Professional Documents
Culture Documents
Syntax 1
Syntax 1
Syntax 1
Part 1
UNED
ÍNDICE
Introduction
This course is an introduction to syntax within the framework of generative grammar.
The emphasis is on description, with some analysis and argumentation. The goal is
twofold: to become familiar with core facts of English syntax, as well as with issues and
techniques central to syntactic theory, and to understand typological adequacy as an
important tool for linguistic research. In the following pages we will also spell out the
role of binary branching theory and then discuss its implications for some well-known
syntactic puzzles.
• Summarize and apply arguments for constituent and phrasal structures for
phrases, clauses and sentences.
• Summarize the arguments for Noun and Verb Phrase Structures.
• Analyse verb movement as an instance of head-movement.
• Identfy the properties of passives and Raising as instances of DP-movement.
• Identify a second instance of transformation: Wh-movement and each of their
idiosyncratic properties.
Contents
The first part of this course includes two chapters. First, we deal with the linguistic
approach we will be working with (Chapter 1) and, second, we examine the basic
notions of syntactic theory according to the X-bar theory (Chapter 2).
As a whole, this first part of the syllabus constitutes an introduction to the fundamentals
of syntactic theory and more particularly to generative grammar, which is the theoretical
paradigm we will be using in this course book. We will first address the methodological
underpinnings of the model, and will provide a preliminary introduction to syntactic
representation in terms of tree diagrams and bracketing. In addition, we will examine
fundamental relations that hold between syntactic units such as dominance, precedence,
c-command. Later on we will work with the X-bar theory.
Chapter 1
The more recent history of linguistics has been permeated by a debate between
what has been called the formal and the functional paradigms 1. GG is a formal model
which has often been contrasted to the functional models. Generativist and functional
models are based on radically different conceptions of the nature of language.
Generalizing a bit, GG maintains that language is an abstract object, and grammar is
thought to be an attempt at defining this object in terms of a set of abstract rules and
principles, which are postulated without making any reference to the uses and meanings
of the constructions described. Consider the following passage:
The study of generative grammar has been guided by several fundamental problems,
each with a traditional flavour. The basic concern is to determine and characterize the
1
A number of different linguistic paradigms emerged under the cover term ‘functional’:
S.C. Dik’s Functional Grammar, Van Valin’s Role and Reference Grammar, Halliday’s
Systemic Functional Grammar, Givón’s Typological Grammar, Hopper’s Emergent
Grammar etc Following Nichols (1984), functional models could be classified in
conservative, moderate and extreme depending on how they hold with respect to the
issues of the autonomy of syntax and the autonomy of grammar. We refer the reader to
the subject Modelos teóricos descriptivos de la lengua inglesa, where a detailed
description of these two major traditions is given. In this regard, Butler (2003) and
Gonzálvez-García and Butler (2006) offer an impressive account of both functional and
cognitive models.
linguistic capacities of particular individuals. We are concerned, then, with states of the
language, which we understand to be some array of cognitive traits and capacities, a
particular component of the human mind/brain. The language faculty has an initial state,
genetically determined; in the normal course of development it passes through a series
of states in early childhood, reaching a relatively stable steady state that undergoes little
subsequent change, apart from the lexicon. To a good first approximation, the initial
state appears to be uniform for the species. Adapting traditional terms to a special
usage, we call the theory of the state attained its grammar and the theory of the initial
state Universal Grammar (UG). (Chomsky, 1996: 14).
In contrast, functional approaches share the assumption that language - and more
in particular its morphosyntactic structures - should be explained with reference to its
semantic and communicative functions. Suffice the following passage which
exemplifies this view:
should be explained and understood within the wider communicative context in which
these occur. Thus, pragmatics is then the all-encompassing framework against which
syntax and semantic must be studied.
As discussed in Mairal and Gil (2003, 2006), the 20th Century witnessed a
renewed impetus in the search for properties common to all languages by linguists
within the functional tradition as well as by generative linguists. This in fact gave rise in
the eighties to numerous descriptive and comparative studies of a broad range of
languages. Linguistic models are now revised or formulated with the aim of providing
grammars for typologically divergent languages. In this regard, as Van Valin and La
Polla affirm (1997:14) Role and Reference Grammar grew out as an attempt to answer
the question ‘what would linguistic theory look like if it were based on the analysis of
Lakhota. Tagalog and Dyirbal, rather than on the analysis of English?’ This signifies
that a theory of grammar should be able to provide descriptions for a wide number of
different languages. This is in fact a very challenging and ambitious enterprise which is
taking a lot of effort given that a lot of the languages in the world are becoming extinct
without any steps being taken to remedy this loss. Thus, from now on you must bear in
mind that your syntactic theory should not only be valid to describe and explain the
inventory of syntactic structures that define English but also the inventory of syntactic
structures of typologically divergent languages.
First, a fundamental distinction is that of Language and language, which has also
been referred to as I-language and E-language in the generative literature 2. Chomsky
believes that GG must “render” explicit the implicit knowledge of the speaker. His
model of GG begins with an axiom and a set of well-defined rules to generate the
desired word sentences. One of his goals with linguistics is to create an explanatory
theory of Language, or else, Internal language (I-language), which has been
contrasted with language , or else External Language (or E-language). Generativists
consider language (Spanish, English, or any language) to be a specific state of the
language faculty. This state is known as I-language (i.e. Language) and represents the
grammatical knowledge of competent speakers, who activate it as an objective and
2
Note that while we use the notions I-language and E-language, the textbook refers to
these two notions as capital L and l.
3
See Mairal and Gil (2003) for an extensive discussion of this issue.
prescriptive and descriptive rules is introduced. This course will deal with descriptive
rules, that is, those rules that are part of the native speaker’s competence, or are an
integral part of Language.
Human beings are born equipped with some internal unconscious knowledge of grammar:
Universal Grammar (UG). UG is a set of universal principles of language, some of which
are rigidly fixed, some of which parametrized. Via the input of the experience of one
particular language this knowledge can be implemented. The acquisition process is
“triggered” by the exposure, the child’s linguistic experience. Exposure will also enable
the child to learn the vocabulary of the language
As argued in Mairal and Gil (2003, 2006), Chomsky (1965) reasons that children
are born with a certain knowledge about their native language and this knowledge,
which they cannot have acquired through experience, must, therefore, be attributable to
innateness and what is innate must be universal, in the sense that it cannot vary
drastically from one person to another. So, UG would be the initial state of the language
faculty and it will develop towards a final stage which is the grammar of a certain
language, obtained thanks to the interaction with experience around the individual,
making possible the learning of one language or another. Consider the two schematic
representations of the generative view of language acquisition extracted from Van
Valin and LaPolla (1997:chapter 1) and Haegeman (1994: 16-17):
(Haegeman, 1994:16)
Current work in the generative tradition puts forward the view of language
acquisition in terms of parameter setting. Recall that out of the successive version that
GG has undergone we advanced that Principles and Parameters (hereafter, PP) would
constitute the theoretical framework of this course. According to the PP model, the child
only has to learn those grammatical properties which are subject to parametric variation
across languages. As Radford (1997:21) argues:
Hence, Chomsky reacts against Skinner´s conductism and rejects the idea that a
child is born “with a white blackboard”. Rather, children are born with a genetic
predisposition to structure their acquisition of linguistic knowledge in a very specific
manner. UG, in that sense, provides speakers with a series of principles that can be
applied to all languages. In contrast, we know that linguistic variation is very wide
ranging, and in order to explain it, generativists postulate that the universal principles of
UG undergo parametric variation. .In this sense, interlinked sets of principles and
parameters constitute an authentically universal core grammar in which options for
variation are reflected in the model by means of parameters. For example, the wh-
parameter accounts for the difference between languages which permit the fronting of a
wh-constituent (interrogative pronouns and adverbs) in questions, and those languages,
such as Chinese, which do not. There is also the pro-drop parameter, which regulates
variation in languages which allow the omission of the subject (e.g. Spanish, Italian…)
and those which do not (e.g. English) 4:
4
For an exhaustive treatment of the topic of linguistic universals we refer the reader to
Mairal and Gil (2003, 2006), and the works contained therein.
No Pro-drop languages:
c. John writes a letter [English]
c’. * Writes a letter
d. Jean écrit une lettre [French]
d’. *Écrit une lettre
(Mairal & Gil, 2003:40)
Finally, we claim that a theory of grammar must meet that three levels of
adequacy: 5 (1) observational adequacy, a grammar should be able to predict which
sentences are grammatical and those which are not; (2) descriptive adequacy, the
grammar must be observationally adequate and develop the general principles and
processes that capture the native speaker’s intuitions about the structure of the sentences
and that allow his /her to decide on the grammaticality of sentences; (3) explanatory
adequacy, the grammar is descriptively adequate and is an integral component of a
larger theory which provides an account of ‘how these facts arise in the mind of the
speaker-hearer’ (Chomsky, 1994: 368). Let us exemplify these notions with the
following examples taken from Haegeman (1994:6):
5
This section namely follows Van Valin and LaPolla (1997: section 1.2.) and Haegeman
(1994:2.1.).
6
Recall that all these examples as well as the line of argumentation behind these two
notions are based on Haegeman (1994: 6).
1.3 Constituency
Before we actually begin drawing a syntactic tree, we need to identify the chapters that
we will be using as well as the metalanguage involved in the theory. One basic claim is
that the sentences of a language consist of words and some words go together to create
larger units. For example, consider a sentence like (1):
(1) The man on the stage bought a newspaper at the bookstore yesterday
We know that certain words here belong to the same part of speech; for example,
man, stage, newspaper, bookstore are all nouns; on and at are prepositions; a and the
are determiners, etc. In this regard, the textbook (section 1) firstly addresses the issues
of parts of speech and argues that semantic criteria are not sufficient to identify, say, a
noun, a verb etc. That is, we cannot simply accept classical definitions that “a noun is a
person, place, or thing” or that a “verb is an action”; there are in fact many
counterexamples that question the validity of this assumption. Instead, the author
maintains that the combination of semantic and distributional criteria provides a more
coherent picture.
Not only do we know that each of the words in (1) belong to a syntactic category
or part of speech, but also we know that certain words go together to create larger units
of analysis called phrases. For example, the words the and man combine to form a noun
phrase. Moreover, phrases can also be part of larger phrases, i.e. the noun phrase the
stage combines with the preposition on to form the prepositional phrase on the stage.
Thus, word-level categories are grouped into phrase-level categories. Moreover, within
a phrase there is a central element called head or nucleus. Using the example above, a
newspaper is a noun phrase since the nucleus is the head noun newspaper. This
combination of words into a larger unit is called constituent.
This chapter is an introductory chapter to the basic tools of syntactic analysis. Thus, one
of the central notions that the student should be able to handle is that of constituency,
which will be of paramount importance for the rest of the chapters. Besides, the
procedures as well as the techniques and rules associated to tree drawing are explained.
In this sense, the student is expected to work out structural ambiguity in terms of phrase
structure. At the end of this chapter, the student is now ready to draw syntactic
representations in terms of constituency. As shall be seen, this initial representation will
be further developed by adding more complexity and refinement.
In order to determine whether or not a given string of words forms a constituent,
GG exploits the geometrical properties of trees. In this line, GG offers an adequate and
explicit description of the linguistic phenomena by making use of theoretical constructs
which have definable formal properties. As Radford (1997:99) puts it:
The use of formal apparatus (involving a certain amount of technical terminology) may
seem confusing at first to the beginner, but as in any other serious field of enquiry (e.g.
molecular biology) no real progress can be made unless we try to construct formal models
of the phenomena we are studying. It would clearly be irrational to accept the use of
formalism in one field of enquiry (e.g. molecular biology) while rejecting it in another
(e.g. linguistics): hence our excursus on the formal properties of phrase-markers.
Thus, as any other branch of science linguistic models develop their own formal
mechanisms for the representation of linguistic phenomena. In this course, we will
study the GG proposal for the hierarchical representation of the clause in terms of a
universal system.
N N N
examination book woman
7
Note that these rules are incomplete since more syntactic categories form part of both an
NP and a PP, as shown in the textbook (page 79). However, they suffice for the purposes
that concern us here.
VP
V NP
open
D N
the door
NP NP (trees)
D N D N
the examination one book
PP PP (trees)
P NP P NP
with on
N D N
pleasure the floor
In the examples above the syntactic distribution of nouns and prepositions forming
parts of NPs or PPs is marked by lines from top to down or the other way round. Recall
that the textbook discusses the two ways to draw a tree, i.e. from top-to-bottom and
from bottom-to-top. In a similar fashion, phrase structure rules also account for the
combination of subject and predicate in both simple clauses (TP) and embedded clauses
(CP) as illustrated by the following examples:
Once you have learnt to identify phrase structure rules you will also be allowed to
account for a number of issues such as structural ambiguity (cf. Section 3). Let us
explain this with a couple of examples. In (9a) there is a double interpretation as shown
by the two readings in (9b) and (9c):
These two interpretations are captured in terms of two different syntactic trees
which have the following format:
(10) VP
V NP
tickled
D N PP
the boy
with a feather
(11) VP
V NP PP
tickled
D N with a feather
the boy
In much the same way, the ambiguity of the nominal phrase in (12) is captured by
proposing a different analysis for women: i) as part of one constituent together with men
as represented by the tree in (13a), or ii) as part of one constituent together with young,
as represented by (13b):
(12) young women and men
(13) a. NP
AP NP
young
women and men
“young women and men”
b.
NP
NP Conj NP
and men
young women
“men and young women”
Looking at both trees in (13), we can conclude that we have two different
structures. The nominal phrase represented by the tree (13a) has a structural similarity
with nominal phrases like “blonde hair and eyebrows” where the adjective “blonde”
modifies both “hair and eyebrows”. On the other hand, the nominal phrase represented
by the tree in (13b) has a structural similarity with the conjoined nominal phrases “short
arms and big shoulders” where we have two different modifiers, one for each noun.
a) Distributional tests
Preposing: Only phrasal constituents may undergo preposing
8
This section is a summary of Radford (1988: Section 2.6.) although we have left out
some of the tests this linguist proposes, i.e. adverb insertion, gapping. For a more detailed
description, we refer the reader to this source which contains a brilliant explanation of the
notion of constituency as well as the tests for constituency.
(16) S
NP VP
VP NP
Det N
N V your
I bought AP N
boat
A
old
(19) a. S
NP VP
VP PP
N V
b. S
NP VP
VP NP
N V P
b) Syntactic processes
Ordinary coordination: Only constituents of the same category can be conjoined.
(20) a. John saw a lion and a tiger.
b. Do you sing in the kitchen or in the bathroom?
c. She speaks very slowly but very articulately.
goes even further and stipulates that only constituents of the same syntactic category
can be coordinated, something which explains the oddity of (22c) and (22d) below:
In the first two examples two PPs and two NPs are coordinated, while the last two
examples coordinate constituents of a different syntactic caegory, an NP and a PP, and a
PP and an NP, thus violating the constraint.
(24) S
NP VP
VP PP
N V
John / Mary walked / ran up the hill
What is important here is that there are restrictions on this type of coordination,
as illustrated by the ungrammaticality of (25).
The reason why this example is ungrammatical has to do with the fact that the
sequence up Mary’s sister is not a constituent of either the clauses containing it (John
ran up Mary’s sister and Harry picked up Mary’s sister). In this case the two VPs are
formed by a V + a particle given that both are phrasal verbs. Then, the preposition up
does not form a constituent with the NP Mary’s sister. The following tree in (26)
illustrates why share coordination of the type in (25) is not possible.
(26) S
NP VP
VP NP
N V P
John/ rang/
Mary picked up Mary’s sister
Finally, note that we can have more than one constituent as the shared part:
Mary bought, and Jane rented, [a car] [in Paris]. What we can’t do is treat as a
constituent something that is not a constituent or to break apart a single constituent:
*John talked to my (and John wrote to his) mother [N’], intermediate constituents
cannot be shared.
Ellipsis: Only VPs may undergo ellipsis (under appropriate discourse conditions)
(27) a. John won’t help me with the dishes but his brother will.
b. Mary wants to close the shop, but I don’t want to.
c. Fetch me an apple, if you can.
d. He may come home early, but then again he may not.
All these examples show that only VPs are elided. Let us represent one of these
structures to see whether this is really true or not. Consider sentences (28a) and its
tree representation in (28b) below:
(28) a. John won’t help me with the dishes but his brother will.
b. S
NP VP
N Aux V NP PP
As this tree shows, the string of words help me with the dishes forms a VP
constituent, thus they can be elided. The same type of argumentation can be applied to
the rest of the examples. As shall be seen in the following chapters, this test shows us
that will, to, can, and may are not part of the VP, because if they were, they should have
been elided necessarily (at least, it should have been possible).
(30) S
NP VP
VP NP NP
N Aux V
He will give his mother a book
(31) S
NP VP
VP NP
N Aux V P
He can give up a book
Chapter 2
Structural Relations and X-bar Theory
First, as for the parts of a tree, the student should be able to identify the following
(cf. Section 1):
• Branch
• Node
• Label
• Root node
• Terminal node
• Non-terminal node.
(1) S
NP VP
Det N NP
V
*The saw man an elephant
Next, Section 4 addresses one of the fundamental notions that the student should
have a clear idea of how it functions, i.e. the notion of c-command, which is retaken in
the following chapter to explain the distribution of anaphora, pronouns and R-
expressions (cf. chapter 5). At this stage, the student is expected to identify the set of
nodes that a specific node c-commands. Let us consider the following representation:
(2) The wealthy man from Los Angeles ate lobster with greenish sauce.
S
NP 1 VP
D AP 1 N1 PP 1 V NP 2
the ate
A1 P1 NP 4 N2 PP 2
wealthy man from Los Angeles lobster
P2 NP 3
with
AP 2 N3
A2 sauce
greenish
Now, if we are asked to identify the nodes that, say, the subject NP 1 c-
commands we will propose the following list:
• VP, V, NP 2 , N 2 , PP 2 , P 2 , NP 3 , AP 2 , A 2 , N 3 .
For further practice, we refer the student to exercises 1-8 of the textbook.
Subjects
The subject of a sentence is not always an Agent, or even the ‘topic’ of the sentence:
(3) a. The baby is asleep.
b. The book was sent by my uncle.
c. It was raining hard.
d. There was a duck in the garden.
2. Case: If replaced with a pronoun, the subject usually has nominative case.
Objects
Objects also cannot be identified semantically.
This part is perhaps one of the most important chapters in the textbook since the author
introduces the basic tools that define the overall syntactic framework that we will be
using in the rest of the book. This chapter then deals with X-bar theory and the
methodological motivations behind it.
So far we have been dealing with a number of phrase structure rules with the following
format:
(1) Phrase structure rulefor NPs
NP - (D) (AP+) N (PP+)
Then, if we want to account for the structure of a sentence like (2a), we would
propose the following syntactic tree in (2b):
b. NP
D AP N PP PP
The wealthy student of Geometry from North Carolina
However, this representation, called a flat structure, does not capture the
hierarchical organization of constituents within a tree: they seem to be on the same
level. So, a more elaborate and articulated theory is required if we want to account for
the hierarchical structure of the clause and the phrases involved. With this in mind, it is
shown that this type of representation, which deals with two major categories N and NP,
is insufficient to account for the contrast between the following structures 9:
(3) a. John met the student from England, and I met the one from France.
b. *Jack met the student of philosophy, and I met the one of Greek.
9
The examples (3a) and (3b) are taken from Poole (2002:37), who includes a very nice
presentation of X-bar Theory. Besides, we have also used this source for the
argumentation that follows and completes this section.
Returning to the examples above, let us see how we can explain the contrast by
firstly providing a syntactic binary representation. Consider the tree in (4) below:
(4) NP
DET N’
The N’ PP
N
student from England
In the first example (3a), one substitutes for a constituent that is both an N and
an N’. If we compare its corresponding tree structure in (4) with the representation of
the (3b) example below in (5), we notice that this time one cannot substitute for student
given that student is not both an N and an N’.
(5) NP
DET N’
The
N PP
student
of philosophy
Then, all seems to indicate that the proform one can substitute for an
intermediate level category, N’, but not an N level category. This is the reason why (b)
is ungrammatical; student is an N category, while student of philosophy is an N’
category. In this sense, note that one could substitute for the entire phrase as shown in
(6) 10.
(6) Jack met this student of philosophy, but I met that one.
In sum, the one replacement test shows the existence of an intermediate level
category. Following this line of thought, the textbook provides argumentation and data
that confirm the existence of this intermediate category across the different phrases: VPs
(section 1.1.), APs and ADVs (section 1.2.), PPs (section 1.3.). This means that all
phrases consist of three projections, i.e. XP or X’’, X’ and X or X0. Suffice the
following chart:
(7) XP
Spec X’
X YP
So, this new three level context leads us to the reformulation of our initial phrase
structure rules. Section 2 of the textbook examines the new format of these rules. Two
important notions are in order: the notion of head and the Principle of Endocentricity.
Let us consider the general format of one of these rules, that for NPs:
(8) NP (D) N’
N’ (AP) N’ or N’ (PP)
N’ N (PP)
10
Poole (2002:39-ff) looks at the other tests for constituency to find evidence that backs up this
intermediate category, i.e. coordination and movement.
If we look back at the tree in (7), X is called the head of the phrase. X' and X'' are
called projections of X. Typographically, these projections are marked by one or more
primes (X' and X''), called bars. Thus, X' is pronounced X-bar; X'', X-double-bar etc..
The head is called the zero projection (also written as X0). The topnode X'' (or XP) is
called the maximal projection of X. All other projections between the head and the
maximal projection are called intermediate projections. Then, one of the central theses
is that all phrases have heads, i.e. every NP has a head N, every AP has a head A, and
every VP has a head V. There are no cases attested where a VP has a N head, or an NP
has an A head. In other words, the theory does not generate rules of the type V’ A
(XP) or NP A’ (XP). This is due to what is called the principle of endocentricity,
according to which phrases inherit their category properties from the head.
(9) XP (Spec) X’
X’ X’ (YP)
X’ X (ZP)
(10) XP
X’
X’ ADJUNCT
X COMPLEMENT
(12) NP
DET N’
The
N’ PP
N PP
director
of films from Paris
The X-bar analysis predicts that complements will always occur closer to their heads
than Adjuncts. Thus, in the examples that follow if the complement phrase and the
adjunct phrase co-occur within the same XP, the complement will precede the Adjunct,
otherwise this will result in an ungrammatical structure:
• Recursivity
The adjunct rule is said to be recursive, which means that an infinite number of X’s can
be generated:
This does not occur with the complement rule, which is not recursive.
• Conjunction
11
See Radford (1988: Section 4.4.) for a complete discussion of the criteria involved to
distinguish complements from adjuncts.
Recall that only constituents of the same syntactic category can be coordinated. This
signifies that only an adjunct plus another adjunct or alternatively a complement plus
another complement can be coordinated, thus blocking out the generation of a structure
with two coordinated sequences of complement plus adjunct. Consider the following
examples:
(17) The students of Geometry from Denmark and from Japan [Adjunct + Adjunct]
• Passivization:
Notice that the NPs can only correspond to an active Complement, not to an active
Adjunct. Some illustrative examples follow:
The distinction complement and adjunct also applies in premodifier positions (cf.
pp. 123-126). For example, let us consider the following example:
DET N’
A
AdjP N’
German NP N
computer salesman
In much the same way as in postmodifier position, complements are closer to their
head than adjuncts in premodifier positions, thus the oddity of the following example:
(25) * The gas Spanish company (cf. The Spanish gas company)
This general framework can also be extended to other categories, VPs, APs, PPs,
ADVPs etc. Let us analyze the following examples and the trees that appear after each
of them:
Spec Adj’
so
Adj’ AdvP
today
Adj PP
conscious
P’
P NP
of
Det N’
the
N
truth
Let us firstly provide some theoretical background before getting into the exact details
of the DP Hypothesis. The structure for noun phrases most commonly adopted in GG
until the mid 80's assumed the noun (N) to be the head of the Noun Phrase (NP); N
could take a complement and had a specifier position hosting determiners and
quantifiers. Adjectives were attached between the determiner and the noun position by a
recursive `adjunction' rule which `stretched' the NP to allow an unlimited number of
adjectives in the representation of (1) of the NP: "The wet, red, soft, spongy ... ball". 12
12
This example has been extracted from the paper "Introduction to Layers in DP" by
Roberto Zamparelli, May 3, 1996 (see website:
http://www.cogsci.ed.ac.uk/~roberto/layers/dp-hyp.html). We refer you to this source to
get a more in-depth account of the DP Hypothesis.
(1) NP
The N’
wet N’
red N’
soft N’
spongy ... N’
N (complement)
ball
However, as Carnie (p. 198-ff) observes, this rule fails to capture the basic X-bar
theory axiom which stipulates that all non-head material must be phrasal. Thus,
determiners should be regarded as phrasal categories. According to the DP-hypothesis,
the most important element in the noun phrase is not the noun, but the category of
articles ("the", "a" in English, "Il/Lo/La/I", "Un/Uno/Una" in Italian, absent in Russian,
Chinese and several other languages) and quantifiers ("Every", "Some", "Three",
"Many", "Most", etc., present to some extent in all human languages). They can be
collectively referred to as determiners. The maximal projection of the determiner, or
Determiner Phrase (DP) is the topmost category of the noun phrase; the determiner
takes the NP as its complement as in the representation of (2), below:
(2) DP
Spec D’
the NP
Spec N’
N (Complement)
ball
In this section, Carnie only examines the behaviour of genitive (possessive) NPs
and, more in particular, identifies the position of the initial possessor phrases (e.g.
"John('s)" in "John's books") in the specifier of DP, where they receive genitive Case
from the ‘s-genitive’ unit found in the D position. Suffice the following representation:
(3) DP
John’ D’
s books
DP
Spec D’
D NP
In what follows we provide you with the most important nominal structures and
their corresponding representations of the English DP using brackets (Castillo 2003: 67-
69):
(4) a) [DP [D’ [D the] [NP [N’ [N director] [PP [P’[P of ] [DP the film] ] ] ] ] ] ]
b) [DP [D’ [D 0 ] [NP [N’ [AP [A’ [A fresh] ] ] [N’ [N water] ] ] ] ] ]
c) [DP [D’ [D that] [NP [N’ [AP [A’ [A black] ] ] [N’ [N cat ] ] ] ] ] ]
d) [DP [D’ [D 0] [NP [N’ [N men] [PP [P’[P from ] [DP Asia] ] ] ] ] ] ]
e) [DP [D’ [D 0] [NP [N’ [N Mary] ] ] ] ]
f) [DP [DP John] [D’ [D‘s] [NP [N’ [N winning ] [DP [D’[D the ] [NP [N’ [N
lottery ] ] ] ] ] ] ]
g) [DP [DP John] [D’ [D ‘s] [NP [N’ [N murder ] [PP [P’[P of ] [NP [N’ [N
Mary ] ] ] ] ] ] ]
h) [DP [QP All ] [D’ [D your] [NP [N’ [N children] ] ] ] ] ]
(5) Peter thinks that they will finally arrive tomorrow [Complement Clause]
(6) For Peter to win the race surprised everyone [Specifier Clause]
(7) The book I bought yesterday was very interesting [Adjunct Clause]
We can further subdivide complement clauses into finite and non-finite clauses.
The former are also called tensed clauses while the latter tenseless clauses. Since we
understand that these two concepts are clearly explained in the textbook, we are not
going to discuss them here any longer.
With this terminological background in mind, we will deal with the following
two rules, which, as they stand, do not fit X-bar theory:
(8) S’ (C) S
(9) S NP (T) VP
(10) SP
Spec S’
Aux S
This representation has two main advantages: (i) it obeys binary-branching and
has a hierarchical structure; (ii) the subject may easily be placed in the Spec position,
following Chomsky’s intuition that all sentences have subjects:
(11) SP
Subject S’
Aux S
Yet, the main problem with this configuration is that the sentence S cannot be the
head of any phrase in X-bar theory because it is not a lexical item or word. Moreover,
we know that some verbs select sentences like in the case of Mary said that it was
raining, but so far we cannot identify the syntactic category of the complement that it
was raining subcategorized for by the verb said.
(12) VP
V’
V S’
said COMP S
However, the tree diagram above does not tell us whether the complementizer
(COMP) is a specifier or a head. In connection with this, Carnie introduces two new
syntactic categories, CP and TP 13. Let us firstly review the S’ rule above. If we want to
adapt (8) to an X-bar format, we will simply have to stipulate C as a maximal phrase
13
We refer to Radford (1988) for further evidence supporting the assumption of the existence of these
two clausal nodes CP and TP
since it is the only head material which is not phrasal. Then, we will be dealing with a
new syntactic category, a CP phrasal category. The new configuration would have the
following format:
(13) CP
Spec C’
C S
This new format presupposes that all main clauses have a complementizer,
something which is debatable since, for example, in English main clauses are not
headed by a complementizer. There are three strong arguments that support the
existence of a C category to introduce clause structures: (i) the existence of lexical
material representing complementizers that introduce main clauses in many different
languages; (ii) the fact that in languages like English the complementizer is reserved for
one particular element in embedded clauses; and (iii) the lexical restrictions of certain
verbs in languages like English that subcategorize for one specific complementizer. Let
us discuss each in turn.
First of all, there is extensive and compelling typological evidence that proves the
occurrence of such category as introductory maker of main clauses, both interrogative,
exclamative and declarative. Suffice the following examples taken from Radford
(1988:296)
Secondly, in close connection with this and as Carnie shows (p.208-209), there is
empirical evidence that suggests that while there are languages which make use of
special morphemes for yes / no questions (cf. the examples above), there are other
languages which use other mechanisms , i.e. subject / auxiliary inversion. In fact, he
shows that the two procedures are in complementary distribution and consequently there
are no languages in which complementizers and auxiliaries occur, that is, we could
assert that both are mutually exclusive. A case in point is English, which lacks special
morphemes for yes / no questions, and then instead it uses auxiliaries. So, the C position
cannot be occupied by both an auxiliary and a complementizer. Consider the following
example:
In the yes/no question example (16a), it is easy to assume that the auxiliary will
occupies the C position because the auxiliary is to the left of the subject. But can we
also assume the same for the auxiliary in the example (16b)? In order to answer this
question, we have included the example (16c). Recall that one of the constituency tests
included coordination and only constituents of the same syntactic category can be
conjoined. If our first assumption was correct and auxiliary will occupies the C position,
we can only conjoin it with another C. In sum we have to conclude that both (16a) and
(16b) are CP examples, as illustrated by the following configuration in (17):
(17) CP
CP Conj CP
but
you are tired will you attend the meeting
Thirdly, we have argued before that the verb said subcategorizes for a particular
clausal complement, that it was raining, but we know that the complementizer that,
which introduces this complement, may be absent in a sentence like Mary said it was
raining. This contrasts with the subcategorization restrictions of other verbs, i.e.
“wonder” as shown in the examples in (18). In such case we may conclude that verbs
like “wonder” not only have to take a clausal complement introduced by a
complementizer as illustrated by the contrast between sentence (18a) and (18b), but
they also choose which complementizer the embedded clause must have, something
which explains the ungrammaticality of the examples (18c) and (18d) in contrast with
the grammatical (18a).
(18) a. Mary wonders whether she will arrive on time
b. *Mary wonders she will arrive on time
c. *Mary wonders that she will arrive on time
d. *Mary wonders for her to arrive on time
CP
Spec C’
What
C S
may
she forget
Continuing with this discussion, if we look back to the structure in (9) above,
which we reproduce here as (20), we still have to deal with the node S.
(20) S NP (T) VP
If we agree that COMP is a head, what is the status of its complement S? Can we
define it as a phrase in X-bar terms? As a first approximation we could agree that T is
the only category that is not phrasal, so, according to X-bar theory format, all non head
material must be phrasal. This signifies that a maximal projection TP could be posited
to replace the S node:
(21) TP
(Specifier) NP T’
T VP (Complement)
Note that the subject NP sits in the specifier of TP. Then, Carnie goes on to
discuss the nature of T. In the first place, it is very plausible to assume that the T node
has verb morphology in the form of Tense and Agr features.
(22) T
[±Tense, ±Present]
[±Agr]
(24) Mary [T' will buy the tickets], but [T' may forget the reception].
In the two grammatical cases of sentential complements of the verbs “say” and
”wonder” discussed above, we have claimed that the complementizer may be either
“that” or “whether”. These two complementizers similarly introduce as their sentential
complements finite sentences whose verb is inflected and therefore finite. On the other
hand, there are other verbs that subcategorize for other sentential complements whose
verb is non-finite or tenseless. A case in point is a verb like “want”, as shown in the
examples in (25). Example (25a) illustrates the fact that this verb subcategorizes for a
sentential complement headed by the complementizer “for”. The contrast of
grammaticality between examples (25a) and (25b) shows that the embedded verb cannot
be inflected. Further the contrast of grammaticality between the examples (25a) and
(25c) crucially indicates that the particle “to” is required when the complementizer is
“for”:
Spec C’
C TP
DP T’
Mary
T V’
-s
V CP
wonder
C’
C TP
whether
she will arrive on time
When “for” in the example I want for her to come acts as a complementizer as in
the tree diagram (27) below, it must introduce a tenseless clause which is headed by the
particle “to” in the T position. In other words, “for” in C subcategorizes for “to” in T.
Unlike in other previous cases, here T doesn’t contain any verb morphology. As a result
its specifier must be represented by an accusative pronoun. Case assignment is one of
the major concerns of Part III of the textbook.
V CP
want
C’
C TP
for DP
her T’
T VP
to
V’
V
come
The assumption here is that the modal “will” in example (28a) starts out from a T
position and moves into C in order to pronounce a [+Q] feature as represented by the
following tree:
(29) CP
C’ TP
C
Will DP T’
you
T VP
---
V’
V DP
attend
the meeting
The endings –ed/-s in the example (28b) are also initially inserted in the T position
but cannot be pronounced there in isolation and need to move downwards to attach to
the verb, as illustrated by the following representation:
(30) TP
DP T’
Mary
T VP
-ed
-s V’
laugh
In sum, the view that X-bar theory can be extended to clausal constituents does
not entail a mere syntactic exercise but captures a number of intriguing facts such as the
position of subjects, the order of auxiliaries and verbal inflection in English, among
other things. We encourage you to do all the exercises and practice the set of rules that
define the X-bar schema. A good summary of this very important Chapter is
exemplified in the final chart provided in Section 5 of the textbook.
14
Note that the examples in (31) as well as the tree diagrams in (32) have been taken from
Haegeman and Guerón (1999: 140-ff), with the difference that we have used the TP node
instead of their IP node. We refer the reader to this source for a more exhaustive
description.
(32) a. CP
C’
C TP
NP T’
You
T VP
will
V’
V NP
invite
Mary
b. CP
Spec C’
When
C TP
will
NP T’
you
T VP
V’
V NP
invite
Mary
c CP
Spec C’
C TP
that
NP T’
you T VP
will
V’
V NP
invite
Mary
d. CP
Spec C’
When
C TP
NP T’
you
T VP
will
V’
V NP
invite
Mary
e. CP
Spec C’
C TP
for
NP T’
you
T VP
to
V’
V NP
invite
Mary
f. CP
Spec C’
When
C TP
NP T’
ec
T VP
to
V’
V NP
invite
Mary
EXERCISES
Now you can do the exercises found at the end of Chapters 1-4 in the textbook. See the
platform for the key to these exercises.