). Thus we see that sound and meaning are
connected, but this connection is not direct; it is mediated by other modules.

In Chapter 1 we studied the structure of the LEXICON, which is the first module at the
top right. We saw that the Lexicon had two components: a Catalog and an Extended
Catalog. All the morphemes necessary to create words are stored in the Catalog.
Generative Morphology creates words combining these morphemes in the appropriate
order. The result of this is that the Lexicon will provide the other modules in (1) with
the words that are necessary to create structures. As we saw in the previous chapter,
only tense (present and past) and possessive –'s morphology are not attached in the
Lexicon. They will be attached in the syntax.

Where do the structures come from? They come from the first module on the left: X'-
syntax.

3.2 Introduction to X'-syntax: The basic architecture of syntactic structures

X'-syntax is the module that will provide us with the structures that we will use to
create phrases and sentences. X'-syntax has a very simple model or template to create
structures. X'-syntax will distinguish three functions: complement, modifier and
specifier. For structures that have no complements, modifiers or specifiers we will use
the structure in (2):

(2) minimal structure, with no complements, modifiers or specifiers

X in (2) can represent any category: verb, noun, adjective, preposition, etc. We will
show how this works with the individual categories in the next chapters.

We say that X is the head of XP. We call XP the maximal category. X' is the
intermediate category.

For structures that have complements, complements will be merged or generated (i.e.,
added) under X'. If a structure has more than one complement, they will all be merged
under X', as shown in (3b):
 English Syntax & Universal Grammar 45

(3)
a. Structure with one complement b. Structure with two complements

As we can see in (3), if a head takes one complement, we merge it under X'. If a head
takes two complements, both complements are merged under X'.

Structures may also include modifiers. To distinguish modifiers from complements, we

will add one additional intermediate X' for each modifier, as shown in (4). When the
structure contains no modifiers, we will only include one intermediate X', as shown in
the structures in (3).

(4)
a. Structure with one modifier b. Structure with 2 modifiers

Notice that we have not attached anything to the lowest X'. This is because this structure
has no complement. In (4b), each modifier is merged under a separate X'. As we saw
in (3), when the structure has no modifier, we do not project any additional intermediate
X'-categories.

If our structure were to have one complement and two modifiers, we would have the
structure in (5):

(5) Structure with one complement and two modifiers

46 Héctor Campos & Bùi Huỳnh Thủy Thương

Finally, structures may also contain specifiers. Specifiers are merged directly under XP:

(6) Structure with one specifier

A specifier may be added to any of the structures we have shown above.

Summarizing, the basic structure that we will assume throughout this book is shown in (7):

(7) Basic X'-syntax schema

In (7) we see that complements are merged under the lowest X'. If more than one
complement, they will all be merged under the lowest X'. We can see that modifiers are
optional. We only project an additional intermediate category if we have modifiers.
Different from complements which are all generated under the first X', modifiers are
generated in additional X' projections, one modifier per additional X'-projection.
This is what the asterisk "*" means in (7). Think of modifiers as being "fat". Thus only
one fits in the additional X'-projection that we create for them when they are present.
Finally, specifiers are generated immediately under XP.
 English Syntax & Universal Grammar 47

PRACTICE 1

Diagram the following structures. Just replace X by the appropriate category using
the correct diagram:
a. a noun phrase with a specifier
b. a noun phrase with a specifier and a modifier
c. a verb phrase with two complements
d. a prepositional phrase with a complement and a specifier
e. an adverb phrase with no complements, modifiers or specifiers (a bare adverb)

3.3 From the Lexicon and X'-syntax to D-Structure and S-Structure

Going back to our diagram in (1) we can see that the two modules at the top interact to
create a level of syntactic representation that we call D-Structure. Models previous to
the Principles and Parameters model that we are studying here used to have a level of
deep structure and another level of surface structure. These levels, however, were
revised considerably in Chomsky (1981)'s Lectures on Government and Binding, so it is
better to refrain from using that old terminology. We could have called them "Level 1"
and "Level 2" since they are so different from the original levels, but we will use the
names that are typically used in the syntactic literature.

How do we form D-Structures? The Lexicon creates the words and X'-syntax creates the
structures. So we must insert our words from the Lexicon into these structures created
by X'-syntax. The principle that regulates this insertion is called the Lexicalization
Principle:

(8) Lexicalization Principle

Insert words created in the Lexicon into appropriate structures created by X'-syntax,
respecting all syntactic specifications of the lexical entry.

Incorporating this principle into our modular system we would get (9):
48 Héctor Campos & Bùi Huỳnh Thủy Thương

(9)
X’-Syntax Lexicon

Lexicalization Principle

D-Structure

S-Structure

PF LF

D-Structureswill be "transformed" into S-Structures by different operations. As we

will see in the following chapters, these transformations will allow us to perform only
three operations: insert, move and delete. These three basic operations are also known
as AFFECT ALPHA, where AFFECT can be any of the three aforementioned
operations (insert, move or delete), and ALPHA can be any element in the structure as
long as that element is a unit or constituent. In this text we will use the general term
TRANSFORMATIONS to refer to these three operations. We will see that by applying
these basic operations to our D-Structures we will be able to generate many
constructions in English and in other languages. We will study these transformations or
operations in the following chapters. Some of the constructions we will study in detail
are: indirect objects, pre- and post-nominal complements, adjective modification,
adverbial modification, modal and auxiliary verbs, the dummy verb do, phrasal and
prepositional verbs, specific questions, wh-questions, relative clauses, etc. (10) shows
the basic structure of our model.

(10) X’-Syntax Lexicon

Lexicalization Principle

D-Structure

Transformation: Insert, Move, Delete

S-Structure

PF LF
 English Syntax & Universal Grammar 49

3.4 Constituents

Before we start exploring the structure of phrases, you must learn to distinguish the
different groups of words that form larger units called constituents. You already have
this knowledge. Consider (11):

(11) The teacher wrote the example on the board.

How many units larger than a word (constituents) can you find in (11)? Try to do this
exercise before you read on!

In (11) we would have at least 7 constituents. Some of them are noun phrases (NP),
others are prepositional phrases (PP), others are verb phrases (VP) and finally we have a
sentence (S).

(12) Constituents of the sentence in (11)

[NP the board] [PP on the board] [NP the example] [VP wrote the example]
[VP wrote the example on the board] [NP the teacher]
[the teacher wrote the example on the board]

How do we know that the elements in brackets in (12) are in fact constituents?
Constituents are independent units. As such we are able to move them around, to
replace them with pronouns, to replace them with interrogative words. Consider the
examples in (13):

(13) a. The teacher wrote the example on it. It = [NP the board]
b. Where did the teacher write the example? Where = [PP on the board]
c. [NP The example], the teacher wrote on the board.
c. Who wrote the example on the board? Who = [NP the teacher]
d. What did the teacher do on the board? What = [NP wrote the example]
e. What did the teacher do? What =[VP wrote the example
on the board]
f. What happened? Happen= [the teacher wrote
the example on the board]

Sometimes its possible for a constituent to be analyzed as being part of different

constituents:
50 Héctor Campos & Bùi Huỳnh Thủy Thương

(14) a. She sent postcards from Spain.

b. She sent [NP postcards [PP from Spain]]
c. She sent [NP postcards] [PP from Spain]

As we can see in (14), the PP [PP from Spain] can either be part of the NP with postcards
(so what she sent were postcards from Spain) or modify the verb send (the country
where she sent the postcards from was Spain). We will look at these structures in more
detail in the next couple of chapters.

PRACTICE 2

I. Watch the video by Dr. Caroline Heycock, "Generative Syntax 1.2: On

constituency" ([Link] and then answer
the following questions:

1. What is the main point of the video?

2. What is a "polar question"? Give two original examples.
3. What is a constituent?
4. What argument does Dr. Heycock use to prove that sentences have internal
structure?
5. Show all the constituents in the following sentences:
a. Khuong read the book.
b. Nguyen bought a nice present for his girlfriend.
c. Phuoc sent flowers to his mother yesterday.

II. Watch the video by Dr. Caroline Heycock, "Generative Syntax 1.3: Constituent
Recursion" ([Link] and then answer
the following questions:
1. What is the main point of this video?
2. What is recursion?
3. Give an original example of recursion where a sentence is recursive.
4. Give an original example of recursion where a noun phrase is recursive.
5. Why can't we find sentences that have infinite recursion?
 English Syntax & Universal Grammar 51

3.5 Summary

In this chapter we explored the basic architecture of our syntactic model. We saw that
our model is modular. Each module has its own characteristics and rules. We saw that
the Lexicon and X'-syntax interact (via the Lexicalization Principle) to produce D-
structures. These D-Structures are transformed into S-Structures by three basic
operations (usually referred to as "Affect α"): insert, move and delete. S-Structure then
maps onto Phonetic/Phonological Form, which interfaces with our sound
production/perception system (also called Sensory-Motor system) and, at the same
time, it maps onto Logical Form, which interfaces with our mental system that
assigns meaning (called the Conceptual-Intentional system). This is the way that our
model connects sound with meaning.

We also saw the basic architecture that our structures will have. We saw that the
minimal structures will have a minimum of three levels: a head, an intermediate
category and a maximal category, which we represented as X, X' and XP,
respectively. We saw that complements will be merged under X'. If a structure has
modifiers, they will be merged in additional X' projections, one per modifier.
Finally, specifiers are merged directly under XP.

X in X'-syntax stands for any of our syntactic category. In the next chapters we will
show that the basic template or schema in (7) can be applied to the different categories
to form phrases and, eventually, full sentences.

CHALLENGE

I. Watch the video by Dr. Caroline Heycock, "Generative Syntax 2.1: Substituting
Constituents" ([Link] and then answer
the questions.

1. What is the main point of this video?

2. Go to googletranslate ([Link] and translate the following
sentences from English into Spanish:

1. Juan's brother is tall. 4. John's sisters are tall.

2. John's sister is tall. 5. John's brothers and sisters are tall.
3. John's brothers are tall. 6. John's sisters and brothers are tall.
52 Héctor Campos & Bùi Huỳnh Thủy Thương

(a) Explain the different forms that the adjective "tall" takes in Spanish.
(b) The definite articles in Spanish are EL, LA, LOS, LAS. Explain how these are
used in the examples above.
(c) What are the main differences between English and Spanish regarding possessive
constructions?
(d) Are Vietnamese possessive constructions similar to English, Spanish or neither?

3. Consider the sentence:

Flying planes can be dangerous.

(a) Is this sentence ambiguous? Explain how. Translate into Vietnamese.

(b) How can verbal agreement show that the sentence above is ambiguous?

4. Consider the three following sentences:

a. I know [the girl who is going out with Quan].

b. I know her.
c. I know the girl who is going out with him.

What phrases do the pronouns him and her replace? Explain why this is possible.

5. Consider the following sentence:

Tuoi came to Saigon in 2008.

Replace the two prepositional phrases by a pro-form. What does this replacement
show?

6. a. Adjectives can be attributive or predicative. Give an example of each type.

b. For some speakers "so" can replace an adjective. Give an original example.
Does this test apply to both types of adjectives you described in (a)? Justify your
answer.

7. What does the form "do so" replace? Give your own examples.

II. Watch the video by Dr. Caroline Heycock "Generative Syntax 2:2: Moving
Constituents" ([Link] and then
answerthe questions.
 English Syntax & Universal Grammar 53

1. What is the main point of this video?

2. What is topicalization? What is it used for? Give an original example of
topicalization.
3. Give an original example of a phrase that cannot be replaced by a pro-form but
which can be topicalized.
4. Give original examples of NP, PP, AP and VP that have been topicalized.
5. Give an original example in English where you have applied replacement and
displacement for NP, AP, PP and VP. What does this operation prove?
54 Héctor Campos & Bùi Huỳnh Thủy Thương

CHAPTER 4

THE VERB PHRASE

In Chapter 2 we saw that there are different syntactic categories: nouns (N), determiners
(D), adjectives (A), verbs (V), adverbs (ADV), prepositions (P), auxiliary verbs (AUX),
modal verbs, negation (NEG), quantifiers (Q), etc. In this chapter we will learn how
verbs combine or merge with other categories. Let us first learn how to combine verbs
with their complements.

4.1. Complements of the verb phrase

In this chapter we will consider 5 types of verbs: (a) intransitive verbs, (b) transitive
verbs, (c) prepositional verbs, (d) verbs that take NP-PP complements and (e)
ditransitive verbs. Other types of verbs will be considered later in this book.

4.1.1. Intransitive verbs. An intransitiive verb is one that takes no complements, as in

the examples in (1). The corresponding structures for (1a) and (1b) are shown below.
Notice that we do not include tense with the verb in the trees below. In the tree
structures below we incorporate verbs in the simple infinitive form (the infinitive form,
but without the particle TO; thus for the verb to sleep, we enter sleep).

(1a) John sleeps. (1b) Tom worked.

4.1.2. Transitive verbs. Transitive verbs are those that take a direct object, also
called a direct complement. As we saw in Chapter 2, direct objects typically bear the
semantic or thematic role of PATIENT. The structures for the verbs in (2a, 2b) are
shown below. We will draw a triangle to represent the internal structure of a phrase that
we have not discussed yet. As in the structures in (1), notice that we do not include
tense in the VP:
 English Syntax & Universal Grammar 55

(2a) Mary baked a cake. (2b) Susan studied the lesson.

PRACTICE 1

Draw the structure for the VP in the following sentences:

(a) Mary opened the door.
(b) Susan sneezed.

***Remember to write the verbs in their simple forms.

4.1.3. Prepositional verbs. Consider prepositional verbs next. These are verbs that
take prepositional complements. Consider the examples in (3a, 3b):

(3a) Alice talked about linguistics. (3b) We complained about the work.

4.1.4 Verbs with NP-PP complements. Some verbs may take two complements.
Consider verbs that take NP + PP complements, as those shown in (4). Notice that all
complements are merged (or generated, or inserted) under V':

(4a) Quan withdrew money from the bank. (4b) Ha put the books on the table.
56 Héctor Campos & Bùi Huỳnh Thủy Thương

PRACTICE 2

Diagram the VPs for the following sentences:

a. Mark dreamt about his girlfriend.
b. Susan took the letter to the post office.
c. Tom lives in Saigon.

4.1.5. Ditransitive verbs. Ditransitive verbs are those that take two NP-complements,
a direct object and an indirect object. Typically the complements may appear in either
order:

(5) a. John sent some flowers to his girlfriend.

(direct object) > (indirect object)

b. John sent his girlfriend some flowers

(indirect object) > (direct object)

In (5) we can see that the verb SEND takes two complements: a direct object (the
PATIENT, what gets sent) and an indirect object (the RECIPIENT, the one who
actually "receives" the direct object). Since these constructions typically involve a
recipient, they appear with verbs of transfer. But there are two strange facts about the
examples in (5). First, we see that the direct object may precede (5a) or follow (5b)
the indirect object. Second, we also see that when the indirect object follows the
direct object, as in (5a), the indirect object is preceded by the "word" TO.

The question we need to tackle first is whether the "word" TO that we see in front of the
indirect object in (5a) is a preposition or not. If it turns out not to be a preposition, we
need to decide what it is. Compare sentence (5a), repeated here as (7a), with sentence
(6a), which contains the preposition TO:

(6) a. I went [to Paris].

b. I went [there].
(7) a. John sent some flowers [to his girlfriend].
b. *John sent some flowers [there].

As we can see in (6), the prepositional phrase [to Paris] may be replaced by the adverb
[there]. This is the typical behavior of the preposition TO when it implies movement or
direction. The preposition TO adds meaning to the sentence. Although TO in (7a) may
also imply some sort of movement or direction since the indirect object is typically a
 English Syntax & Universal Grammar 57

RECIPIENT, this replacement of the TO-phrase by the adverb THERE is not possible in
(7), as shown (by the asterisk in front of the sentence) in (7b), thus showing that the
indirect object in (7a) is not a prepositional phrase, as the one in (6a). So, if TO in
(6a) is not a preposition, what is it?

We will assume that the c-selection for the lexical entry for SEND is [ _ NP NP]. This
c-selection captures the two constructions in (5), which have the same meaning. We do
not want to claim that the indirect object is a PP complement in (5a) but an NP-
complement in (5b). Ideally, the indirect object should have the same category in both
(5a) and (5b)!

To understand the analysis that we will propose here, we need to consider a language
like Japanese. Japanese is a language that "marks" syntactic functions distinctly: -WA
is used to mark the subject (SUBJ) or topic of a sentence, -O is used to mark the direct
object (DO) of a sentence and -NI is used to mark the indirect object (IO) of a sentence.
Apart from marking those particular syntactic functions, these markers do not have any
meaning or further role in the syntax of Japanese:

(8) John-wa Mary-ni tegami-o okurimas-ita

John-SUBJ Mary-IO letter-DO send-PAST
'John sent a letter to Mary.'

We will analyze the word TO that appears in front of indirect objects as an INDIRECT
OBJECT MARKER, similar to the Japanese markers shown in (8). As we can see in
(5), this marker appears when the indirect object does not appear adjacent to the verb. If
the indirect object follows the verb directly, then TO does not appear.

We will say that ditransitive verbs always take two NP-complements, whether the
direct object precedes or follows the indirect object. We can see this in (9a) and (9b)
below. In English, if the indirect object does not immediately follow the verb, the
indirect object is marked by TO. If the indirect object is adjacent to the verb, no
marker is necessary.

In Chapter 2 we saw that insertion is one of the transformation operations that

transform D-Structures into S-Structures. We will claim that TO is the result of an
insertion operation in English. We say that the TO that precedes an indirect object is a
pseudo-preposition. Prepositions form PPs and they appear as such at D-
Structure. Pseudo-prepositions do not form PPs, they do not appear at D-Structure
and they are inserted into the structure. In the case of the pseudo-preposition TO,
TO is inserted to mark that a noun phrase is an indirect object. Thus we claim that
58 Héctor Campos & Bùi Huỳnh Thủy Thương

TO is an INDIRECT OBJECT MARKER. The structures for (5), repeated here as

(9a, 9b) are shown below. We will revisit these constructions in Chapter 5, after we get
more background on the licensing of noun phrases. For the moment, we will indicate
this insertion process by putting the marker in parenthesis in the structure and by
inserting a double arrow under it to indicate that this element has been inserted:

(9a) John sent some flowers to his girlfriend.

D-Structure S-Structure

(9b) John sent his girlfriend some flowers.

D-Structure and S-Structure:

Japanese would have the same structure as English, with the exception that in Japanese
the verb projects its complements to the left. We say that Japanese is a verb-final or
an SOV language since the order is typically subject (S) – object (O) – verb (V), with
the verb at the end of the verb phrase. As in English, at D-Structure, Japanese would
have no markers. Markers are inserted at the level of S-Structure. The structure for
Japanese is shown in (10):

(10) John-wa Mary-ni tegami-o okurimas-ita

John-SUBJ Mary-IO letter-DO send-PAST
'John sent a letter to Mary.'
 English Syntax & Universal Grammar 59

D-Structure: S-Structure:

We see that at D-Structure, Japanese has no object markers (just as English 9a, 9b).
Object markers are inserted at S-Structure. When discussing structures in this book, we
will typically use the S-Structure representation since it is relatively easy to read off
both the D- as well as the S-Structure from it. The arrows indicate the elements that
have been inserted at S-Structure. If we eliminate those elements we can read off the D-
Structure. If no elements are inserted (as in 9b), that means that the D- and S-Structure
are the same.

[Link]. Types of indirect objects in English. In (9a) we saw an example of an indirect

object preceded by the marker TO. The examples in (11) show some verbs that take TO
in front of their indirect object. Verbs that take TO as a marker for their indirect objects
will be classified as belonging to CLASS 1. Notice that the verbs in (11) also allow for
the order Indirect Object (IO) > Direct Object (DO):

(11) Class 1: Indirect object takes TO as a marker. IO > DO order is possible

a. John sent some flowers TO his girlfriend./ John sent his girlfriend some flowers.
b. Mary gave a present TO her boyfriend./ Mary gave her boyfriend a present.

Some indirect objects in English cannot be marked with TO; they need to be marked
with FOR, as shown in (12). We will say that these verbs belong to CLASS 2:1

(12) Class 2: Indirect object takes FOR as a marker. IO > DO order is possible
a. Thomas baked a cake FOR/*to his girlfriend. / Thomas baked his girlfriend a cake.
b. Denisse found a job FOR/*to her son. / Denisse found her son a job.

TO is by far the most productive marker for indirect objects in English. Thus those
verbs that take FOR as a marker for indirect objects will have to be marked as

1
Some verbs that take FOR with indirect objects: book, buy, catch, choose, cook, fetch, find, get, make,
order, pour, save, etc.
60 Héctor Campos & Bùi Huỳnh Thủy Thương

such in the lexicon. This means that the child (and students of English as a foreign
language) will have to memorize these verbs that take FOR rather than TO. The two
structures for (12a) are shown in (13):

(13)

Some verbs may take either TO or FOR in front of the indirect object, as shown in (14).
We will classify these verbs as CLASS 3:

(14) Class 3: Indirect objects with either TO or FOR as a marker. IO > DO order is
possible.2

a. I brought the money TO/FOR my friend. / I brought my friend the money.

b. I read the essay TO/FOR my teacher. / I read my teacher the essay.

The two structures for (14a) are shown in (15):

(15)

There is a CLASS 4 of verbs. These are verbs that can take either TO or FOR with an
indirect object but which cannot switch the order of their complements, that is, they will
always appear in the order DIRECT OBJECT > INDIRECT OBJECT:

2
Some verbs with either TO or FOR: bring, leave, pay, play, post, read, sell, send, sing, take.
 English Syntax & Universal Grammar 61

(16) Class 4A: Verbs that take TO but which cannot appear in the order IO > DO3
a. The teacher explained the lesson TO the students.
b. *The teacher explained the students the lesson.

(17) Class 4B: Verbs that take FOR but which cannot appear in the order IO > DO4
a. He repaired the car FOR his friend. / *He repaired his friend the car.
b. She collected stamps FOR her brother. / *She collected her brother stamps.

The structures for (16a) and (17a) are shown in (18).

(18)

Finally, there is a CLASS 5 of verbs that only allow the order IO > DO, that is, they
cannot be used with either TO or FOR:5

(19)
a. I envy John his lifestyle. / *I envy his lifestyle to/for John.
b. They refused him entrance to the club. / *They refused entrance to the club to/for
him.

The structure for (19a) is shown in (20):

(20)

3
Verbs that take TO as an indirect object marker and do not allow the order IO > DO: admit, announce,
demonstrate, explain, introduce, mention, point out, prove, report, say, suggest, etc.
4
Verbs that take FOR as in indirect object marker and do not allow the order IO > DO : collect, mend,
repair.
5
Other verbs that do not allow TO or FOR and only appear with the order IO > DO are: allow, ask,
cost, deny, forgive, guarantee, permit, refuse.
62 Héctor Campos & Bùi Huỳnh Thủy Thương

Summarizing, we have seen that there are five classes of verbs when we study the
behavior of their indirect objects. The following chart represents the different classes

DO > IO IO > DO
Class 1 DO to-IO
Class 2 DO for-IO
Class 3 DO to/for-IO
Class 4 DO to-IO *
DO for-IO *
Class 5 *

The most productive class of verbs is Class 1. With these verbs we use TO as an indirect
object marker and we allow for the indirect object to precede the direct object. These
properties need not be marked in the Lexicon as they will be assumed to be the default
(or unmarked) case for indirect objects. All the other classes, however, will have to be
marked in the Lexicon (under "Special Properties") as such. Unless a verb is marked as
bearing a special class of indirect object (any class other than Class 1), the syntax will
assume that the indirect object marker is TO and will also assume that the indirect
object may precede the direct object. Thomas Payne (2011), in his book Understanding
English Grammar, notes that, as a rule of thumb, only monosyllabic verbs allow the
order Indirect Object > Direct Object.

PRACTICE 3

I. Classify the following verbs according to the classes above: find, offer, order,
sell, permit, introduce, fix, allow, read, collect, lend. Give examples to prove that
they belong to that class.

CHALLENGE

(A) Consider the following examples from Vietnamese:

a. Tôi đã đưa một quyển sách cho John.
b. Tôi tặng một quyển sách cho John.
c. Cô Tươi giảng bài cho sinh viên.

All three verbs appear in the order DO>IO. Do these verbs also allow the order
IO>DO?
 English Syntax & Universal Grammar 63

(B) Are the verbs that allow the two possibilities DO > IO and IO > DO the same in
English and Vietnamese? Give examples.

(C) How would you go about teaching these structures to students who are trying
to learn English, on the one hand, and Vietnamese, on the other?

4.2. Modifiers in the verb phrase

In section 1 we saw that complements are projected under V'. In this section we will
explore how to incorporate modifiers into our VP structure.

Complements are typically obligatory elements. In examples where we omit them, our
minds still have a generic complement replacing that complement. Take for instance the
verb read. I can say both I am reading a book and I am reading. The verb read c-selects
an NP-complement. The complement is present in the first example, but it is missing in
the second example. But even though it is missing, I still understand that I am reading
something. Thus my mind fills in a complement in spite of the fact that the complement
is missing in the second example.

Different from complements, which are usually obligatory and if absent, still
understood, modifiers are optional. While complements are necessary to complete
the meaning of the verb, modifiers just give additional information making
reference to time, place, manner, etc.

4.2.1 Postverbal modifiers.

To incorporate modifiers into the VP structure discussed in section 1, we will add them
as additional projections of V', as shown in (21):

(21) a. John did the homework. b. John did the homework in his room.
64 Héctor Campos & Bùi Huỳnh Thủy Thương

Notice that in (21b) we have created an additional V' to incorporate the PP-modifier.
What do we do if the verb phrase contains more than one modifier phrase? Consider
(22):

(22) John did the homework in his room in the morning.

As we can see in (22), the second modifier is incorporated as a second additional V'.
If there are any additional modifiers, they will also be incorporated as additional V'. We
call these additional V's adjunctions, that is the reason why modifiers are also referred
to as adjuncts. Consider now (23) which contains three modifiers.

(23) John did the homework in his room in the morning at 7 o'clock.

Notice that the tree in (23) has 4 V bars (V'). The lowest one is reserved for the direct
object (the homework); each modifier appears in its own V'-projection.

Special attention must be paid when we modify an intransitive verb. Remember that the
lowest V' is "reserved" for complements. Thus if a verb takes no complement, that V'
 English Syntax & Universal Grammar 65

level must be left empty and the modifier must be projected under an additional
V', as shown in (24):

(24) John slept. John slept in his room.

If a verb does not include or take any complements, modifiers must be projected in
additional and separate V's. Notice that we do not attach any category to the lowest V'
since this space is reserved for complements and intransitive verbs do not take
complements:

(25) John slept in his room at night at 11 o'clock.

[Link] Order of postverbal adverbs. When adverbs or prepositional phrases are

incorporated into the VP, are they incorporated or merged in any particular order? The
answer seems to be yes, although the order may be changed when we want to emphasize
a particular adverb. Typically we incorporate adverbs or prepositions of manner, then
place, then time :

MANNER > PLACE > TIME

66 Héctor Campos & Bùi Huỳnh Thủy Thương

(25) He worked quietly at home in the morning.

Manner > Place > Time

We can see that we have merged the adverbs in the order MANNER > PLACE > TIME.
This is the most natural and unmarked order for adverbs.

Verbs of movement typically take complements that indicate place, so with these verbs
the order is usually PLACE > MANNER > TIME

(26) I went to Cambodia by bus in the summer.

PLACE > MANNER > TIME

4.2.2 Preverbal modifiers

We can see in (25) that adverb and prepositional phrases can appear postverbally as
modifiers. Adverb phrases may also appear preverbally:
 English Syntax & Universal Grammar 67

(27) Ha always visits Quan.

Adverbs that typically appear in preverbal position are:

(28) a. degree adverbs:6 Thomas hardly visits me.

b. adverbs of frequency: 7 Angel usually visits me.
c. adverbs that indicate order:8 I first met them in 2015.

PRACTICE 4

Form VPs using the verb and the indicated adverb or prepositional phrases. Then
diagram the corresponding phrases.

a. suffer (from headaches, always, at night)

b. differ (significantly, from his brothers)
c. go (by bus, to school, in the morning)

CHALLENGE

Diagram the VPs for the following sentences. Use Δ for categories other than VP.
Remember that TREES REPRESENT CATEGORIES, NOT FUNCTIONS!

6
Other degree adverbs that appear in VP: almost, hardly, nearly, quite, rather, scarcely, etc. These may
also appear postverbally:
(i) I greatly appreciate your help.
(ii) I appreciate your help greatly.
7
Other adverbs of frequency that appear in VP: always, hardly ever, never, often, rarely, regularly,
seldom.
8
Other adverbs that indicate order: first, last, next. These can also appear postverbally, and if they co-
occur with a phrase indicating time, they typically precede the time expression:
(i) I first met her in 2015.
(ii) I met them first in 2015.
68 Héctor Campos & Bùi Huỳnh Thủy Thương

a. Ha bought wine from Chile at the supermarket.

b. Co Chi borrowed some Russian books from the library yesterday.
c. Trieu sent a bouquet of flowers to his girlfriend for her birthday last week.
d. Thuong bought a very nice present for her son on Saturday.
e. They finally decided on the boat. (ambiguous)
f. Ha cooked Chinese food for her boyfriend yesterday. (ambiguous)
g. The syntax professor invited the students of Linguistics to his house on the
beach. (ambiguous)

4.2.3 An important note on NPs that function as adverbs.

In English (and in many languages) we have NPs that have an adverbial function.
However, you must remember that TREES REPRESENT CATEGORIES, NOT
FUNCTIONS. Thus these NPs will have to be represented as NPs and not as ADVPs.

(29) a. Khiem will go to Europe next year.

b. Khiem calls Quan every day.

The structure for (29a, b) is shown in (30):

Modifiers like yesterday and today, on the other hand, will be treated as adverbs:

(30) a. We went to the movies yesterday.

 English Syntax & Universal Grammar 69

4.3 Phrasal versus Prepositional verbs

In section 1.3 we discussed prepositional verbs. We saw that prepositional verbs like
(31) take a prepositional phrase as a complement. There is a second class of verbs called
phrasal verbs that look like prepositional verbs and should not be confused with
prepositional verbs:

(31) Prepositional verbs

a. Khiem talks about syntax all the time.
b. Tuoi believes in ghosts.

(32) Phrasal verbs

a. Khiem looked up the word in the dictionary.
b. Tuoi put on a beautiful ao dai for the party.

In (31-32) we can see that both prepositional and phrasal verbs exhibit the sequence V-
P-NP. (33) shows the structure we will propose for these two classes of verbs:

(33) a. talk about syntax b. look up a word

Prepositional verb Phrasal verb

As we can see in (33a) talk takes a PP complement. In (33b) we see that the verb is
actually look-up and this verb takes NP as a complement.

We could break up the phrasal verb look up into its component parts: the verb (V) look
and the particle/preposition (P) up. As we can see in (34) below, look up is still analyzed
as a word since it has the category V.

(34) full representation for a phrasal verb

70 Héctor Campos & Bùi Huỳnh Thủy Thương

How can we distinguish these two classes of verbs?

A very simple diagnostic test is to try to put the noun phrase between the verb and the
preposition:

(35) a. Prepositional verb:

Khiem always talks about syntax. *Khiem always talks syntax about.

b. Phrasal verb:
Khiem always looks up words. Khiem always looks words up.

As we can see in (35), only phrasal verbs like look up allow us to put the noun phrase
between the verb and the particle/preposition.

Another simple diagnostic test is to replace the NP by a pronoun, as shown in (36):

(36) a. Prepositional verb:

Khiem always talks about syntax. Khiem always talks about it.

b. Phrasal verb:
Khiem always looks up words. *Khiem always looks up them.
Khiem always looks them up.
As we can see in (36), prepositional verbs allow for the pronoun to go after the
preposition (36a). Phrasal verbs must put the pronoun between the verb and
theparticle/preposition, as shown in (36b).

As we can see in (33), the P and P form a unit or constituent (PP) only with
prepositional verbs. Thus we expect P and NP to be able to move as a unit. This is what
we observe in (37):

(37) a. Prepositional verb:

[About syntax] Khiem always talks.

b. Phrasal verb:
*[Up words] Khiem always looks.
As we can see in (37), only prepositional verbs allow P to move along with the NP.
Phrasal verbs do not allow this movement because P and NP do not form a constituent.
We can clearly see this difference in (33).

Finally we can try the "shared constituent" test. Only constituents can be shared. Thus
we get a difference between the two verbs:
 English Syntax & Universal Grammar 71

(38) a. Prepositional verb:

Khiem always talks and Quan also talks [about syntax].
b. Phrasal verb:
*Khiem always looks and Quan also looks [up the new words].

Only in (38a) is the P + NP a constituent and thus can be shared. This is exactly what
structure (33) predicts.

PRACTICE 5

I. Identify the following verbs as Phrasal or Prepositional. Then draw the

corresponding structures for the VP.
a. Ha [thought about the problem].
b. Khiem [took off his shirt].
c. Toan [ate up all the cake].
d. Tuoi [put on a beautiful ao dai].
e. An [wrote down the structures].
II. Diagram the following VPs. Pay attention as to whether the verbs ought to
be analyzed as phrasal or prepositional verbs. Also be careful distinguishing
complements from modifiers.
a. She threw away her old books last week.
b. They took off their jackets before the meeting.
c. Thuong always looks after her baby.
d. She turned in her homework late.
e. She turned up the volume at the party.
f. She turned up late for the party.
g. They found out the truth last night.
h. I called off my subscription to the Washington Post.

RESEARCH PROJECT

Consider the examples in (f) and (g) below. Propose a structure for them and
motivate your structure.

a. Quan cannot [put up with the noise].

b. Khiem will not [keep up with the reading].
72 Héctor Campos & Bùi Huỳnh Thủy Thương

CHALLENGE 1
Consider the following data for the invented verb glew up. Would you analyze
it like a phrasal or like a prepositional verb? Use the following data to write as
many arguments as you can.
1. She glew up the book.
2. She glew the book up.
3. She glew it up.
4. *She glew up it.
5. *Mary glew and I also glew up the book.
6. Mary glew up and I also glew up the book.
6. -Did Mary glew up the book?
-*Up the book? You're crazy!!!
7. The book, I don't think that Mary actually glew up.
CHALLENGE 2
Explain what the phrase did so stands for in the following sentences. Then
explain why the last sentence is ungrammatical.
1. John did the homework at the library and Mary did so too.
2. John did the homework at the library and Mary did so at home.
3. John sent money to his parents and Mary did so too.
4. *John sent money to his parents and Mary did so to her grandparents.

4.4 Summary

In this chapter we have seen that complements are merged at the lowest V' projection
while modifiers are merged at additional V' projections. We have seen that modifiers
may appear either to the left or to the right of the verb. The general skeleton for VP is
shown in (39):
(39)
 English Syntax & Universal Grammar 73

We argued for 5 different classes of indirect objects in English, depending on whether

they take TO or FOR as a marker, and depending on whether they allow the order IO >
DO or not. We argued that indirect object markers are not present at D-Structure;
indirect object markers only appear at S-Structure as a result of an insertion operation.
To account for SOV languages, we claimed that the head of VP can also be projected to
the right of the complements. In the next chapter we will pursue a more sophisticated
analysis for TO-insertion. We will try to understand why indirect objects need to be
marked in English since English, unlike the Japanese (see example 10), is not typically a
language that marks functions.

We saw that while complements are merged into the lowest V', modifiers are merged
into additional V'-projections. Modifiers do have a natural order in which they are
merged: they typically follow the order Manner > Place > time, unless we have a verb
whose complement indicates a place, in which case the order of the first two adverbs is
inverted.

Finally we discussed phrasal and prepositional verbs. We argued that prepositional

verbs take a PP-complement. In phrasal verbs the verb and the preposition (usually
referred to as a particle) form a unit. In prepositional verbs, the preposition is the head
of a PP; in a phrasal verb, the preposition is just a P (there is no PP) and forms a word
with the verb. Phrasal verbs (V + P) can take a noun phrase as a complement, as in
(33b), or a PP as a complement, as in examples (a) and (b) in the Research Projectfor
this chapter.
74 Héctor Campos & Bùi Huỳnh Thủy Thương

CHAPTER 5

THE STRUCTURE OF THE NOUN PHRASE

In this chapter we will apply X'-theory to noun phrases.

5.1 The structure of noun phrases

The simplest noun phrases are those where the noun appears by itself (bare nouns).
Examples can be found in (1). Following the X'-syntax model we introduced in Chapter
2, these nouns would have the minimal structure in (1)1:

(1) Mary, apples, icecream

5.2. Determiners as specifiers of NP.

Nouns may appear with a determiner (D). Definiteand indefinite articles, possessives
and demonstratives will be considered determiners:

1
There are languages where proper names like John or Mary appear with a determiner. Take Greek, for
instance:
(i) o Giannis, i Maria
the John, the Mary
'John, Mary'
Other languages, like French, may have special determiners to mark indefinite plurals (PL-INDEF):
(ii) des gateaux, des enfants
PL-INDEF cakes, PL-INDEF children
'cakes, children'
 English Syntax & Universal Grammar 75

(2) the girl a boy

my house those books

5.3 NP modifiers. Adjectives typically modify nouns. As we did with verbs, we will
include all modifiers in additional N' projections, as shown in (3):

(3) a. a smart girl b. a beautiful smart girl

Notice that in the examples in (3), each adjective is merged in a separate N'-projection
on top of the lowest N'. The lowest N' contains no projection since these examples do
not include complements.

Is there an order in the way we merge modifiers to the noun? Yes, there is, and in
general, we can posit that subjective adjectives precede objective adjectives. In our
example in (3b) we have two adjectives: beautiful and smart. Beautiful precedes smart
since beautiful may be considered more subjective than smart.

We can take the following template as a guide for merging prenominal adjectives:
76 Héctor Campos & Bùi Huỳnh Thủy Thương

(4) Opinion > size > age > shape > color > origin > material > used for/about > N
A lovely small vold round red Vietnamese wooden storing box
A big black school bag
A spectacular red Italian evening dress

All the examples with adjectives we have discussed so far are prenominal modifiers.
This is by far the most common position of adjectives in English. However, adjectives
may also appear in postnominal position. This is typically the case when the adjective
appears before a to-infinitive or with a prepositional phrase which is a complement of
the adjective itself:

(5) a. a program [intended to facilitate learning]

b. the minister [responsible for health services]
c. a movie [suitable for all audiences]

In these cases, we project the adjective phrase to the right of the noun. The structure for
(5b) is shown in (6):

(6)

PRACTICE 1

I. Diagram the following noun phrases:

a. Susan
b. the teacher
c. the American teacher
d. the coffee with milk
e. the victims involved in the accident
f. the visible stars, the stars visible (what is the difference?)

II. Order the following noun and adjectives. Then diagram the phrases:
 English Syntax & Universal Grammar 77

a. bedroom (elegant, big, modern)

b. rug (expensive, woolen, Persian, antique)

III. The following phrase is ambiguous. Explain the ambiguity and propose two
structures to account for the ambiguity.

the cold coffee with milk

5.4 Complements of NP: PP-complements. Consider the following verbal phrases

and their corresponding deverbal nouns(i.e., nouns that derive from verbs) in (7):

(7) a. They traveled to Cambodia. → the travel to Cambodia

b. They argued about money. → an argument about money
c. They applied for the job. → an applicant for the job

The structure for the deverbal nouns in (7a), (7b) is shown in (8):

(8)

Notice that the deverbal nouns in (7) maintain the same preposition as the
corresponding verb.

We observe a perfect parallelism between the VP and the corresponding NP.

(9) They argued about money an argument about money

In (9), both the verb and its derived nominal take the same PP complement. This
structural parallelism is called isomorphism.
78 Héctor Campos & Bùi Huỳnh Thủy Thương

Isomorphism simplifies the process of language acquisition. Once the child learns
that the verb takes a PP complement, the child may assume that the corresponding
deverbal noun will take the same prepositional phrase as its complement. Our model
should try to preserve isomorphism where possible.

5.5 NP complements: non-PP complements. Let us now look at complements of

nouns. Nouns that derive from verbs (deverbal nouns) may typically appear with
complements. As with modifiers, complements in English may appear prenominally or
postnominally.

(10) Verbal phrase Prenominal complement Postnominal complement

a. He studies chemistry. the chemistry student the student OF chemistry

b. She studies languages. the language student the student OF languages

As we can see in (10), complements may appear either pre- or postnominally in

English. If the complement appears prenominally, notice that the noun loses its plural
form (as seen in (10b)). On the other hand, if the complement appears postnominally,
OF precedes the complement. How can we account for these two facts? Let us take one
problem at a time.

5.5.1 Prenominal complements. Why do we need a singular form in (10b) when the
complement appears prenominally? Facts like the language student seem to suggest that
we may be outside the domain of syntax proper. Language student can be analyzed as a
compound word, which is the domain of Generative Morphology (see Chapter 2).
Other examples where the plural has been lost in the compound word can be found in:
an opener for bottles → a bottle opener, a trap for mice → a mouse trap, etc. Notice
that in these compound words, not only does the first noun appear in the singular form,
but stress always falls on the first noun: a BOttle opener, a MOUSE trap. The same is
observed with our phrases in (10): the CHEmistry student, the LANguage student,
where the syllable in capital letters is the syllable that bears the phrase stress. Thus there
is evidence that prenominal complements of nouns should be treated
morphologically rather than syntactically, thus accounting for the fact that these
complement are always singular.

Notwithstanding these facts, we will treat these constructions syntactically, projecting

the complement of N to the left of N, in prenominal position (11b). Since we know the
internal structure of NP now, we can project the structure both for student and
chemistry, as shown on the diagram in (11c):
 English Syntax & Universal Grammar 79

(11) a. the chemistry student

(11b) (11c)

One argument in favor of treating prenominal complements as syntactic rather than

morphological complements is the fact that prenominal nouns may be modified, as
suggested by the phrase the modern chemistry student, shown in (11d).

(11d) d. the modern chemistry student

If chemistry student were a word, we should not be able to modify part of the word.
Take, for instance, the word football. Imagine there is a special ball for people who kick
with their left foot. We would never be able to have a left-football. Thus we are justified
in treating constructions like (11) as constructions containing a prenominal syntactic
object. However, notice that in (11a) stress falls on CHEmistry, while in (11b), it falls
on STUdent.

PRACTICE 2

The following phrases are ambiguous. Discuss their ambiguity and propose
structures to explain the respective readings:
a. the foreign car importers
b. a European book publisher
80 Héctor Campos & Bùi Huỳnh Thủy Thương

5.5.2 Postnominal Of-complements. Let us consider now the more complex case of
complements of nouns in postnominal position: the student of chemistry. To understand
why ‘of’ appears in these construction we need to have a brief discussion about Case.

English only shows Case with pronouns, as shown in (12)

(12)

SUBJECT OBJECT OF OBJECT OF POSSESSIVE POSSESSIVE

V P DETERMINER PRONOUN
(Nominative Case) (Objective Case) (Oblique Case) (Genitive Case)
I me me my mine
you you you your yours
he him him his his
she her her her hers
it it it its its
we us us our ours
you you you your yours
they them them their theirs

NOMINATIVE Case is typically used with subjects2; OBJECTIVE Case is used with
complements of the verb; OBLIQUE case is used with complements of prepositions
and GENITIVE case is used to indicate possession. We can see in the table in (12) that
pronouns with OBJECTIVE and OBLIQUE case bear the same form in English.

The examples in (13) show the different forms that the pronoun SHE may take when it
replaces Mary, which appears in different functions in (13): as a subject in (13a), as a
direct object in (13b), as object of a preposition in (13c), as a possessive in (13d,e):

(13) a. Mary speaks English well. → She speaks English well.

b. I know Mary very well. → I know her very well.
c. I always study with Mary. → I always study with her.
d. I always go to Mary's house. → I always go to her house.
e. This is my house and that is Mary's.→ This is my house and that is hers.

2
A better name for this Case would have been SUBJECTIVE Case since Nominative Case is just one
of the possible realizations of Case. Thus many languages also exhibit Ergative or Absolutive Case
with subjects.
 English Syntax & Universal Grammar 81

We can see that while Mary does not change in form in (13). If we replace Mary with
the corresponding pronoun, the pronoun must take the appropriate Case depending
on the function that Mary has in the sentence. We say that nouns in English are not
declined for Case, only pronouns are, as shown in the table in (12). In (13) we see the
different declensions for the pronoun she.

In other languages, nouns and not just pronouns may be inflected for case. We can see
this example with Albanian in (14). Notice the different forms that Martin may take
depending on the different functions it plays in the sentence:

(14) a. Martini flet shqip. Martin speaks Albanian.

b. Unë kam ftuar Martinin. I have invited Martin.
c. Unë i kam dërguar Martinit një letër. I have sent Martin a letter.
d. Unë jetoj pranë Martinit. I live near Martin.
e. Kjo është shtëpia e Martinit. This is Martin's house.

In (14a) Martin is a subject, in (14b) a direct object, in (14c) an indirect object, in (14d)
the object of a preposition and in (14e) a possessor. As evidenced by the data from
pronouns in English, nouns and pronouns from Albanian in (14) and in many other
languages, we will thus assume that all noun phrases must bear Case. We will call
this the CASE CONDITION:

(15) THE CASE CONDITION

All noun phrases must bear Case.

We will assume that this condition is universal. Some languages may have highly
inflected nouns and pronouns (like Albanian); others may have more restricted systems,
like English, where we only see inflection in pronouns. In yet other languages, like
Vietnamese, we may see no inflection on nouns or pronouns at all. In those forms where
Case is not obvious or transparent, we will say that they have zero-inflection.

Let us go back to (10). Now we have a way to explain why OF appears with
postnominal complements. Compare (16a) with (16b):

(16) a. He studies chemistry.

b. the student OF chemistry.
82 Héctor Campos & Bùi Huỳnh Thủy Thương

(17) a. b.

Objective Case

(17a) and (17b) show perfect isomorphism: both the verb study and the noun student
take the NP chemistry as its complement. In (17a), the verb is able to assign Objective
Case to its complement chemistry. Thus we are able to say (he) studies chemistry. But
we are not able to say *the student chemistry. Instead we must say the student OF
chemistry. We will claim that OF is a pseudo-preposition which is inserted into the
structure in (17b) so that chemistry is able to receive (Oblique) Case to satisfy the Case
Condition in (15). The structure for (16b) is shown in (17c):

(17c)

Oblique Case

In (17c), the pseudo-preposition OF assigns oblique case to the noun phrase chemistry
and thus chemistry satisfies the Case Condition in (15).

If OF assigns Oblique Case to chemistry in (17c) thus helping chemistry satisfy the Case
Condition in (15), how does chemistry in the phrase the chemistry student satisfy the
Case Condition in (15)? We suggested above that the chemistry student may be outside
the realm of syntax proper, the compound chemistry student being formed in the
morphology rather than in the syntax. If so, chemistry (and prenominal complements in
general) may be exempt from the Case Condition in (15). In general, we will say that
 English Syntax & Universal Grammar 83

prenominal complements are exempt from the Case Condition. These prenominal
complements are at the morphology-syntax interface.

5.6. PP-modifiers. Prepositional phrases may also modify nouns. When they do,
they always appear postnominally, as shown in (18):

(18) the students of linguistics [from Saigon] the student [from Saigon]
*the [from Saigon] student

However, the PP-modifier in (18) may also be realized as a prenominal noun phrase
modifier:

(19) the [Saigon] students of linguistics the [Saigon] student

Notice that in (19) the preposition has disappeared. This may suggest again that we may
be in the realm of morphology rather than syntax since syntax does not allow the
deletion of elements that have meaning. Furthermore, notice that the noun phrase
Saigon in (19) escapes the Case Condition since nothing seems to assign Case to it.
Once again, facts like this suggest that we may be in the realm of morphology rather
84 Héctor Campos & Bùi Huỳnh Thủy Thương

than syntax or at the interface morphology-syntax. Trying to keep these structures in the
realm of syntax, we need to claim that prenominal complement- and modifier- NPs
are allowed to escape or constitute an exception for the Case Condition in (15).

As expected, both the complement and the modifier may appear in prenominal position:

(20) The Saigon linguistics student

PRACTICE 3

Diagram the following NPs:

a. the syntax teacher from America
b. the American teacher of syntax
c. the taxi driver from Hanoi
d. the Hanoi taxi driver
e. a cold coffee with ice

5.7. Explaining some ambiguous noun phrases. Now we have enough knowledge
about structures to start explaining some basic ambiguities within the noun phrase.
Consider the NP in (21), which is ambiguous between the readings in (21a) and (21b):

(21) an English teacher

a. a teacher who teaches English
b. a teacher from England

We can represent this ambiguity with the two following structures:

 English Syntax & Universal Grammar 85

(22) a. an English teacher b. an English teacher

(a teacher who teaches English) (a teacher from England)

We can see that in (22a), English is a noun phrase and it is merged as a (prenominal)
complement of N. In (22b), English is an adjective phrase and it is merged as a modifier
of N.

PRACTICE 4

I. The following phrase is ambiguous. Diagram two structures that correspond

with each of the readings:
the Vietnamese teacher
a. the teacher teaches Vietnamese
b. the teacher is from Vietnam

II. The following phrase is at least three times ambiguous. Diagram three
structures that correspond with each of the indicated readings:
the old Vietnamese teacher
a. the teacher is old and is from Vietnam
b. the teacher is old and teaches Vietnamese
c. the teacher teaches old Vietnamese

The word old may also mean "former". Would we have a different structure for this
particular reading? Justify your answer. Would this be considered a syntactic or a
lexical/semantic ambiguity?
86 Héctor Campos & Bùi Huỳnh Thủy Thương

CHALLENGE

I. Diagram the following structures. Briefly explain your structures justifying

your choice of "complement" versus "adjunct".
a. a possible missile attack on our bases in Korea
b. a generous property bequest to charity for the elderly
c. a spy exchange with the Russians

II. The following noun phrases are ambiguous. For each reading propose a
structure and briefly explain the reading.
a. the house in the forest near the school
b. a toy factory
c. a brass button holder
d. the tall girl in jeans

5.8. A note on quantifiers. In section 1.1 we saw that articles (definite as well as
indefinite), possessives and demonstratives merge as specifiers to the NP. Some
quantifiers also seem to behave like determiners since they cannot co-occur with
determiners:

(23) any students, each student, either student, every student, neither student, no
student, some students

Similar to the way we treated articles, demonstratives and possessives in section 1.1, we
will treat these quantifiers as determiners (D):

(24) some students every student

Numerical quantifiers, however, will be treated like modifiers. As we can see in (25),
they may co-occur with determiners:
 English Syntax & Universal Grammar 87

(25) my two sons the first chapter

RESEARCH PROJECT 1

1. Consider the quantifiers MANY, FEW, MUCH, LITTLE. Would you analyze
them as determiners similar to some, every, etc. in (24) or similar to the numerical
QPs discussed in (25)? Justify your answer.

2. How would you analyze the universal quantifier ALL in the phrase all the
students?

5.9. Partitives

Nouns can be classified into [+count] (count nouns) and [-count] (mass nouns). Count
nouns are those that may be counted and may appear inflected for plural: dogs, books,
cities, feet, etc. Mass nouns do not have a plural form, but we can use partitive phrases
to measure them:

(26) a drop of water, some slices of bread, a word of advice, a pair of scissors, a gallon
of gas, etc.
88 Héctor Campos & Bùi Huỳnh Thủy Thương

As we can see in (26), the partitive phrase takes the [-count] noun as a complement.
The partitive phrase itself is the head of the NP.3 The mass noun may also (but not
always) appear in prenominal position:

(27) a water drop, some bread slices, but *an advice word, *a scissors pair, ...

PRACTICE 5

I. Diagram the following NPs which contain partitive phrases:

a. my two pieces of luggage
b. those two large servings of strawberry icecream

II. Diagram the following NPs. Do they have the same meaning? Justify your
answer.
a bottle of wine, a wine bottle

5.10. Non-deverbal nouns with complements.

In the previous sections we discussed deverbal nouns and we saw that they typically
maintain the c-selection of the verb from which they derive. There are a few nouns
which do not have a corresponding verb from which to be derived and yet they can still
bear a complement. In these cases, the complement is typically understood as a
PATIENT, receiving the action or effect of the head noun. As we can see in (28), the
complements may be preceded by OF, as in (28a) or by a preposition, as in (28b). A
couple of structures are shown in (28):
3
We can see that the partitive phrase is the head of the whole NP when we look at agreement:
(a) One slice of bread IS rotten.
(b) Two slices of bread ARE rotten.
As we can see in the examples above, the verb agrees with slice(s) and not with bread, which shows
that actually slice(s) is the head or nucleus of the NP.
 English Syntax & Universal Grammar 89

(28) a. the President of the country, the Dean of the School, the news of his accident,
the rumor of her resignation, the idea of a promotion, etc.

b. the impact on society, the impression on the people, the need for help, the
opposition to the regime, the involvement in the war, etc.)

a. the President of that country b. the impact on society

In (28b) the preposition on assigns Oblique Case to the NP society. In (28a), however,
the noun President cannot assign Case to its complement. Thus OF-insertion is needed
to save the structure from violating the Case Condition in (15). We say that OF-
insertion is a last resort operation; OF-insertion takes place to save the structure from
collapsing (we usually say, save the structure from"crashing").

PRACTICE 6

I. Diagram the VPs indicated in brackets. Expand all the VPs and NPs.
a. The university [invited the author of that famous novel].
b. I [saw a picture of the organizers of the symposium in the newspaper].

II. For the two following VPs try to expand the PPs following X'-syntax. Project
the complement of P under V'. Pay attention to the inserted elements.
a. The President [called the Dean of the School of Languages to his office on
the third floor].
b. We [read a book about the possibility of a nuclear war]
90 Héctor Campos & Bùi Huỳnh Thủy Thương

CHALLENGE

Explain what the phrase “the one” replaces in the following structures. Then explain
why structures (2) and (4) are ungrammatical.
1. the student from Washington and the one from Saigon
2. *the student of chemistry and the one of physics
3. the driver from Saigon and the one from Hanoi.
4. *the driver of the taxi and the one of the truck

RESEARCH PROJECT 2

Does Vietnamese have an operation similar to OF-insertion in English? Look for

examples that have the pseudo-preposition OF in English and translate them into
Vietnamese. What do you think is happening in Vietnamese?

Hint: Try translating these two phrases: Mr. Hai's room and Mr. Hai's new room .
What interesting phenomenon do you observe? What happens if you omit the noun?
What happens if you use the possessive predicatively as in This room is Mr. Hai's.
Describe the full phenomenon as thoroughly as possible.

5.11. Summary

In this chapter we have expanded X'-syntax to the domain of the noun phrase. We have
seen that noun phrases can appear bare (1), with a complement (7), with a specifier (2)
or with a modifier (18). We saw that complements of nouns, unlike what we observe in
verb phrases, can be merged both pre- and postnominally. The same can be said of
modifiers.

We have seen that there is a perfect isomorphism between the verb and the
corresponding derived or deverbal noun. We saw that there are nouns which exhibit a
pseudo-preposition OF in front of their complement. We argued that this pseudo-
preposition is inserted. To account for the presence of OF in these constructions we
invoked the Case Condition, introduced in (15), repeated here as (29):

(29) Case Condition

All noun phrases must bear Case.
 English Syntax & Universal Grammar 91

When and where is the pseudo-preposition inserted? To answer this question, we need
to review the model we introduced in Chapter 2, repeated here as (30):

(30) D-Structure

General Principles: Transformations:

-Projection Principle INSERT (OF, ...)
-Extended Projection Principle
-Theta-Criterion S-Structure

Phonetic/Phonological Form Logical Form

As we can see in (30), INSERTION occurs between D-Structure and S-Structure. In

Chapter 3 we considered insertion of indirect object markers and in this chapter we
studied insertion of the pseudo-preposition OF. Consider (17) again, repeated here as
(31):

(31) a. He studies chemistry b. student of chemistry

D-Structure:

Objective Case

(31) represents the D-structure of the VP and NP, respectively. It is at this level that we
find isomorphism. VP and NP have the same structure with the exception of the
category that differentiates them (V versus N). In (31a) the verb study is able to assign
Objective Case to the NP chemistry, but the noun student is not able to assign Case to
92 Héctor Campos & Bùi Huỳnh Thủy Thương

this NP. Thus at S-Structure, we need to insert the pseudo-preposition OF to assign

(Oblique) Case to the NP chemistry in (31b). This is shown in (32):

(32) a. He studies chemistry b. student of chemistry

S-Structures:

Objective Case
Oblique Case

As we can see in (32b), the pseudo-preposition OF is inserted in front of the

complement of the nounto assign Oblique Case to the NP-complement chemistry.
No insertion is necessary in (32a) because the verb study is able to assign Objective
Case to its complement chemistry. Thus we see that the D-Structure and the S-
Structure for the VP are the same in (24a) and (25a). In the case of the student of
chemistry, there is no pseudo-preposition at the D-Structure in (24b). The pseudo-
preposition is inserted between D- and S-Structure, as shown in (30). This
"emergency insertion" (remember that insertion of the pseudo-preposition is
necessary in order to avoid a violation of the Case Condition in 29) is considered a
last resort operation and takes place so that the NP complement of student in (29) can
receive Case in order to satisfy the Case Condition. Unless the pseudo-preposition is
inserted, the derivation would crash! This is why we say that insertion of the pseudo-
preposition OF is a "last resort" operation.
 English Syntax & Universal Grammar 93

APPENDIX: REVISITING INDIRECT OBJECTS

The previous analysis, whereby we insert OF in front of the direct object in order to
save the structure (last resort operation), may lead us to rethink our analysis of TO-
insertion in Chapter 4. We argued that TO was inserted as an indirect object marker
when the indirect object is far from the verb. In chapter 4 we argued that TO is a
pseudo-preposition, it is not there at D-Structure and it is only inserted at S-Structure
to mark the indirect object. This analysis is very similar to what we claimed for the
pseudo-preposition OF in this chapter. While the analysis of TO as an indirect object
marker can account for the facts we observed in Chapter 4, we may wonder why
English needs to mark the indirect object (and not the direct object, for instance!). We
saw that Japanese marked subjects, direct as well as indirect objects. Thus we may want
to explore whether TO does more than just mark the indirect object in English. This is
what we explore in this Appendix.

In English a noun phrase needs to be adjacent to the verb in order to receive Case
from the verb, as we can see in (1):

(1) a. I put the books on the table.

b. * I put on the table the books.
c. Ha did her homework at the library.
d. *Ha did at the library her homework.

Thus we observe the following generalization for English, which is found in some, but
not all, languages:

(2) Case Adjacency Principle

In English, a noun phrase that receives Case from a verb must be adjacent to
the verb.

Consider now examples with indirect objects, as we discussed in Chapter 4:

(3) a. John sent some flowers TO his girlfriend.

b. John sent his girlfriend some flowers.

As we can see in (3), TO appears in front of the indirect object only when the
indirect object is far away from the verb. Invoking the Case Adjacency Principle, we
may argue that a verb can only assign Case to its indirect object when it is close to it (as
in 3b). When the indirect object is far from the verb, the verb is unable to assign it Case.
94 Héctor Campos & Bùi Huỳnh Thủy Thương

Thus the pseudo-preposition TO is inserted to assign Case to the indirect object.

Consider the structures in (4):

(4) John sent some flowers to his girlfriend.

a. D-Structure b. S-Structure

Objective Oblique
Case Case

In (4a), the direct object some flowers is able to receive Case from the verb send since it
is adjacent to it. But the verb is unable to assign Case to the indirect object his girlfriend
because this NP is too far from the verb (invoking the Case Adjacency Principle in 2).
Thus we insert TO as a last resort operation in front of the indirect object. TO, as a
pseudo-preposition, is able to assign Oblique Case to the indirect object. At S-
Structure the noun phrase his girlfriend is able to satisfy the Case Condition since the
pseudo-preposition has assigned Oblique Case to it.

We thus see that TO appears in front of indirect objects for two reasons:4

(a) it marks the indirect object when it is far away from the verb
(b) it assigns Case to the indirect object, so that this NP may satisfy the Case
Condition.

Now that we can explain why TO appears in front of indirect objects when the indirect
object is far from the verb, you may be wondering how we assign Case when we find
the reverse order, that is, when the indirect object precedes the direct object, as in (3b),
repeated here as (5):

(5) John sent his girlfriend some flowers.

4
The same argument applies for indirect objects preceded by FOR, as discussed in Chapter 4.
 English Syntax & Universal Grammar 95

a. D-Structure b. S-Structure-1 c. S-Structure-2

OBJ OBJ OBJ OBJ

from D-Structure

F
As we can see in (5a), the direct object some flowers receives Objective Case from the
verb at D-Structure because it is adjacent to the verb. But the second NP, the indirect
object, fails to receive Case in (5a) because it is too far from the verb. At this point we
have two options: We either insert TO as a last resort, as in (4b), or we move the
indirect object next to the verb so it is able to receive Objective Case from the verb. The
final result for this operation is seen in (5c). In (5c) both NPs receive Objective Case:
the direct object some flowers receives Objective Case from the verb at D-Structure, as
in (5a), and the indirect object his girlfriend receives Objective Case from the verb after
it moves, at S-Structure, as shown in (5c). Thus in (5c), as in (5b), both the direct and
the indirect object receive Case and satisfy the Case Condition.

There is evidence that the indirect object in (5c) receives Objective Case. The indirect
object when it precedes the direct object in ditransitive verbs actually behaves like a
direct object (which typically receive Objective Case). Similar to a direct object, for
example, the indirect object can be passivized, as we can see in (6)-(7) below:

(6) a. The professors sent flowers to the Dean.

b. Flowers were sent to the Dean by the professors.

(7) a. The professors sent the Dean flowers.

b. The Dean was sent flowers by the professors.

In Chapter 4 we argued that indirect objects of Class 4 do not allow the order IO > DO,
as shown in (8):

(8) a. The teacher explained the lesson to the students.

b. *The teacher explained the students the lesson.
c. The lesson was explained to the students by the teacher.
d. *The students were explained the lesson by the teacher.
96 Héctor Campos & Bùi Huỳnh Thủy Thương

As we see in (8), the indirect object cannot precede the direct object when we use the
verb explain. In (8a), the students receives Oblique Case from the pseudo-preposition
TO. Since the passive operation requires an objective complement to become a subject,
we predict that only the lesson could be a possible subject for a passive construction.
This is borne out, as we can see in (8c) and (8d).

Thus we conclude that the pseudo-preposition that appears in front of indirect

objects in English has two functions, namely, (a) to mark the indirect object when
it is not adjacent to the verb and, (b) to assign Case to the indirect object so it can
satisfy the Case Condition.

As we can see in (4) and (5), an indirect object has two possible ways to receive Case
and both ways are last resort operations: (a) it may be assigned Case by the inserted
pseudo-preposition TO, as in (2) or (b) it may move next to the verb, where it will be
adjacent to receive Objective Case from the verb, as in (5).

The analysis presented in this Appendix allows us to understand and explain why TO
appears in front of indirect objects when these are far away from the verb. This account
provides us with a deeper analysis of indirect objects than the analysis presented in the
previous chapter.
 English Syntax & Universal Grammar 97

CHAPTER 6

THE STRUCTURE OF AP, PP AND ADVP

In this chapter we continue to apply the X'-syntax model to adjectival, prepositional and
adverbial phrases. We will show that the simple model we proposed in Chapter 2 can be
easily applied to these categories as well.

6.1. The structure of AP

The simplest AP will be those adjectives phrases that contain a bare adjective:

(1) pretty smart

Adjectives may be modified by adverbs. We will take adverbs to be the specifiers of

AP:

(2) very pretty so smart

Can adjectives take complements? Indeed they can! And they are typically PP, as we
can see in (3):

(3) absent from school, accustomed to the weather, addicted to work, adjacent to the
bank, clever at math, impressed with the situation, busy with my homework, crazy
about her, etc.
98 Héctor Campos & Bùi Huỳnh Thủy Thương

(4)

Where the adjective is related to or derives from a verb, notice the perfect isomorphism
between verb and adjective:

(5) a. John concentrates on the problem → concentrated on the problem

b. Susan succeeded in business → successful in business

(6)

In spite of the perfect isomorphism observed in (6), notice that OF appears in front of
some adjectives derived from verbs:

(7) a. John fears heights. → fearful OF heights

b. John ignores the results. → ignorant OF the results
c. John understands the situations. → understanding OF the situation

We can see that in (7) the verbs take their complement directly, while in the
corresponding derived adjective form require the pseudo-preposition OF, similar to
what we observed with derived nouns in the previous chapter. As with deverbal nouns,
we need OF-insertion for these constructions to be well-formed:
 English Syntax & Universal Grammar 99

(8)
a. (John) fears heights b. fearful of heights

Objective case
Oblique case

We saw in Chapter 4 that only verbs and prepositions assign Case. The complement
heights may receive Case from the verb fear in (8a), but it cannot receive Case from the
adjective fearful in (8b). OF-insertion is a last-resort operation necessary to avoid
violating the Case Condition. Thus OF-insertion saves our AP structure in (8b) from
crashing. We will thus assume that whenever OF appears in front of a complement of an
adjective, OF is an inserted pseudo-preposition. Some if these adjectives are listed in
(9):

(9) capable of, conscious of, afraid of, ashamed of, fond of, frightened of, ignorant of,
understanding of, regardless of, shy of, sick of, tired of, proud of, etc.

(10) ashamed of his grades proud of her parents

In the previous discussion, we saw that adjectives can appear bare (1), they can be
modified by an adverb (2) and they can appear with a complement, as in (6) and (10).
Can adjectives contain a modifier?
100 Héctor Campos & Bùi Huỳnh Thủy Thương

Adjectives may contain a modifier. This is the structure we find with resultative
constructions, where the sentence expressing result is a modifier to the adjective. This
is a correlative structure, since the sentence cannot appear without the adverb in the
structure so A that... For the moment we will represent the sentence as a CP structure,
where CP stands for a "complementizer phrase".

(11) so tired that he cannot move

PRACTICE 1

I. Diagram the following APs:

a. quite acquainted with the culture

b. pretty certain of the date
c. amused with his funny jokes
d. so busy that I forgot to call her.

II. Diagram the following NPs which contain AP. Expand all nodes:

a. the girl proud of her mother

b. the people busy with their work
c. the corn fields covered with snow
d. a young man angry at his syntax teacher

III. Diagram the VPs in brackets. Expand all nodes.

a. She [invited the students present].

b. She [discussed the present situation with her boss].
c. I [need a list of the doctors available for consultation].
d. The teacher [answered the questions for the students curious about syntax].
 English Syntax & Universal Grammar 101

6.2. The structure of PP

English has over 100 prepositions.1 Prepositions are typically transitive, that is, they
take a complement:

(12) near the house, after the class, under the table, on the floor, etc.

Prepositions may be modified by an adverb. As with adjectives, we take this adverb to

be a specifier of PP:

(13) very near the house, right at the corner, immediately after the bank

Some prepositions are complex prepositions: they take another PP as a complement:

(14) close to the bank, away from home, contrary to her wishes, far from your house,
etc.

1
See [Link]
102 Héctor Campos & Bùi Huỳnh Thủy Thương

Some prepositions may function as adverbs. Typically, when used intransitively, they
are considered adverbs:

(15) a. She lives close.

b. They went very far.

Traditional English grammar considers these forms adverbs. Other grammars consider
them as prepositional adverbs and yet others as adverbial prepositions. Given our early
condition that trees represent categories, not functions, we will assume that these
forms are actually prepositions. When used intransitively, as in (15a), they function as
adverbs. But in both instances in (16) near is a preposition:

(16) a. She lives [near]. b. She lives [near the house].

So far we have seen prepositions that take specifiers (13) and complements (12, 13, 14).
Can prepositions take modifiers? As with adjective phrases, PP may include a modifier
with resultative clauses:
 English Syntax & Universal Grammar 103

(17) so far from school that he cannot walk

PRACTICE 2

I. Diagram the structures in brackets. Expand all categories.

a. She is [very worried about the problem with her parents].
b. My sister will [visit her boyfriend right after class].
c. She is [quite afraid of big black spiders].
d. I [put the book right on the table near the window].
e. He [appeared from behind the big black curtain].

II. The following sentences are ambiguous. Propose structures to account for the
different readings and briefly explain the readings.
a. She sent postcards from Spain.
b. She waved to the girl with her scarf.

6.3. The structure of ADVP

Bare adverbs are very common in English. Adverbs may be modified by another
adverb as in (18b). As adverbs modifying adjectives, we will take adverbs modifying
other adverbs to be specifiers, as you can see in (19):

(18) a. Ha speaks French [well].

b. Ha speaks French [very well].
104 Héctor Campos & Bùi Huỳnh Thủy Thương

(19) a. well b. very well

A logical question to ask at this point is whether the ADV in the specifier position in
(19b) (as well as the adverb that appears as a specifier for AP in examples 2 and 13) can
be expanded as a full adverb. The same question may be asked for the determiner in a
noun phrase like [the girl], where we have labeled the determiner as D, not following
X'-Syntax. We will discuss determiner phrases in Chapter 8 and we will show that
determiners can be made to follow X'-syntax. No such claim has been made for adverbs
in specifier position yet, so we will continue to assume that (19b), (2) and (13) is the
correct structure for adverb specifiers for the moment.

Applying X'-Syntax to adverbs, the next logical question to ask is whether adverbs can
take complements? The answer is positive:

(20) a. This Faculty operates [separately from the university].

b. This device will operate [ independently of your computer skills].

(21) a. separately from the university b. independently of your computer skills

Oblique case

Oblique case
 English Syntax & Universal Grammar 105

As with nouns and adjectives, which do not have the ability to assign Case, we will
assume that the OF that appears in (21b) is the result of OF-insertion necessary to assign
(Oblique) Case to the NP your computer skills. In (21a), the NP university can get
Oblique Case from the preposition, thus no OF-insertion is necessary.

We have seen that adverbs can appear bare, can take complements and specifiers. Can
adverbs also take modifiers? Again, the answer is affirmative when we consider
resultative clauses, as in (22):

(22) She was riding her bike so carelessly [that she fell down].

As with resultative clauses in adjectives, this is a correlative construction. The structure

for (22) is shown in (23):

(23)

We thus see that our X'-syntax model that we introduced in Chapter 3 can be easily
applied to adverb phrases as well.

In (20) – (23) we have discussed adverbs that modify VPs. What about adverbs that
modify other adverbs, as in (19b), or adverbs that modify prepositions, as in (13) and
(17), or adverbs that modify adjectives, as in (2)? If you check the structures assigned to
these structures, you can see that we generated these adverbs as heads (ADV) in their
respective specifiers. To make them comply with X'-syntax we will expand them fully
as intransitive ADVPs. The respective structures are shown in (24):
(24) a. very pretty b. very near the house
106 Héctor Campos & Bùi Huỳnh Thủy Thương

c. so far from school that he cannot walk

As you can see in the structures above, now all the categories comply with X'-syntax.
The only exception is D in (24b), since we have D without having a DP, thus violating
X'-syntax. We will return to the structure of Determiner Phrases in Chapter 8.

6.4. Summary

We have seen that our model of X'-syntax can easily be extended to AP, PP and ADVP.
All three categories may appear bare, they may take a complement, a specifier or a
modifier. We have observed the isomorphism that holds between the verb and the
derived adjective. This isomorphism motivated the insertion of the pseudo-preposition
OF when it appears in front of the complements of an adjective (see 8b). OF-insertion is
also used as a last resort operation by adverbs to save their structures from crashing (see
21b). OF-insertion guarantees that the complement of nouns, adjectives and adverbs, all
categories that do not assign Case, can appear with a complement that satisfies the Case
Condition.2

2
For a review of the basic principles of X'-syntax, I highly recommend viewing Caroline Heycock's
video "The X-bar Schema" in:
[Link]
TFwO&index=7
108 Héctor Campos & Bùi Huỳnh Thủy Thương

CHAPTER 7

MAKING SENTENCES : THE STRUCTURE OF IP

We have now applied our X'-syntax model to VP, NP, AP, PP and ADVP. In this
chapter we will expand our system to the sentence level.

7.1. Subject and Predicate.

In traditional grammar, sentences are said to consist of a subject and a predicate. The
predicate provides information about the subject while the subject is necessary to make
the predicate complete. Consider the sentence in (1):

(1) Jack loves Susan.

The sentence in (1) has two parts: the subject Jack and the predicate love Susan. The VP
[loves Susan] is the predicate. The head of this predicate is the verb LOVE. But this
predicate is not complete since LOVE needs two arguments, an Experiencer (to do the
loving) and a PATIENT or THEME (to receive this love). We say that Susan is the
internal argument of LOVE in (1). Internal arguments are found inside the VP.
However, the VP [loves Susan] cannot form a sentence without the external argument
and, at the same time, the external argument cannot express a whole sentence unless it
has a predicate to combine with. Once we manage to put the NP [Jack] and the VP [love
Susan] together, we can say that the NP [Jack] is the external argument of [love
 English Syntax & Universal Grammar 109

Susan] since it is outside the VP. Notice also that the verb is missing the tense and the
agreement in the two structures shown above.

How shall we put this subject and predicate together in our model? What is the element
that can "connect" the subject and the predicate in (1)? The answer is inflection. The
morpheme -s is what connects the two elements. On the one hand -s indicates that the
sentence is in the present tense, tense being an integral part of the verb. On the other
hand, -s is also the realization of the agreement triggered by the subject Jack. English
has a poor system of subject-verb agreement. In languages like Spanish, for instance, we
see the endings vary as the subject varies:

(2) yo cant-a-ba-Φ I used to sing

tú cant-a-ba-s you used to sing
él/ella cant-a-ba-Φ he/she used to sing
nosotros cant-á-ba-mos we used to sing
vosotros cant-a-ba-is you used to sing
ellos/ella cant-a-ba-n they used to sing

In (2) we can see that each pronoun triggers a different personal agreement (-Φ, -s, -
Φ, -mos, -is, -n). -Φ represents the zero-morpheme. In English, on the other hand, every
person has the same zero-agreement in the past, although we do get a morpheme for
the past tense, namely -ed. In Spanish, the morpheme -ba- is the morpheme for the past
tense. The morpheme -a- is called the thematic vowel. As we discussed in chapter 1, in
Spanish there are three classes of verbs and each class bears a different thematic vowel.
Cant- is the root of the verb. The stem of a word is the basic form to which further
affixes can be attached. The thematic vowel is an integral part of the verb. The root
and the thematic vowel together are called the stem of the verb: this is the part of the
verb that precedes all inflectional elements. The tense morpheme as well as the personal
agreement endings will vary depending on the tense and person that we combine the
verb with. We say that the morpheme -ba- and the personal agreement constitute the
inflection of the verb. The different morphemes of the Spanish form cantabas 'you used
to sing' are shown below:

(3) cant –a – ba –s 'you used to sing'

root thematic vowel tense marker personal agreement
STEM INFLECTION

In the case of English, we just have a stem/root and inflection. Notice that tense and
agreement are represented by a single morpheme in English and not as two separate
morphemes, as in Spanish:
110 Héctor Campos & Bùi Huỳnh Thủy Thương

(4) a. (he/she) sing -s '(he/she) sings'

stem/root tense marker + personal agreement
b. (they) sing -Φ '(they) sing'
stem/root tense marker + personal agreement
c. (he/she) work -ed '(they) worked'
stem/root tense marker + personal agreement

In our model we will use inflection as the glue that puts together the subject and the
predicate. As we mentioned above, inflection contains crucial information about the
predicate (the tense) and about the subject (the nominal features that would trigger
agreement on the verb). Applying X'-syntax to inflection, inflection joins the subject to
the predicate as follows:

(5)

Applying X'-syntax to the structure of the sentence, we generate the subject as the
specifier of IP while we generate the predicate as a complement of I.

7.2. Diagramming sentences

The structure for our sentence in (1) is shown in (6):

(6) John loves Susan

 English Syntax & Universal Grammar 111

A Spanish example is given in (7):

(7) Los chicos tocaban la guitarra

The boys used-to-play the guitar

In languages where there is no inflection (as well as in English in forms that bear no
visible inflection) we will assume a zero-morpheme, which we will represent as –Φ.
Consider the example:

(8) You love me

SOV languages like Japanese will project the head of IP to the right instead of to the
left. The same applies to complements of the verb, which are also projected to the left of
the verb. Remember that Japanese has special markers to indicate the subject (-wa) and
the object (-o) of the sentence. As we claimed in Chapter 4, these elements are inserted
into our original structure:
112 Héctor Campos & Bùi Huỳnh Thủy Thương

(9) Watashi-wa bunshoo-o kakimas-ita

I-SUBJ homework-OBJ write-PAST

7.3. Putting the verb and inflection together

Now that we have merged the subject and the verb in the same tree, we need to make
sure that the verb and the inflection get together. The inflectional affixes in (6) – (9)
cannot be left without an attached verb. As we saw in chapter 2, inflectional
morphemes in English are bound verbal affixes and as such, they need a verb to
attach to.

There is a principle that prevents bound morphemes from being left unattached. This is
the Stranded Affix Condition shown in (10):

(10) STRANDED AFFIX CONDITION

A bound affix must be attached to the appropriate root or stem.

Consider the tree in (4), repeated here as (11):

(11) John loves Susan.
 English Syntax & Universal Grammar 113

We have two possibilities to satisfy the Stranded Affix Condition in (10):

(a) either the verb moves up to the inflection (we call this movement "V to I
movement" )
(b) or the inflection moves down to the verb (we call this movement "I to V
movement")

How can we decide which is the correct strategy that English uses to satisfy the
Stranded Affix Condition? To decide, it is useful to compare English with other
languages.

Let us compare English with Spanish. We separate inflection from the root for our
examples:

(12) a. Mary always do-es her homework. (English: ADV>V)

b. Los estudiantes hace-n siempre la tarea. (Spanish: V>ADV)
The students do-PRES,IIISG always the homework
c. *Mary does always her homework. (English: *V>ADV)
d. *Los estudiantes siempre hacen la tarea. (Spanish: *ADV>V)
The students always do the homework

Notice that in English the adverb must precede the verb while in Spanish the verb must
precede the adverb. In English the adverb may not appear between the verb and its
direct object but this is possible in Spanish. How can we explain this important
difference between English and Spanish?

We will assume that English and Spanish have the same starting point at D-
Structure, with the adverb merged as a modifier to the left of the verb:

(13) D-Structures for English and Spanish:

114 Héctor Campos & Bùi Huỳnh Thủy Thương

We see that at D-Structure, English and Spanish have exactly the same syntactic D-
structure: I precedes the VP and the ADVP, a modifier merged to the left of V,
precedes the verb. How can we get the verb and the inflection together so as to satisfy
the Stranded Affix Condition in (12)?

We saw in (12a) that in English the adverb precedes the inflected verb (i.e, the verb
plus the inflection), while in Spanish it is the inflected verb which precedes adverb. We
can get the correct word order for both languages if we assume that inflection moves to
the verb in English while the verb moves to inflection in Spanish, as shown in (14):

(14) S-Structure for English and Spanish

Mary always do-es her homework Los estudiantes hace-n siempre la tare

Remember that in the model we are pursuing in this book there are two levels of
representation: a D-structure level, shown in (13), or the level before the movement
takes place and an S-Structure level, shown in (14) above, the level after the movement
has taken place. Movement is one of the operations that connect the two levels to derive
the final verb forms in (14). This movement of I to V in English and of V to I in Spanish
is a type of head-movement, since we are moving the head of IP (the I-node) or VP
(the V-node). To indicate that it is the head that moves we put a circle around the I or
the V head that is moving and use an arrow pointing to the place to which it is moving.
(15) shows the model we have been building so far. In this book we will use S-
Structures like those in (14) to represent final structures, since both D- as well as S-
Structure can be easily read off from them. The structures in (14) represent the S-
Structures; if we eliminate all the elements marked as inserted (with a double arrow
underneath) and all the elements marked as moved (marked with an arrow indicating
the origin and destination of the movement), we can easily read off the corresponding
D-Structures.
 English Syntax & Universal Grammar 115

Our system will also allow for deletion although we have not seen any cases yet. We
summarize our model in (15) below:

(15)

X’-Syntax Lexicon
SyntaxX’-
Syntax Lexicalization Principle
X’-Syntax
D-Structure
X’-Syntax
Transformation: Insert, Move, Delete
X’-Syntax
S-Structure
S-Structure
PF LF
LF

PRACTICE 1

I. Diagram the following sentences in English and in Spanish:

a. The students seldom come to class on time.
b. Los estudiantes llega-n siempre a clase a tiempo.
The students arrive-PRES, IIIPL always to class on time

II. Consider the following sentences in French and in Italian. Do we have

movement of V to I or I to V in these languages? Justify your answers. Show the
corresponding trees. TV indicates the thematic vowel; PRES, present tense, and
IIIPL third person plural:

(a) French:
Les étudiants parl-e-nt toujours avec le professeur.
The students speak-TV-PRES, IIIPL always with the teacher

(b) Italian:
Gli studenti parl-a-no sempre con il professore.
The students speak-TV-PRES,IIIPL always with the teacher'
116 Héctor Campos & Bùi Huỳnh Thủy Thương

RESEARCH PROJECTS

1. Consider now Vietnamese. We said before that languages with no verbal inflection
will be claimed to have zero-morphology. What type of movement do we get in
Vietnamese to satisfy the Stranded Affix Condition? How do your facts support the
claim that Vietnamese has zero-morphology for tensed verbs?

2. Consider now Japanese. Remember that Japanese is an SOV language. How does
Japanese satisfy the Stranded Affix Condition?
Seito-wa tsuneni sensei to hanashimas-ita.
Students-SUBJ always teacher with speak-PAST,IIIPL
'The students always speak with the teacher.'

7.4. Incorporating Modal Verbs into IP.

In the previous section we learned how to analyze and diagram sentences that contain
just one main verb. But there are constructions in English that may contain more than
one verb. We will start analyzing those constructions here.

English has a series of modal verbs that are used to express different modalities like
doubt, possibility, obligation, recommendation, ability, request, necessity, prohibition,
duty, criticism, logical assumption, permission, suggestion, advice, etc.

The English modal verbs are listed in (16):

(16) MODAL VERBS IN ENGLISH

can – could, may – might, will – would, shall – should, must

In the next section we will learn how modals are different from main verbs:

7.4.1 Differences between modal and main verbs. Modal verbs differ from main
verbs in several respects. Here we will list only 8 differences (there are more!), to show
that main verbs vastly differ from modal verbs:
 English Syntax & Universal Grammar 117

(17)

MAIN VERBS MODAL VERBS

1. Main verbs are inflected in the third 1. Modal verbs never show subject-verb
person singular in the present tense: agreement in the present tense:
a. She speak-s a. *She cans/wills speak English
b. *She speak-Φ b. She can/will speak English.
2. A main verb may be followed by another 2. Modal verbs must be followed by a verb in
verb provided that this verb is in the the simple infinitive form (the infinitive
infinitive, the simple infinitive or the without the to):
progressive form:
c. Ha likes to speak English. c. Quan can speak English very well.
d. Ha lets Quan cook for her. d. Quan will visit America soon.
e. Ha enjoys speaking English. e. Quan should look for a job.
3. Main verbs require do-support to form 3. Modal verbs do not need do-support to
negative sentences: form negative sentences. The negative
form –n't can attach directly to the modal:
f. Khiem doesn't like to swim. f. Khiem can't do the front crawl stroke.
4. Main verbs require do-support to form 4. Modal verbs do not require do-support to
interrogative sentences: form interrogative sentences:
g. Does Khiem know Ha? g. Will Ha visit Quan this summer?
h. Where did Khiem meet Ha? h. Where should Quan work next year?
5. Main verbs have infinitive forms: 5. Modal verbs do not have infinitive forms:
i. to speak, to like, to enjoy, to meet i. *to can, *to will, *to should
6. Main verbs have progressive forms: 6. Modal verbs lack progressive forms:1
j. speaking, liking, enjoying, meeting j. *canning, *willing, *shoulding
7. The past tense form on a main verb 7. A past tense form on a modal verb does
typically indicates past time: not necessarily indicate past time:
k. Ha liked Khiem's performance k. Ha might/should go to Nghe An next
yesterday/*tomorrow. year.

1
Because modal verbs lack infinitive and progressive forms, we say that they are defective verbs.
118 Héctor Campos & Bùi Huỳnh Thủy Thương

8. We may get any number of verbs in 8. We may only have ONE modal verb per
sequence provided that are in the right sentence:2
form: l. *Quan will can get a job soon.
l. Ha asked Khiem to tell Quan to go m.*Quan will must go to the doctor soon.
fishing with her.

We thus conclude that modal verbs form a separate class from main verbs. Our
syntactic model must thus treat modal verbs and main verbs differently.

7.4.2 A syntactic representation for modal verbs. In many languages modals are
bound morphemes found inside the verb, as shown in the following examples from
Spanish:

(18) Yo cantaré, tú cantarás, él cantará; yo cantaría, tú cantarías, ...

I will sing, you will sing, he will sing I would sing, you would sing, ...

We see in (18) that the future morpheme WILL is a free morpheme in English while it
is a bound morpheme in Spanish. Both in English and in Spanish the future morpheme
(represented as a modal verb or as bound morpheme) will be in Inflection.

Finally, notice that when there is a modal the verb bears no tense:

(19) a. I will go to Hanoi next weekend.

b. *I will went to Hanoi next weekend.

We will thus propose that modals are merged directly in inflection. The structure for
the two sentences containing modal verbs is shown in (20):

(20) a. Ha should go to Nghe An. b. Quan will go to Dalat.

2
Some American dialects allow for might to co-occur with could:
(i) I might could go.
There is evidence that in these dialects might-could may be a unit, a complex modal. Some arguments
support this claim: they will always appear in this order, these are the only two modals that can co-
occur, no adverb may appear between the two modals.
 English Syntax & Universal Grammar 119

In the structures in (20), the modal verb is a free morpheme. In Spanish, on the other
hand, it is a bound morpheme. As we saw before, the verb would move to inflection in
Spanish to satisfy the Stranded Affix Condition:

(21) Quan i-rá siempre a Dalat.

Quan go-will always to Dalat
'Quan will always go to Dalat'

The fact that the adverb siempre 'always' is between the verb and the PP complement a
Dalat 'to Dalat' suggests that the V has moved to I in Spanish.

7.4.3 A note on semi-modal verbs. Some verbs are considered semi-modals in

English because they share some properties both of modal and main verbs. Some semi-
modals are need, dare, have got to, ought, had better, used to, etc.

Take need, for instance. Similar to other modal verbs, it does not require do-support to
form the negative form, the negative morpheme can be attached directly to it, as shown
in (22a). Neither do we need do-support to form questions, as shown in (22b):

Need as a semi-modal Need as a main verb

(22) a. You needn't say that. a'. You don't need to say that.
b. Need you go now? b'. Do you need to go now?

Different from real modal verbs, however, the semi-modal need can also be used as a
main verb, as shown in (22a') and (22b') without much difference in meaning. In fact,
these are the structures most commonly used in American English. Different from real
modal verbs, the semi-modal verb need is rarely used in affirmative sentences. For
120 Héctor Campos & Bùi Huỳnh Thủy Thương

affirmative sentences, we typically resort to a main verb structure, as shown in (23),

where the double question mark (??) indicates that this sentence is not typically said by
speakers:

(23) a. ??We need go to the supermarket.

b. We need to go to the supermarket.

To use it in the past tense, we resort to the auxiliary have, similar to what happens with
real modals, as shown in (24):

(24) a. Peter must be at home now. Peter must have been at home last night.
b. You needn't have called me last night.

We thus see that need has a mixed behavior.

Consider next ought. Different from modals, ought requires an infinitive, but similar to
modal verbs, it requires the auxiliary have to express obligation/regret in the past:

(25) a. She ought to come tomorrow.

b. She ought to have come yesterday.

To form a question, we typically do not use ought, but should:

(26) a. ??Ought I to go to her party?

b. Should I go to her party?

Ought is typically not used in the negative form; we typically use shouldn't. If it is used,
it is typically used with a simple infinitive or with the unreduced negative form:

(27) a. *You oughtn't to call her./ You shouldn't call her.

b. You oughtn't leave./ You ought not to leave.

We will not include semi-modals into our analysis and will just concentrate on modal
and auxiliary verbs.

PRACTICE 2

Diagram the following sentences. Expand all nodes.

a. The syntax professor from a foreign university teaches linguistics in the morning.
b. The professor from Chile should teach generative syntax in the afternoon.
 English Syntax & Universal Grammar 121

RESEARCH PROJECTS
1. We mentioned that the verb dare is also a semi-modal. Can you show why it is
typically not considered a modal verb?
2. Had better is sometimes listed as a modal verb, sometimes as a semi-modal. Can
you summarize its behavior and defend whether it should be considered a modal
or a semi-modal verb? Or does it exhibit a mixed behavior?
3. Discuss the behavior of used to.

7.5. Auxiliary Verbs in English

English has three auxiliary verbs: have (+ past participle) to show perfective aspect,
be (+ progressive form) to show progressive aspect and be (+ past participle) to show
passive voice.

Auxiliary verbs share some properties with modal verbs:

1. Auxiliary, like modal verbs, do not need do-support to form the interrogative or
negative form:

(28) a. Have you seen the movie Angry Birds yet?

b. Phuong's not planning to go to the movies with Quan.

Notice that this property also distinguishes auxiliary verbs from main verbs. Main verbs
would require do-support in interrogatives and negatives.

2. Auxiliary verbs do not need do-support to form a question tag:

(29) a. You have seen the movie Angry Birds, haven't you?
b. Ha is not planning to go to the movies with Khiem, is she?

This property also distinguishes auxiliary verbs from main verbs. Main verbs require
do-support for their question tags.

3. Auxiliary verbs have infinitive and progressive forms: to have gone, to be going, to
be built. In other words, unlike modal verbs and more like main verbs, they are not
defective verbs.

But auxiliary verbs also differ from modal verbs in other respects:
122 Héctor Campos & Bùi Huỳnh Thủy Thương

4. We saw above that modal verbs are typically followed by verbs in their simple
infinitive form (the infinitive form without to). The perfective auxiliary HAVE is
always followed by a past participle, the progressive auxiliary BE is always followed by
a progressive -ing form and the passive auxiliary BE is always followed by a past
participle.

5. We saw that we can typically get one modal verb per sentence, but we can get up to
three auxiliaries in the same sentence. Notice that the auxiliaries need to appear in a
specific order, should all three auxiliary verbs co-occur: the perfective have precedes
the progressive be and, in turn, the progressive be precedes the passive be:

(30) That house has been being built for over 10 years.
PERF> PROG> PASS

6. Auxiliary verbs precede main verbs, but they follow modal verbs:

(31) That house must have been being built for over 10 years.
MODAL > PERF>PROG>PASS

The sentence in (31) clearly shows that auxiliary verbs cannot be analyzed as modal
verbs since only one modal verb is allowed per sentence. The fact that auxiliaries
always precede main verbs and appear in a fixed order makes us conclude that they are
generated as separate categories from modal and main verbs. We will explore their
structure in the next section.

Consider the example in (32). We will posit that auxiliary verbs are generated as AUXP.
(32) shows an example with the perfective auxiliary have. We will generate past
participles directly in VP.

(32) Toan should have called his girlfriend.

 English Syntax & Universal Grammar 123

An example with the progressive auxiliary be is shown in (33). The progressive -ing
form is also generated directly in VP:

(33) Toan must be calling his girlfriend.

We saw above that auxiliary verbs, unlike modal verbs, may co-occur. Consider
example (34) with both a perfective and a progressive auxiliary. We will ignore passive
auxiliaries here as we will discuss passives in chapter 13.

(34) Toan must have been calling his girlfriend.

124 Héctor Campos & Bùi Huỳnh Thủy Thương

What happens when we have an auxiliary without a modal verb? Consider example
(35):

(35) Toan is calling his girlfriend.

We will claim that auxiliary verbs move to I. We know that auxiliary verbs move to I
because the auxiliary will keep on moving up the sentence by itself when we form a
question: Is Toan calling his girlfriend?

When we have two co-occurring auxiliaries, the highest auxiliary (the one closest to the
inflection) is the one that moves:

(36) Toan has been calling his girlfriend:

 English Syntax & Universal Grammar 125

In 4.1 we summarized the main differences between modal and main verbs. The chart in
(37) shows the differences between modal and auxiliary verbs:

MODAL VERBS AUXILIARY VERBS

[Link] verbs do not take agreement: 1. Auxiliaries agree with the subject:
a. I/We must call Thuong. a. Nu Anh has called Trieu.
b. Thuong must/*musts call me. b. Sang is thinking about Hoa.
c. Khuong may/*mays visit the US. c. The teacher was invited to a banquet.
2. Modal verbs do not have infinitival or 2. Auxiliaries have infinitival and gerundive
gerundive form. There is an alternative form:
form to express those forms: a. To have invited Quynh was a good idea.
a. *to can, *canning b. Having ridden to school with Phuoc,
a’. to be able to, being able to the teacher arrived on time.
b. *to must, *musting c. Being invited to present, Trieu talked
b’. to have to, having to about language acquisition.
3. Modal verbs are always followed by a 3. Each auxiliary takes a verb in a particular
verb in its bare (or simple) infinitive form:
form: a. Nhu has learned many new dances.
a. Phuc must sing very well. b. Khoi is learning many new dances.
b. Binh should read the new book. c. Hanh was invited to teach syntax.
c. Huong may take a vacation next
week.
4. Modal verbs cannot co-occur. To express 4. Auxiliaries can co-occur. They must always
the idea we use an alternative form: appear in the order: Perfective (have +pp) >
a. *I will can go to the party. Progressive (be + V-ing), Passive (be + pp):
b. I will be able to go to the party. a. It has been being built for 10 years.
PERF > PROG > PASS
5. Modal verbs precede auxiliaries: 5. Auxiliaries follow modal verbs:
a. Hoang must have studied Spanish. a. Khuong may have lived in Cuba.

(38) summarizes the way that English satisfies the Stranded Affix Condition in (10):

(38) a. If there is a modal verb, no movement happens. The main verb appears in its
simple infinitive form. (Toan should call his girlfriend)
b. If there is no auxiliary verb, I moves down to the main verb in VP. (Toan
always calls his girlfriend)
126 Héctor Campos & Bùi Huỳnh Thủy Thương

c. If there is an auxiliary verb, the auxiliary moves up to I. (Toan has called his
girlfriend/Toan is calling his girlfriend)
d. If there are two or three auxiliary verbs, the highest auxiliary verb moves
up to I. (Toan has been calling his girlfriend)
PRACTICE 3

Diagram the sentences in (35) (repeated here as a-e below) to see review how
English satisfies the Stranded Affix Condition in (10):
a. Toan should call his girlfriend.
b. Toan always calls his girlfriend.
c. Toan has called his girlfriend.
d. Toan is calling his girlfriend.
e. Toan has been calling his girlfriend.

CHALLENGE

Diagram the following sentences. Indicate all the inserted or moved elements.
Where a sentence is ambiguous, briefly discuss the ambiguity of each structure.
Make sure you do not leave verbal morphology stranded to avoid a violation of the
Stranded Affix condition!
a. The journey from Washington to New York took us 5 hours.
b. The assistant will always offer valuable help to the desperate students of
generative syntax.
c. The experienced city guide should gladly have shown the King of Spain the
most important sights of her beautiful city.
d. The President of that country in South America has been ordering the arrest
of the advocates of the abolition of the death penalty.
e. i. O daskalos dida-kse tin ili s-tous foitites.
(Greek)
The teacher teach-past the-acc material-acc to-the-acc students-acc
ii. O daskalos dida-kse tous foitites tin ili.
The teacher teach-past the-acc students-acc the-acc material-acc
f. They ordered three concert tickets for them. (2 times ambiguous)
g. She is a student of high moral principles. (2 times ambiguous)
h. We saw the discussion of the riots in the bar. (3 times ambiguous)
 English Syntax & Universal Grammar 127

7.6. Summary

In this chapter we saw that our model of X'-Syntax can be extended to the structure of
the sentence assuming that inflection phrase (IP) is what holds the subject and the
predicate together. The specification for Tense and Agreement can be found at I, the
head of IP. Tense in I specifies the tense in which the verb appears and Agreement
specifies the Person (I, II or III) and the Number, singular (SG) or plural (PL) of the
NP in the specifier of IP. In our system the process of agreement is simply feature
matching: the person and number features of the NP must be the same as the features of
person and number in I. When the features match and all the principles and conditions
are obeyed, we say that the sentence converges. A convergent sentence is
grammatical. If there is a feature mismatch or we violate any of the principles or
conditions we have postulated, then we say that the (derivation of the) sentence crashes.
Derivations that crash give us ungrammatical sentences.

We showed that inflection moves to the verb in English, while in Spanish it is the verb
that moves to inflection. Inflection contains the features for tense and subject
agreement. We argued that inflection also contains modals in English. We showed that
modals behave syntactically different from auxiliary verbs. We merge auxiliary verbs as
independent categories, following modals but before main verbs. While only one modal
can be found per sentence, we saw that we can have all three auxiliary verbs at the same
time.
 English Syntax & Universal Grammar 127

CHAPTER 8

FROM NP TO DP

In this chapter we explore whether our model of X'-Syntax can be extended to DP.

8.1 Some basic properties of X'-syntax model

So far we have been implemented an X'-system to account for the structure of VP, NP, AP,
PP, ADVP and IP. The basic structure we have been trying to apply to all these categories is
known as X'-syntax and the basic structure is shown in (1):

(1)

Several properties are hidden in the structure in (1):

1. Every phrase in X'-syntax is endocentric, which means that the head or nucleus of a
phrase (X) and the maximal category (XP) must be of the same type. Thus a verb head (V)
must be under a verb phrase (VP), a noun head must be under a noun phrase (NP), etc.

(2) Phrases must be endocentric

This also means that phrases with structures such as the following are ruled out:
128 Héctor Campos & Bùi Huỳnh Thủy Thương

(3) Impossible structures

Although adjectives may turn into nouns (the rich, the poor, etc.) and sentences may appear
in places where noun phrases typically appear (for instance, in subject position: That Quan
studies hard is a fact), structures such as (3) are ruled out because they violate the law of
endocentricity imposed by X'-syntax.

2. Every tree will have a minimum of 3 levels:

(4)

(4) shows the minimal structure for every phrase. As shown in (1), all complements will be
merged under X'; modifiers (also called adjuncts) will be projected as additional X's.
Specifiers always branch off to the left of XP.

3. Every tree must obey the following architectural rule:

(4) Xn → Xm or Xn → Xm-1 (where n = m)

This means that when we build a tree we are allowed only two options: the lower node must
either contain the same number of bars (Xn = Xm), as shown in (5a), or the lower node must
be one bar lower than the upper node (Xn → Xm-1), as shown in (5b):

(5) a. lower node contains b. lower node is one bar lower

the same number of bars in value than the upper node
 English Syntax & Universal Grammar 129

We have already found structures like these in our trees: the one in (5a) is what we have
proposed for modifiers, the ones in (5b) is what we find in all our basic trees without

adjuncts. The second tree under (5b) satisfies our rule in (4) because XP is another way of
representing X", thus we have a node with two bars over a node with a single bar.

Our rules in (4) suggest that there may be two more possible structures, as shown in (6):

(6) a. a head over another head b. an XP over another XP

As we saw in chapter 2, the structures in (6a) are widely used to represent morphological
structures. (6a) may be adopted for our analysis of phrasal verbs (chapter 4), since we
argued that the verb and the particle form a word. Prepositional verbs have a PP-
complement, as we argued in chapter 4:

(7) a. look up the word b. think about the word

8.2 Revisiting our structure for NP

These properties now force us to reconsider the structure of NP. Take the structure in (8):
130 Héctor Campos & Bùi Huỳnh Thủy Thương

(8) the syntax student with long hair

If we follow the NP from its maximal projection (NP) to the head (N) we can see that
student is the head of this phrase. Syntax is a complement and as such, it is projected in the
lower N' projection. PP is a modifier and we project it as an adjunct to N'. If you check the
structures of all the phrases (NP, PP, AP) you will see that we have obeyed every rule of X'-
syntax. With one exception, though: the determiner on the left side of the structure!

Notice that in (8) we have projected a determiner as a D (a head), not as a phrase (DP). This
violates the second rule we discussed above requiring that every phrase have a minimum of
three levels.

Two possibilities come to mind to incorporate D into our X-syntax framework. We could
project DP following X'-syntax, leaving DP as the specifier. Or alternatively, we could
project DP following X'-syntax and we make NP the complement of D. The two options are
shown in (9):
(9) a. the car b. the car

In the next section we will show that (9b) is the correct structure for DP.
 English Syntax & Universal Grammar 131

8.3 Revising our structure for NP

Assuming the DP structure in (8b), our structure for (8) would be (10):
(10) the linguistics student with long hair

Check every phrase XP in the tree above and you will see that every principle of X'-syntax
discussed above has been respected. When there is no visible determinerfor D, we will
include the zero-determiner Φ.

8.4 Why is our DP structure better than our old NP structure?

A natural question to ask is why are (10) and (9b) better than our old structures?

8.4.1 Possessive constructions. Our old structure could not explain without major
complications the structure of possessive phrases:

(11) John's car

132 Héctor Campos & Bùi Huỳnh Thủy Thương

(11a) (11b)

The only way that our old structure could account for a possessive structure like (11) is by
positing the structure in (11a). But in our old system, we analyzed John as a noun phrase, as
shown in (11b). Ideally, we want the noun phrase John to always have the same basic tree
in both examples!

In addition, notice that the possessor may also include a determiner, which is a clear proof
that we are dealing with a NP possessor:

(12) a. my friend's car b. my friend

In fact, the possessor could even be a very long noun phrase:

(13) [the syntax student with long hair]'s car

It is clear from these three cases that our old structure could not generate possessive
structures in a coherent way. How can our new system generate these structures?

We will assume that the possessor is the specifier of the DP while the possessee is the NP
complement of D, as shown in (14):

(14) John's car

 English Syntax & Universal Grammar 133

In (14) John is the possessor and we merge it as the specifier of DP. The head of DP is the
possessor morpheme –'s. This structure easily allows us to include any noun phrase as a
possessor, including a possessor phrase which may contain a determiner, as in (15):

(15) my friend with long hair's car

Notice that in (14) and (15) the possessive form –'s is an affix. According to the Stranded
Affix Condition we cannot have unattached affixes, thus we must move this –'s to attach to
the adjacent DP, as shown in (16):
134 Héctor Campos & Bùi Huỳnh Thủy Thương

(16) a. My friend with long hair's car

In (16) we see that the possessor morpheme –'s attaches to the whole DP and not to the
noun hair, contrary to what we observe in the orthography of English. The possessor is not
really hair, but rather, the whole DP [my friend with long hair]. Thus we say that the
possessive morpheme –'s is a DP-affix.

Applying the same logic to our structure in (14), we would get the following structure:

(16) b. John's car

 English Syntax & Universal Grammar 135

PRACTICE 1

Diagram the following structures:

a. Ha's new tablet

b. The syntax students' old book
c. Ha's mother's new computer

8.4.2 Deverbal nouns and their subjects. Another reason why this structure is better
than our old structure comes from analyzing deverbal nouns, that is, verbs that are derived
from verbs. Consider the structures in (17):

(17) a. The enemy destroyed the city.

b. The enemy's destruction of the city

In (17a) we have a sentence. The internal argument (the direct object) is merged inside the
VP, as shown in (18a), while the external argument (the subject) is merged as a specifier of
an external category, IP, which is on top of VP, as shown in (18):

(18a) destroyed the city (18b) The enemy destroyed the city.

Consider now the deverbal noun destruction in (19):

136 Héctor Campos & Bùi Huỳnh Thủy Thương

(19a) destruction of the city (19b) the enemy's destruction of the city

(18b) and (19b) represent the D-Structure for the examples in (17). We need to ensure that:
(i) all nouns receive Case to satisfy the Case Condition and
(ii) we also need to make sure that we satisfy the Stranded Affix Condition.
Let us consider (18b) first, repeated here as (20):
(20) The enemy destroyed the city.

a. D-Structure b. S-Structure

In (20a), the direct object the city receives Objective Case from the verb. Thus the direct
object DP the city satisfies the Case Condition.1 What about the subject the enemy in (20)?
How does it satisfy the Case Condition? We will assume that a subject receives
Nominative Case from I when I contains Tense and Agreement. Thus in (20a) the
enemy is able to receive Nominative Case from I since it contains both Tense (PAST) and
Agreement (III SG). We can thus see that the two noun phrases in (20), namely the direct

1
Since NP has been replaced by DP in our system, the need to reformulate the Case Condition as in (i):
(i) The Case Condition
Every DP needs to bear Case.
 English Syntax & Universal Grammar 137
object the city and the subject the enemy, satisfy the Case Condition because both of them
are able to receive Case in (20a/20b). At D-Structure in (20a) we have a stranded affix (-ed).
Our Stranded Affix Condition forbids us from having unattached affixes. Thus at S-
Structure we must lower it to V so as not to violate the Stranded Affix Condition. This is
shown in (20b). We thus see that in (20b) our two well-formedness conditions have been
met: (i) all DPs receive Case and (ii) there are no unattached affixes in our tree.2

Consider now the DP structure for the deverbal noun in (21):

(21) the enemy's destruction of the city

a. D-Structure b. S-Structure

In (21) we can see the D- and the S-Structure for the DP the enemy's destruction of the city.
The DP complement DP the city in (21a) is not able to receive Case from destruction in
(20a) because nouns do not have the ability to assign Case, only verbs and prepositions
assign Case. Thus at S-Structure in (21b), we insert the pseudo-preposition OF to ensure
that the complement DP the city gets Oblique Case. We will assume that D with a
POSS(ESSIVE) feature is able to assign Genitive Case to its specifier. Thus the DP the
enemy can get Genitive Case both in (20a) and (20b). The Genitive Case morpheme -'s is a
DP-affix and in order to satisfy the Stranded Affix Condition it attaches to DP in (21b).

PRACTICE 2

Diagram the following structures. For each structure show both the D- and the S-
Structure.
a. The student summarized the story.
b. The student's summary of the story

2
See the Appendix to this chapter for a more detailed discussion of Nominative and Genitive Case
assignment within IP and DP, respectively.
138 Héctor Campos & Bùi Huỳnh Thủy Thương

c. Khiem applied for a visa at the American Embassy.

d. Khiem's application for a visa at the American Embassy

8.4.3 Postnominal determiners. Our new DP structure also provides us with an elegant
way to account for postnominal determiners. In this section we will consider postnominal
determiners in Romanian and in Vietnamese.

In Romanian the definite article follows the noun, as shown in (22):

(22) om-ul bun

man-the good
'the good man'

We can assume the following D- and S-structures for Romanian:

(23) a. D-Structure b. S-Structure

In Romanian, the definite article -ul is a nominal affix (that is, it needs an affix to attach to).
As such, we cannot leave it unattached in (23a). Thus at S-Structure, the head noun om
'man' moves to attach to -ul, forming omul 'the man'. We thus get a postnominal determiner
in Romanian.

Let us compare now two basic phrases in English and in Romanian, as in (24):
 English Syntax & Universal Grammar 139

(24) a. the good man b. omul bun

As we can see in (24), English and Romanian share the same D-Structure. The only
difference between the two languages is the fact that in Romanian the definite article –ul
'the' is an affix while in English the definite article the is a free morpheme. Thus in
Romanian, but not in English, we get movement of N to D to satisfy the Stranded Affix
Condition.

Finally consider postnominal determiners in Vietnamese. Vietnamese does not have definite
articles, but it does have demonstratives. Demonstratives typically follow the noun in
Vietnamese:

(25) giáo viên này

teacher this
'this teacher'

(26) a. *này giáo viên b. giáo viên này

This structure in (25) would have been impossible to generate with our old system since
specifiers are universally always projected as the left-most element of the phrase. The
structure in (25) can be easily represented with our new structure since the demonstrative is
no longer a specifier but the head of DP. All we need to do is project the DP backwards, as
shown in (26b).
140 Héctor Campos & Bùi Huỳnh Thủy Thương

There is further evidence that our structure in (26b) is correct. As we can see in (27a), a
complement of N would also precede the demonstrative, as would a modifier, as shown in
(27b). Notice that if both the complement and the modifier co-occur, as in (27c), then they
both precede the demonstrative, as our model predicts. Our model also predicts that if both
the complement and the modifier should co-occur, the complement should precede the
modifier. This is indeed what we find in (27c):

(27) a. giáo viên tiếng Anh này b. giáo viên trẻ này
teacher English this teacher young this
'this English teacher' 'this young teacher'

c. giáo viên tiếng Anh trẻ này

In section 6 we will reconsider the analysis of này 'this' as a determiner.3

3
Notice that tiếng Anh 'English' appears a direct object of the noun giáo viên 'teacher'. The alert reader
may wonder what Case this direct object has. Other languages show the same phenomenon as
Vietnamese, where no (inserted) pseudo-preposition is found, contrary to what we observe in English.
Instead, the observe that the complement of the noun appears in Genitive Case:
(i) a. enas dhaskalos tis ellinikis glossas (Greek)
a teacher the-genitive Greek-genitive language-genitive
'a teacher of (the) Greek (language).'
b. mësues të matematikës
 English Syntax & Universal Grammar 141

8.5 What elements can appear under D?

We saw in Chapter 5 that the category Determiner (D) includes definite and indefinite
articles, possessives and demonstratives. Now that we have a separate phrase for DP, they
will appear as the head of DP, as shown in (28):

(28) a. the house b. a house

c. my house d. that house

Count nouns must always appear with a determiner, while mass nouns may appear without
one:

(29) a. I buy sugar.

b. *I buy book.
c. I buy a book.

teacher the-genitive mathematics-genitive

'a teacher of math'
To accommodate facts like these which are present in many languages, we can assume that nouns (as
well as adjectives) assign Genitive Case. In English this Genitive Case is realized via the pseudo-
preposition OF (which is inserted) while in other languages it may be a particular ending on the noun
(and adjective, determiner, etc) that compose the DP. As it is usually the case in Vietnamese, Case is not
marked morphologically. Thus we will assume that the complement of giáo viên 'teacher' has Genitive
Case in all the examples in (27).
142 Héctor Campos & Bùi Huỳnh Thủy Thương

We will assume that DPs with no determiner have a silent determiner, which we will
represent as 'Φ'. This zero-determiner has the feature "indefinite", thus these DPs will be
interpreted as indefinite.

(30) a. buy sugar b. buy a book

Count nouns may appear without a determiner when used in the plural, when we want to
express an indefinite quantity. In this case, we have a zero determiner, which we also
represent as 'Φ' :

(31) I buy apples.

There are languages where this indefinite determiner is pronounced. French is one of such
languages, where the indefinite plural determiner in (32b) is called a "partitive":

(32) a. *J'achète pommes.

I buy apples
 English Syntax & Universal Grammar 143

b. J' achète despommes.

I buy "des" apples
'I buy apples.'
c.

PRACTICE 3

Diagram the VPs in brackets. Indicate all the inserted elements.

a. We [bought tickets for the Saturday game].
b. Il [écri-t des lettres sa copine]. (French)
He write-s "des" letters to his friend
'He writes letters to his friend.'
c. Ajo [vizit-on mësues-in shqiptar të matematikës]. (Albanian)
She visit-pres teacher-the Albanian the math
'I visit the Albanian math teacher.'

8.6 QP and classifiers in Vietnamese

We mentioned in Chapter 2 that nouns can be classified as count nouns (you can count
them) or mass nouns (you cannot count them).

(33) a. three books (count noun)

b. *three breads (mass noun)

In order to count mass nouns in English, we need to use measure words, as in (34):

(34) three loaves of bread, three pounds of bread, three slices of bread, etc.
144 Héctor Campos & Bùi Huỳnh Thủy Thương

The structure for the indefinite noun phrase (33a) is shown in (35b). This structure is similar
to the equivalent definite noun phrase in (35a), with the exception that the indefinite DP
has a zero-determiner. We will analyze numbers as QP's thatmodify the NP, as shown
in (35). A definite DP may be introduced by a definite article (the books), a possessive (my
three books) or a demonstrative (these three books). An indefinite DP is always
introduced by the indefinite article "a/an" or by a zero determiner (three books, some
books, many books, etc.) which bears the feature "indefinite":

(35) a. the three books b. three books

c. some/many books

In languages like Vietnamese, most nouns seem to behave like mass nouns in English.
Thus, in order to count nouns in Vietnamese, we need a special word, called a classifier,
which allows us to count or individualize the noun. Vietnamese has no definite article, but a
noun phrase can be made definite by adding a demonstrative or a qualifying phrase:
 English Syntax & Universal Grammar 145

(36) a. nhà
b. *một nhà, *một nhà đẹp
c. một căn nhà, mấy căn nhà, căn nhà này, căn nhà ở cuối phố
'a house, some houses, this house, the house at the end of the street'
We can express a definite noun phrase by using a demonstrative, as shown in (37).
Different from English, where the QP modifies the NP, we will assume that in Vietnamese
the quantifier modifies the classifier phrase CLP.
(37) a. Tôi thích căn nhà này. b. Tôi thích ba căn nhà này.
I like CL house this I like three CL houses these
'I like this house.' 'I like these three houses.'

a'. căn nhà này b'. ba căn nhà này

CL house this three CL house these
'this house' 'these three houses'

We see in (37b') that the quantifier (QP) modifies the classifier phrase (CLP). This analysis
captures a very important fact of Vietnamese grammar: number quantifiers almostalways
appear with a classifier.4At the end of this section we will return to explain the fact that
the form này 'this' also requires a classifier.

The structure in (37a') shows a singular noun phrase with a demonstrative. To express a
corresponding plural form we can use a numberhigher than one, as shown in (37b'). We can
also express a plural form resorting to the plural marker những. We will generate these
plural markers as quantifiers, thus capturing the fact that the plural marker những and
number quantifiers are in complementary distribution (we can use one or the other):

4
In Vietnamese there are a few nouns that may appear without classifiers. Some examples include giáo
viên 'teacher', học sinh 'student', phụ nữ 'woman', etc.
146 Héctor Campos & Bùi Huỳnh Thủy Thương

(38) a. những căn nhà này b. căn nhà này

Plural CL house this CL house this
'these houses' 'this house'

As in English, an indefinite noun phrase may be expressed without a determiner or a

demonstrative, as shown in (39). We will assume that these constructions contain a zero-
determiner, which in this case corresponds to a zero-indefinite demonstrative since
Vietnamese has no articles per se. We will represent this as 'Φ-indef'. This zero-determiner
is equivalent to the English Φ determiner that we saw in (35b), with the difference that in
Vietnamese the head D of DP is found to the right of the NP and the NP includes a CLP as
well:

(39) a. Cô ấy mua ba căn nhà. b. Cô ấy mua mấy căn nhà.

'She bought three houses.' 'She bought some houses'

a'. ba căn nhà b'. mấy căn nhà

three CL houses some CL houses
'three houses' 'some houses'
 English Syntax & Universal Grammar 147

The quantifiers vài and một vài may also be used to express an indefinite plural noun phrase
(vài căn nhà, một vài căn nhà 'some houses'). Notice that in these examples, the presence of
the classifier is obligatory. This follows from the fact that quantifiers modify CLP in
Vietnamese, while QPs may modify NP directly in English.5

We saw in (29a) that we can use a bare NP (an NP with no determiner) when the noun is
interpreted as generic or as a class. We observe the same phenomenon in Vietnamese:

(40) a. I like to read interesting books.

b. Tôi thích đọc sách hay.
I like read books interesting
'I like to read (interesting) books.'

We can see that both in English and Vietnamese we can use a bare NP to express the
general idea of books as a group. It is interesting to observe that in these cases, Vietnamese
does not use a classifier. The structure for (40b) is shown in (41). In this case, we will
assume that in these generic noun phrases, the adjective modifies the NP directly in
Vietnamese:

(41)

5
Diferent from the quantifiers mấy and vài, the quantifier nhiều 'many' may be used with or without the
classifier (nhiều căn nhà, nhiều nhà). We will return to this fact later in the discussion.
148 Héctor Campos & Bùi Huỳnh Thủy Thương

As we mentioned above, the quantifier nhiều 'many' is different from the quantifiers
equivalent to 'some' (vài, một vài, and mấy) discussed above: the latter always require a
classifier, however nhiềumay appear with or without a classifier, as noted in (42):

(42) a. Ở đây bán nhiều sách hay.

Here they-sell many books interesting
'They sell many interesting books here.'
b. Ở đây bán nhiều quyển sách hay.
Here they-sell many CL books interesting
'They sell many (kinds of) interesting books here.'

Some speakers report no difference between (42a) and (42b), but others note a very subtle
difference that may be explained through our analysis. For some speakers, "books" is
considered as a unit or group in (42a), while in (42b) we want to emphasize the individual
books (which are many). This intuition can be explained if the quantifier nhiều 'many' may
modify either the NP or the CLP, as shown in (43):

(43)
a. nhiều sách hay b. nhiều quyển sách hay
many books interesting many CL books interesting
'many interesting books' 'many (kinds of) interesting books'
 English Syntax & Universal Grammar 149

The "unit or mass reading" that some speakers report is explained by (43a) since both the
QP and the AP modify NP. The NP has not been partitioned by a classifier. Thus the noun
phrase is taken as a unit, it has not been broken or separated. But in (43b), both the QP and
the AP modify CLP, thus the CLP is active, explaining why speakers feel like they are
talking about individual or different kinds of books.

Unlike English, Vietnamese seems to have a silent definite determiner (a silent

demonstrative, to be more exact, since Vietnamese does not have any articles) that is used
with definite noun phrases. However, to license this definite determiner, the noun needs to
be modified by a phrase that can make it definite. In these cases, the classifier must still be
used, as it is used with demonstratives:

(44)
a. Căn nhà [ở cuối phố] rất đẹp. b. Ba căn nhà [ở đấy] rất đẹp.
The house (at the end of the street) is The three houses (there) are very pretty.
very pretty.

(45)
a. căn nhà ở cuối phố b. ba căn nhà ở đây
CL house at end street three CL houses at here
'the house at the end of the street' 'the three houses here'
150 Héctor Campos & Bùi Huỳnh Thủy Thương

We will assume that in these constructions there is a definite zero-determiner, which we

will represent as "Φ-def", as shown in (45). This zero-determiner is the equivalent of a
demonstrative, as in (37a'). In fact, speakers usually tend to add a demonstrative like này
'this', kia 'there' or đó 'there' to these constructions to make them sound more natural.

Similar to the construction in (38a), we can use the plural marker những to make the plural
of (37a). This is shown in (46):
(46)
a. những căn nhà ở cuối phố b. căn nhà ở cuối phố
plural CL houses at end street CL house at end street
'the houses at the end of the street' 'the house at the end of the street'
 English Syntax & Universal Grammar 151

If the context is sufficiently clear, these constructions may also be found without the
specifying PP, in which case we would say that the context licenses the definite determiner:

(47) Context: (-Quyển sách bạn mua hôm qua đâu rồi?)
('Where is the book you bought yesterday?')
a. -Quyển sách ở trên bàn. b. quyển sách
'The book is on the table.' CL book
'the book'

Finally, consider the universal quantifier all in English, which may precede a definite noun
phrase:

(48) a. all the books

b.

As we can see in (48), the universal quantifier all in English, different from the number and
indefinite quantifiers discussed before, can precede the full DP. The same fact can be found
in Vietnamese:
152 Héctor Campos & Bùi Huỳnh Thủy Thương

(49) a. Tôi mua tất cả những quyển sách này.

I bought all Plural CL book these.
'I bought all these books.'
b.

This section does not pretend to present a complete analysis of classifiers and quantifiers in
Vietnamese. It is just a prelimimary approach that can explain some crucial differences
between English and Vietnamese. Some of the main differences are listed below:

(a) Vietnamese DP projects to the right. English DP projects to the left.

(b) Vietnamese DP includes a classifier phrase CLP. English does not have classifiers.
The closest elements in English equivalent to Vietnamese classifiers are the partitive
constructions we discussed in chapter5, section 9.

(c) Numbers and plural marker những are analyzed as QPs and they modify CLP. This
explains why every time we use a number quantifier or a plural marker in
Vietnamese, a classifier is obligatory with most nouns. In English number
quantifiers modify the NP directly.

(d) The universal quantifiers all in English and the equivalent tất cả in Vietnamese are
also analyzed as QP, but they take DP as a complement. This explains why they may
occur with a determiner in English or with a demonstrative in Vietnamese.
 English Syntax & Universal Grammar 153
(e) While English only has an indefinite zero-determiner, Vietnamese has both an
indefinite as well as a definite zero-determiner.

Before we close this section, let us explore if our analysis can be simplified. Looking at all
the examples that include classifiers, we notice that a classifier is used every time we
modify CLP, whether by a QP or a modifier of some sort. The only exception to this
generalization seems to be the obligatory use of a classifier with a demonstrative in (37a').
These cases can be unified if we follow (some) Vietnamese traditional grammar and assume
that the demonstrative determiner in Vietnamese is actually an adjective phrase (AP) or an
adverb phrase (ADVP). If we do not analyze demonstratives as determiners (as we have
been doing), but we analyze them as adjectives or adverbs, then we can easily explain why
the use of a classifier is required with a demonstrative. Furthermore, if this analysis is
correct, then we no longer need to assume that DP projects to the right, as shown in (50b)
since the only motivation we had was the fact that the demonstrative appeared as the
rightmost element in the DP. The revised structure for the Vietnamese noun phrase is
shown in (50):

(50) a. căn nhà này

CL house this
'this house'
b.

8.7 Summary

In this chapter we have expanded our model of X'-syntax to the DP. We have argued that
noun phrases have the minimal structure shown in (28):
154 Héctor Campos & Bùi Huỳnh Thủy Thương

(51) Minimal structure for all noun phrases

For those noun phrases that do not have realized determiners, we will posit a zero-
determiner. We have seen that some languages like Romanian allow for affixal
determiners while other languages like Vietnamese allow for the head of the DP to be
generated to the right rather than to the left of the NP.

Finally we have seen that IP and DP share a certain parallelism. IP serves to introduce the
subject of VP. An I-head that contains Tense and Agreement is able to assign Nominative
Case to the DP in its specifier position. Similarly, a DP also serves to introduce the subject
of deverbal nouns as well as possessors, in general. A D-head with the feature
POSS(ESSIVE) will assign Genitive Case to the DP in the specifier of DP.

The realization of both I (with Tense and Agreement) and D (with Possessive) are affixes.
However, while the inflectional morpheme with the features of Tense and Agreement is a
verbal affix, the head D with the feature Possessive is a nominal affix. In English, the I affix
attaches to V while the D affix attaches to DP.
 English Syntax & Universal Grammar 155

APPENDIX: SOME PARALLELISMS BETWEEN DP AND IP: NOMINATIVE

AND GENITIVE CASE

In previous chapters we have observed the beautiful parallelism that exists among
categories regarding the distribution of complements. In this appendix we will discuss one
further parallelism between DP and IP, which will lead us to reanalyze the position where
the subject is generated.

There are some interesting data suggesting that subjects are generated in VP. This is known
as the VP-internal subject hypothesis. Consider the following data:
(1) a. The students will all pass this course.
b. My friends have all seen "Angry Birds".

It is clear that the quantifier all in (1) refers back to the students and my friends,
respectively. We can see this connection better in languages which show agreement, where
the quantifier all will agree in gender and number with the subject. Take a language like
Spanish:
(2) a. Mis amigos están todos estudiando español.
My friends-masc are all-masc studying Spanish
b. Mis amigas están todas estudiando español.
My friends-fem are all-fem studying Spanish
We can see in (2) that the quantifier and the noun agree in gender (masculine) and number
(plural). It is interesting to observe that we can also say:
(3) a. [All my friends] are studying Spanish.
b. [Todos mis amigos] están estudiando español.
c. [Todas mis amigas] están estudiando español.
To explain the position of the quantifier in (2) and (3), we could assume that [all my
friends] and [todos mis amigos] start as a unit (a QP) inside the VP and then the DP moves
to the specifier of IP leaving the quantifier behind. We will assume the following structure
for (4a):
(4) a. My friends will all pass this course.
156 Héctor Campos & Bùi Huỳnh Thủy Thương

b.

In (4), the DP my friends moves to the specifier of IP, leaving the quantifier behind. This
phenomenon is called quantifier float because the quantifier is left behind "floating" in the
specifier of the VP.

Structures like (4) show that the external argument of a verb (i.e., the subject of a
sentence) is merged in the specifier position of VP. Thus we extend the structure in (4) to
all verbs, regardless of whether the subject includes a quantifier or not. Consider the
example in (5):
(5) a. The enemy may destroy the city.
 English Syntax & Universal Grammar 157

b.

In (5b) the enemy is generated in the specifier of VP and then it moves to the specifier of IP.
Why does the subject move from the specifier of VP to the specifier of IP?

In (5b), "the city" receives Objective Case from the verb. However, "the enemy" is not able
to receive Case in the specifier of VP.

We will assume that Nominative Case is assigned to the specifier of IP whenever IP

contains a modal or inflection for tense and agreement ([+TNS, +AGR]).

Thus in (5b), the subject "the enemy" moves from the specifier of VP to the specifier of IP
to receive Nominative Case.

Consider now the corresponding nominalized structure for (5), shown in (6). By
isomorphism we could posit (6b) as the underlying (D-Structure) for (6a):
(6) a. the enemy's destruction of the city
158 Héctor Campos & Bùi Huỳnh Thủy Thương

b.

In (6b) both the Agent ("the enemy") and the Patient ("the city") are generated inside the NP
(as they are both generated inside the VP in 5b). In (6b) both DPs fail to receive Case since
nouns do not assign (Objective or Nominative) Case. As a last resort, we insert the pseudo-
preposition OF in front of the object DP thus assigning OBLIQUE Case to the DP "the
city".

Just as the specifier of IP is assigned Nominative Case when it contains a modal or

inflection for Tense/Agreement, we will assume that the specifier of DP is assigned
Genitive Case when D bears the feature POSSESSOR, which in English is realized as –'s.
This is shown in (7) below:

(7)

Even though the derivations shown in (5b) and (7) are correct for nominative and genitive
subjects, respectively, up until Chapter 13, we will continue to generatenominative subjects
directly in the specifier of IP, as in (18b), where they receive Nominative Case. Similarly,
we will continue to diagram the possessive subjects of NP in the specifier of DP, as in
(19b). We do this to simplify our structures and discussion. But we must remember that
 English Syntax & Universal Grammar 159
subjects are actually generated in the specifier of VP and then move to the specifier of IP to
receive Nominative Case.6

6
To review some of the concepts you learned in this chapter, I recommend viewing Caroline Heycock's
video "The DP hypothesis" in:
[Link]
TFwO&index=10
160 Héctor Campos & Bùi Huỳnh Thủy Thương

CHAPTER 9

THE STRUCTURE OF CP

So far we have applied our X'-syntax model to VP, NP, AP, PP, ADVP, IP and DP. In the
last chapter we saw that DP was an extended projection of NP. In this chapter we will
pursue the hypothesis that CP is an extended projection of IP. We will also explore the
structure of yes/no questions as well as of negative sentences.

9.1 Types of sentences

So far our system is able to generate declarative affirmative sentences, but we still lack
structure to generate interrogative sentences since we do not have a place to move the
modal to in (1b):

(1) a. Quan will visit Ha.

b. Will Quan visit Ha?
c.

To account for interrogative sentences like (1b), we will posit that there is a higher phrase
on top of IP. This category is the complementizer phrase or CP. We can actually see the
realization of a C head in embedded sentences in English, as shown in (2):

(2) a. I know [CP that [IP Ha loves Quan]].

b. I wonder [CP if [IP Quan loves Ha]].

The complementizers that and if in (2) serve to introduce the embedded sentential
complements in (2). That introduces a declarative embedded clause while if introduces an
interrogative embedded clause.
 English Syntax & Universal Grammar 161

9.2 Types of C

In the previous section we have discussed two types of sentences: affirmative and
interrogative. A declarative sentence is shown in (3a). There are two types of interrogative
sentences: we have simple questions or yes/no questions like (3b) and we also have
specific questions or wh-questions as in (3c):

(3) a. Quan will visit Ha. (declarative)

b. Will Quan visit Ha? (simple question)
c. When will Quan visit Ha? (wh-question)

To account for these three types of sentences, we will propose two features, [±Q] and
[±WH]:

(a) [±Q]:
[+Q] indicates that the sentence is a question,
[-Q] indicates that the sentence is not a question.

(b) [±WH]:
[+WH] indicates that the sentence contains a WH-word (who, what, when, how,
why, how many, how much, how old, whether, etc.).

We use [-WH] when the sentence does not contain a WH-word.

Let us apply those features to our examples in (3), repeated here as (4):

(4) a. Quan will visit Ha. (declarative, [-Q, -WH])

b. Will Quan visit Ha? (simple question, [+Q, -WH])
c. When will Quan visit Ha? (wh-question, [+Q, +WH])

We will explore the fourth possible combination of these features, namely [-Q, +WH], in
the next chapter.

9.3 CP structures: Embedded clauses

Let us start with embedded sentences since they are easier to analyze due to the fact that in
English wemay pronounce the complementizer. Consider the examples we discussed in (2),
repeated here as (5):

(5) a. I know [that[IP Ha loves Quan]].

b. I wonder [if[IP Quan loves Ha]].
162 Héctor Campos & Bùi Huỳnh Thủy Thương

(5a) (5b)

As we can see in (5), the complementizer that is the spell-out of the features [-Q, -WH]
while if is the spell-out of the features [+Q,-WH].

An embedded declarative sentence in English may also be introduced by the zero-

complementizer Φ, as shown in (6):

(6) a. I know [CP Φ[IP Ha loves Quan]].

b.

PRACTICE 1

Diagram the VPs in brackets in the following sentences. Expand the embedded IP
fully:
a. Jack [said that he liked icecream].
b. Susan [said she should go to Hoi An this weekend].
c. I [wonder if we will study other types of questions].
 English Syntax & Universal Grammar 163

RESEARCH PROJECT 1

How does Vietnamese realize the features [-Q, -WH] and [+Q, -WH]?

9.4 CP Structures: Declarative main clauses

Declarative main clauses [-Q, -WH] always appear with no complementizer in English, as
shown in (6). We will thus say that the spell-out of the complementizer features [-Q, -WH]
in main clauses is always a zero complementizer (Φ) in English:

(7) Our teacher loves Vietnam.

9.5 Interrogative main clauses: Simple questions

Before we explore the structure of simple questions in English, let us consider the structure
of simple questions in other languages.

Consider first Japanese. Remember that Japanese is an SOV language:

(8) a. Anata-wa nihongo-o hanashimasu.

You-SUBJ Japanese-OBJ speak
'You speak Japanese.'
b. Anata-wa nihongo-o hanashimasu ka?
You-SUBJ Japanese-OBJ speak Q?
'Do you speak Japanese?

As we can see in (8), KA is used when we form a yes/no question in Japanese. We say that
KA is the realization of [+Q, -WH] in Japanese, as shown in (9):
164 Héctor Campos & Bùi Huỳnh Thủy Thương

(9)

We find a similar realization in Vietnamese, although in Vietnamese there may be two

morphemes to indicate a simple question:

(10) a. Em nói tiếng Việt.

You speak Vietnamese.
b. Em (có) nói tiếng Việt không?
You Qspeak Vietnamese Q
'Do you speak Vietnamese?'
c.

Both in Japanese and in Vietnamese, the head of CP is projected to the right. Consider now
Catalan, which projects C to the left, just as English does:

(11) a. Tu parles català.

You speak Catalan
'You speak Catalan.'
b. Que parles tu català?
Q speak you Catalan?

As we can see in (11b), the morpheme QUE is the realization of [+Q,-WH] feature in
Catalan. However, this morpheme is a bound interrogative morpheme. In Catalan the
verb moves to I to pick up the tense and agreement features, and then moves on to C to
 English Syntax & Universal Grammar 165

support the question marker QUE, to avoid a Stranded Affix Filter violation. This is shown
in (12):

(11) c.

We will adopt a similar analysis for English. The only difference will be that in English the
realization of [+Q,-WH] is always a silentzero-question morpheme, a bound
morpheme –Φ, which will also be responsible for the subject-verb inversion we observe
in questions in English. We will explore simple questions systematically in the next
section.

9.6 Simple yes/no questions in English.

In this section we will discuss the way we form simple yes/no questions with modals (6.1),
with auxiliaries (6.2) and with main verbs (6.3).

9.6.1 Simple questions with modals. To form a question with a sentence that contains a
modal, we must invert the modal as shown in (12):

(12) a. Quan will visit Ha.

b. Will Quan visit Ha?
166 Héctor Campos & Bùi Huỳnh Thủy Thương

c.
D-Structure S-Structure

As we can see in (12a), the zero bound morpheme representing [+Q,-WH] is unattached.
In order to satisfy the Stranded Affix Condition, we move the modal to C, as shown in
(12b). Thus we obtain subject-modal inversion observed in (12b).

9.6.2 Simple questions with auxiliaries. As we can see in (13), we also get subject-verb
inversion with auxiliaries.

(13) a. Quan has visited Ha.

b. Has Quan visited Ha?
c.
D-Structure S-Structure
 English Syntax & Universal Grammar 167

As we can see in (13c), there are two stranded affixes, namely the –s in inflection and the
zero-morpheme –Φ in C. To satisfy the Stranded Affix Condition, we move AUX to I and
then I to C. At the S-Structure in (13c), both affixes have attached to a verb, and thus the
Stranded Affix Condition is satisfied thus deriving (13b).

PRACTICE 2

I. Diagram the following yes/no questions:

a. Should we practice structures every day?
b. Are the Saigon taxi drivers lowering their fares next week?
c. Will you have been studying syntax for one month?

II. We get subject-verb inversion in main sentences, but we do not get it in embedded
sentences. Explain why.
a. Will she come?
b. *I wonder will she come?
c. I wonder if she will come.

III. In Irish English it is possible to say (IIb). How does Irish English differ from
Standard English?

9.6.3 Simple questions with main verbs. Simple questions with main verbs require the
appearance of the auxiliary DO. This operation is called DO-SUPPORT. Consider the
sentences in (14):

(14) a. Jack loves Susan.

b. Does Jack love Susan?

The structures for the declarative and the interrogative sentences in (14) are shown in (14c):

(14) c. Jack loves Susan

168 Héctor Campos & Bùi Huỳnh Thủy Thương

i. D-Structure ii. S-Structure

In (14c.a) we see one stranded affix, namely the –s,which spells out the features of I,
[PRES, IIISG]. As we have shown before, I moves down to Vin English as in (14c,b) so
that the stranded affix is no longer stranded and thus satisfies the Stranded Affix Condition.

Consider now the equivalent interrogative sentence in (15):

(15) Does Jack love Susan?

a. D-Structure b. S-Structure
 English Syntax & Universal Grammar 169

As we can see in (15a), there are two stranded affixes, namely the –s which spells out the
features of I [PRES, IIISG] and the interrogative affix in C, which is the spell out of the
[+Q, -WH] features. If we were to lower I to V, as we did in (14c,b), we would still have a
stranded affix: the interrogative affix in C, which would then violate the Stranded Affix
Condition. Thus Universal Grammar makes available a last resort operation, an
emergency operation, which allows us to satisfy the Stranded Affix Condition. By inserting
the dummy auxiliary DO in I, we provide the affix –s with a verb to attach to, and
furthermore, we also provide a verb for –Φ to attach to and thus satisfy the Stranded Affix
Condition. As we can see in (15b), both affixes are now attached to the inserted element
DO. Thus do-support is the rescue or last resort operation that helps us avoid a violation
of the Stranded Affix Condition when we form simple yes/no questions with main verbs in
English.

PRACTICE 3
I. Diagram the following sentences:
a. Shall we go to the movies this weekend?
b. Have you been studying syntax for a long time?
c. Did your Linguistics Professor publish a new book this year?

II. We have observed that do-support is obligatory for main verbs in main sentences,
as shown in (a). However, in embedded sentences, do-support is ungrammatical, as
seen in (b,c). The only grammatical construction for embedded interrogative yes/no
sentences is (d):
a. Do you study syntax?
b. *I wonder if do you study syntax.
c. *I wonder do you study syntax.
d. I wonder if you study syntax.
How can we account for the facts in (a) – (d) above?

III. In Irish English we can form embedded questions as shown in (a) and (b):
a. I wonder if you study syntax.
b. I wonder do you study syntax.
How would you account for the two sentences in (a) and (b)? How does Irish English
differ from Standard English?

IV. Consider the following sentence from Shakespeare's Julius Caesar. How would
you derive (a)?
170 Héctor Campos & Bùi Huỳnh Thủy Thương

Wrong I mine enemies?

'Do I wrong my enemies?'

9.7 Negative sentences

To conclude this chapter, we will study the structure of negative sentences. As we did when
we studied interrogative sentences, we will consider negative sentences with modal verbs
(7.1), with auxiliary verbs (7.2), and with main verbs (7.3). We will conclude this section
with a comparison between the reduced negative –n't, which is the most common form
found in modern spoken English, and the full negative form not, which is found more
commonly in formal speech or literary texts.

Before we develop our structure for negative sentences, let us observe a negative sentence
which contains a modal, an auxiliary, and a main verb together with negation, as in (16):

(16) Tom shouldn't have gone to bed so late last night.

MODAL –n't AUX MAIN VERB

We can see the negative –n't appears after the modal verb and before the auxiliary and the
main verb. We will assume that negation constitutes its own phrase NEGP, with –n't as the
head. As we can see in (16), -n't is an affix, and as such, it requires a verbal head to attach to.
We will thus pursue an analysis of negative sentences with the structure shown in (17):

(17)

9.7.1 Negative sentences with modal verbs. In (16) we saw that modal verbs precede the
negative element, which in turn, precedes VP. We will thus pursue the following
architecture for negative sentences with modal verbs.

(18) a. Quan shouldn't go to bed so late.

 English Syntax & Universal Grammar 171

b.

As we can see in (18b) above, -n't is an affix, as such, it must attach to a verbal element to
satisfy the Stranded Affix Condition. In (18b), the negative affix attaches to the modal
should, thus deriving (18a).

If the head C in (19b) were to contain the features for a simple yes/no question, then both
the negative affix as well as the modal would move to C to satisfy the Stranded Affix
Condition:

(19) a. Shouldn't Quan study harder?

b.

Structures like (19) show that –n't has moved to I and then the complex [modal + -n't] has
moved higher to attach to the interrogative affix in C.
172 Héctor Campos & Bùi Huỳnh Thủy Thương

Example (19b) is important for our analysis. Typically we assume that –n't is the
phonologically-reduced form for not and that they are one and the same element.
However, notice that we cannot substitute -n't by not in (19b), as shown in (20). Instead, the
full form not must precede the verb, as shown in (20c):

(20) a. Shouldn't Quan study harder?

b. *Should not Quan study harder?
c. Should Quan not study harder?

Facts like (20) suggest that we should analyze the reduced form –n't and the full form not as
different elements. We will pursue such an analysis in section 7.4.

9.7.2 Negative sentences with auxiliary verbs. We saw in (16) that the negative affix –
n't precedes the auxiliary verb. We will thus assume that NEGP precedes AUXP, as shown
in (21):

(21) a. Ha hasn't visited France.

b.

In (21b) we have two affixes: the negative –n't and –s, the spell-out of [PRES, IIISG].
These affixes need to be attached to a verbal element in order to satisfy the Stranded Affix
Condition. As we saw in Chapter 8, the auxiliary verb is able to move in English. In so
doing in (21b), the auxiliary have moves first to NEG and then to I. In (21b) both affixes are
attached to the auxiliary thus satisfying the Stranded Affix Condition.1

1
Notice that we don't get the form haven't, but hasn't. When all the features of I are spelled out at PF,
Morphology will put these affixes in order: AUX-INFL-n't (hasn't).
 English Syntax & Universal Grammar 173

If we turn (21) into a negative question, then the auxiliary, the reduced negative as well as
the inflection will all move to C, as shown in (22).

(22) a. Hasn't Ha visited France?

b.

As we observed with modal verbs, the reduced negative –n't cannot be replaced by the full
negative not in (22a):

(23) a. Hasn't Ha visited France?

b. *Has not Ha visited France?
c. Has Ha not visited France?

Facts like (23) add to our claim that the reduced form –n't and the full form not cannot be
analyzed as the same element. We will elaborate on this claim in section 7.4 below.

9.7.3 Negative sentences with main verbs. As every student of English as a foreign
language knows (having suffered learning this!), main verbs require do-support to form a
negative sentence.

(24) Toan doesn't like icecream.

Let us try to explain why we need do-support in (24). Compare the declarative sentence in
(25a) with the corresponding negative sentence in (25b):
174 Héctor Campos & Bùi Huỳnh Thủy Thương

(25) a. Toan likes icecream.

b. Toan doesn't like icecream.

The D-structures for (25a,b) are shown in (26a,b), respectively:

(26) a. Toan likes icecream. b. Toan doesn't like icecream.

The only difference between (26a) and (26b) is the NEGP that appears in (26b) between IP
and VP. In (26a), the inflection -s is able to lower to V and attach to the verb like. However,
in (26b) inflection cannot lower to the verb because NEG blocks movement of I to V.
Why doesn't inflection lower to NEG and then lower to V? Because the affix –n't does not
attach to another affix, it requires a verbal head to attach to. Thus we see that the only
option to satisfy the Stranded Affix Condition is to insert DO in I (we refer to this operation
as do-support) and then move the negative affix to I.

The S-Structure for the sentences in (26a,b) is shown in (27a,b), respectively:

(27)a. Toan likes icecream. b. Toan doesn't like icecream.

 English Syntax & Universal Grammar 175

S-Structure:

As we can see in (27a), I lowers to V in declarative sentences. In the negative sentence

(27a), I is blocked by NEG from moving to V. Furthermore, I is not able to attach to NEG
because the affix –n't only attaches to verbal elements and the affix under I in (27) is not a
verbal element. Thus, as a last resort operation, we insert DO under I. In chapter 9 we
saw a similar operation when we formed yes/no questions with main verbs. In (27b), the
negative affix –n't raises to I and attaches to DO and the inflection, yielding the desired
doesn't form in (27b).

If we were to form the negative question for (27b), we would get (28a). Once again notice
that the reduced form –n't and the full form not are not interchangeable, as observed in
(28b):

(28) a. Doesn't Toan like icecream?

b. *Does not Toan like icecream?
c. Does Toan not like icecream?

Facts like (28), as those in (20) and (23) show that –n't and not must be analyzed as
different elements to account for these facts. We will pursue such an analysis in the next
section.
176 Héctor Campos & Bùi Huỳnh Thủy Thương

PRACTICE 4

Diagram the following sentences:

a. The foreign taxi driver can't speak English well.
b. You won't see me again until next week.
c. Couldn't the English teachers call the department chair before the meeting?
d. They didn't accept the job offer.
e. Didn't your friend from highschool study English at the American Institute?

RESEARCH PROJECTS

I. Consider the following sentences in Vietnamese:

a. Ngày mai Hà sẽ không đi làm.
b. Hà đang tự hỏi liệu Quân có yêu Hà không.
c. Khiêm sắp không học ở đại học nữa.

In traditional Vietnamese grammar, the words sẽ, đang, sắp are analyzed as "phụ từ"
or adverbs. Is it possible to analyze them as "modal verbs" or "modal particles",
generated under I? Justify your answer.

II. We have seen that verbs should and must are best analyzed as modal verbs in
English, generated as Inflection. Consider the equivalent verbs in Vietnamese, nên and
phải. Can these verbs also be analyzed as "modal verbs" under I in Vietnamese?
Justify your answer. Try to use the evidence you found in problem I to support your
claim.

III. Consider the following negative sentence in Vietnamese:

a. Khiêm không học ở trường nữa.
Is this sentence problematic for the claim we made in a previous chapter, namely that I
lowers to V in Vietnamese just as it does in English?

IV. Consider the following sentences in Japanese :

a. Khiem-wa gohan-o tabemas-ita.
Khiêm-SUBJ rice-OBJ eat-PAST
' Khiêm ate rice.'
b. Khiêm -wa gohan-o tabemas-en des-ita.
Khiêm-SUBJ rice-OBJ eat-not des-PAST
'Khiêm didn't eat rice.'
 English Syntax & Universal Grammar 177

Diagram the structures in (a) and (b). Propose an analysis for the verb DES- 'to be'
that appears in the negative sentence in (b).

9.7.4 Exploring the syntax of NOT. We have shown above that the reduced form –n't
and the full negative form not cannot be analyzed the same way. In this section we will
explore a way to analyze sentences with NOT that will be able to account for the
differences noted in the previous sections.

[Link] History of negation in [Link] will start with French. French has a double
negative, as shown in (29a). It is interesting to observe that in Old French negation was
formed with only one negative, as shown in (29b):

(29) a. Tom ne parle pas français. (Modern French)

Tom not speak not French
'Tom doesn't speak French.'
b. Tom ne parle français. (Old French)
Tom not speak French
'Tom doesn't speak French.'

How can we account for the evolution of French from (29b) to (29a)? We will assume that
ne- was the head of NEGP, as shown in (30):

(30) Stage 1: Ne-, a preverbal suffix, is used as a marker for negation.

Jean ne parle français. (Old French)

178 Héctor Campos & Bùi Huỳnh Thủy Thương

As we can see in (30), ne- is an affix (a prefix, to be more accurate) and it must attach to a
verbal element. In French, as in Spanish, the verb moves to inflection to satisfy the Stranded
Affix Condition. Thus in (30), V moves to NEG, picks up ne- on its way, and then moves to
I, thus deriving the negative sentence in (29b, 30).

Around the 15th Century, French vowels were reduced to schwa, so ne- became [nə], with a
barely perceptible vowel. Thus a reinforcer was introduced in order to make negation more
audible. Many adverbs competed to become the official negative reinforcer. Pas (meaning
'step') came to become the reinforcer. This reinforcer appeared in the specifier of NEGP, as
shown in (31), thus deriving the structure we find in Standard Modern French. As we can
see in (29a), pas appears between the verb and its direct object:

(31) Stage 2: The negative marker ne- is reinforced by the adverb pas.

Jean ne parle pas français. (Modern Standard French)

We have direct evidence that the verb parl- moves through NEG on its way to I since the
adverb pas appears between the verb and the direct object.

Modern French has continued to evolve and now, at the present time, in modern colloquial
French, ne- is not pronounced any more. Thus we find the structure in (32):

(32) Stage 3: The original negative marker ne- is no longer pronounced (i.e, it has
become a silent affix). Pas becomes the main (visible/audible) marker for negation.
 English Syntax & Universal Grammar 179

Jean parle pas français. (Modern colloquial French)

As we can see in (32), the affix ne- has become silent (represented as -Φ), thus pas has
become the main indicator that the sentence is negative.

[Link]. History of negation in English. Consider now the evolution of negation in

English. In Old English ne- was also used as a negative marker, as shown in (33a). In
Middle English this marker came to be reinforced by nawiht (which later became naught
and then not). In Early Modern English, the preverbal negative marker disappeared thus
giving us a postverbal negative marker as in (33c). Notice that in (33a,b,c) the verb moved
to inflection. In Modern English the verb stops moving to inflection, thus do-support
appears, as shown in (33d). Finally –n't appears, as shown in (33e).

(33) Old English

a. Ic ne seah. (Old English)
I not saw
b. I ne saugh nawiht/naught. (Middle English)
I not saw not
c. I saw not. (Early modern English)
I saw not
d. I did not see. (Modern formal English)
e. I didn't see. (Modern spoken English)

The structure in (33b) is similar to the structure we find in modern formal French. As we
can see, English has gone two steps further than French. In (34) – (38) we represent the
180 Héctor Campos & Bùi Huỳnh Thủy Thương

different stages found in the evolution of negation in English. We will use modern
orthography to understand the changes better:

(34) Stage 1: Ne-, a preverbal affix, is used as a negative marker.

I ne saw. (Old English)

(35) Stage 2: The negative reinforcer naught/not is added to Stage 1:

I ne saw not. (Middle English)

(36) Stage 3: Original negative marker ne- is not pronounced any more (i.e., it becomes a
silent prefix). Not becomes the main marker for negation.
 English Syntax & Universal Grammar 181

I saw not. (Early Modern English)

Stages 2 and 3 are parallel to stages 2 and 3 in French. However, English goes further than
French in its evolution. We witness two more stages.

(37) Stage 4: The verb stops moving to inflection, do-support becomes necessary to
satisfy the Stranded Affix Condition.

I did not see. (Modern formal/literary English)

(38) Stage 5: Not, in its reduced phonological form –n't, is reanalyzed as the head of
NEGP.
Do-support is necessary to satisfy the Stranded Affix Condition.
182 Héctor Campos & Bùi Huỳnh Thủy Thương

I didn't see. (Modern spoken English)

As we can see, both Stage 4 and Stage 5 co-exist in Modern English, although Stage 4 is
a more formal and literary construction than the one found in Stage 5. The result of this
coexistence is that we can use both –n't and not to indicate negation. These two forms are
not interchangeable in all cases, as we have seen above.

Now we can go back to our original claim that the reduced form –n't should be analyzed
differently from the full form not. We will claim that the full form not is in the specifier
position of NEGP, as shown in (37), while the reduced form –n't is in the head of NEG,
as shown in (38).

Let us take the interrogative question for (38), shown in (39):

(39) a. Didn't you see?

b. *Did not you see?
c. Did you not see?

As we can see in (39), negative interrogative questions allow two options: either we use –n't
on the inverted modal or auxiliary, as in (39a), or we use the full form not in front of the
main verb, as shown in (39c). These facts receive an elegant explanation from our analysis.
Consider first (39a,b). Why is only the reduced form possible in negative interrogative
sentences? The structure for (39a) is shown in (40):
 English Syntax & Universal Grammar 183

(40) Didn't you see?

*Did not you see?

In (40), DO has been inserted under I. This dummy verb serves as support both for the past
morpheme –ed and the reduced negative –n't, which raises to I. The negative dummy
auxiliary then raises to C so that the zero-interrogative affix in (40) does not violate the
Stranded Affix Condition.

In (40), -n't cannot be replaced by the full form not because, as we claimed above, not is in
the specifier position of NEGP and not in the head position. Only a head like –n't is able
to move to C with an auxiliary or a modal in inflection. A full adverb specifier like not,
will stay in its place and thus always appears in front of the main verb:

(41) Did you not see?

*Did not you see?
184 Héctor Campos & Bùi Huỳnh Thủy Thương

As we can see in (41), NOT is in the specifier of NEGP while –n't in (40) is in the head
of NEGP. NOT does not move from the specifier position, thus NOT always appears in
front of the verb in negative sentences and can never front along with the auxiliary and
modal verbs when these move from I to C. Only –n't is able to move along with auxiliary or
modal verbs.

PRACTICE 5

I. Diagram the following sentences:

a. She didn't come to the party yesterday.
b. She did not come to the party yesterday.
c. Will you not return the book to the library?
d. Won't you return the book to the library?
e. Aren't they making a new movie now?
f. Are they not making a new movie now?

II. Diagram the following sentences in French:

a. Ha n'aim-e pas Khiem.
Ha not-love-s not Khiem
'Ha does not love Khiem.'
b. Ne regard-es tu pas la télé?
Not watch-PRES,IISG you not the TV
´Don't you watch TV?'

CHALLENGE

I. Diagram the following sentences. Indicate all the inserted and moved elements. Make
sure you do not leave verbal morphology stranded.
a. The foreign taxi driver admitted that he couldn't speak English very well.
b. Tuoi said that she wouldn't buy her boyfriend chocolates any more.
c. Could the foreign language teachers not have called the Department Chair before the
meeting?
d. Did the announcement that the Principal had called off the meeting surprise the busy
teachers at that elementary school?
e. I don't know if my little sister has told my parents that she will marry that ugly man
from Nghe An next week.
 English Syntax & Universal Grammar 185

f. Les étudiants de linguistique generative n'aim-ent pas leur professeur. (FR)

The students of linguistics generative not-love-PRES not their professor
'The generative syntax students do not like their syntax teacher.'

II. Consider the following sentences from Croatian, a Slavic language:

(1) Studenti pi-ju pivo.
The-students drink-3pl beer
'The students drink beer.'
(2) Pijuli studenti pivo?
'Do the students drink beer?'
(3) Dali studenti piju pivo? (colloquial in Croatian, but OK in Serbian)
'Do the students drink beer?'
(A) Diagram structures (1) - (2) and briefly describe the way you form affirmative
and interrogative sentences. For each structure, clearly indicate which elements are
inserted.
(B) Discuss sentence (3). How do Serbian and Croatian differ?
(C) Serbian speakers accept (2) and (3). You are curious about the structure in (4)
and you ask a native speaker if you can say (4):
(4) Dali piju studenti pivo?
Your speaker accepts (4). Is this example problematic for the analysis you proposed
above? Justify your answer. What hypothesis could you make about the probable
history of "dali"?

III. Consider the following data from Hiberno-English:

(1) a. Is Mary working?
b. I wonder if Mary is working.
c. I wonder is Mary working.
How does Hiberno-English differ from Standard English? Diagram (1a), (1b), (1c).

Consider now (2):

(2) a. *Tom said was Mary working?
b. Did Tom say was Mary working?
c. Tom didn't say was Mary working.

How would you account for the data in (2)? Are these sentences problematic for your
analysis for the sentences in (1)?

IV. Translate the following sentences into Vietnamese:

(1) a. Tuoi is not a student.
186 Héctor Campos & Bùi Huỳnh Thủy Thương

b. Tuoi is not tall.

c. Tuoi does not teach in Hanoi.

Briefly explain the Vietnamese system of negation.

V. Vietnamese can use the negative adverb đâu to make an emphatic negation. Give
examples and explore the different positions where đâumay appear

9.8 Sentence adverbs

In chapter 4 we discussed the way to incorporate adverbs into the VP. We call those
adverbs verbal adverbs.

(42) a. Quan always comes to class on time.

b. Thuong never misses a class.

There is a second class of adverbs that seem to modify the whole sentence. These adverbs
typically appear at the very front or at the very end of a sentence:

(43) a. Clearly, Quan loves Ha.

b. Sadly, Ha doesn't love Quan.
c. She is ignoring him, obviously.
d. They have never gone out on a date, curiously.

Notice that these adverbs include the speaker's opinion. In chapter 8 we discussed the basic
properties of X'-syntax and we studied all the possible configurations that X'-syntax allows.
We saw that X'-syntax allows for the recursion of category levels, and we have already seen
examples where X may appear over X (morphological units or phrasal verbs, chapters 2, 4),
X' over X' (modifiers, chapters 4-6). We will assume that sentencial modifiers are instances
of CP over CP. The structures for (43a) and (43b) are shown in (44):
 English Syntax & Universal Grammar 187

(44)

There is evidence that these sentential modifiers are adjunctions to CP. Notice that the
adverb may precede a question and we know that in questions I moves to C:2

(45) a. Frankly, do you think they can have a future together?

b. Curiously, why do people hate you?

The structure for (45b) is shown in (46):

(46)

2
We will see in the next chapter that wh-words like why in (45b) move to the specifier of CP, thus
giving further evidence that sentential adverbs are adjunctions to CP. As we can see in (47b), sentence
adverbs may also appear at the end of the sentence.
188 Héctor Campos & Bùi Huỳnh Thủy Thương

Notice that some of these adverbs may also function as verbal adverbs. In this case they
modify the VP rather than CP:

(47) a. She asked curiously.

b. He answered frankly.

The two structures for (48a,b) are shown in (48a',b'), respectively:

(48)
a. She smiled strangely. b. She smiled, strangely.
a'. b'.

9 Summary

In this chapter we saw that CP contains information as to the type of sentence we have. We
use the features [±Q] and [±WH] to differentiate three types of sentences so far. [±Q]
indicates whether the sentence is a question or not while [±WH] indicates whether the
sentence contains a wh-element or not. A C with the features [-Q,-WH] is a declarative
sentence, a C with the features [+Q,-WH] is a simple yes/no question and a C with [+Q,
+WH] is a wh-questionor a specific question.

We saw that main and embedded clauses realize these features in different ways:
 English Syntax & Universal Grammar 189

Type of CP Main Clause Embedded Clause

C[-Q,-WH] Φ that, Φ
declarative sentence (free silent interrogative morpheme)
C [+Q, -WH] -Φ If
yes/no question (bound silent interrogative morpheme)

Declarative embedded clauses can take two complementizers in English: that and Φ.
Main declarative sentences, on the other hand, always appear with a Φ complementizer.
Non-wh embedded interrogative clauses always appear with the complementizer if.
Main non-wh interrogative clauses always appear with a silent affix as a complementizer
–Φ. This silent complementizer explains why the verb must move to C in interrogative
sentences, thus producing subject/verb inversion.

We will study WH-questions in the next chapter. We will claim that wh-questions are the
realization of C [+Q,+WH].

In this chapter we also analyzed negative sentences and showed that the reduced negative
–n't and the full negative not cannot be analyzed in the same way. We claimed that –n't
is best analyzed as the head of NEGP while not is best analyzed as an adverb phrase in
the specifier of NEGP. We explored the evolution of both French and English negation
and we showed how our model can clearly explain the different stages in the evolution of
negation providing us with evidence for the analysis of the negative not in modern English.
We saw that English and French show a surprising similarity in the evolution of negative
sentences. Many languages have followed the same path of evolution. This path is known
as Jespersen's cycle. The following quote is from Jespersen's (1917) Negation in English
and other languages:

The history of negative expressions in various languages makes us witness the

following curious fluctuation: the original negative adverb is first weakened,
then found insufficient and therefore strengthened, generally through some
additional word, and this in turn maybe felt as the negative proper and may
then in the course of time be subject to the samedevelopment as the original
word.3

3
For a review of IP and CP structure, we highly recommend that you view Caroline Heycock's
"Sentence Structure" in
[Link]
TFwO&index=8
190 Héctor Campos & Bùi Huỳnh Thủy Thương

However, the idea of a negative cycle, had been observed thirteen years before by Alan
Gardiner when he was describing Ancient Egyptian and Coptic, comparing it with French.
As he noted:

These words form the Latin passum and punctum, were originally adverbal
accusatives placed at the end of negative sentences for the purpose of
emphasis; just like the English "not a jot", "not a straw"… Pas and point, and
like them the Demotic xxx. Copticⲁⲛ, next lose their emphasizing force, and
become mere adjuncts of negative words. Last of all, they come themselves to
be looked upon as negative words. (Gardiner 1904:134)
 English Syntax & Universal Grammar 191

CHAPTER 10

WH-MOVEMENT: WH-INTERROGATIVES

In this chapter we will study the syntax of specific questions, also called wh-questions,
since they typically involve a wh-word (who, what, where, when, why, how, ...).

10.1 Wh-questions

In the previous chapter we studied the syntax of simple or yes/no questions. An example
and a structure are shown in (1):

(1) a. Will Quan visit Ha?

b.

In Chapter 9 we proposed that a C head with the features [+Q,-WH] is realized as the zero-
affix -Φ. This zero-affix is responsible for the subject-verb inversion we observe in yes/no
questions.

We observe subject-verb inversion with wh-questions as well, as shown in (2):

(2) a. What will you study?

b. What are you studying?
c. What do you study?

From the data in (2) we can make two big generalizations regarding wh-questions:
(a) there is subject-verb inversion
(b) the wh-word moves to the front of the sentence
192 Héctor Campos & Bùi Huỳnh Thủy Thương

We want our analysis to capture these properties directly. In the next sections we will
consider wh-movement in main clauses (1.1) and then wh-movement in embedded clauses
(1.2).

10.2 Wh-questions in main clauses. The structure for (2a) is shown in (3):

(3) a. What will you study?

b. D-Structure:

We will assume that question words like who, what, which are of category DP. On the
other hand, wh-words like where, when, why, how are analyzed as ADVP. In (3b) the
wh-word what has been merged as a complement of the verb study. This follows from the
Projection Principle, which requires that if a verb c-selects a complement, that
complement must be present at all levels of syntactic structure. In (3b) there is a stranded
affix, -Φ, in C. As with yes/no questions, we can move I to C so this stranded affix can
satisfy the Stranded Affix Condition. This operation is shown in (4):

(4) a. *Will you study what?

 English Syntax & Universal Grammar 193

b. S-Structure 1:

In (4b), the modal will has moved to C to support the interrogative affix –Φ. But notice that
the sentence is still ungrammatical. The sentence remains ungrammatical until the wh-word
moves to the front of the sentence as in (5).

(5) a. What will you study?

b. S-Structure 2:

c. [CP whati willj [IP you tj study ti]]

(4c) shows the tree structure for (4a). Since our structures will be getting longer from this
chapter on, we will introduce an alternative type of structure representation. This is known
194 Héctor Campos & Bùi Huỳnh Thủy Thương

as bracket structure. (4c) is exactly equivalent to (4b), but it has the advantage that it
shows the final position of the moved elements. All moved elements leave behind a
coindexed trace, as we can see in (4c). We can see in (4b) and (4c) that the wh-element has
moved to CP.

To account for the movement of the wh-word to the front of the sentence (CP) we will
propose a Wh-condition also known as the Wh-Criterion:

(5) WH-CRITERION
a. A wh-phrase must move to the specifier position of a CP which contains the
feature [+wh].
b. A C-head that has the feature [+wh] must host a wh-word in its specifier.

The Wh-Criterion thus guarantees that the wh-word will move to the specifier of CP.

How do we know that it is the whole DP that moves in (5b)? We can consider another wh-
word that contains a NP, as in (6):

(6) a. Which book will you study?

b.

c. [CP [which book]i willj [IP you tj buy ti]]

 English Syntax & Universal Grammar 195

In general we will say that Wh-movement is phrasal movement. That means that we must
move the whole phrase. The whole phrase moves to the specifier position. The structures in
(6b,c) show the two kinds of movement found in the syntax of world languages: head-
movement, exemplified by the movement of I to C, and XP-movement or phrasal
movement, exemplified by the movement of the wh-phrase to the specifier of CP.

(7) CONDITIONS ON MOVEMENT

a. AnX-head must move to another head position.
b. An XP-phrase must move to a specifier position.

The sentence in (8) shows an example of an adverb wh-word:

(8) a. Where did you buy that book?

b.

c. [CP wherei didj [IP you tj buy that book ti]]

In (8b) we have two affixes: the affix representing [PAST, IISG], which we label as -ed and
the affix for [+Q, +WH], which we represent as -Φ. As we argued in the previous chapter,
do-support is the last resort operation which saves the structure from having two stranded
affixes. The wh-word where moves to the specifier of CP to satisfy the Wh-criterion in (6).

What happens when a wh-word is the complement of a preposition in English? Consider the
examples in (9):
196 Héctor Campos & Bùi Huỳnh Thủy Thương

(9) a. Tuoi cooked dinner with her roommate.

b. Who did Tuoi cook dinner with? (preposition stranding)
c. With whom did Tuoi cook dinner? (pied-piping)

In (9b) we have replaced the complement of the preposition by the corresponding wh-word.
In (9b) we have fronted the wh-word leaving the preposition behind. This phenomenon of
leaving the preposition behind is called preposition stranding and is very common in
spoken English. In (9c), which represents a more formal or literary style (also an older
construction), the preposition has moved along with the wh-phrase. Notice also that we
must use the form whom, an older Objective-Case-marked form of who, which indicates
that the question word is an object and not a subject. The operation of moving the
preposition along with the wh-word in (9c) is called pied-piping, remembering the story of
the pied-piper of Hamelin (the pied-piper played his flute and the rats that were infecting the
town followed him). In our example, the wh-word whom moves to CP and the preposition
with follows behind. This is what we observe in (9c).

We can derive (9b) directly from our system, without any more further assumptions:

(10) a. Who(m) did Tuoi cook dinner with?

b.

c. [CP whoi didj [IP Tuoi tj cook dinner [PP with ti] ]]
 English Syntax & Universal Grammar 197

In (10a), DO is inserted as a last resort to make the affixes –ed in I and –Φ in C comply
with the Stranded Affix Condition. WHO moves to the specifier of CP to satisfy the WH-
Criterion in (5).

As we saw in (9c), there is yet another alternative. The preposition may pied-pipe or be
moved along with the wh-word:

(11) a. With whom/*with who did Tuoi cook dinner?

b.

c. [CP [PP with whom]i didj [IP Tuoi tj cook dinner ti]]

In (11b,c) the preposition has moved along with the wh-word whom. We say that the wh-
feature of DP has percolated to PP, thus making the whole PP a wh-phrase. Since PP is
now a wh-phrase, it needs to satisfy the Wh-Criterion, and thus the whole PP moves to the
specifier of CP. Notice that when pied-piping takes place, only the inflected pronoun whom
can be used.

PRACTICE 1

Diagram the following sentences:

a. Will you still love me tomorrow?
b. Where will we go tomorrow?
198 Héctor Campos & Bùi Huỳnh Thủy Thương

c. Who did you go to Spain with last summer?

d. Which flowers did the boy from Dalat send to his girlfriend from America?
e. How long have you been studying syntax?

RESEARCH PROJECT

Consider the sentences below:

a. Quan will visit Ha in Dalat.
b. Who will Quan visit in Dalat?
c. Quan loves Ha.
d. Who loves Ha?
e. *Who does love Ha?

(i) Diagram sentences (a) through (d).

(ii) Propose a structure for (d). How does (d) differ from the wh-questions we have
been studying in this chapter?

10.3 Wh-questions in embedded clauses

In the previous section we considered questions in main clauses. Consider the examples in
(12):

(12) a. Tuoi will go to America next year.

b. Where will Tuoi go next year?
c. [CP wherei willj [IP Tuoi tj go next year ti]]

We see that (12b,c), as wh-questions in main clauses in general, display two properties:

(i) the subject and the verb invert

(ii) the wh-word appears at the front of the sentence

We accounted for these two properties with two basic principles:

(A) The Stranded Affix Condition, which requires that all affixes be attached. A
crucial assumption we made was that a C [+Q, +WH] contained a silent –Φ affix.
This affix, as in yes/no questions, was responsible for the subject/verb inversion
we observe in both types of questions in main clauses. This accounts for
property (i).
 English Syntax & Universal Grammar 199

(B) The Wh-Criterion, which required the wh-word to be in a specifier of a CP which

contained a [+WH] feature. This accounted for the fact that wh-words move to the
front of the sentence. This explains property (ii).

Let us now consider wh-questions in embedded questions comparing them with embedded
yes/no questions. Consider the embedded yes/no questions in (13):

(13) a. I wonder if Quan will visit Ha this summer.

b. *I wonder if will Quan visit Ha this summer.
c. *I wonder will Quan visit Ha this summer.

The structure for (13a) is shown in (14):

(14) a. I wonder if Quan will visit Ha this summer.

b.

c. [CP Φ [IP I wonder [CP if [IP Quan will visit Ha this summer]]]]

The tree structure in (14b) shows the structure from the main VP down, while the bracket
structure in (14c) shows the structure for the whole sentence. As we can see in (14b,c),
there is no subject/verb inversion in embedded yes/no questions. This follows from the
fact that the embedded complementizer is spelled out as the complementizer if, which is a
free morpheme in English and as such, does not require a verb to attach to it.
200 Héctor Campos & Bùi Huỳnh Thủy Thương

Consider now embedded wh-questions:

(15) a. I wonder when Quan will visit Ha.

b. *I wonder when will Quan visit Ha.

(16) a. I wonder why Quan is learning Spanish.

b. *I wonder why is Quan learning Spanish.

(17) a. I wonder where Quan works.

b. *I wonder where does Quan work.

As we can see in (15) – (17) above, there is no subject/verb inversion in embedded

questionsin English. And this restriction applies to constructions with modal verbs (15),
with auxiliary verbs (16) and with main verbs (17). Furthermore, notice that when a wh-
word is used in the embedded sentence, the complementizer is necessarily the zero-
complementizer.

The structure for (15a) is shown in (17). We only include the structure from the main VP:

(17) a. I wonder when Quan will visit Ha

b.

c. [CP Φ [IP I wonder [CP wheni Φ [IP Quan will visit Ha this summer ti]]]]
 English Syntax & Universal Grammar 201

As we can see in (17b,c), only the wh-word moves to CP in the embedded sentence, there is
no subject/verb inversion. The auxiliary will does not move as there is a free morpheme (the
zero-morpheme) occupying this position. This morpheme is not a bound morpheme
(contrary to what we find in interrogative main clauses).

The table in (18) summarizes the different realizations of the complementizer in C

depending on its [±Q] and [±WH] values in main and embedded sentences:

(18) Complementizer selection in English depending on [±Q] and [±WH] values

Main clause Embedded clause

C [-Q, -WH] Φ Φ, that
(declarative)
C [+Q, -WH] -Φ if
(yes/no question)
C [+Q, +WH] -Φ Φ
(wh-question)

As we can see in (18), only main clause CPs that have the value [+Q] have a –Φ
complementizer. Only in these constructions do we find subject verb inversion. Thus we
only find inversion in main clauses when we have a question, whether a yes/no question or
a wh-question.

PRACTICE 2

I. Diagram the following sentences. Diagram the full sentences:

a. Do you ever wonder why you are studying English?
b. I don't know if the author of that book will give a public lecture.
c. I don't know when the author of that book will give a public lecture.
d. Do you know where the author of that famous book will give a public lecture?
e. What has our syntax professor been saying about our final exam?

II. Consider the following questions in Irish English:

a. I wonder if he will come.
b. I wonder when he will come.
c. I wonder when that he will come.
d. *I wonder when if he will come.
e. I wonder when will he come.
f. Will he come?
202 Héctor Campos & Bùi Huỳnh Thủy Thương

g. *If he will come?

h. *If will he come?

Diagram the grammatical sentences in Irish English. How does Irish English differ
from Standard English? Make a chart similar to (18) summarizing the complementizer
system in Irish English.

10.4 Long distance wh-questions

In all the previous questions in this chapter we have been considering structures where the
wh-word moves to the closest CP. Can a wh-word move further than the closest CP?
Consider (19):

(19) Who did Khiem say that Quan loves?

In (19) the wh-word who has moved from the complement position of loves to the front of
the sentence as it is asking for the person that Ha loves. Consider the D-structure for (19),
shown in (20):
(20) a. Who did Khiem say that Quan loves?
b. D-Structure:

c. [CP–Φ[IP Khiem –ed say [CP that [IP Quan –s love who]]]]
 English Syntax & Universal Grammar 203

As we can see in (20b), trees take a lot of space, especially when we start including
embedded sentences. This is where bracket structures like the one in (20c) become much
more handy.

In (20b,c) we have three stranded affixes: the lower –s, spelling out the features of the lower
I [-Q,-WH]; the –ed, spelling out outthe upper I [PAST,IIISG], and the interrogative affix –
Φ, which is the spell-out of the C [+Q,+WH]. Note that the lower CP has the features [-Q,-
WH], which are spelled out as the complementizer that (they could also be spelled out as Φ,
given our complementizer realization chart in 18). Unless these affixes attach to a verb, they
will violate the Stranded Affix Condition. Thus the lower affix in I lowers to the verb love
forming loves, and DO is inserted into the upper I, forming DID, which subsequently moves
to C to attach to the stranded affix –Φ. These processes are shown in (21):

(21) a. Who did Khiem say that Quan loves?

b. S-Structure: Part 1: Inflection movement

c. [CP didi[IP Khiem ti say [CP that [IP Quan tj love-sj who]]]]
d. [CP didi[IP Khiem ti say [CP that [IP Quan loves who]]]]
204 Héctor Campos & Bùi Huỳnh Thủy Thương

In the bracket structure in (21c), we have shown the two moved elements: the dummy
auxiliary do that moved to C leaving behind ti and the –s which spells out the features
[PRES, IIISG], which leaves behind tj. In general, we will not indicate inflection movement
unless I moves to C. Thus (21d) would be the representation for the structure in (21b).

In (21b,c,d) all affixes satisfy the Stranded Affix Condition. But we are still in violation of
the Wh-Criterion as the wh-word who is not in a specifier of a CP which contains the wh-
feature, nor does the higher CP, which contains a wh-feature, have a wh-word in its
specifier. Thus we must move who to the higher CP. The question is how we move to the
higher specifier. We have two choices:

(a) we move directly in a one-fell-swoop, or

(b) we move in two steps, first to the lowest CP and then to the higher CP

There is evidence that we must choose the alternative in (b), that is wh-words move CP to
CP (typically we say Comp to Comp), they cannot move directly to their final CP
destination.

(22) a. Who did Khiem say that Quan loves?

b. S-Structure: Part 2: Structure after inflections and wh-words have moved

c. [CP Whoj didi[IP Khiem ti say [CP tj that [IP Quan loves tj]]]]
 English Syntax & Universal Grammar 205

Why doesn't the wh-word move directly to the higher CP? These restrictions on movement
are studied in a more advanced syntax course, so we will not deal with them here. Just
remember that if you move a wh-word long-distance, you must move it Comp to Comp.
Why doesn't the wh-word stay in the intermediate CP? Because the intermediate CP does
not have the feature [+WH] necessary to make the wh-word satisfy the Wh-Criterion.
Furthermore, if the wh-word were to stay in the intermediate or lower CP, both the wh-
word as well as the higher CP would be in violation of the Wh-Criterion. That is why the
wh-word must keep on moving until it finds a CP with the feature [+WH], where both of
them will satisfy the Wh-Criterion.

PRACTICE 3

I. Diagram the following sentences. Then show the corresponding bracket structure:

a. What did John buy?

b. What did you say that John bought at the bookstore?
c. When will Quan buy a new motorcycle?
d. When do you think that Quan will buy a new motorcycle?
e. When do you think that Quan said that he would buy a new motorcycle?

II. The following sentence is ambiguous. Show two structures to explain the
ambiguity.
When did the professor say that we would have our final exam?

CHALLENGE 1

Diagram the following sentences. Show all the moved and inserted elements. Show both
the tree and the bracket structure.
a. When will the Dean of the School of Languages meet the new generative syntax
professor in his office?
b. Did José wonder if the assassin had cleaned his knife carefully with a handkerchief
after the brutal murder of the old French actress?
c. The young syntax teacher from Arizona said the first year students of generative syntax
are indeed very proud of their solid analysis of the Vietnamese data.
d. Why did your department chair buy a bottle of whiskey for his wife for her birthday?
e. Did she say if the new president of that famous university on the west coast would be
giving the tenured professors an increase in their salary?
206 Héctor Campos & Bùi Huỳnh Thủy Thương

f. ¿Cuándo public-ó tu buen amigo de Italia muchos artículos de investigación

When publish-ed your good friend from Italy many articles of research
sobre el léxico en famosas revistas internacionales?
about the lexicon in famous journals international
'When did your good friend from Italy publish many research articles about the
lexicon in famous international journals?'
g. Stella said that she had picked a very hard problem for the final project in her syntax
class.
h. Who gave Tom's brother's wife a new car?

CHALLENGE 2

Consider the following sentences:

a. Quan said that Ha went to Dalat.

b. Quan wonders why Ha went to Dalat.

Replace "to Dalat" by the question word "where" and form an interrogative sentence.
Is it possible to ask the question in both cases? Explain your results. Do your results
help prove the hypothesis that wh-movement happens via Comp to Comp and not
directly? Explain how so.

10.5 What about Vietnamese and other languages that don't move wh-words?

Now that you understand how English and other languages that move wh-words to the front
of the sentence work, you may wonder what happens in languages where wh-words do not
move. Vietnamese is one example:

(23) a. Anh yêu ai?

You love who
'Who do you love?'
b. Em đang làm gì?
You PROGR do what
'What are you doing?'
 English Syntax & Universal Grammar 207

As we can see in (23), wh-words do not move in Vietnamese. Does that mean that C does
not contain the feature [+WH] or does it mean that Vietnamese, like many other languages,
does not have to obey the Wh-Criterion?

Remember that our system had different levels of representation:

(24)

X'-Syntax Lexicon

D-Structure

Transformations: Insert, Move

Principles: Stranded Affix Condition, Wh-Criterion

S-Structure

PF LF

We have seen that movement (movement of I to V, Aux to I, I to C, wh-movement) as well

as insertion (do-support, OF-insertion, TO/FOR-insertion) in English occurs between D-
and S-Structure. This is why we can see the effects of movement in word order at the
Phonetic/Phonological level. But movement may also take place between S-Structure and
Logical Form. This movement only occurs in our mind and does not have a reflex in word
order. We will say that Vietnamese and languages that do not move wh-words satisfy
the Wh-Criterion at Logical Form (LF). Again, this is a topic for a more advanced syntax
book, so we will not discuss the issue further here.

To conclude this section, notice that English may also move wh-words after S-Structure, as
shown by this multiple wh-questions:

(24) a. Who loves who?

b. What did you buy where?
208 Héctor Campos & Bùi Huỳnh Thủy Thương

We can see that the second wh-word has not moved to CP and thus is in violation of the
Wh-Criterion. There is evidence (that we will not discuss here) that these wh-words also
move to CP, but between S-Structure and Logical Form (LF). The study of wh-movement
at the level of Logical Form is the subject of a more advanced syntax course, so we will not
dwell on this topic any further.

10.6 What about subject wh-questions?

In our discussion of the examples in (2) we saw that specific questions exhibit two
properties:

(a) there is subject-verb inversion

(b) the wh-word moves to the front of the sentence

We accounted for the two properties above with two principles. The Stranded Affix
Condition explained property (a) assuming that [+Q,+WH] is realized as a zero-affix which
requires a verbal element to attach to. The Wh-Criterion accounted for property (b), by
requiring all wh-words to move to the specifier of CP.

Consider now the examples in (25) and (26):

(25) a. Who did you visit?

b. D-Structure:
[CP -Φ [IP you –ed visit who]]

c. S-Structure:
[CP whoj didi [IP you ti visit tj]]

(26) a. Who visited you?

b. *Who did visit you?

c. D-Structure:
[CP -Φ [IP who –ed visit you]]

d. S-Structure:
*[CP whoi-Φ [IP ti visited you]]

In (25b) the wh-word does not satisfy the Wh-Criterion if it stays in place. Thus in (25c),
who moves to the specifier of CP to satisfy the Wh-Criterion. In (25b) we have two
stranded affixes: the inflection –ed and the zero-affix which is the spell-out of the
 English Syntax & Universal Grammar 209

interrogative C. DO-support is necessary to make both affixes satisfy the Stranded Affix
Condition. Thus, both the Wh-Criterion and the Stranded Affix Condition are satisfied in
(25c) and the sentence converges.

Consider now subject extraction (i.e. movement of a subject which is a wh-word) in (26).
In (26d), we have moved the wh-word to the specifier of CP to satisfy the Wh-Criterion.
But how do we satisfy the Stranded Affix Condition? As you can see in (26b), we cannot
use do-support as we did in (25c), since do-support gives us a wrong result. As we saw in
Chapter 4, verbs do not move to C in English, with the exception of modal and auxiliary
verbs. So, in (26) the verb visited cannot move to C to satisfy the Stranded Affix Condition.
So our structure in (26d) violates the Stranded Affix Condition and cannot be the correct
representation for our sentence in (26a).

To understand (26a) a little better, we need to consider embedded interrogative sentences.

When we extract an embedded object, we can use either that or Φ as a complementizer. But
when we extract an embedded subject, we can only use Φ as a complementizer. This
contrast is shown in (27) and (28):

(27) a. Khiem said that/Φ Ha loved cooking.

b. Whati did Khiem say that/ Φ Ha loved ti?

(28) a. Khiem said that Ha loved cooking.

b. *Whoi did Khiem say that ti loved cooking?
c. Whoi did Khiem say Φ ti loved cooking?

We can assume that the zero-complementizer that allows for embedded subjects to be
moved has special properties. We will label this complementizer ΦS. ΦS is a special
complementizer that allows for subjects to be extracted. In many languages this
complementizer has a special form. In French, for instance, we use the complementizer qui
when we extract a subject wh-word, but we use the complementizer que for all other cases.

In a more advanced syntax textbook we would study that the traces of moved elements
need to be licensed. ΦS would provide us with the appropriate licensing in English. Trying
to understand these licensing conditions for traces is a topic for a more advanced syntax
book. Therefore, to account for (28c) as well as for (26d), we will assume here that there is
a special complementizer ΦS that we use when subjects are extracted in English, as
shown in (29). We will extend our analysis from embedded clauses in (29a) to our original
problem in (26c). We show the derivation for subject extraction in the main clause in (29b):

(29) a. [CP whoi did [IP Khiem say [CP tiΦS [IP ti loved cooking]]]]
210 Héctor Campos & Bùi Huỳnh Thủy Thương

b. [CP whoiΦS [IP ti loves cooking]]

As we can see in (29), ΦS is a free morpheme and as such, it does not require a verb to
move to C so it can attach to it. Thus we get no subject verb inversion in (26c, 29b). The
wh-word moves to the specifier of CP to satisfy the Wh-Criterion.

Subject extraction in the different languages in the world was a major area of research in the
80s and early 90s. All you must learn from our discussion here is that when we extract a
wh-subject we need a special complementizer that we have labeled ΦS. This
complementizer is a free affix, and thus does not require a verb to move to C. This special
complementizer will be able to license the trace in subject position.

10.7 What about whether?

When you learned English, you probably learned that if and whether are interchangeable.
This is correct in constructions like (30), but not so in examples like (31):

(30) a. I wonder if Tuoi is in love this time.

b. I wonder whether Tuoi is in love this time.

(31) a. I wonder whether to go to Dalat (or not).

b. *I wonder if to go to Dalat (or not).

There is no noticeable difference between (30a) and (30b), so from these data we may
conclude that both if and whether are complementizers. However, as we can see in (31), the
complementizer if is not compatible with infinitives, while whether is. The
complementizer if requires that the embedded verb in the lower IP bear some tense.
As we saw in chapter 2, only heads can establish restrictions on their complement(s). This
suggests that if is a head while whether is not.

Since whether is in front of the embedded sentence, we can assume that it is in CP.
However, since it is not a head (because it does not set the tense restriction that if does), we
can assume that it is a specifier. This suggests that whether may not be a complementizer
per se as complementizers (that, if, Φ, as we have seen before). In addition, whether looks
like a wh-word by the fact that it is used in indirect questions, just like other question words
like who, what, where, etc. Similar to other wh-words, it may appear with tensed or
untensed verbs. Notice the similarity of the behavior between when and where in (32). We
see that wh-words as well as whether, unlike if, are insensitive to tense in the embedded
clause:
 English Syntax & Universal Grammar 211

(32) a. I'm wondering when Co Chi rests.

b. I wonder when to rest.
c. I wonder whether Co Chi needs a rest.
d. I wonder whether to rest (or not).

Whether can also be conjoined with other question words which are adverbs, thus
suggesting that they share the same category, namely ADVP.

(33) a. The question of whether and when to become a parent will be debated tonight.
b. We cannot predict whether and how those proposals will be implemented.
c. *We cannot predict whether and who will come.
d. *We cannot predict whether and what she will do.

We saw in this chapter that wh-words usually move to CP, they are not generated or merged
there. So where is whether moving from to appear in front of the sentence?

Before we go on, it must be noted that whether… or not is the interrogative form for
either… or, whose negative is neither… nor. Thus there is an implicit conjunctive structure,
a Conjunctive Phrase or ConjP. Unfortunately, the internal structure of conjunctive
structures is beyond our level of comprehension at this point, so we will simplify our
discussion by just labeling it ConjP. Consider (33):

(33) a. I know [CP that [IP he is [ConjP either a doctor or an engineer]]].

b. I know [CP that [IP he is [ConjP neither a doctor nor an engineer]]].

(33a) is an affirmative conjunctive structure, while (33b) is a negative conjunctive structure.

This conjunctive structure can also be interrogative, as shown in (34a). However, since
whether is a wh-word, it must satisfy the Wh-Criterion, thus it moves to the specifier of CP,
as shown in (34b).

(34) a. I don't know [CP Φ[IP he is [ whether a doctor or an engineer]]].

b. I don't know [CP whetheriΦ[IP he is [ tia doctor or an engineer]]].
c. I don't know [CP whetherΦ [IP he is [ ti a doctor or not (a doctor)]]].
d. I don't know [CP whetherΦ [IP he is [ ti a doctor (or not a doctor)]]].

As shown in (34c,d), the second conjunct need not be expressed. If it is not mentioned, we
interpret the second conjunct as the negation of the first conjunct. We may or may not use
212 Héctor Campos & Bùi Huỳnh Thủy Thương

or not as shown in (34c,d). In (34c,d) we see that the second conjunct has been elided.
Ellipsis is a topic for more advanced syntax that we cannot tackle here.

We thus see that we can treat whether as a wh-word, which originates in a Conjunction
Phrase and moves to the specifier of CP in order to satisfy the Wh-Criterion, similar to the
other question words that we have encountered. Although in many constructions whether
and if are interchangeable, we have argued that if is a complementizer and is merged in C,
while whether is an adverbial wh-word and it moves to the specifier of CP, just like other
wh-words.

10.8 Summary

In this chapter we have studied wh-question both in main and in embedded sentences. We
have also studied short- and long-distance movement. We have seen that all cases of wh-
movement in interrogative sentences can be explained assuming two basic principles: the
Stranded Affix Condition, which we motivated in a previous chapter, and the Wh-
Criterion. The Stranded Affix Condition was the motivation to move I to C, thus obtaining
subject/verb inversion, a phenomenon which is found in main clauses, but not in embedded
clauses in standard English. The Wh-Criterion guaranteed that the wh-word would be
moved to the front of the sentence. These two basic principles are able to explain the basic
behavior of most wh-interrogatives in English with the exception of wh-subjects. We
showed that extraction of wh-subjects requires a special complementizer that we have
labeled ΦS. This complementizer must be used when we extract a subject both from a main
as well as from an embedded clause.
 English Syntax & Universal Grammar 213

CHAPTER 11

WH-MOVEMENT: RELATIVE CLAUSES

In Chapter 9 we studied the syntax of yes/no questions. We proposed that these

interrogative structures include a CP with the features [+Q,-WH]. In Chapter 10 we
analyzed specific or wh-questions. We suggested that these structures contain a CP with the
features [+Q, +WH]. In this chapter we will consider CP with the features [-Q,+WH]. The
features indicate that these are not questions but yet, they contain a wh-element. We will
argue below that this is the specification for relative clauses.

11.1 Restrictive and Non-restrictive relative clauses

Relative clauses function like adjectives, they are used to modify the noun. That is why in
traditional grammar, relative clauses are also called adjectival clauses. A relative clause
construction has three parts, as we can see in (1):

(1) [the girl] [ who I met yesterday]

antecedent relative pronoun
[___________relative clause_____]

A relative clause construction has an antecedent: this is the noun we are modifying with a
clause. The clause we use to modify the adjective is the relative clause. Finally we need a
pronoun to connect the antecedent with the relative pronoun. This pronoun is called the
relative pronoun.

There are two kinds of relative clauses: restrictive relative clauses (also called specifying
relative clauses) and non-restrictive relative clauses (non-specifying relative clauses).
Consider the examples in (2):

(2) a. the students who study hard will get an A in this course. (restrictive)
b. the students, who study hard, will get an A in this course. (non-restrictive)

To understand the difference between restrictive and non-restrictive let us consider the
following situation. Imagine that we have a group of students that we have classified into
three groups: those who study hard, those who do not study hard and those who do not
214 Héctor Campos & Bùi Huỳnh Thủy Thương

come to class. In (2a) we are saying that only those who study hard will pass the course.
This is represented in the diagram in (3):

(3)

(2a) is a restrictive relative clause because we are restricting, specifying or delimiting the
set of students, as shown in (3).

In a non-restrictive relative clause, on the other hand, we do not delimit the set. We refer to
the entire set. So in (2b) we are saying that all the students study hard and that all the
students will get an A in the course. This is shown in diagram (4):

(4)
Students who study hard

As we can see in (4) we are referring to the entire set of students. As we can see in (2), we
distinguish restrictive from non-restrictive relative clauses by writing a comma in the
latter but not in the former. The syntax of restrictive and non-restrictive relative clauses is a
bit different, so we will study them separately.

11.2 Restrictive relative clauses.

Let us start with restrictive relative clauses. We can think of the relative pronoun as the
"representative" of the antecedent in the relative clause. As we will see below, the
antecedent, via its representative (the relative pronoun), may have different functions within
the relative clause.
 English Syntax & Universal Grammar 215

Let us consider HUMAN antecedents first:

(5) a. the professor [(who)/that [___ teaches linguistics]] (subject)

b. the professor [(who)/(that) [we met ___]] (direct object)
c. the professor [to whom [we sent a letter ___]] (indirect object)
c'. the professor [(who)/(that) [we sent a letter to ___]] (indirect object)
d. the professor [with whom [we study syntax ___]] (object of a preposition)
d'. the professor [(who)/(that) [we study syntax with _]] (object of a preposition)
e. the professor [whose wife [we met ___]] (possessor, direct object)

Notice that in all the examples in (5) the antecedent is HUMAN. Human antecedents may
take who as a relative pronoun. As we can see in (5), they may also appear with the
conjunction that. Notice that in all constructions where the conjunction that is possible, it is
optional, with the exception of the subject position in (5a). On the other hand, the relative
pronoun who(m) is also optional, except in cases where it is preceded by a preposition (5d)
or a pseudo-preposition (5c). If the (pseudo-)preposition is left stranded, then the relative
pronoun becomes optional, as we can see in (5c') and (5d').

When the antecedent is NON-HUMAN, we may get the relative pronoun which or the
conjunction that, as shown in (6):

(6) a. the book [(which)/that [___ costs $50]] (subject)

b. the book [(which)/(that) [we bought ___ at the bookstore]] (direct object)
c. the book [with which [we study ___]] (object of a preposition)
d. the book [(which)/(that) [we study with ___]] (object of a preposition)
e. the book [whose cover [my friend designed ___]] (possessor, direct object)

As we can see in (6), non-human antecedents may use the relative pronoun which or the
conjunction that to introduce the relative clause. Where the conjunction that is possible, it is
always optional, with the exception of the subject in (6a). The relative pronoun which is
obligatory when it is preceded by a preposition.

Let us consider these constructions in detail. We will assume that restrictive relative clauses
have the structure in shown in (7):
216 Héctor Campos & Bùi Huỳnh Thủy Thương

(7)

As we can see in (7), we will analyze relative clauses as a CP. Because relative clauses
modify or restrict the noun, we will merge them as modifiers, in a second N', above the
node N' reserved for complements. We will look at the structure of CP in more detail in the
sections that follow.

11.2.1 Restrictive relative clauses with Who or Which. Consider the two examples in
(8):

(8) a. the professor who teaches linguistics

b. the book which we bought at the bookstore.

The S-Structure for (8a) is shown in (9a,b):

(9) a.

b. [DP the [NP professor [CP whoiΦ [IP ti teaches linguistics]]]]

 English Syntax & Universal Grammar 217

In (9) we can see that relative clauses have a special combination of features in C, namely

[-Q,+WH]. [-Q] indicates that the clause is not a question, [+WH] indicates that the clause
contains a WH-element. As usual, inflection lowers to the verb in (9) to satisfy the Stranded
Affix Condition. The relative pronoun, a wh-word, on the other hand, must move to the
specifier of CP to satisfy the Wh-Condition.

Consider now (8b), shown in (10):

(10) a. the book which we bought at the bookstore

b. [DP the [NP book [CP whichiΦ [IP we bought ti at the bookstore]]]]

As we can see in (10a), I lowers to V to satisfy the Stranded Affix Condition and the
relative pronoun which moves to the specifier of CP to satisfy the Wh-Criterion.

When we have an indirect object or an object of a preposition, either the relative pronoun or
the pronoun together with the (pseudo-) preposition move to the specifier of CP, as shown
in (11) and (12) below. We only represent the bracket structure; you should try drawing the
tree on your own:
218 Héctor Campos & Bùi Huỳnh Thủy Thương

(11) a. the professor who(m) he studies linguistics with

b. [DP the [NP professor [CP who(m)iΦ [IP he studies linguistics [PP with ti] ]]]]

In (11b) the wh-word moves to the specifier of the CP to satisfy the Wh-Criterion. The
tense/agreement inflection –s lowers to the verb to satisfy the Stranded Affix Condition.

Alternatively, the wh-feature of the DP whom may percolate all the way up to PP, thus
turning PP into a wh-word, forcing it to move to the specifier of CP to satisfy the Wh-
Criterion:

(12) a. the professor with whom he studies linguistics

b. [DP the [NP professor [CP [PP with whom]iΦ [IP he studies linguistics ti]]]]

As in structures above, the wh-word moves to the specifier of CP to satisfy the Wh-
Criterion. At the same time, the –s inflection lowers to the verb to satisfy the Stranded Affix
Condition.

Consider now (6e), repeated here as (13):

(13) a. the book whose cover my friend designed last year

b.
 English Syntax & Universal Grammar 219

c. [DP the [NP book [CP [whose cover]iΦ [IP my friend designed ti last year]]]]

As we can see in (13), the relative pronoun whose turns the whole DP into a wh-word DP.
As such, it must move to the specifier of CP to satisfy the Wh-Criterion. As usual, I lowers
to V to attach to the verb.

It should be clear from the preceding discussion that we can generate restrictive relative
clauses with the following minimal assumptions:

14) Properties of Relative Clauses

(a) Relative clauses are CPs with the features [-Q,+WH].

(b) CP relative clauses are merged as modifiers to the noun, under a second level of N'.

(c) Relative clauses include a wh-element which serves as a relative pronoun. This wh-
element or relative pronoun, being a wh-word, is subject to the Wh-Criterion. Thus
relative clauses will always involve the movement of a wh-element.

(d) In our system that is analyzed as a complementizer, unlike traditional grammar

which considers it to be a relative pronoun. Notice that in English relative pronouns
are sensitive to [±HUMAN]. That shows no such restriction.

There is evidence for our claim in (d), namely, that that is a complementizer and not a
relative pronoun. In Middle English as well as in some Germanic dialects we may get a
relative pronoun and a complementizer co-occurring: thy freend which that thou has lorn'
your friend which you have lost' (Middle English), de Hund der wo gester d'Katz bissn hod
'the dog that bit the cat yesterday' (Bavarian). Thus we will assume that the that which
appears in relative clauses is a complementizer and not a relative pronoun.

PRACTICE 1

I. Diagram the following noun phrases which include restrictive relative clauses:
a. the syntax book which the professor recommended
b. the boy who I met at Ha's party last night
c. the Vietnamese professor with whom I will study syntax next semester
d. the Vietnamese professor who I will study syntax with next semester
e. the library where we do research
f. the day when my brother bought his first motorcycle
g. the student who everybody thinks that the professor likes
220 Héctor Campos & Bùi Huỳnh Thủy Thương

II. Diagram the following sentences which include restrictive relative clauses:
a. Do you think I should buy the book which the professor recommended last class?
b. Have you studied with the professor who the English Department just hired?
c. I don't know when I will meet the man who my father works for.

11.2.2 Restrictive relative clauses without a pronounced relative pronoun. In the

previous section we studied restrictive relative clauses that included visible wh-words. In
this section we will study relative clauses with silent wh-words. These constructions
typically appear with the relative complementizer that or the zero-complementizer as shown
in (15).

(15) a. the professor [that [ ___ teaches linguistics]] (subject)

b. the professor [(that) [ we met ___]] (direct object)
c. the professor [(that) [we sent a letter to ___]] (indirect object)
d. the professor [(that) [we study syntax with ___]] (object of a preposition)

In (15) we repeat the examples we mentioned in (5) and which may appear with that. All
these examples include a human antecedent. As we mentioned above, the relative
complementizer that is optional in all cases except when the antecedent is the subject of
the relative clause, as in (15a).

The same restriction is observed when the antecedent is non-human:

(16) a. the book [that [ ___ is on the table]] (subject)

b. the book [(that) we bought ___ at the bookstore]] (direct object)
c. the book [(that) we study with ___]] (object of preposition)

We will assume that relative clauses that do not contain a pronounced wh-word (who,
which, where, when, how, etc.) contain a silent wh-word. We will use the symbol wh-OP
(for 'wh-operator') to represent these silent wh-phrases. (17) shows the structure for (15a),
repeated here as (17):

(17) a. the professor that teaches linguistics

 English Syntax & Universal Grammar 221

b.

c. [DP the [NP professor [CP wh-OPi that [IP ti teaches linguistics ]]]]

As we can see in (17b), inflection lowers to V to avoid violating the Stranded Affix
Condition. The wh-operator (a silent wh-word) must move to the specifier of CP to
satisfy the Wh-Criterion, similar to what we observed with pronounced wh-relative
pronouns in the previous section. Remember that when the antecedent is the subject of the
relative clause and we are using a silent wh-operator, the relative complementizer (the
head of the CP which introduces the relative clause) must be realized as that. This is
observed both for human (15a) as well as for non-human antecedents (16a).

In all the other cases in (15) and (16) the use of the relative complementizer that is optional.
In this case the complementizer may be realized either as that or as a zero-relative
complementizer. In the following examples, we show the bracket structure. You should do
the tree structure on your own:

(18) a. the professor that we met

b. [DP the [NP professor [CP wh-OPi that [IP we met ti ]]]]

In (18), similar to what we observed in (17), inflection lowers to the verb in IP and the silent
wh-operator moves to the specifier of CP to satisfy the Wh-Criterion. In (18b) the relative
complementizer has been realized as that. As shown in (19), the relative complementizer
may also be realized as the zero-relative complementizer, Φ:

(19) a. the professor we met

b. [DP the [NP professor [CP wh-OPiΦ [IP we met ti ]]]]
222 Héctor Campos & Bùi Huỳnh Thủy Thương

We thus see that (18) and (19) have parallel derivations. The only difference between (18)
and (19) is that the relative complementizer may be spelled out as that or Φ when we use a
silent wh-operator. As we saw in (17), the only exception to this rule would be when the
wh-operator starts in subject position, in which case the relative complementizer must
always be realized as that.

Finally consider antecedents that are interpreted as objects of prepositions and which
contain a silent relative pronoun, as in (20):
(20) a. the professor that I studied with
b. [DP the [NP professor [CP wh-OPithat [IP we studied with ti ]]]]
c. the professor I studied with
d. [DP the [NP professor [CP wh-OPiΦ [IP we studied with ti ]]]]

As we can see in (20), objects of prepositions can also be antecedents and they can be
replaced by a silent wh-operator which serves as a relative pronoun. The only difference
between (20a) and (20c) is on whether we pronounce the relative complementizer or not. In
both cases inflection is lowered to the verb and the silent wh-operator is moved to the
specifier of CP to satisfy the Wh-Criterion.

In this section we have shown that relative clauses without pronounced relative pronouns
have a silent wh-operator which functions as the relative pronoun. Since this wh-operator is
a wh-word (albeit silent!), it must obey the Wh-Criterion and must move to the specifier
position of CP which contains the features [-Q, +WH]. Thus relative clauses with overt
relative pronouns and those with silent relative pronouns receive a parallel analysis in our
model.

Table (21) summarizes the different realizations that C may have depending on its [±Q] and
[±WH] values:
Main clause Embedded clause
(21)
C [-Q, -WH] Φ Φ, that
(declarative)
C [+Q, -WH] -Φ If
(yes/no question)
C [+Q, +WH] -Φ Φ
(wh-question)
C [-Q, +WH] * Φ, that
(relative clauses)
 English Syntax & Universal Grammar 223

As we can see in (21), relative clauses can only appear in embedded clauses. As we have
shown in this and in the previous section, the relative complementizer may be realized
either as Φ or that, similar to what we observe in declarative embedded clauses.

The chart in (22) summarizes the type of relative pronoun and complementizer we may find
in restrictive relative clauses:

(22) Relative clauses: relative pronounsand relative complementizers

Relative
Antecedent is Relative pronoun
complementizer
SUBJECT
a. human a. who (9) a. Φ
a'. wh-OP (17) a'. that
b. non-human b. which b. Φ
b'. wh-OP c. that
DIRECT OBJECT
a. human a. who(m) a. Φ
a'. wh-OP (18,19) a'. Φ, that
b. non-human b. which (10) b. Φ
b'. wh-OP b'. Φ, that
INDIRECT OBJECT/OBJECT OF P
a. human a. P whom (12) a. Φ
a'. who(m) ... P (11) a'. Φ
a''. wh-OP .. .P (20b,c) a''. Φ
b. non-human b. P which b. Φ
b'. which ... P b'. Φ
b". wh-OP ... P b". Φ
POSSESSOR
human/non-human whose+N (13) Φ

PRACTICE 2

I. Diagram the following noun phrases that contain a restrictive relative clause:
a. the bus I took this morning
b. the company I work for
c. the teacher to whom I gave my homework
d. the teacher that I gave my homework to
e. the teacher I gave my homework to
224 Héctor Campos & Bùi Huỳnh Thủy Thương

II. Motivate an analysis for each of the following constructions:

a. the day you were born
b. the way she behaves
c. the reason she came

III. Propose an analysis for each of the following relative clause constructions in
Spanish. How is Spanish different from English?
a. la chica con quien yo estudi-o
the girl with whom I study-present
'the girl with whom I study'
b. el chico con quien yo estudi-o
the boy with whom I study
'the boy with whom I study'
c. la chica con la cual yo estudi-o
the girl with the-fem,sg which I study-pres
'the girl with whom I study'
d. el chico con el cual yo estudi-o
the boy with the-masc,sg which I study-pres
'the boy with whom I study'
e. la chica con la que yo estudi-o
the girl with the-fem,sg that I study-pres
'the girl with whom I study'
f. el chico con el que yo estudi-o
the boy with the-masc,sg that I study-pres
'the boy with whom I study'

IV. Diagram the following sentences:

a. Have you seen the new film that is showing at the Galaxy?
b. When will the teacher correct the sentences which we did last night?
c. I bought the book she said she enjoyed a lot.
d. I studied with the teacher whose husband works for a foreign company.
e. The tourists that went to Phan Thiet wrote of review of the hotel where they stayed.
 English Syntax & Universal Grammar 225

CHALLENGE

Diagram the following structures. Expand all nodes. Comment on any structure that
you feel deserves a comment. Indicate all the moved and inserted elements.
a. When did you say that Stella would go to China? (ambiguous)
b. How much did that Pakistani taxi driver charge the tourists?
c. How much sugar do you think Annie put in Masha's lemonade?
d. The fact that some students come to class late surprises the teacher who teaches that
syntax class for first year students.
e. The professor from whom she took generative syntax criticized the sources which
Eva used in the final paper for that class.
f. Where did you put the book we bought?
g. Do you know the girl with whom Jonathan went out?1
h. ¿Conoce-s tú a la chica con quien Jonathan sal-ía?
Know-pres,2sg you "a" the girl with whom Jonathan went-out-past,3sg
´Do you know the girl with whom Jonathan went out?´
i. ¿Conoce-s tú al chico con el que Tris sal-ía?
Know-pres,2sg you "a"'-the boy with the-masc,sg that Tris went-out-past,3sg
'Do you know the boy that Tris went out?'

RESEARCH PROJECT

Translate the following sentences from English to Vietnamese:

a. the man that works at that company
b. the man that I saw
c. the man for whom I work

Compare the structure of the sentences above with the equivalent sentences in
Vietnamese. Propose an analysis for the sentences in Vietnamese.

11.3 Non-restrictive relative clauses

We can see in (22) that the system for restrictive relative clauses is quite complex. The
choice of who/which is sensitive to the feature [±HUMAN]. The choice of relative
complementizer is dependent on whether we have a pronounced or a non-pronounced
1
Cf. -Where is Jonathan? -He went out.
226 Héctor Campos & Bùi Huỳnh Thủy Thương

relative pronoun. In this section we will study non-restrictive relative clauses. As you can
see below, the system is much simpler than that found in restrictive relative clauses. Our
examples below parallel those in (5) and (6) so you can observe the differences better.

Let us consider HUMAN antecedents first:

(23) a. the students, [who [ ___ study linguistics]], ... (subject)

b. the students, [who(m) [we met ___]], ... (direct object)
c. the students, [to whom [we sent a letter ___]], ... (indirect object)
c'. the students,[who(m) [we sent a letter to ___]], ... (indirect object)
d. the students, [with whom [we study syntax ___]], .. (object of a preposition)
d'. the students, [who(m) [we study syntax with ___]], ..(object of a preposition)
e. the students, [whose wives [we met ___]], ... (possessor, direct object)

As we can see in (23), non-restrictive relative clauses with a human antecedent only use
WHO/WHOM as a relative pronoun. The wh-operator option that we studied in the
previous sections is not used with non-restrictive relative clauses. In other words, the
relative pronoun must always be pronounced in non-restrictive relative clauses.

When the antecedent is NON-HUMAN, we only get the relative pronoun which, as shown
in (24). For possessives, as in (24d), we still use whose:

(24) a. the books, [which [___ are on the table]], ... (subject)
b. the books, [which [we bought ___ at the bookstore]], ... (direct object)
c. the books, [with which [we study ___]], ... (object of preposition)
c'. the books, [which [we study with ___]], ... (object of preposition)
d. the books, [whose cover [my friend designed ___]], .. (possessor, direct object)

Non-restrictive relative clauses are seldom used in conversation but rather, they belong to a
more formal or literary style. That is the reason why constructions with preposition
stranding, like the examples in (23c', d') and (24 c'), are less common in a formal or literary
style than the corresponding structures where pied-piping has applied.

As we mentioned in section 1, a non-restrictive relative clause does not restrict or limit the
antecedent, it just gives extra information about the antecedent. That is the reason why
when we have an antecedent that is unique, we can only use a non-restrictive relative clause
to modify it. Compare the two sentences in (25):

(25) a. my father, who is 69 years old, lives in California.

 English Syntax & Universal Grammar 227

b. my father who is 69 years old lives in California.

In (25a), I am giving extra information about my father and I am also saying that I only
have one father. In (25b), on the other hand, I am saying that I have more than one father
and only the one who is 69 years old lives in California. We will assume that non-restrictive
relative clauses modify DP rather than NP, as shown in (26):

(26)

In fact, we can even have a DP which includes both a restrictive and a non restrictive
relative clause, as we can see in (27):

(27) a. [The students who are smart, who study very hard], get excellent grades.

b.

In (27) we are dividing the group of students into those who are smart and those who are
not and I am talking only about the group of those students who are smart. I am saying that
all those students who are smart also study very hard.
228 Héctor Campos & Bùi Huỳnh Thủy Thương

The syntax of non-restrictive relative clauses is pretty much the same as the syntax we
discussed for restrictive relative clauses. As we can see in (27), the only difference will be
that the non-restrictive relative clause merges as a modifier to DP rather than as a modifier
to NP.

Consider the example in (28):

(28) a. my classmates, who work very hard, ...

b.

c. [DP my [NP classmates] [CP whoiΦ[IP ti work very hard ]]]

In (28b) the affix in the embedded I lowers to the verb work to satisfy the Stranded Affix
Condition. The wh-word who moves to the specifier of CP to satisfy the Wh-Criterion. In
the bracket structure shown in (28c) we see that the relative clause CP is outside the NP,
thus intended to modify the DP my classmates.

We have already explored one important difference between restrictive and non-restrictive
clauses, namely the fact that non-restrictive clauses cannot be constructed with the silent
wh-operator. Another difference is the non-restrictive relative pronoun which may take a
previous sentence as an antecedent:
 English Syntax & Universal Grammar 229

(29) Quan kissed Ha, which surprised Khiem.

In (29), what surprised Khiem was the fact that Quan kissed Ha, so the whole sentence is
the antecedent for the relative pronoun which. The structure for (29) is shown in (30):

(30)

As we can see in (30), the non-restrictive relative clause [which surprised Khiem] modifies
the whole previous sentence. The more detailed relative clause in (30) is shown in (31):

(31) a. Quan kissed Ha, which surprised Khiem.

b.

In (31) the relative pronoun which replaces the whole sentence inside the relative clause.
Since which is a wh-word, it must move to CP to satisfy the Wh-Criterion.
230 Héctor Campos & Bùi Huỳnh Thủy Thương

PRACTICE 3

Diagram the following sentences. Pay attention to the restrictive and non-restrictive
relative clauses.
a. The professor, who has taught that syntax class many times, loves the students who
take his class.
b. She never did the exercises the professor assigned, which explains why she failed the
class.
c. My parents, who I talk to on the phone every week, sent me a present they bought for
me in Spain.
d. Where did you buy the book that the professor recommended for our final paper?
e. Don't you remember the time when you lived with your parents in Hanoi?

CHALLENGE

I. Diagram the following noun phrases that contain a relative clauses in Vietnamese.
Is the use of "mà" obligatory? Then discuss the differences you observe between
English and Vietnamese:
a. cuốn sách mà tôi đang đọc
b. người mà anh gặp hôm qua
c. trường đại học mà ba ấy giảng dạy
d. năm mà tôi tốt nghiệp đại học
e. người đến đây hôm qua

II. Diagram the following sentences and discuss the differences you observe between
English and Vietnamese:
a. Em rất thích cái ví mẹ tặng cho ba.
b. Em có thích cái ví mà mẹ tặng cho ba không?

11.4 Summary

In this chapter we studied the syntax of relative clauses. We divided relative clauses into
restrictive and non-restrictive relative clauses. We analyzed the former type as NP
modifiers while we analyzed the latter type as DP modifiers. We showed that in both
instances the relative clause should be analyzed as a CP. We invoked two basic
principles to motivate the different structures, principles that we had already invoked in the
previous chapter for Wh-questions, namely the Stranded Affix Condition and the Wh-
 English Syntax & Universal Grammar 231

Criterion. The Wh-Criterion guarantees that the wh-relative pronoun gets moved to the
front of the CP, next to the antecedent, while the Stranded Affix Condition guarantees that
no verbal affix is left unattached. We saw that the system for restrictive relative clauses is
much more complex than that for non-restrictive relative clauses, requiring the postulation
of a silent relative pronoun, wh-operator, which moves just like other wh-words so as to
satisfy the Wh-Criterion.

One natural question to ask is whether wh-movement can apply in any context (context-
free) or if there are any constraints to this kind of movement. As we can see in (31) – (33)
below, the movement of the wh-word, both in wh-questions and in relative clauses is not
context free. Movement of the wh-word is blocked in the following examples:

(31) a. I wonder whether Quan loves Ha.

b. *Who do you wonder whether Quan loves?

(32) a. [That Quan loves Ha] is obvious.

b. *Who that Quan loves is obvious.

(33) a. I met the girl who John invited to the party.

b. I met the girl that I know that John invited to the party.
c. *I met the girl who I wonder why John invited to the party.

As we can see in the previous examples, who cannot replace the noun Ha in (31) and (32) in
the contexts provided. Nor can we build a long distance relative clause such as (33c),
although we can build a relative clause over the context in (33b).

We will not study restrictions on wh-movement in this book. This is the topic for a more
advanced syntax course.
232 Héctor Campos & Bùi Huỳnh Thủy Thương

APPENDIX: RELATIVE CLAUSES WITHOUT A SILENT ANTECEDENT

As we have seen in this chapter, a relative clause construction has three elements that are
crucial for our analysis:

(1) [the girl] [CP whoi Φ [IP I met ti yesterday]

antecedent relative pronoun relative complementizer
[___________relative clause__________________________________]

We discussed relative clause constructions where the relative pronoun is silent. We argued
that in those cases we had a silent wh-operator that moved to CP, exactly like an overt wh-
relative pronoun. We also saw that the relative complementizer could be realized as zero
orthat.

There is yet another type of relative clauses that we have not analyzed. These are the so
called free relatives. These free relatives are shown in brackets in (2):

(2) a. Ha is [who always knows the answer].

b. Chocolates is [what Quan bought yesterday].
c. That store is [where she works].
d. Tomorrow is [when I will take my final exam].

The constructions in brackets are called free relatives because they seem to be lacking an
antecedent if they are to follow the structure shown in (1). It is interesting to observe that we
can replace the relative pronouns in (2) by an appropriate noun phrase, as in (3):

(3) a. Ha is [the person that always knows the answer].

b. Chocolates is [the thing that Quan bought yesterday].
c. That store is [the place she works at].
d. Tomorrow is [the day that I will take my final exam].

We will thus assume that these constructions contain a silent noun which is interpreted
generically as PERSON, THING, PLACE, TIME, etc. This silent noun is modified by the
relative clause. Thus free relatives have the same syntax as regular relative clauses. The
structure for the free relative in (4a) is shown in (4b):

(4) a. [What she bought] was a sweater.

 English Syntax & Universal Grammar 233

b.

As we can see in (4), free relatives have the same structure as restrictive relative clause
constructions, the only difference being that the antecedent in (4) is silent and interpreted
generically as [THING]. We still have a relative pronoun what, which moves to the
specifier of CP to satisfy the Wh-Criterion.

Now that you know the structure of free relatives, try diagramming the other examples in
(2) for some extra practice.
234 Héctor Campos & Bùi Huỳnh Thủy Thương

CHAPTER 12

NON-TENSED CLAUSES

In Chapter 7 we learned how to combine tensed declarative sentences, as in (1a). In that

chapter we also learned how to embed an interrogative sentence (1b). Then in Chapter 10
we learned how to embed wh-questions, as in (1c):

(1) a. I know [CP that/Φ [IP Khiem likes linguistics]].

b. I wonder [CP if [IP Khiem likes linguistics]].
c. I wonder [CP whatiΦ [IP Khiem likes ti]].

In this chapter we will learn to embed infinitival clauses. We will distinguish four types of
infinitival clauses:

(2) a. Khiem wants to study advanced syntax.

b. I want Khiem to study advanced syntax.
c. I told Khiem to study advanced syntax.
d. I made Khiem study advanced syntax.

We will study these types of sentences in the following sections.

12.1 Infinitival clauses without an overt subject: subject-controlled PRO

Example (2a), repeated here as (3a) shows an example of a sentence with an infinitive
without an overt subject. The structure for the embedded infinitive is shown in (3b):

(3) a. Khiem wants to study syntax.

 English Syntax & Universal Grammar 235

b.

Let us concentrate on the important differences between a tensed sentence and an infinitival
sentence.

(a) CP: As we can see in (3b), we will assume that infinitival sentences without an
overt subject are CPs. In (3b) the embedded CP has the features [-Q,-WH]. These
features are pronounced as a zero-complementizer in English. As we can see in
(1a), in an embedded tensed sentence, C [-Q,-WH] can be either pronounced as that
or as Φ.

(b) IP: Let us look at the embedded IP in (3b). Infinitival clauses do not have tense,
thus we will label them as [-TENSE] ([-TNS]). Neither do they have agreement;
thus we will label them as [-AGREEMENT] ([-AGR]). Thus infinitival clauses
will always be [-TNS, -AGR]. In English the combination of these features [-
TNS, -AGR] is spelled out as TO, which is the morpheme that marks infinitives in
English. Tensed clauses, as those in (1), always have either tense or a modal in I.

(c) The subject of the infinitive:The Extended Projection Principle, shown in (4)
below, requires that every sentence have a subject.

(4) Extended Projection Principle (EPP)

Every sentence (IP) must have a subject.
236 Héctor Campos & Bùi Huỳnh Thủy Thương

From the Principle in (4) it follows that we need to have a subject in the specifier position of
IP in (3b). We will assume that this subject of an infinitive clause without an overt
subject is a silent pronoun, which we will represent as PRO (typically called "big
PRO"). Notice that PRO bears a subindex "i". If you examine the structure in (3b)
carefully, you will see that Khiem also bears a subindex "i". This indicates that PRO takes
Khiem as its antecedent. In other words, PRO refers to Khiem. We say that PRO and
Khiem are coreferent. This coreference is indicated with the same subindex. We can use
different letters to indicate coreference, as long as the same letter links PRO to its
antecedent. In example (3), we say that we have an instance of subject-controlled PRO
since it is a subject that gives reference to -or controls- PRO. In section 3 we will study
cases where PRO in controlled by an object rather than by a subject.

In Chapter 7 we motivated the Case Condition, repeated here as (5a). (5b) shows the
different elements that can assign Case in our system:

(5) a. The Case Condition

Every DP must receive Case.
b. Case-assigning elements:
i. Verbs assign Objective Case.
ii. Prepositions assign Oblique Case.
iii. [+TNS, +AGR] assigns Nominative Case to the subject.

As we can see in (5b), only three elements assign Case in our system: verbs, prepositions
and a tensed inflection, that is, an inflection head that has both a specification for tense and
agreement. Let us go back to (3b) and check Case on every DP. The subject of the main
clause Khiem is assigned Nominative Case by I [PRES, IIISG] since inflection contains
both Tense and Agreement. The embedded object syntax receives Objective Case from the
verb since it is its complement. What about PRO, the subject of the infinitival clause? The
embedded inflection is [-TNS, -AGR], thus it is not able to assign Case. Nor is PRO the
complement of the main verb want. Thus we will assume that PRO has inherent Null
Case. This means that PRO comes with Null Case from the Lexicon. Because it has
inherent Null Case, PRO can only appear in positions where Case is not assigned. This
is borne out. PRO will appear only as a subject of infinitival or gerundive clauses
where there is no tense or agreement.
 English Syntax & Universal Grammar 237

PRACTICE 1

I. Diagram the following sentences:

a. The students intend to study hard.
b. My parents decided to go to Europe next summer.
c. What movie do you want to see this weekend?
d. When are they hoping to leave?
e. The syntax students expect to pass the final exam with honors.

II. The following sentence is ambiguous. Propose two structures to account for the
ambiguity.
When did thy promise to leave?

12.2 Infinitival clauses with overt subjects

In section 1 we saw infinitival verbs that appear without an overt subject. We claimed that
the subject of an infinitival clause without an overt subject is a silent pronoun, which
we labeled PRO, as in (6).

(6) Theyi promised [CP Φ [IP PROi to come on time]]

However, there are verbs that allow realized subjects in the infinitival clauses. There are
two types of subjects with infinitival clauses: those that appear preceded by for and those
that appear without for, as shown in (7):

(7) a. They arranged for [Susan to stay in Washington].

b. They want [Susan to stay in Washington].

We will discuss these two constructions in the two following sections.

12.2.1 Verbs that take an infinitival clause with a For complementizer. Some verbs,
but not all, allow a For complementizer. FOR is called a prepositional complementizer.
As a complementizer, ‘for’ introduces an embedded sentence. As a preposition, it is able
to assign Oblique Case to the following DP. This is why constructions like (7a) allow for a
subject preceding the infinitive. Although the infinitive is not able to assign Case to the
embedded subject, the prepositional complementizer can do so, in spite of the fact that the
subject of the infinitive is not its complement. The structure for the embedded infinitive
clause is shown in (8):
238 Héctor Campos & Bùi Huỳnh Thủy Thương

(8) a. They arranged [CP for [IP Susan to stay in Washington]].

b.

Oblique Case

In (8b) the DP Susan is able to receive Oblique Case from the prepositional
complementizer, thus it satisfies the Case Condition in (5). As we mentioned above, the DP
Susan is not a complement of the prepositional complementizer for. However, for is able to
assign Case across the IP. For this to be possible, it must be the case that IP does not block
Case assignment by the prepositional complementizer. We say that IP is not a barrier to
Case assignment from the prepositional complementizer. In other words, the IP is
invisible for (or does not block) Case assignment in (8b).

Notice that if we do not include the prepositional complementizer in (8a), the sentence
becomes ungrammatical, as shown in (9a):

(9) a. *They arranged Susan to stay in Washington.

b. *They arranged [for [PRO to stay in Washington]]
c. *They hope for Susan to stay in Washington.

(9a) is ungrammatical because the DP Susan fails to get Case because the infinitive cannot
assign Case to it, nor is there a prepositional complementizer present to assign Oblique Case
to it. (9b) is ungrammatical because PRO would be assigned Case by the propositional
complementizer for, but as we claimed in the previous section, PRO has inherent Null Case,
thus it cannot be assigned a second Case. In (9c) we can see that not every verb can be
followed by a prepositional complementizer for. Thus while the verb arrange can take a
complementizer for, the verb hope cannot.

This special property of being able to take a for-prepositional complementizer needs to

be specified in the lexicon as a special property of those verbs. Other verbs that can take
 English Syntax & Universal Grammar 239

the prepositional complementizer are: arrange, advertise, campaign, long, plan, wait, etc.
In fact, not many verbs allow for this option.

PRACTICE 2

Diagram the following sentences:

a. The landlord threatened to evict them from their apartment.
b. They long for their son to visit them every weekend.
c. They said that Rania would visit France next summer.
d. What country does Rania want to visit next summer?
e. What country did they arrange for Rania to visit next summer?

RESEARCH PROJECT 1

1. Does Vietnamese have infinitives? How would you say the following sentences in
Vietnamese?
Matt wants to come to Vietnam.

2. Does Vietnamese have the equivalent of a 'for'-complementizer?

12.2.2 Verbs that take an infinitival clause with an overt subject and without For.
Some verbs allow an infinitival clause with an embedded subject without the for-
complementizer:

(10) a. They want [Susan to stay in Washington].

b. They expect [the students to graduate next summer].
c. They would like [John to bring the beer].
d. They need [the teachers to teach more courses].

All the verbs in (10) take an infinitival sentence as a complement, they want/expect/would
like/need "something". That "something" that they need is what the infinitival clause
specifies. It is interesting to observe that if we replace Susan or the students in (11) by a
corresponding pronoun, we get the Objective form of the pronoun:
240 Héctor Campos & Bùi Huỳnh Thủy Thương

(11) a. They want [Susan to stay in Washington].

↓
a'. They want [her to stay in Washington].

b. They expect [the students to graduate next summer].

↓
b'. They expect [them to graduate next summer].

The question is where the Case for Susan in (11a) or the students in (11b) is coming from.
To understand these constructions better, let us compare these constructions with the
constructions with a PRO subject that we discussed above.

Consider the structures in (12):

(12) a. Theyi want [CP Φ[IP PROi to stay in Washington]]

b. They want [CP Φ[IP Susan to stay in Washington]]

Both sentences must satisfy the Case Condition because both sentences are grammatical!
Grammaticality derives from the fact that all Principles and Conditions are respected.

The Case Condition in (5a) applies to all DPs, whether they are pronounced or not.
Thus both PRO and Susan in (12) must satisfy the Case Condition. PRO and Susan cannot
receive Case from the infinitive because infinitives do not assign Case, only inflection that
contains both Tense and Agreement does. So how do PRO and Susan satisfy the Case
Condition in (12)?

We could claim that (12b) contains a silent pseudo-preposition for which assigns Case to
Susan. However, this claim makes the wrong prediction for (12a). We claimed above that
PRO has inherent Null Case. Thus PRO in (12a) has Case and satisfies the Case
Condition in (5a) automatically. A silent prepositional complementizer in (12), however,
would predict that (12a) is ungrammatical since PRO would be assigned a second Case. So
we conclude that there is no silent pseudo-preposition in (12). So how does Susan get
Case in (12b) if there is no no silent pseudo-preposition? The only possibility left is that
Susan actually gets Case from the verb want in (12b). However, if Susan can get Case from
the verb want in (12), why don't we say the same for PRO and avoid positing an inherent
Null Case for PRO? Thus we need to motivate the claim that PRO has inherent Null Case.

The evidence for inherent Null Case for PRO comes from the contrast observed in (13):

(13) a. [CP Φ[PRO to stay in Washington]] is a good idea.

b. *[CP Φ[Susan to stay in Washington]] is a good idea.
 English Syntax & Universal Grammar 241

Assuming that every DP must have Case to satisfy the Case Condition in (5a), PRO
satisfies it in (13a) since PRO has inherent Null Case. However, Susan in (13b) cannot
satisfy the Case Condition since the infinitive cannot assign Case, nor is there a
prepositional complementizer. Susan fails to receive Case in (13b), hence the
ungrammaticality of (13b). So, in subject infinitival clauses we are only allowed to have
PRO as a subject. The claim that PRO has inherent Null Case explains why we can have
PRO in (13a), but not Susan in (13b). Why can we have both PRO and Susan in (12),
repeated here?

(12) a. Theyi want [CP Φ[IP PROi to stay in Washington]]

b. They want [CP Φ[IP Susan to stay in Washington]]

The only difference between the structures in (13) and the structures in (12) is that the
infinitival clause appears next to the verb want. So Susan must be getting Case from the
higher verb want in (12b). However, now we face a new dilemma, if Susan can get Case
from want in (12b), how do we exempt PRO from receiving Case from want? We do not
want PRO to receive two Cases!

In order to avoid the paradox presented by the sentences in (12) and (13), we say that verbs
like want and those in (11) are Exceptional Case Marking verbs (ECM verbs). In what
sense are they exceptional? They are exceptional in that they allow the deletion of CP.
Allowing the deletion of CP, there is only an IP between the verb and the DP subject,
similar to what we observed with prepositional complementizers in (8b). Below we show
the D- as well as the S-Structure for sentence (12b), shown here as (14a). As shown in (14b)
we can "delete" the CP by scratching it out of the structure. Alternatively, we can just have
IP as a complement of want, as in (14b'):

(14) a. D-Structure: [CPΦ [IP They want [CP Φ [IP Susan to stay in Washington]]]]
b. S-Structure: [CPΦ [IP They want [CP Φ [IP Susan to stay in Washington]]]
b'. S-Structure: [CPΦ [IP They want [IP Susan to stay in Washington]]]

By deleting the CP in (14b, b'), there is only an IP between the verb want and Susan. In
(14b,b'), the verb is able to assign Objective Case to Susan even though Susan is not its
complement since IP is not a barrier for Case assignment, similar to what we saw in (8b).
When there is PRO subject, as in (12a), we cannot delete the CP since PRO would then
receive a second Case from the verb want, in addition to the inherent Null Case that it has.
So the structure for the infinitival clause without a pronounced subject is (15):

(15) D-Structure and S-Structure:

[CPΦ [IP They want [CPΦ [IP PRO to stay in Washington]]]]
/---------/ CP+IP = barrier for Case assignment
242 Héctor Campos & Bùi Huỳnh Thủy Thương

As we can see in (15), CP+IP together "protect" PRO from getting Case from the main
verb. When there is only an IP, as in (14b, b'), the verb is able to assign Case to the
embedded infinitival subject. We say that IP alone is not a barrier to Case assignment.
However, when IP has a CP on top of it, then it constitutes a barrier for Case assignment.
The notion of "barrier" is the topic of a more advanced syntax course, so we will just
stipulate that IP alone is not a barrier for Case assignment, but CP+IP constitutes a
barrier for Case assignment. Thus we need to make sure that whenever we have PRO as
the subject of an infinitive, we must have CP+IP to protect PRO from getting a second
Case. When we have an overt subject, as in (7), we need to make sure that one of these two
conditions applies:

(a) we have a prepositional complementizer FOR that is able to assign Case to the
subject of the infinitival clause, as in (7a), or

(b) the infinitival clause is the complement of an Exceptional Case Marking verb. An
exceptional Case Marking Verb allows for the CP of the embedded clause to be
deleted, after which the subject of the embedded clause is able to receive Objective
Case from the upper verb. An example is shown in (14b,b').

If we want to represent ECM structures in a tree structure, similar to what we did in (14b.
b'), we can indicate the "deletion" of CP by scratching it out, as in (16a), or we can just
generate IP as a complement of the ECM verb, as in (16b). (16) shows the way we represent
infinitival complements of Exceptional Case Marking verbs in general:1

(16) a. b.

1
Some ECM verbs are listed below:
FACTUAL = {announce, assume, believe, consider, declare, expect, feel, find, imagine, know,
presume, proclaim, pronounce, report, reckon, take, understand, ...}
CAUSE = {appoint, cause, elect, name, vote,...}
 English Syntax & Universal Grammar 243

To account for Exceptional Case Marking verbs our system now allows for DELETION.
We incorporate this deletion operation to our system, as shown in (17):

(17)

X'-Syntax Lexicon

General Principles: D-Structure

-Projection Principle Transformations: Insert, Move, Delete
-Extended Prejection Principle Principles: Stranded Affix Condition, Wh-Criterion,
-Theta-Criterion Case Condition
S-Structure

PF LF

As we can see in (17), our system now allows for three basic operations/transformations:
movement, insertion and deletion. (18) shows the elements that have undergone these
operations in our system so far:

(18) a. MOVEMENT:
a1. Head Movement: I to V, AUX to I, I to C, AUX to NEG, NEG to I
a2. XP-movement: Wh-movement
b. INSERTION: TO/FOR (indirect objects), OF (complements of N, A, P), do-
support
c. DELETION: CP-deletion for Exceptional Case Marking verbs

The three operations called TRANSFORMATIONS in our chart above are also referred to
as AFFECT ALPHA, where "affect" can be any of the three operations mentioned above
and "alpha" refers to any head (X) or phrase (XP).

PRACTICE 2

I. Diagram the following sentences:

a. The servant at John's house refused to open the door.
b. The professor always waits for the students to arrive.
244 Héctor Campos & Bùi Huỳnh Thủy Thương

c. The Chair of the Department has arranged for the faculty to meet with the textbook
company representative.
d. The teacher didn't allow the students to use their dictionaries during the test.
e. I would like our teacher to explain this chapter again.

II. For each of the sentences above, discuss how all the DPs are assigned Case.

RESEARCH PROJECT 2

Consider the following sentence in Vietnamese:

a. Anh ấy muốn tôi về nhà.
b. Anh ấy tưởng tôi về nhà.

Propose an analysis for the sentences (a) and (b). Discuss how each DP is assigned
Case. Can we argue that they have the same structure? To claim that they have the
same structure we need to prove that they behave exactly the same way under different
operations.

12.3 Infinitival clauses without an overt subject: Object-controlled PRO

In Section 1 we analyzed examples where the subject of the infinitive is coreferential with
the subject of the main clause, as shown in (19):

(19) Tuoii wants [CP Φ [IP PROi to go to Washington]].

In this section we will study constructions where PRO is coreferential with an object rather
than with the subject. Some examples can be found in (20):

(20) a. Quani told Tuoij [CP Φ [IP PROj to invite him to dinner]]
b. Tuoii advised Haj [CP Φ [IP PROj to look for another boyfriend]]

As we can see in (20), PRO takes Tuoi in (20a) and Ha in (20b) as its antecedent. This is
shown by having PRO and its coreferents take the same subindex, "j" in this case. The
structure for (20a) is shown in (21):

(21) a. Quan told Tuoi to invite him to dinner

 English Syntax & Universal Grammar 245

b.

As we can see in (21b), the verb tell takes two complements: an indirect object DP (Tuoi)
and a sentential complement (CP) since we must tell someone something. The indirect
object Tuoi is coreferential with the subject of the infinitival clause. We indicate this by
having the two DPs share the same index, "j" in this case, since "i" is the index assigned to
Quan. Just as in the infinitivals discussed in Section 1, this infinitival CP takes a zero-
complementizer which is the realization of [-Q,-WH]. [-TNS,-AGR] is spelled out as "to".
As we mentioned in the previous section, PRO must always be protected by IP + CP. This
is also what we observe in (21b).

While the structure in (21) is easy to understand, there is a problem in identifying whether a
verb takes the structure in (21) or the structure in (16):

(22) a. Quan told Tuoi to invite him to dinner.

b. Quan wants Tuoi to invite him to dinner.

In (22), both constructions look exactly the same, with the exception of the main verb. As
we have seen above, (22b) is an Exceptional Case Marking construction, as shown in
246 Héctor Campos & Bùi Huỳnh Thủy Thương

(23b), while the construction in (22a) contains an object-controlled PRO, as shown in

(23a):

(23) a. [CP Φ [IP Quani told Tuoij [CP Φ [IP PROj to invite him to dinner]]]]
b. [CP Φ [IP Quan wants [IP Tuoi to invite him to dinner]]]

In both examples in (23), Tuoi is the one who may invite Quan to dinner, but Tuoi is the
subject of the infinitive only in (23b). In (23a) we see that Tuoi is a complement of tell (see
structure 21) and the infinitival clause contains a PRO coreferential with Tuoi, thus Tuoi
comes to be interpreted as the subject of the infinitive. In (23b), on the other hand, there is
no CP (remember that Exceptional Case Marking verbs delete CP or take IP directly as
their complement) nor is there a PRO. Tuoi is the subject of the infinitival clause. So in
spite of the fact that the sequence DP + infinitive looks similar in (22), the structures are
very different, as we see in (23).

Other verbs that take object-controlled PRO, as in (23a) are: advise, allow, ask, beg,
direct, encourage, exhort, forbid, force, implore, instruct, invite, order, persuade,
recommend, remind, request, require, teach, urge, warn, etc.

So how can we distinguish the two classes of verbs? There are some tests that we can
perform to distinguish these two classes of verbs. However, it must be observed that the
tests do not always work.

Test 1: The "something" test

As you can see in (23), tell takes two complements (DP – CP) while want takes only one
(IP). Thus we are able to replace the IP by "something", but we are not able to replace (DP
CP) by "something" since we need two complements to satisfy the Projection Principle:

(24) a. *Quan told "something" (Object-controlled PRO)

b. Quan wants "something" (ECM)

Test 2: The "that" test

Consider (25):

(25) a. Does Quan want Tuoi to invite him to dinner? (ECM verb)
b. Yes, he wants "that".

c. Did Quan tell Tuoi to invite him to dinner? (Object-controlled PRO)

d. *Yes, he told "that"
 English Syntax & Universal Grammar 247

As we can see in (25), "that", as "something" in (24), is able to replace only one
complement. Thus it will only be accepted as an answer with ECM verbs, but not with
object-controlled verbs since these verbs require two complements.

Test 3: Tensing the CP

Sometimes we are able to tense the embedded CP. This is not possible with ECM verbs
since they do not contain CP:

(26) a. Quan told [DP Tuoi] [CP PRO to invite him to dinner] (object-controlled PRO)
b. Quan told [DP Tuoi] [CP that she should invite him to dinner]

c. Quan wants [IP Tuoi to invite him to dinner]. (ECM verb)

d. *Quan wants [that Tuoi invites him to dinner].

However, notice that if you try this test with allow, invite, etc. the test does not work.

Test 4: Adverb modification

An adverb may appear after the DP preceding the infinitive when we have an object-
controlled verb, but it may not appear with ECM verbs, as we can see in (26);

(27) a. Quan has told Tuoi repeatedly to invite him to dinner. (object-controlled PRO)
b. *Quan wants Tuoi sincerely to invite him to dinner. (ECM verb)

These tests prove that ECM verbs and DP + infinitive constructions like those in (22) have
very different structures. We will assume the structure in (21b) to be the correct structure
for object- controlled PRO.

PRACTICE 3

Diagram the following sentences:

a. Do the linguistics students expect to finish their work on time?
b. Will the teacher expect his students to finish their work on time?
c. They didn't actually promise to arrive on time.
d. They promised the teacher to arrive on time.
e. That teacher hasn't allowed the students to come to class late.
f. Parents always remind their children to call them on the phone.
g. My sister who lives in Australia convinced my parents to visit her.
h. My neighbors advised me to lock my door at night.
248 Héctor Campos & Bùi Huỳnh Thủy Thương

RESEARCH PROJECT 2

Propose a structure for the following sentences in Vietnamese. Motivate your

structures.
a. Tôi mong gặp Hà ngày mai.
b. Tôi mong bạn gặp Hà ngày mai.
c. Tôi sẽ giúp bạn gặp Hà ngày mai.
d. Tôi cấm Quân gặp Hà ngày mai.

12.4 Arbitrary PRO

In sections 2 and 3 we discussed a PRO that may be controlled by the subject or by the
object. We call this Control PRO because it needs an antecedent to be interpreted. An
example for each kind is shown in (28):

(28) Control PRO

a. Hai wants [CP Φ [IP PROi to bake a cake for Quan]]
b. Quani told Haj [CP [IP PROj to bake a cake for him]]

There is a second kind of PRO that is not controlled. It is called Arbitrary PRO. This PRO
has a generic reference. It may refer to you, me, the people, everyone, etc. Examples are
given in (29):

(29) Arbitrary PRO

a. When we study syntax, it is important [CP Φ [IP PROto do the exercises]].
b. It's necessary [CP Φ[IP PRO to come to class on time]].

As we can see in (29), we understand the referent for PRO to be we, the students, etc.
Control PRO (examples in 28) bears the subindex of its antecedent, arbitrary PRO bears no
subindex. This is why it is important to include the subindex in our structure when we
have a control PRO. Subject or Object Control is a lexical property of verbs. As such, it
must be indicated as such in the lexicon.

12.5 A brief look at Small Clauses

In the previous sections we learned how to merge different types of infinitival clauses. In
general, we identify these infinitival clauses by the simple fact that the embedded verb is
introduced by TO, which as we claimed, is the realization of [-TNS, -AGR].
 English Syntax & Universal Grammar 249

There are other tenseless constructions which, in spite of being tenseless, appear without
TO:

(30) a. I saw [Tuoi smile].

b. I heard [Khiem laugh].
c. I made/had [Quan make icecream].
d. I don't let [my students use their phones in class].

It should be clear from the examples in (30) that the embedded clauses do not contain
tensed verbs since we lack subject-verb agreement in (30). We cannot tense the embedded
verb either. So, for instance, we cannot say I saw Tuoi smiled. The question is whether the
clauses in (30) are infinitival clauses where [-TNS, -AGR] has not been pronounced or
whether they have an entirely different structure.

It is also clear that the structures in brackets in (30) are constituents since they behave like
units:

(31) a. [Tuoi smile] is what I saw.

b. I saw [Tuoi smile] and [Quan cry].
c. Did you see [Tuoi smile]? Yes, I saw "that".

To understand these constructions, we need to look at other constructions that show a

similar structure. We will call these constructions "small clauses".

Consider the following examples. We will consider these structures Small Clauses (SC), as
they are indeed clauses, but they do not contain a verb:

(32) a. I consider [SC Quan [AP very smart]].

b. I want [SC Khiem [PP out of my life]].
c. They appointed [SC Ha (as) [DP class monitor]].
d. I saw [SC Tuoi [VP write the report for her class]].

We can prove that the small clauses above are constituents since they can appear as
sentence fragments (33a), in coordination structures (33b), and as shared constituents (33c).

(33) a. -The teacher considers [Quan very smart].

-[Quan smart]? But he is always playing games on his telephone!
b. The teacher considers [Quan smart] and [Ha hard-working].
c. The teacher considers and I also consider [Quan smart].

There has been much debate in generative syntax since the 80's regarding the structure of
these constructions. Here we will analyze them as predicate structures, as in (34):
250 Héctor Campos & Bùi Huỳnh Thủy Thương

(34) a. I consider Quan/him very smart.

b.

In (34b), Quan receives Objective Case from the verb consider. This is why we can replace
Quan by the objective pronoun him in (34a).

As we can see in (30d), repeated here as (35), the Predicate head can sometimes be
pronounced:

(35) a. They appointed Ha as class monitor.

b.

We will extend this analysis to constructions that include a simple infinitive (an infinitive
without TO):

(36) a. I saw Tuoi write the report for her class.

 English Syntax & Universal Grammar 251

b.

We will thus analyze simple infinitive clauses as PredP rather than as infinitives with a
deleted TO. This explains why these constructions cannot be tensed, negated, etc.

Perhaps you may wonder why we do not analyze structure (37b) as an infinitival clause
with the verb TO BE and then delete TO BE, as shown in (37a) below.

(37) a. The doctor found the child to be sick.

b. The doctor found the child sick.

We could mark to be deletion as a special property for those verbs that take small clauses.
Although this is a possible analysis, there are two arguments against this claim.

The first argument is that we would lose the parallelism with the constructions that we
found in (32). We saw that small clauses (PredP) can actually be found across
categories, not just with adjective phrases. A structure like (35b), which includes PredP,
unifies all these constructions. Furthermore, not all of the constructions in (32) allow the
copulative verb to be.

While the first argument is theory internal (having such an analysis makes our theory
"more elegant"), the second argument is empirical. Consider the two examples in (38).

(38) a. The doctor found the children to be sick.

b. The doctor found the children sick.

If (38b) were to be derived from (38a) by deleting the verb TO BE, we would not expect
meaning to change. However, (38a) and (38b) do not have the same meaning. (38a) implies
that the doctor actually examined the children and then diagnosed them as sick. In (38b) the
doctor just found the children in a particular condition, namely sick. Constructions like (38)
prove that small clauses like the one in (38b) do not derive from an infinitival clause after
deleting the copular verb to be.
252 Héctor Campos & Bùi Huỳnh Thủy Thương

PRACTICE 4

Diagram the following sentences. Indicate all the moved, inserted or deleted elements.
a. The professor wants Quan to format the article he wrote for that journal.
b. Did you hear who the new President will appoint as his Secretary of Defense?
c. Tuoi wants Quan to let her cook dinner for her boyfriend from Cambodia.
d. Tuoi told Quan to help her cook dinner for her boyfriend.
e. Khiem said that he had never seen Ha cry like that.

CHALLENGE

Diagram the following sentences. Indicate all the moved, inserted or deleted elements.
a. Does Quan's sister's friend really want to publish an article on current issues on
morphological theory?
b. Isn't it true that Sasha expects Stella to ask the students from the English Department
to buy the famous Georgetown cupcakes for our last class?
c. Which linguists believe language acquisition to be a very important subject of study?
d. Do you not think the party's chairman's speeches in undecided districts will help the
Democratic candidates for the House win the November election?
e. I wonder if she hadn't heard the rumor that the Chair of the Linguistics department
gave the graduate students a copy of Chomsky's latest book on small clauses.
f. For school officials to have sent the graduate students this letter by certified mail
very early yesterday shows that the President of the University considers the issue
of plagiarism of utmost importance.

12.6 Summary

In this chapter we considered four different structures with tenseless or non-finite verbs,
i.e. verbs that do not contain tense:

(39) a. [The students]i want [CP Φ[IP PROi to study advanced syntax]].
b. The teacher wants [CP Φ [IP the students to study advanced syntax].
b'. The teacher wants [IP the students to study advanced syntax].
c. The teacher told [the students]i [CP Φ[IP PROi to study advanced syntax]]
d. The teacher made [PredP Khiem study advanced syntax].
 English Syntax & Universal Grammar 253

Structures (39a-c) are infinitival clauses, structure (39d) is a simple infinitive clause.
Structures like (39a) include a subject-controlled PRO. We can see that these
constructions include a CP complement whose complementizer is realized as Φ. The
example in (39b,b') contains an Exceptional Case Marking (ECM) verb. Exceptional
Case Marking verbs like want delete the CP of the embedded clause and are thus able to
assign Case to the subject of the infinitive. We can use either the representation in (39b) or
in (39b') for Exceptional Case Marking verbs. The sentence in (39c) is an example of an
object-controlled PRO. In contrast with (39b,b'), both subject- as well as object-
controlled verbs take a CP complement. The subject of the embedded infinitival clause in
(39c) is controlled or coreferent with the object of the main clause. Finally we considered
structures like (39d), where the complement is a clause that contains a simple infinitive (an
infinitive without TO). We argued that this construction constitutes a Predicate Phrase and
is parallel to similar constructions found in other categories, as shown in (32), repeated here
as (40):

(40) a. I consider [PredP Quan [AP very smart]].

b. I want [PredP Khiem [PP out of my life]].
c. They appointed [PredP Ha (as) [DP class monitor]].
d. I saw [PredP Tuoi [VP write the report for her class]].

The appearance of PRO in (39a,c) is forced by the Extended Projection Principle (EPP),
repeated here as (41):

(41) Extended Projection Principle

Every sentence must have a subject.

We also assumed that PRO had inherent Null Case. This inherent Case guaranteed that
PRO could satisfy the Case Condition in (42):

(42) Case Condition

Every DP (silent or non-silent) must have Case.

Because PRO bears inherent Case it must be blocked from receiving Case from an upper
element. This is why whenever we find PRO in (38), we always find a CP + IP dominating
that PRO. We said that CP+IP constituted a barrier or blocked Case assignment from a
higher verb.
254 Héctor Campos & Bùi Huỳnh Thủy Thương

In ECM constructions (39b.b') as well as in Small Clause constructions (39d), the subject of
the embedded clause is able to receive Case from the higher verb. This follows assuming
that IP as well as PredP are not barriers to Case assignment. As we mentioned above,
barriers is a topic we leave for a more advanced course in syntax.
 English Syntax & Universal Grammar 255

CHAPTER 13

DP-MOVEMENT

In the previous chapters we considered two kinds of movement: head movement and
phrasal movement. Head movement occurs when a head moves to another (the closest!)
head. Examples of head movement in English are the movements of I to V, AUX to I, NEG
to I, and I to C. We also saw one type of phrasal movement. In phrasal movement, the
whole phrase, rather than just the head, moves. Wh-Movement is an instance of phrasal
movement since the whole phrase moves to the specifier of CP.

In this chapter we will study a second type of phrasal movement. We will consider
structures where the DP moves to the specifier position of IP in search for Case. Since this
operation is restricted to DPs, it is called DP-Movement. We will consider the following
structures:

(1) a. Quan is smart. (copulative verb)

b. Ha seems to enjoy syntax. (subject raising verb)
c. Khiem was invited to the party. (passive construction)
d. There are many students in this class. (there-construction)

We will end this chapter by revisiting the original merge position of subjects.

13.1 Copulative verbs

In the previous chapter we argued for the existence of small clauses or predicate phrases.
We proposed (2b) as the structure for (2a):

(2) a. They consider Tuoi smart.

b. They consider [PredP Tuoi/her[APsmart]]

A condition that will be central to our discussion in this chapter is the Case Condition in (3):

(3) The Case Condition

Every DP must receive Case.
256 Héctor Campos & Bùi Huỳnh Thủy Thương

In the structure in (2b), Tuoi satisfies the Case Condition in (3) by receiving Case from the
main verb consider. We can prove that Tuoi receives Objective Case by replacing it by the
corresponding pronoun.

We are now ready to understand the syntax of copulative structures in English. Although
this was probably the first structure you studied when you were learning English, you had to
go through 12 chapters of this book to understand their syntax! Consider the example in (4).
Since we have an adjective and a DP that is predicated on that adjective, we can assume a
small clause or PredP, as we argued in the previous chapter:

(4) a. Tuoi is smart.

b.

As we can see in (4b), the verb to be (as well as all copulative verbs) takes a PredP as its
complement. We analyze copulative be as a main verb, not an auxiliary verb. As such,
it is generated as a verb under VP, not under AUXP. However, similar to auxiliary verbs, it
does have the ability to move to I. Evidence that be moves to I in (4b) comes from the fact
that in questions, we do not get do-support, rather the verb be itself moves to C. Notice that
if there is a negative element like –n't on the way to I and C, the verb carries it along, as
shown in (5b):

(5) a. Is Tuoi smart?

b. Isn't Tuoi smart?

Why does Tuoi move to the specifier of IP in (4b)? Because Tuoi needs to get Case to
satisfy the Case Condition in (3); it cannot get Case in the specifier position of PredP.
Remember that only (transitive) verbs, prepositions, and [+TNS,+AGR] have the
 English Syntax & Universal Grammar 257

ability to assign Case. The specifier of IP is empty and thus available in (4). Therefore,
Tuoi moves to the specifier of the tensed IP to satisfy the Case Condition, which requires
that every DP have Case.1 We call this kind of operation DP-movement.

Other copulative verbs are: to seem, to appear, to look, to sound, to smell, to taste, to feel, to
become, to get, etc.

PRACTICE 1

Diagram the following structures.

a. The students at this university are very smart.
b. Why aren't the literature students happy with the exam results?
c. Shouldn't they be proud of their results?
d. This pho tastes delicious.
e. You shouldn't be feeling so tired.

13.2 The verb to seem

As we saw above, to seem is also a copulative verb. Consider the structure in (6):
(6) a. Ha seems happy.
b.

1
You may wonder how we get the correct agreement in (4b) if Tuoi is not in the specifier of IP. In this
model, it is assumed that we can generate any features we want. If the agreement features of I and
those of the DP in the specifier do not match, then the sentence becomes ungrammatical. This is what
happens if we say *The boys is happy. If the features match, then we will get a well-formed sentence,
as in The boy is happy. We will tackle this problem more in depth in section 4.
258 Héctor Campos & Bùi Huỳnh Thủy Thương

As a raising verb, the verb to seem also takes a PredP as a complement. Different from the
verb to be in (4b), however, I lowers to V to satisfy the Stranded Affix Condition. As we
saw above, Ha moves to the specifier of IP because in that position it will be able to receive
Nominative Case from I, which contains tense (PRES) and agreement (IIISG).

A natural question to ask at this point would be: Why doesn't Ha receive Case from the verb
seem in (6b)? There are two reasons. The first reason is that if Ha were to receive Case from
seem, it would receive Objective Case (remember that verbs assign Objective Case).
However, Ha receives Nominative Case since we say She seems happy rather than *Her
seems happy. The second reason has to do with a universal condition regarding Case
assignment. This Condition is known as Burzio's Generalization. Burzio was the linguist
who discovered this generalization in 1981, and ever since the generalization has carried his
name.

(7) Burzio's Generalization

A verb assigns Objective Case if and only if it has an external subject.

As we can see in (6b) (as well as in 4b), copulative verbs do not have their own subject.
It is always the case that a DP moves into the IP-specifier of a copulative verb. Thus we
say that copulative verbs have derived subjects because at D-Structure these
constructions do not have a subject in the specifier of IP, but rather, the subject moves to
that position.

Summarizing the syntax of copulative verbs we see:

(a) Copulative verbs take a PredP as a complement.

(b) Copulative verbs do not have a D-Structure subject; the specifier of PredP moves
to the specifier of their IP to become the subject of the copulative structure.
(c) The copulative verb is not able to assign Objective Case to the specifier of the
copulative structure because it lacks an external subject.
(d) The DP in the specifier position of the PredP moves to the specifier of IP of the
verb in order to receive Case from inflection, which bears [+TNS, +AGR] and
thus satisfies the Case Condition.

We saw in (6) that the verb seem does not have its own subject; it has a derived subject.
We must preserve this property when the verb seem is followed by an infinitival clause, as
in (8):

(8) a. Quan seems to love Ha.

 English Syntax & Universal Grammar 259

b.

In (8) the verb seem takes a CP complement (an infinitival clause). To satisfy the Stranded
Affix Condition, I moves to V in the main clause. In the embedded clause no such
movement is necessary since the I, which contains the features [-TNS,-AGR], is spelled out
as "to". Quan in (8b) cannot receive Case from the lower I since it does not contain tense or
agreement. Nor can it receive Case from the main verb seem since this verb, in accordance
with Burzio's Generalization in (7), lacks an external subject. However, the specifier
position of IP in the main clause is empty and available, and Nominative Case can be
assigned to that position since I contains tense [PRES] and agreement [IIISG].

We saw in the previous chapter that CP+IP constitutes a barrier for Case assignment.
CP+IP also constitutes a barrier for movement. Thus, if we want to move Quan to the
specifier of the higher IP, we must delete the intermediate CP, an operation that we already
applied to Exceptional Case Marking (ECM) verbs in the previous chapter. By deleting this
CP, we ensure that there will be no barrier for Quan to move to the higher IP. Deleting the
CP is our last resort to save the structure in (8) from crashing or not converging (i.e., not
getting a good grammatical sentence). As with ECM verbs we can assume that this CP is
deleted (as in 9b below) or that seem takes an IP, rather than a CP, as a complement, as
shown in (9c). In either case, (9c) would be the final structure if the CP is deleted:
260 Héctor Campos & Bùi Huỳnh Thủy Thương

(9) a. Quan seems to love Ha.

b. S-Structure with CP-deletion:

[CPΦ[IP Quani seems [ CP Φ [IP ti to love Ha]]]

c. S-Structure with IP as a complement:

 English Syntax & Universal Grammar 261

[CPΦ[IP Quani seems [IP ti to love Ha]]]

To represent the S-Structure of these constructions using bracket structures, we use traces
(t) to indicate the positions from which we move or we have moved through. These traces
are coindexed with the original element that moved since they are coreferential with
their antecedents. We call verbs like to seem (subject) raising verbs because the subject of
a raising verb raises from an embedded position, thus raising verbs always have a derived
subject.

Consider now the example in (10):

(10) Quan seems to be smart.

In (10) we have two raising verbs: to seem and to be. Thus we need to make sure that both
verbs start with an empty subject at D-Structure, as shown in (11):

(11) D-Structure:
262 Héctor Campos & Bùi Huỳnh Thủy Thương

a.

b. [CPΦ[IP Δ–s seem [CP [IP Δ to be [PredP Quan Φ smart]]]]]

As we can see in (11), neither seem nor be has an external subject. Quan is not able to
receive Case in the specifier of PredP because the head Pred is not a Case assigner. The
verb to be cannot assign Case to it either because to be does not have its own external
subject either. Notice that Quan would be able to receive Case at the specifier position of
the higher IP, but to move there, just as long-distance wh-movement, it must move first to
the lower IP and then to the higher IP. Let us take one step at a time.

(11) c. Quan seems to be smart.

 English Syntax & Universal Grammar 263

S-Structure (Step 1):

d. [CP [IP Δ–s seem [CP [IP Quani to be [PredP ti Φ smart]]]]]]

In (12), Quan has moved from the specifier position of PredP to the specifier position of the
lower IP. By moving there, Quan helps the lower IP satisfy the Extended Projection
Principle by providingthe lower IP with a subject. However,Quan cannot get Case in that
position either because the lower I with the features [-TNS,-AGR] is not able to assign
Case. Nor can the main verb seem assign Case because it lacks an external subject.
Therefore, Quan will keep on moving to the specifier of the higher IP, which will be able to
assign it Case because it has both tense and agreement. To go from the lower specifier of IP
to the higher specifier of IP, there is a CP+IP barrier that would block the movement.
Raising verbs are able to delete the embedded CP (or we can say that they take an IP
complement), and thus Quan is able to move on.

(13) a. Quan seems to be smart.

264 Héctor Campos & Bùi Huỳnh Thủy Thương

S-Structure: Step 2

b. [CP [IP Quani seems [IP ti to be [PredP tiΦsmart]]]]

In the specifier position of the higher IP (as subject of the verb seem), Quan is now able to
get Nominative Case from I [PRES, IIISG] and thus finally satisfies the Case Condition.

As you can see from our previous discussion, both movements in (13) are motivated by
Quan's need to satisfy the Case Condition. In moving to the higher available IP-specifiers
Quan also helps those IPs satisfy the Extended Projection Principle, which requires every
IP to have a subject. Notice that this condition may be satisfied either with a pronounced DP
(as in the higher IP), or with a trace (as in the lower IP).

PRACTICE 2

Diagram the following sentences:

a. Our syntax teacher seems to like Italian icecream.
 English Syntax & Universal Grammar 265

b. The students seem to be happy with their results on the exam.

c. Tuoi seems to like to go to the beach with her friends.
d. Does your syntax teacher seem to like to give quizzes on Friday?

13.3 Passive Constructions

In this section we will analyze passive constructions. Consider the examples in (14). (14a)
shows the construction in active voice and (14b) the corresponding construction in passive
voice.

(14) a. Tuoi visited Quan. (active)

b. Quan was visited by Tuoi. (passive)

Comparing the active with the passive construction in (14) we can see 4 main differences:
(a) The direct object of the active construction has become the subject of the passive
construction.
(b) The subject of the active construction has become the object of the preposition by.
(c) There is an additional passive auxiliary (to be), which takes a (passive) past
participle as a complement.
(d) The theta-roles of the DPs have not changed. In both sentences, Quan is the AGENT
and Tuoi is the PATIENT.

We will assume that passive morphology has one important effect in our structure: passive
morphology absorbs the external subject (or argument) of the verb. This argument can
optionally be realized with an adjunct by- phrase as in (14b). We will thus assume that (15)
is the underlying structure (or D-Structure) for passive constructions:

(15) a. Quan was visited by Tuoi.

266 Héctor Campos & Bùi Huỳnh Thủy Thương

b. D-Structure:

c. [CP [IPΔ–ed be [VP visited Quan by Tuoi]]]

As we can see in (15b), the passive auxiliary be (which is merged as AUXP) has
absorbed the external subject, leaving the position empty. The external subject is realized
as an adjunct in the VP. As an adjunct, it is optional. Let us now make sure that all DPs
satisfy the Case Condition in (15b). Tuoi is assigned Oblique Case by the preposition by.
Quan, however, is not able to receive Case in (15b) because the verb visit, due to the
passive morphology, has lost its external argument. Thus the verb visit is no longer able to
assign Case to Quan, as it did in the active construction. Nonetheless, the specifier of the IP
is available and I has both tense and agreement, so Quan moves there to receive Case. In
(15b) we also have –ed, a stranded affix which needs to satisfy the Stranded Affix
Condition. All conditionsare thus satisfied at S-Structure:

(16) a. Quan was visited by Tuoi.

 English Syntax & Universal Grammar 267

b. S-Structure:

c. [CP [IP Quani was [VP visited ti by Tuoi]]]

As we can see in (16b,c), the DP Quan moves to the specifier of IP in search of Case. In this
new position, it will be assigned Case by I, which contains tense [PAST] and agreement
[IIISG]. At the same time, the passive auxiliary be rises to I to satisfy the Stranded Affix
Condition.

PRACTICE 3

Diagram the following structures:

a. That house in the woods was built by a famous architect.
b. That class has been being taught for 3 years now.
c. That painting seems to have been done by Picasso.
d. Wasn't that school founded by French priests?
e. When will our building be painted again?

13.4 There-constructions

Let us consider now constructions with ‘there’:

(17) a. There is one student in this class.

b. There are many students in this class.
268 Héctor Campos & Bùi Huỳnh Thủy Thương

These constructions with there + be are also called presentational or existential

constructions.

We observe two curious facts about there-constructions:

(a) The grammatical subject is the pronoun there. This is the element that appears in
the canonical subject position.

(b) The verb agrees with the underlined DPs in (17).

In section 1 we analyzed copulative be as a verb with a derived subject. We will maintain

this analysis here and assume that the verb be in (17) takes a PredP complement:

(18) a. There is one student in this class.

b.

b. [CPΦ[IP Δ –s be [PredP [DP one student] Pred [PP in this class]]]]

In (18b) copulative be takes PredP as a complement. The DP one student fails to receive
Case in the specifier position of PredP because Pred is unable to assign Case and because
the verb be cannot assign Case since it lacks a subject. The sentence in (18b) is in violation
of the Extended Projection Principle as well, since it lacks a subject. We have two options
at this point, as we can see in (19): (a) either the DP one student raises to the specifier of IP,
or (b) we insert there to make the structure comply with the Extended Projection Principle,
which requires every sentence to have a subject.
 English Syntax & Universal Grammar 269

(19) a. One student is in this class.

b. There is one student in this class.

Let us consider each of these options in turn. First (19a), which is parallel to what we did
with copulative structures in section 1. This structure, however, is far less common than
structure (19b):

(20) a. One student is in this class.

b.

c. [CPΦ[IP [DP one student]i is [PredP tiPred [PP in this class]]]]

When the DP moves to the specifier of IP, it satisfies two conditions for the sentence: it
provides IP with a subject, thus satisfying the Extended Projection Principle, and one
student is able to receive Case since I contains tense [PRES] and agreement [IIISG].

Consider now (19b), repeated here as (21):

(21) a. There is one student in this class.

270 Héctor Campos & Bùi Huỳnh Thủy Thương

b.

c. [CPΦ[IP there is [PredP [DP one student] Pred [PP in this class]]]]

In (21b,c) we have inserted THERE in the specifier of IP to satisfy the Extended

Projection Principle, which requires every sentence to have a subject. We call a pronoun
like THERE an expletive pronoun. Expletive pronouns are the result of last resort
insertions. They are inserted only for the purpose of saving the structure in (21b,c) from
crashing by providing it with the subject it needs to satisfy the Extended Projection
Principle.

Up until this chapter we have generated agreement freely in I to matchthe features of the
DP in the specifier of IP and the features on the head of I. Agreement checking/
matching occurs between the head I and the element in the specifier of IP. The features
PERSON (PN) and NUMBER (NR) of the DP and those on I must "agree" or "match" for
the sentence to converge; otherwise, the structure crashes. Although this mechanism works
for all the examples we have discussed in previous chapters, where we generate subjects in
the specifier of IP, it does not work for examples like (21), which have an expletive
pronoun there in the IP specifier. Structures like (21) raise two questions:

(i) What allows the DP one student to agree with the verb in (21)? Notice that one
student is not in the specifier of IP!
 English Syntax & Universal Grammar 271

(ii) How does the DP one student, sitting in the specifier of VP, get Case in order to
satisfy the Case Condition?

To answer these questions, we will assume that Inflection (I) bears a valued feature for
TENSE (PRESENT or PAST) but has unvalued features for PERSON (PN) and
NUMBER (NR). On the other hand, the DP one student has features for PERSON (III)
and NUMBER (SG), but it has an unvalued feature for Case. We indicate unvalued
features with the sign "?". The structure is shown in (22):

(22) [CPΦ[IP Δ INFL be [PredP [DP one student] Pred [PP in this class]]]]
[TNS: PRES]
[PN: ?] [III, SG]
[NR: ?] [Case:?]

Inflection will "probe" down the structure looking for a "target", i.e., a DP with the
features it needs: PERSON and NUMBER. Once it finds those featureson the first DP it
encounters, it will value (that is, copy) those features, as shown in (23a,b). The whole
mechanism is shown in (23), where features are valued and Case is assigned:

(23) a. Inflection "probes" for features to fill its unvalued features:

[CPΦ[IP Δ NFL be [PredP [DP one student] Pred [PP in this class]]]]
[TNS: PRES]
[PN: ?] [III, SG]
[NR: ?] [Case: ?]

b. Inflection finds the necessary features in the closest DP and values those
features:
[CPΦ[IP Δ is [PredP [DP one student] Pred [PP in this class]]]]
[TNS: PRES]
[PN: III] [III, SG]
[NR: SG] [Case: ?]

We see that Inflection in (23b) now has acquired the features of the DP one student in the
specifier of PredP and can now spell-out the respective agreement on the verb. This
generates the form "is". If a DP is able to value the unvalued features of I, it gets
NOMINATIVE CASE from Inflection in return. Thus the DP in the specifier of PredP
now is marked with Nominative Case:

c. Inflection assigns NOMINATIVE Case to the DP that valued its unvalued

features:
272 Héctor Campos & Bùi Huỳnh Thủy Thương

[CPΦ[IP is [PredP [DP one student] Pred [PP in this class]]]]

[TNS: PRES]
[PN: III] [III, SG]
[NR: SG] [Case: NOMINATIVE]

In order to satisfy the Extended Projection Principle, we insert there in the specifier of IP
and our derivation is complete.

d. Insert expletive THERE in subject position to satisfy Extended Projection

Principle:
[CPΦ[IP THERE is [PredP [DP one student] Pred [PP in this class]]]]

TNS: PRES]
[PN: III] [III, SG]
[NR: SG] [Case: NOMINATIVE]

This probing system (also called AGREE) provides us with a new way to implement
agreement and assign Nominative Case to DPs that agree with the verb, but which are
not in the specifier of IP. The question now is whether this system can be generalized to all
cases of subject-verb agreement. We will pursue this hypothesis in the next section.

PRACTICE 4

Diagram the following sentences. Briefly explain how agreement is checked.

a. There is a problem in this argument.
b. There are several problems in this argument.
c. There seems to be a problem in this argument.
d. There seemto be several problems in this argument.

13.5 Where is the subject of a sentence merged?

We saw in section 2 that raising verbs as well as passive constructions have a derived
subject, that is, these constructions do not have a subject at D-Structure. A subject moves to
that position in search of Case.

As we saw in the appendix to Chapter 8, there are some interesting data suggesting that
actually all subjects, and not just subjects of copulative verbs, may be derived subjects. We
review those facts here and extend the analysis of the previous sentence to those cases.

Consider the following data with quantifier float.

 English Syntax & Universal Grammar 273

(24) a. The students will all pass this course.

b. My friends have all seen "Angry Birds".

It is clear that the quantifier all in (24) refers back to the students and my friends,
respectively. We can see this connection better in languages which show agreement, where
the quantifier all will agree in gender and number with the subject. Take a language like
Spanish:

(25) a. Mis amigos están todos estudiando español.

My friends-masc are all-masc studying Spanish
b. Mis amigas están todas estudiando español.
My friends-fem are all-fem studying Spanish

We can see in (25) that the quantifier and the noun agree in gender (masculine) and number
(plural). It is interesting to observe that we can also say:

(26) a. [All my friends] are studying Spanish.

b. [Todos mis amigos] están estudiando español.
c. [Todas mis amigas] están estudiando español.

Thus it seems that [all my friends] starts as a unit, and then the DP moves leaving the
quantifier behind. We will assume the following structure for (24b):
(27) a. My friends will all pass this course.
b.
274 Héctor Campos & Bùi Huỳnh Thủy Thương

In (27), the DP my friends moves to the specifier of IP in order to get Case, leaving the
quantifier behind. This phenomenon is called quantifier float because the quantifier is left
behind "floating" in the specifier of the VP.

Structures like (27) show that the external argument of a verb (i.e., the subject of a
sentence) is merged in the specifier position of VP or we could not explain the structures
containing a floating quantifier. Thus we extend the structure in (27) to all verbs. Consider
the example in (28). We will now apply the system that we introduced in the previous
section:

(28) a. Ha bakes delicious cakes.

b.

(29) a. D-Structure: The subject of the sentence starts as the specifier of VP.
Inflection (INFL) is specified only for TNS. The Person (PN) and Number
(NR) features are unvalued. The DP starts with PN and NR features specified,
but it has an unvalued Case feature:
[CPΦ[IP Δ INFL [VP [DP Ha] bake delicious cakes]]]
[TNS: PRES] [III,SG]
[PN: ?] [Case:?]
[NR: ?]

b. S-Structure-1: INFL probes down the structure looking for PN and NR

features. It finds the closest DP in the specifier of VP. Ha has the features that
INFL needs to be valued:
[CPΦ[IP Δ INFL [VP [DP Ha] bake delicious cakes]]]
[TNS: PRES] [III,SG]
[PN: ? ] [Case: ?]
[NR: ?]
 English Syntax & Universal Grammar 275

c. S-Structure-2: Features are copied to value PN and NR in INFL:

[CPΦ[IP Δ INFL [VP [DP Ha] bake delicious cakes]]]
[TNS: PRES] [III,SG]
[PN: III ] [Case: ?]
[NR: SG]

d. S-Structure 3: Ha receives Nominative Case for valuing I:

[CPΦ[IP Δ INFL [VP [DP Ha] bake delicious cakes]]]
[TNS: PRES] [III,SG]
[PN: III ] [Case: NOMINATIVE]
[NR: SG]

e. S-Structure-4: Features of INFL [PRES, IIISG] are spelled out as –s; Ha is

realized as a Nominative DP:
[CPΦ[IP Δ -s [VP [DP Ha] bake delicious cakes]]]

f. S-Structure-4: Inflection –s lowers to the verb and Ha raises to the specifier of

IP to satisfy the Extended Projection Principle:
[CPΦ[IP Hai [VP ti bakes delicious cakes]]]

We thus see that the AGREE (or probing) system that we proposed for there-constructions
in section 3 can be easily applied to regular subject agreement cases. We now thus have a
unified agreement system for constructions with the expletive pronoun there and for
constructions without the expletive. We have also improved our system by avoiding
overgeneration of structures with the wrong agreement and then ruling them out because of
a mismatch, as we had been doing before this chapter.

We can thus assume that subjects merge in the specifier of VP. This claim is known as the
VP-internal Subject Hypothesis. From the specifier of VP, a DP is able to value the
unvalued Person (PN) and Number (NR) features of Inflection, and in turn, it is also able to
receive Nominative Case (via AGREE or the probing system) from Inflection. If there is an
expletive pronoun there available for the derivation, it will merge in the specifier of IP to
satisfy the Extended Projection Principle. Otherwise, the DP in the specifier of VP will
move to the specifier of IP to satisfy this Principle.

13.6 Extraposed sentences?

In section 1 of this chapter we learned how to analyze the verb TO BE when it functions as
a copulative verb. Now we are ready to tackle what have been called"extraposed
sentences" as those in (30):
276 Héctor Campos & Bùi Huỳnh Thủy Thương

(30) a. It is easy to learn syntax.

b. It is important to know syntax.

These sentences seem to be equivalent to:

(31) a. [To learn syntax] is easy.

b. [To know syntax] is important.

These sentences in (30) are called "extraposed sentences" because the understood subject
of the clause is the infinitival clause, as shown in (31), and their subject (the infinitival
clause) seems to have been "extraposed" (or moved) to the end of the sentence. To
understand why the infinitival clause is the subject, think of a corresponding structure with a
DP subject, as Syntax is easy, Syntax is important, where syntax is clearly the subject of the
sentence, since something needs to be easy or important.

However, there is evidence to suggest that in the constructions in (30), the infinitival clause
has not been extraposed at all. Compare the following sentences and compare the form (the
mood) of the embedded verb:

(32) a. It is important that he know syntax.

b. It is obvious that he knows syntax.

In (32a) the verb appears in the present subjunctive, while in (32b) it appears in the
present indicative, hence the difference in form (the mood) that we observe. It is clear
from the examples above that it must our choice of adjective that determines the mood
of the embedded clause. This can be captured syntactically only if the embedded clause is
actually a complement of the adjective in the impersonal constructions in (32). We
extend the analysis to the constructions in (30). The structures for (30b) and (32a) are
shown in (33), assuming the analysis for the main verb to be we proposed in section 1 in
this chapter:
 English Syntax & Universal Grammar 277

(33) a. It is important that he know syntax.

b. It is important to know syntax.

278 Héctor Campos & Bùi Huỳnh Thủy Thương

In the structures in (33), the Predicate phrase needs a subject. Since it has none, we insert
the expletive pronoun it. We will assume that the expletive pronoun IT has the features
PN: III and NR: SG, i.e., third person singular features. After the valuing mechanism that
we described in the previous two sections, I becomes PRES, IIISG, as shown in (33). In
order to satisfy the Extended Projection Principle the expletive pronoun moves to the
specifier of IP, thus deriving (30b) and (32a), respectively.

What about the structures in (31) where the infinitive appears in subject position? These
constructions can be derived if the infinitival clause moves to the specifier of PredP and
then to the specifier of IP, as shown in (34):

(34)

In the structure in (34), the CP complement of A moves to the specifier of PredP and then
moves on to the specifier of IP. What about the agreement on the verb? Inflection acquires
the default values of third person singular, thus accounting for the agreement on I.2 It is

2
Notice that even if we conjoin two infinitives, the agreement on the copulative verb is still IIISG:
a. It is important to study and to practice.
b. To study and to practice is important.
Agreement on the verb to be in (a) is third person singular due to the fact that the expletive pronoun it
bears those features. Since there is no expletive pronoun to agree with in example (b), we can
 English Syntax & Universal Grammar 279

important to note that constructions with the expletive, as in (33), are much more common
than the structures where the CP has moved to the subject position, as in (34).

Summarizing this section we see that what have been called "extraposed sentences" have
not really been extraposed at all. In these extraposed constructions the CP appears as the
object of the adjective as shown by the different mood selection we find depending on the
adjective (see examples in (32)). An expletive pronoun IT is inserted in the specifier of
PredP, which then moves to the specifier of IP, thus deriving constructions like (30).
Alternatively, the CP itself may move to the specifier of PredP and then to the specifier of
IP, thus satisfying the Extended Projection Principle for those categories and deriving the
less common constructions shown in (31).

13.7 Tough-constructions

There is another set of constructions which are formed with the main verb TO BE that we
analyzed in section 1 of this chapter. Examples are shown in (35):

(35) a. John is tough to please.

b. That class seems fun to take.
c. That interface looks easy to understand.
d. That car is too expensive to buy.

These constructions are called Tough-construction since they follow the model of the
adjective tough in (35a). The two characteristics of these constructions are listed below:

(a) Tough-constructions are formed with a copulative verb.

(b) The subject of the clause is interpreted as the object of the verb in the infinitival
clause.

The structure for (36a) is shown in (36b) below:

(36) a. John is easy to please.

conclude that the agreement found in (b) is really a default agreement. This is explained by the fact
that, when probing down, Inflection cannot find the features of Person (PN) and Number (NR), thus
acquiring a default value.
280 Héctor Campos & Bùi Huỳnh Thủy Thương

b.

In (36), John is generated in the specifier of PredP. John then moves the specifier of the
upper IP to satisfy the Extended Projection Principle. The infinitival clause contains a WH-
Operator which moves to the specifier of the lower CP to satisfy the WH-Criterion. This
OP, similar to the OP that we find in restrictive relative clauses (Chapter 11, section 2), is
coreferential with the subject of the PredP, similar to the way that OP are coreferential with
their antecedent in relative clauses. The PRO in the infinitival clause is an arbitrary PRO
(see Chapter 12, section 4).

Not all adjectives allow the tough-construction, though. Consider the pair of sentences in
(37):

(37) a. John is easy to visit.

b. John is eager to visit.

In (37a) John gets visited, while in (37b) John is the one who does the visiting. The
structures for (37a, b) are shown in (38a,b), respectively:
 English Syntax & Universal Grammar 281

(38) a. John is easy to visit.

b. John is eager to visit.

282 Héctor Campos & Bùi Huỳnh Thủy Thương

In (38a), similar to what we saw in the tough-construction in (36), the WH-operator takes
John as its referent. The PRO is understood as an arbitrary PRO. John moves to the subject
position to satisfy the Extended Projection Principle. In this example John is the one to be
visited as the WH-Operator coreferent with John is in object position. In (38b), on the other
hand, PRO is controlled by John, hence John is the one to do the visiting. In this example,
the object of visit is understood as a generic object, John is eager to visit someone or some
place; it's not specified in the sentence.

CHALLENGE

A. A very common mistake that advanced students of English who are native speakers
of Vietnamese make is shown (i):

(i) a. *The students are easy to forget the grammar.

b. *The students are difficult to get to school.
 English Syntax & Universal Grammar 283

(a) Briefly explain why these sentences are ungrammatical in English.

(b) What is the equivalent in Vietnamese? How would you analyze those examples?

B. Does Vietnamese have the equivalent of a tough-construction in English? Give two

original examples and translate them into English. Use different adjectives.

C. Propose an analysis for the following sentences in Vietnamese:

(i) Rất khó để người già học ngoại ngữ.
(ii) Người già rất khó học ngoại ngữ.
(iii) Người già rất khó để học ngoại ngữ.

13.8 Summary

In this last chapter we saw four types of constructions that fall under the general operation
of DP-movement. These four construction types are:

(39) a. Quan is smart. (copulative verb)

b. Ha seems to enjoy syntax. (subject raising verb)
c. Khiem was invited to the party. (passive construction)
d. There are many students in this class. (there-construction)

As we saw above, in all of these constructions a DP moves to the specifier of a tensed IP to

find Case in order to satisfy the Case Condition in (3). This DP must move to an empty
specifier position. This movement, apart from obtaining a Case for the moved DP, also
helps us satisfy the Extended Projection Principle, which requires every sentence to have
a subject. We saw that copulative, subject raising verbs, and passive constructions all have
derived subjects, that is, they start with an empty subject at D-Structure, which gets filled
by S-Structure in order to satisfy the Extended Projection Principle. We extended the
analysis to there-constructions, also called existential or presentational constructions.
We argued that THERE was inserted to satisfy the Extended Projection Principle.
Furthermore we proposed a mechanism to account for the agreement found in these
existential or presentational constructions. We then extended this mechanism to all
structures.

We provided evidence in support of the claim that the subject of a sentence merges in the
specifier position of VP. This claim is known in the syntactic literature as the VP-internal
Subject Hypothesis. To account for subject/verb agreement, we implemented the same
probing or AGREE system that we suggested for there-constructions.
284 Héctor Campos & Bùi Huỳnh Thủy Thương

Finally we discussed extraposed sentences and tough-constructions. We argued that

extraposed sentences like (30) are not really extraposed, but have a CP which appears as a
complement of the adjective that c-selects them. The expletive pronoun it is inserted so as to
satisfy the Projection Principle. The pronoun it is responsible for the agreement found on
the copulative verb. As we saw in (33)-(34), these constructions contain PredP. As shown in
(36), tough-constructions are also generated in PredP and they contain a WH-operator that
is coreferential with the subject of PredP. Vietnamese speakers beware! The subject of a
tough-construction needs to be interpreted as the object of the embedded verb.

This chapter introduced a new operation: DP-Movement. Similar to Wh-Movement, it also

moves a maximal category. The chart in (40) summarizes the differences between Wh- and
DP-movement:

(40)
WH-Movement DP-Movement
Ex: Who did you visit? Ex: Quan seems to love Ha.
1. Motivation for movement is the 1. Motivation for movement is the
satisfaction of the Wh-Criterion. satisfaction of the Case Condition and
the Extended Projection Principle.
2. We typically move DP, AP, ADVP, PP, 2. We only move DP.
...
3. We move to the specifier of CP. 3. We move to the specifier of IP.
4. The head of CP has a [+WH] feature 4. The head of IP has the features [+TNS,
that triggers the movement. +AGR] that trigger the movement.
5. If long distance, movement must occur 5. If long distance, movement must occur
Comp to Comp. from specifier of IP to specifier of IP.
6. If a DP moves, it must receive Case 6. The DP does not get Case in its original
before it moves. position; it moves to receive Case.
7. General operation that derives: 7. General operation that derives:
specific questions, copulativeverbs, raising verbs, passive
relative clauses, etc. constructions, there-or existential/
presentational constructions, etc.

We thus see that WH-movement is different from DP-movement. The only common
element they share is that both operations are moving a phrasal element.
 English Syntax & Universal Grammar 285

FINAL PRACTICE

I. Define the following terms briefly. Give original examples where applicable.

1. Lexicon 27. Relative clause, antecedent, relative

2. Catalog pronoun
3. Lexical entry 28. wh-OP
4. C-selection 29. -n't vs. not
5. S-selection 30. Modal vs. auxiliary verb
6. Extended Catalog 31. Phrasal vs. Prepositional verb
7. Generative Morphology 32. Passive constructions
8. X'-Syntax 33. Subject-controlled PRO, object-
9. Complements controlled PRO
10. Modifiers 34. Inserted elements
[Link] 35. Types of movement
12. Pseudo-prepositions 36. Deleted elements
13. Markers 37. Constituent tests
14. Only 3 operations allowed in syntax 38. Syntactic ambiguity
15. Syntactic Category vs. Syntactic 39. Quantifier
Function 40. DP hypothesis
16. Intransitive, Transitive, Ditransitive verbs 41. Projection Principle
17. Copulative verb, raising verbs 42. Extended Projection Principle
18. PREDP, raising verbs, causative verbs, 43. Theta-Criterion
small clauses 44. Case Condition
19. ECM verbs 45. Stranded Affix Condition
20. Nominative, Objective Case 46. Wh-Criterion
21. NP-CP complement 47. Three syntactic Cases
22. Infinitival complement 48. Ambiguity
23. Simple questions 49. Prepositional complementizer
24. Specific questions 50. Restrictive, non-restrictive relative clause
25. Short distance Wh-movement, 51. Extraposed sentence
Long-distance Wh-movement 52. Sentential adverbs, VP-adverbs
26. [+/-WH], [+/-Q], complementizers 53. Tough-sentences
54. Oblique, Genitive Case

II. Diagram the following sentences. Indicate all the inserted, moved and deleted
elements.

1. We met the students that wanted to get a scholarship.

2. Hasn't the man that we sent a message to replied yet?
286 Héctor Campos & Bùi Huỳnh Thủy Thương

3. How much sugar do you think we need to buy for the party?
4. The article you read seems to have been published by that famous political journalist.
5. Why don't you tell the people that are having trouble to call up the police department?
6. The girl who received the award seems to have been a waitress.
7. I can't believe you sold the computer we bought last week.
8. How tall is your son?
9. The teacher encouraged the students he expected to get a good grade on the final
exam to do their best.
10. a. She likes to eat her soup spicy.
b. She likes to eat her spicy soup.
11. Do the linguistics students seem to understand syntax well?
12. Our teacher seems to have been invited to the party by the School Dean.
13. Which book did the media say was published illegally?
14. Weren't the students told to go to a different room for the final exam?
15. I met the student who seemed to enjoy her syntax class.
16. She bought a car that was made in Japan.
17. The students were expected to review for their last syntax class.
18. Lisa wants to be promoted in her job.
19. The graduates from the English Department in this University want to be happy.
20. The students in the English Department at this University seem to be happy.
21. The professor always tells his two assistants to make the students write up their
syntactic arguments carefully.
22. I regard the Dean of Faculty's daughter as my best student in my graduate syntax
class.
23. Do you wonder why Tyler, who is always so secretive, finally let the cat out of the
bag?
24. Who did the professor expect to give up her original plan to spend her Thanksgiving
holiday in Hawaii with her boyfriend?
25. Who do the graduate students of linguistics usually hold responsible for the mistakes
on the weekly syntax assignments they receive?
26. When did you say that Laura wanted to go to France? (ambiguous)
27. a. I met the girl who you bought that syntax book from.
b. I met the girl from whom you bought that syntax book.
c. I met the girl you bought that syntax book from.
28. The girl whom we met at the party said she had enjoyed the syntax class that she took
last year.
29. Haven't you read the article that the professor said that we would discuss in class
next week?
 English Syntax & Universal Grammar 287

30. Have you not read the article whose author the professor said he would invite to our
class next week?
31. Doesn't the Chair of the English Department expect the linguistics teachers to show
up to the end of the semester meeting?
32. Will the articles that were written in French be published by that famous journal?
33. The computer which he bought seems to have been backed up.
34. Our teacher of modern syntax seems to have written the exam we are taking.
35. Were the students told to go to the classroom where the exam was being held?
36. Tôi đã xem một bộ phim có Hoài Linh đóng.
37. Bạn thích xem những chương trình nào trên ti vi?
38. Tôi muốn mở một tài khoản ở ngân hàng gần nhà để gửi tiền tiết kiệm.
39. Shakespeare đã nói rằng thế giới là một sân khấu.
40. Có một số câu hỏi về bài thi (mà) tôi muốn hỏi bạn.
41. Bạn có thể cho tôi biết khách sạn gần nhất ở đâu không?
42. Tôi không đồng ý về cách cư xử của bạn với người lớn tuổi.
43. Chúng ta cần suy nghĩ nhanh để làm kịp bài thi .
44. Tôi sẽ gửi bức thư này cho một người bạn ở Đà Nẵng.
45. Nam hướng dẫn tôi cách dùng Iphone.
46. Thầy giáo yêu cầu sinh viên rời khỏi phòng.
47. Bà ấy có ý định đề nghị ủy ban nên hoãn cuộc thảo luận về vấn đề này.
48. Những công ty có nhân viên đang bãi công sẽ đóng cửa trong hai tuần nữa.
49. Quân đã nói với tôi em ấy có nhiều việc phải làm.
50. Không ai trong văn phòng của chúng tôi còn muốn lái xe đi làm nữa.
288 Héctor Campos & Bùi Huỳnh Thủy Thương

EPILOGUE

In our Introduction we mentioned that the objective of (theoretical) linguistics is to create a

model for a speaker's I-language which makes it possible to explain that speaker's E-
language. This book introduced you to the model known as the Principles and
Parameters theory developed by Chomsky and his followers in the 80's and 90's. In the
mid 90's and early 21st century, this model evolved into what is known as the Minimalist
Program. To be able to understand this later system however, you must thoroughly
understand the previous model.

We saw that the Principles and Parameters theory is a modular system. Each module has
its own structure and its own principles and conditions. The model has six modules, as we
can see in (1):

(1)
X'-Syntax Lexicon

D-Structure

S-Structure

PF LF

In Chapter 2 we explored the structure of the Lexicon. The Lexicon is where words are
created and where all the necessary phonetic, phonological, morphological, syntactic and
semantic information to do our syntax is stored. We saw that the Lexicon has two
components: the Catalog and the Extended Catalog. The Catalog contains all the
morphemes necessary to create words. Generative Morphology uses those morphemes to
create words. These words are stored in the Extended Catalog. All words are created in the
Extended Catalog with the exception of verbal morphology which, as we saw in Chapter 7,
happens between D- and S-Structure. Morphology also applies at the level of PF, where it
spells out verbal morphology.
 English Syntax & Universal Grammar 289

Three principles that are central to our syntactic model are based on information that is
stored in the Lexicon. They are listed in (2) –(4):

(2) The Projection Principle

The c-selection of a verb must be respected at every syntactic level.

The Projection Principle guarantees that a head will always have its complements next to it,
as specified in the c-selection for that element.

(3) The Theta-Criterion

Each argument must receive one and only one theta-role and each theta-role
must be assigned to one and only one argument. (In brief: ARG ↔ Θ-role)

We saw in Chapter 2 that predicates select arguments. We encoded this information in the
s-selection. The Theta-Criterion guarantees that all arguments will be merged in the
syntactic structure. It also requires that each argument be assigned one and only one theta-
role. At the same time it requires that each theta-role that a verb s-selects be assigned to one
and only one argument. Thus theta-roles cannot be shared by arguments and all theta-roles
must be assigned.

(4) Extended Projection Principle

Every sentence must have a subject.

The Projection Principle guarantees that the c-selection of a predicate is respected in the
syntax. The Extended Projection Principle, on the other hand, guarantees that the s-
selection of a predicate is respected. As we saw in Chapter 12, the Extended Projection
Principle and the Theta-Criterion are the principles that forced the existence of PRO, as we
need a subject for infinitives. The Extended Projection Principle also guaranteed the
movement of derived subjects in raising and passive constructions in Chapter 13.

X'-Syntax is the module where our syntactic structures are created. We saw that all
categories share the same basic architecture given by X'-Syntax. This architectural model or
template is shown in (5):
290 Héctor Campos & Bùi Huỳnh Thủy Thương

(5)

From (5) we can see that all structures have a minimum of 3 levels: an X-level called the
head or nucleus of the phrase, an X'-level called the intermediate projection level and
XP, the phrasal level. As shown in (5), the level of modifiers is optional. This additional
X'-level is projected only if there is a modifier. When we encounter more than one modifier
in a sentence, we project each modifier in a separate X'-projection. These extra projections
are called adjunctions. This is why modifiers are also called adjuncts. Complements, on
the other hand, are all projected under the first or lowest X'-projection as sisters to the head
X. (5) requires all categories to be endocentric.

We showed that the architectural template in (5) can be applied to the verb phrase (VP,
Chapter 4), to the noun phrase (NP, Chapter 5), to adjective phrase, the prepositional phrase
and the adverb phrase (AP, PP, ADVP, Chapter 6). We then applied this template to some
functional projections: inflection phrase (IP, Chapter 7), determiner phrase (DP, Chapter
8), complementizer phrase (CP, Chapter 9), auxiliary phrase (AUXP, Chapter 7), negation
phrase (NEGP, Chapter 7). We argued that modal verbs are generated directly in Inflection
(Chapter 7).

The Lexicalization Principle introduced in chapter 3 was responsible for inserting words
into the structures created by X'-Syntax. We thus generated D-Structures. D-Structures
map the c-selection as well as the s-selection of a head into a syntactic structure.

We then studied the three basic processes that allow us to go from D-Structure to S-
Structure. These processes are known as transformations (also known as Affect Alpha).
The only transformation operations allowed in our system are movement, insertion and
deletion. In order to account for subject-verb agreement, we implemented the system of
Probing (also called Agree).

We posited two kinds of movement: phrasal movement, also called X-max movement
(Chapters 10-12) and head movement (Chapter 7).

We argued for two kinds of phrasal movement: Wh-Movement (Chapters 10, 11) and
DP-Movement (Chapter 12). We saw that Wh-Movement allowed us to derive specific
 English Syntax & Universal Grammar 291

wh-questions as well as relative clauses. The motivation for Wh-Movement is the Wh-
Criterion in (6):

(6) WH-CRITERION
a. A wh-phrase must move to the specifier position of a CP which contains the
feature [+wh].
b. A C-head that has the feature [+wh] must host a wh-word in its specifier.

The Wh-Criterion forces the movement of wh-words to CPs that are marked with the
feature [+WH].

DP-Movement is motivated by the need of DP to get Case in order to satisfy the Case
Condition, which we introduced in Chapter 5:

(7) The Case Condition

Every DP must bear Case.

DP-Movement takes place so that a DP, originally in a non-Case-assigning position, can get
Case. As we saw in Chapter 13, this movement also helps the structure satisfy the
Extended Projection Principle in (4).

We argued that different constructions can be argued to involve DP-Movement, namely:

copulative constructions, raising verbs, passive verbs, presentational sentences and
eventually, all verbs, once we accept the VP-Internal Subject Hypothesis.

We saw that only three elements are allowed to assign Case in our system:

(8) Case Assigners

(a) verbs, which assign OBJ(ective) Case
(b) prepositions, which assign OBL(ique) Case
(c) inflections which contains specifications for Tense and Agreement

The Wh-Criterion as well as the Case Condition motivate phrasal movement: Wh-
movement and DP-Movement. The second kind of movement, head movement, on
the other hand, is motivated by the Stranded Affix Condition, which we introduced in
Chapter 7:
292 Héctor Campos & Bùi Huỳnh Thủy Thương

(9) Stranded Affix Condition

A bound affix must be attached to the appropriate root or stem.

This Condition forced the movement of V to I, Aux to I, I to C, Neg to I. It also served as a

motivation for do-support, which is an insertion operation. As we saw in Chapter 7, do-
support saves our structures from crashing by providing our D-Structures with the dummy
verb do to which the violating stranded affixes can attach to.

Insertion adds material to our D-Structures and allows us to go from D- to S-Structure. It

can only insert elements that either mark a particular function, or as a last resort, which
helps us save the structure and avoid crashing. We argued to TO/FOR-insertion is an
example of marking in English (Chapter 7), although we saw that this insertion was also
responsible for assigning Case to the indirect object which was too far away from the verb
(chapter 8, appendix). English has the requirement that elements that are case marked by a
particular head must be adjacent to that head:

(9) Case Adjacency Condition

In English, an XP that receives Case from a head must be adjacent to the that head X.

In (8) we can see that nouns, adjectives and adverbs do not have the ability to assign Case.
In Chapters 5 and 6 we saw that the pseudo-preposition OF is inserted in front of the
complement of nouns, adjectives, and adverbs to guarantee that the object of these
categories gets Case so as to satisfy the Case Condition in (7).

The third and last type of operation is deletion. We argued for deletion in structures that
involved Exceptional Case-Marking verbs (Chapter 12). We showed that these
constructions must delete CP in order for the embedded subject of an infinitive to be able to
receive Case from a higher verb. However, we showed that these constructions could also
be generated directly with an IP complement, in which case, our system can dispense of
deletion completely.

The diagram in (10) summarizes our system with its modules and principles:
 English Syntax & Universal Grammar 293

(10)

X'-Syntax Lexicon

D-Structure

Transformations:
- Insert (to/for, of, do),
- Move (Head, Wh-words, DP),
General Principles: - Delete (CP),
- Projection Principle - Probing
- Extended Projection Principle Principles:
- Theta-Criterion - Stranded Affix Condition,
- Wh-Criterion,
- Case Condition,
- Case Adjacency Condition

S-Structure

PF LF

(10) provides you with a system and theory that allows you to analyze many different
structures not only in English, but in many different languages. As we saw in Chapter 2, the
level of phonetic/phonological form (PF) interfaces with the perceptual/production
system (also called the sensory-motor system) while the level of logical form interfaces
with the meaning system in our mind (also called the conceptual/intentional system).
As we can see in (10), this model associates sound (PF) with meaning (LF) in an indirect
way, via the four previous modules: the Lexicon, X'-Syntax, D-Structure and S-Structure.

The Minimalist Program that follows this model tries to do away with the modules of X'-
Syntax, D-Structure and S-Structure, thus allowing the connection sound-meaning to be
more direct than in our diagram in (10).