Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

Faculty of Letters and Human Sciences,

Mohamed V University,
Department of English,
Rabat.

Semester 5
Group. A, Linguistics
Course: Syntax

X-bar Theory and the structure of phrasal categories in English

Prof. Elhassan Souali

I. Introduction

In this lecture (and the next one), we shall present X-bar theory and show you how it can
account for the structure of phrasal categories in a principled manner. In other lectures, we
shall apply this theory to clausal categories. To gain a satisfactory understanding of this
theory, you should also read the chapters presenting it in your Syntax coursebooks, namely
Radford (1988) and Carnie (2011). I also advise you to read the chapters on this theory in
Haegeman's course book (1994) and other course books you may come across.

X-bar theory is a theory of syntactic structure (phrase structure and clause structure) first
proposed by Chomsky (1970), and later developed by Ray Jackendoff (1977), Chomsky (1981,
1986, ...), and other syntacticians. It can refer either to the theory developed by these scholars
or to the sub-theory of Universal Grammar that consists of universal principles of phrase and
clause structure (See Chomsky 1981).

This theory was developed in the seventies and eighties of the last century to overcome the
weaknesses and drawbacks of Phrase Structure Grammar (Chomsky 1957, 1965, ...), which
made use of many phrase structure rules such as the following, to account for the structure of
each phrasal category and each clausal category:

1) a. S-bar -----------Comp S
b. S ---------------- NP Aux VP
c. NP -------------- Det (AdjP) (NP) N PP S-bar
d. VP ------------- (AdvP) V NP/S-bar) (AdvP)
e. AdjP ------------ (AdvP) Adj (PP)
f. AdvP ------------- (advP) Adv (PP)
g. PP ----------- (AdvP) P NP/S-bar

Among the major drawbacks of this Phrase Structure Grammar, which X-bar
theory sought to overcome, we can mention the following:
- It failed to predict that all phrases (and all clauses), possibly in all human
languages, are underlyingly uniform/similar, and being so, they should be
accounted for by the same universal principles (or metarules (See below)).

- It failed to account for the binary nature of Syntactic structure and for the existence
of what X-bar theory calls "intermediate constituents) (See below).

- It allowed violations of the universal syntactic principles, called the Headedness


Principle, which ensures that all phrases and clauses have an obligatory syntactic
head. The PS rules violating this principle are those accounting for the clausal
categories S and S-bar. In the next lecture on the X-bar Theory account of clausal
categories, we shall show you how this account does not violate the Headedness
Principle.

- It failed to account for first language acquisition, since its account of phrase and
clause structure was not simple and economical, as should be the case, given the
remarkable ease and speed with which all normal children acquire their first
language.

II. A major claim of X-bar theory: Phrase structure (in all languages) is
underlyingly uniform/similar.

As already pointed out above, the major claim of X-bar theory is that all
phrases (and all clauses, as we shall show later) have strikingly similar structure,
which can be accounted for in the same way (i.e., using the same principles or metarules). In
today's lecture we shall limit ourselves to the uniformity of phrase structure, leaving clause
structure to the next lecture. This uniformity concerns the following two major aspects, which
are crucial for your understanding of X-bar Theory and its major claims:

a. All phrases involve the same syntactic functions, namely Head,


Complement, Adjunct and Spec(ifier), which are assigned to the same constituent
type, as we shall show you below.

b. All phrases have an articulate structure involving intermediate


constituents (referred to as X-bar constituents), as we shall show you below.

II.1. All phrases involve the same functions

According to X-bar theory, all phrases involve four syntactic functions, namely:

a. Head

All phrases in all human languages must have one and only one head, as required
by the Headedness Principle. Thus, NP is headed by N, VP is headed by V, AdjP is
headed by Adj, AdvP is headed by Adv, PP is headed by P, etc. According to X-bar
theory, the function Head is assigned to a word-level category, NOT to a phrasal
category, and since word level categories share many syntactic properties (e.g. they
all behave as heads, may assign features such as Case, etc., they are referred to as
the variable X, whose categorial value (N, V, ....) depends on its categorial featur e
specifications. Thus, if the universal category X is specified as being +N, -V, it will be
a noun, if it is specified as –N, +V, it will be a verb, if it is specified as +N, +N, it will
be an adjective, and if it is specified as –N, -V, it will be a preposition (See Chomsky
1981, for details). As we shall see below, the metarules of X-bar theory are
formulated in terms this variable X so as to account for the structure of all phrases
(and clauses, assuming that the latter also have their own heads as we shall see in
the next lecture).

b. Complement

Next to the (universal) function head, which is assigned to the (universal) category
X (with any categorial value), all phrases (in all languages) may involve the
syntactic function "Complement". A complement is an XP constituent (i.e a phrase
or a clause), NOT a word-level constituent (X), which is selected by a specific word
(or head). This kind of syntactic information, technically referred to as
Subcategorization information, is included in the lexical entries of lexical items (or
words). The following are examples of complements of V, complements of N,
complements of Adj, complements of Adv, and complements of P (all put between
brackets):

(i) Complements of V
-NP complement (e.g., I saw [that person])
-PP complement (e.g., He relied [on his father])
-Clausal (S-bar) complement (e.g., I think [that she has left])

II) Complement of N
-PP complement (e.g. The destruction [of the city])
-Clausal (S-bar) complement (e.g., The claim [that he is wrong])

III) Complement of Adj


-PP complement (e.g., She is fond [of her parents])
-Clausal (S-bar) complement (e.g., I am doubtful [whether she understood you]

IV) Complement of Adv


- PP complement (e.g., She made up her mind independently [of him].

V) Complement of P

- NP complement (e.g., He lives in [a big city]


- PP complement (e.g., The baby appeared from [under the table])
- Clausal complement of P (e.g., I am familiar with [what what you said]

C. Adjunct

The thirst syntactic function that any phrasal category (in any language) may
involve is called Adjunct. This syntactic function subsumes such traditional
grammatical functions as premodifiers of N, as in (2a) below, postmodifiers of N, as
in (2b), VP adverbials (Adjuncts and disjuncts), as in (2c,d), modifiers of Adjective,
as (2e) and modifier od Adverb, as in (2f) (See your Syntax course books for more
examples):
2) a. A [very tall] boy.
b. A man [with dark hair]
c. I saw him [in the street]
d. [In fact], he is very smart.
e. The weather is [extremely] cold.
f. He ran [very fast].

The main differences between Adjuncts and complements are (See your
coursebooks for details):

a. Complements tend to be obligatory while adjuncts are always optional.


b. Complements generally have a fixed syntactic position (after the head selecting
them in English), but adjuncts have a relatively free distribution.
c. When a complement and an adjunct occur in the same phrase, the complement
must be closer to the head selecting it than the adjunct (e.g I will see John
tomorrow (not "*I will see tomorrow John").
d. A complement is conjoined with a complement, and an adjunct is conjoined with
an adjunct (e.g., I saw Mary and John yesterday and last week (not " *I saw
Mary, John and last week.).

D) Specifier

The fourth (and last) syntactic function recognized by X-bar Theory is the one
called Specifier. As we shall see, this is the function of the highest constituent in a
phrase or clause. Actually, as we shall see later in this course, the real syntactic
position associated with this function is the one that normally serves as the landing
site for XP movement operations, such as Wh-movement (specifier of CP) and NP-
movement (Specifier of TP). The syntactic function Specifier may also be assigned
to the phrases between brackets in the following examples:

3) a. [These students]' teacher of Physics.


c. [Three miles] under the surface of the ocean.
d. She made [John] leave the meeting.

In (3a), the NP "these students" is the specifier of the whole NP. In (3b), the NP
"three miles is the specifier of the whole Prepositional phrase "Three ....ocean",
headed by the preposition "under". And in example (3c), the NP "John" functions as
the specifier of the VP "John leave the meeting" (referred to as a small clause).

What is also worth noting here is that the functions Complement, Adjunct and
Specifier are all assigned to full phrases or clauses, that is they are assigned to XP
categories, also called maximal projections (See your course books and my
forthcoming lecture for the treatment of clausal categories as XPs).
II.2. All phrases have a binary syntactic structure (They all involve binary
intermediate constituents).

II.2.1. The binary X-bar structure of phrasal categories

Another similarity between all phrasal categories (also shared by clausal categories, as we
shall see in another lecture) is that they all have a binary structure involving the so-called
intermediate constituents (or X-bar constituents, for whose existence we shall argue later).
Thus, according to X-bar Theory, when a head (X), a complement (YP), an adjunct (ZP) and a
specifier (WP) appear in a phrase (any phrase), thy should be merged together in a stepwise
manner. Thus, the head (X) is first merged with its complement to form the lowest
intermediate constituent (or x̄ ) (which is higher than the head and smaller than the full
phrase), then the lowest intermediate constituent is in turn merged with an Adjunct to form a
new intermediate constituent (another higher x̄ ), which is then merged with the specifier to
form the full phrase or maximal projection (XP). These three Merger steps give us the
following X-bar structure, which is clearly binary in nature, since it does not allow more than
two branching nodes at a time:

4) XP

WP (Spec) x̄

x̄ ZP (Adjunct)

X YP (Complement)

The three syntactic operations, which has given us the binary structure/schema in (4) above, are expressed within
X-bar Theory in the form of the following three universal operations or meta-rules (meant to replace the old and
inadequate PS-rules of Phrase Structure Grammar:

5) a. XP ---------- (WP) ; x̄ (The Spec rule)


b. x̄ --------- (x̄ ; ZP) (The Adjunct rule
c. x̄ ---------- X ; (YP) (The Complement rule)

(Where WP is a specifier, ZP is an adjunct and YP is a complement ; and where the semicolon means any
order, to account for phrase structure in different languages, including dead-final languages, where the head
follows its complement, such as Turkish and Dutch). The parentheses mean that the specifier, the adjunct and
the complement are optional, and that the only rule that is optional is the Adjunct rule, the other two rules
being obligatory, that they must operate even in the absence of a complement and an adjunct, as in the
following :

6) XP (Maximale projection)

x̄ (intermediate projection)

Another thing to observe is the only meta-rule that happens to be recursive (i.e.,
applying in an open-ended manner) is the Adjunct rule, which can introduce as many
adjuncts in a phrase as possible. This recursive nature of this rule is crucial given the fact
the number of adjuncts, unlike the number of complements and specifiers, is not limited
in human languages, at least as far as grammatical competence is concerned.

II.2.2 Examples

In what follows, I will give you concrete examples to illustrate the X-bar structure of
phrasal categories in English (NP, VP, AdjP, AdvP and PP). Instead of the bar notation
(x̄), I will use the prime notation (X’), hence I will use N’ for N-bar, V’ for V-bar, Adj’
for Adj-bar, Adv’ for Adv-bar, and P’ for P-bar. I will also use triangles to avoid
irrelevant details.

II.2.2 The Noun Phrase (See Radford (1988) for an excellent chapter on this
phrasal category; please read it and summarize it).

Applying the meta-rules of X-bar theory to the Noun Phrase in English, we get the
following rules accounting for the (binary) syntactic structure of this major phrasal
category :

7) NP ----------- (Spec) ; N’
N’ ------------- (Adjunct ; N’ (or N’ ; Adjunct, if the adjunct follows the noun))
N’ ------------ N ; (Complement)
(Where Spec is NP, for example, Adjunct is AdjP, NP, PP or clause, and
complement is PP or clause ; later we shall give examples involving clauses).

Example 1 : This girl’s honest teacher of History

NP

NP
N’

This girl’s AdjP N’

N PP

Very honest teacher of History

In this noun phrase, the PP complement « of History » is merged with the head noun
« teacher », as a result of the application of the complement meta-rule of X-bar Theory, thus
producing the first/lowest intermediate (N-bar or N’) constituent, which behaves as a unit, as
we shall show you later. Then, this intermediate N-bar constituent is in term merged with the
Adjunct (the AdjP « very honest »), forming the second/higher intermediate (N-bar)
constituent, as a result of the application of the Adjunct meta-rule of X-bar Theory. Finally, this
highest N-bar is merged with the specifier, namely the NP «This girl».
Example 2 : Students of History

NP

N’

N PP

students of History

Example 3 : Students with blue eyes

NP

N’

N’ PP

students with blue eyes

Example 4 : John’s students

NP

NP N’

John’s students

Example 5 : students

NP

N’
N

students

Example 6 : These very clever students of History with dark hair

NP

N’

NP AdjP N’

N’ PP

N PP

This school’s very clever students of History with dark hair

Notice that each adjunct is merged with an x-bar (or X’) constituent to form another X-bar (or
X’) constituent. So the number of X-bar constituents we have in a phrase, without counting the
X-bar constituent resulting from the application of the complement rule, depends on how many
adjuncts we have (and therefore how many times the adjunct meta-rule is activated). If, for
example, we have two adjuncts, as in example (6) above, we should have two X-bar (here two
N-bar) constituents, as shown in the X-bar tree in (6) above. What makes this possible is the
recursive nature of the Adjunct meta-rule, i.e., it is formulated in such a way as to make it apply
as many times as needed (that is as many times as we have adjuncts). Notice that the other two
meta-rules of X-bar theory are not recursive, hence they must apply only once, which is
possible since the number of complements, unlike the number of adjuncts, is fixed in human
language (possibly not more than one complement for each head, in which case the problem
with ditransitive verbs should be resolved; A solution to this problem is dealt with at a more
advanced level of Syntax).

In our next written lecture, we shall deal with the X-bar structure of the remaining
phrasal categories, namely VP, AdjP, AdvP and PP. In this lecture, we shall also provide
linguistic evidence (syntactic, semantic and phonological) in support of X-bar theory and the
binary nature of phrase structure (in English). Needless to remind you that you should read
(and SUMMARIZE) the relevant chapters in your Syntax course books.

Best of luck.

Prof. Elhassan Souali

You might also like