lOMoARcPSD|32882462

Syntax, Arguments for X-bar Theory

English (Université Mohammed-V de Rabat)

Scan to open on Studocu

Studocu is not sponsored or endorsed by any college or university

Downloaded by El anzaoui Abdelilah (abdelilahelanzaoui@gmail.com)
 lOMoARcPSD|32882462

Department of English,
Faculty of Letters and Human Sciences,
Mohamed V University.
Rabat.
Course: Syntax
Prof. Elhassan Souali
2020/21

Lecture 6: Arguments for the existence of X-bar constituents in phrasal

categories

I. Introduction and recapitulation.

Hello, dear students.

I hope you are all well and safe.

In today's lecture, we shall provide some linguistic arguments (or pieces of

evidence) supporting the X-bar theory of phrase structure presented so far in
relation to Phrasal categories (NP, VP, AdjP, AdvP and PP). In a forthcoming
lecture, we shall extend this theory to clausal categories, showing you that they
too have a headed binary syntactic structure, which is determined by the same
metarules/principles of X-bar theory. Before presenting these (conceptual and
empirical) arguments, let me state again the major claim of X-bar theory, namely
that the syntactic structure of all phrases and all clauses is underlyingly
similar/uniform.

This underlying uniformity/similarity concerns two aspects: the first

aspect is that all phrases and all clauses involve the same four syntactic
functions (namely Head, assigned to X (i.e. zero-level/word-level categories),
Complement (assigned to XP categories), Adjunct (assigned to XP categories)
and Specifier (also assigned only to XP categories, according to more recent
versions of X-bar theory, word. The second aspect is that all phrases and all
clauses have a binary syntactic structure involving the so-called
INTERMMEDIATE CONSTITUENTS (or INTERMMEDIATE PROJECTIONS), which
are larger than the Head and smaller than the ordinary phrase (the latter treated

Downloaded by El anzaoui Abdelilah (abdelilahelanzaoui@gmail.com)

 lOMoARcPSD|32882462

as a maximal (or X-double-bar) projection (X'' (or XP). In my previous lecture, I

dealt with these two aspects in detail, limiting myself to phrasal categories.

Concerning the first aspect, i.e. that all phrases and clauses involve the
same functions, I have already provided examples illustrating and supporting it
in the previous lecture, hence I won't argue for it here. Thus, I showed you that
all phrases do indeed involve the same general syntactic functions, which
display the same universal linguistic properties/behavior. For example,
Adjuncts, which subsume both modifiers and adverbials, are all optional and
unlimited in number, unlike complements, which are obligatory (i.e. required by
specific heads/words) and limited in number (two, maximum, if we do not adopt
the Binarity Principle, or one, maximum, if we adopt this principle; but in this
course you are allowed to have two complements in a VP, as is the case with
ditransitive verbs)). See below for other linguistic differences between
complements and adjuncts.

What we need to argue for now is the claim that Phrases do indeed involve
intermediate constituents/projections, e.g. N-bar, V-bar, Adj-bar, Adv-bar, and P-
bar constituents, i.e. that the structure of phrasal categories (we shall show the
same thing for clauses) is binary and therefore more articulate than viewed by
the traditional Phrase Structure Grammar approach that X-bar Theory came to
replace.

II. Linguistic arguments for intermediate X-bar constituents in

phrasal categories

1. Arguments for intermediate N-bar constituents in the Noun

Phrase.

According to X-bar theory, a Noun Phrase like "These very nice students of
Chemistry from Paris" has the following binary syntactic structure involving
hierarchically related intermediate N-bar constituents:

Downloaded by El anzaoui Abdelilah (abdelilahelanzaoui@gmail.com)

 lOMoARcPSD|32882462

1)
NP

N'

N'

N'

Det AdjP N PP PP

These very nice students of Chemistry from Paris

The NP structure in (1) above is the result of the application of the three
metarules of X-bar theory, presented in my previous lecture, and repeated in (2)
below, for convenience. First, the head Noun "students" is merged with its PP
complement "of chemistry", via the complement metarule, to form the
first/lowest N-bar constituent (N'), then this N-bar constituent is in turn merged
with the first Adjunct "very nice" to form another N-bar constituent, via the
Adjunct metarule, then this newly constructed N-bar constituent is merged with
the second Adjunct "from Paris", via the same Adjunct metarule, to form the
highest N-bar constituent in this NP, and finally, and assuming for the moment
that Det is a specifier of NP (as you saw in Grammar 3), this highest N-bar is
merged with this specifier, via the Spec metarule, to form the maximal
projection NP.

2) XP --------> Spec; X'

X' ---------> X'; Adjunct
X' ---------> X ; Complement

When applied to NP, these universal metarules look as follows:

3) NP ----------> Spec ; N'

N' -----------> N' ; Adjunct
N' -----------> N ; Complement

Among the linguistic pieces of evidence (or empirical arguments)

supporting the existence of these intermediate constituents (and therefore X-bar
theory as a whole), there are those that are syntactic in nature, e.g. those
involving the Coordination constituency test and the Pro-form replacement test.

Downloaded by El anzaoui Abdelilah (abdelilahelanzaoui@gmail.com)

 lOMoARcPSD|32882462

Thus, there is evidence that each of the N-bar constituents in English may
undergo coordination with a similar N-bar constituent, as exemplified in (4) and
(5) below, respectively:

4) a. These very nice [N' students of Chemistry] and [N' teachers of

French] from Paris

b. These [N' very nice students of Chemistry] and [N' quite competent
teacher of French] from Paris.

c. These [N' very nice students of Chemistry from Paris] and [N' very
competent teachers of French]

5) a. These very nice [N' students of Chemistry] from Paris are happier
than these very discrete ones from London.

b. These [N-bar very nice students of Chemistry] from Paris are

happier that these ones from London.

c. These [N' very nice students of Chemistry from Paris] are happier than
these ones.

Notice that each of the three intermediate N-bar constituents that

appear in the tree diagram in (1) above may be coordinated with an
identical N-bar constituent, as illustrated in (4a,b,c) above, and may also
be replaced by the proform ONE, as illustrated in (5a,b, c) above. Thus,
both the coordination test and Proform test (ONE-Replacement test)
surely apply to our N-bar intermediate constituents, thus providing us
with linguistic empirical evidence justifying their existence as well as
supporting the adequacy of X-bar theory (See Radford 1988, Haegeman, L.
1994, and Carnie (2006)).

Another syntactic argument concerns the so-called Complement-

Adjunct asymmetries. Thus, the binary articulate structure, where
complements, adjuncts and specifiers occupy different syntactic positions
in the syntactic hierarchy allows us to account for the well-known
differences (or asymmetries) between these syntactic functions, including
Complement/Adjunct asymmetries (or differences), in a principled
manner. Among these differences/asymmetries, we have the following
ones:

Downloaded by El anzaoui Abdelilah (abdelilahelanzaoui@gmail.com)

 lOMoARcPSD|32882462

(i) Complements, unlike adjuncts, have a universally fixed number

(maximum 2, or just 1, if we adopt the Binarity Principle).

It is well known that the number of complements is fixed but the

number of adjuncts is not. For example, while we can say "A very smart
middle class university student with green eyes from London", which
contains four adjuncts, we cannot say "a student of Chemistry of Physics",
since the noun "student" is lexically specified as having only one
complement. This well-known fact can be accounted for by X-bar theory in
a straightforward manner, given the fact that the Complement metarule
applies only once, being non-recursive, whereas the Adjunct metarule can
in principle apply many times (as many times as we have adjuncts), given
the fact that this metarule is recursive (formulated in a recursive manner,
as I have already pointed out).

(ii) Complements are closer to the head selecting them than adjuncts.

Thus, if we force a complement to be introduced in NP structure

after an adjunct, we will get an ungrammatical NP like "*A student from
London of Physics", which cannot be represented by X-bar theory, since
the latter requires that the complement metarule and the adjunct
metarule apply in a universally fixed order, namely Complement metarule
followed by Adjunct metarule (not vice versa).

(ii) Complements must be coordinated with complements, and

adjuncts must be coordinated with adjuncts.

Thus, while we can say "Students of Physics and of Biology from London"
and "Students of Physics from London and from Paris", we cannot say
"*Students of Physics and from London". Thus, while the grammatical
cases can be represented by X-bar theory, the ungrammatical one can't be
so represented.

In addition to purely syntactic arguments supporting X-bar theory

and its claim that the structure of phrases (e.g. NP in the present case) is
binary, involving binary intermediate constituents (e.g., N-bar
constituents), we can add, briefly, both phonological and semantic
arguments or pieces of evidence. Thus, at the phonological level, it is clear
that the so-called intermediate constituents (e.g., N-bar) constituents, are
pronounced as units, which means that they are not only syntactic units
but also phonological units (units of pronunciation). For example, in the
Noun Phrase "[These [[clever [students of Physics]] from Paris]], the
whole NP as well as each N-bar constituent are pronounced as a unit, as
you can verify for yourselves. So a well-formed pronunciation should

Downloaded by El anzaoui Abdelilah (abdelilahelanzaoui@gmail.com)

 lOMoARcPSD|32882462

faithfully respect the syntactic structure of the linguistic construction to be

pronounced, which appears to be the case in all human languages.

At the semantic level, we can also claim that our intermediate

constituents are semantically active as well, hence their existence is
motivated on semantic grounds. The most important semantic aspect that
clearly involves these constituents is semantic modification (or semantic
scope). For example, in our Noun Phrase "These very nice students of
Chemistry from Paris", dealt with above, the demonstrative "These"
modifies (hence has semantic scope over) the highest N-bar constituent
"very nice students of Chemistry from Paris", the AdjP "very nice" modifies
(hence has semantic scope over) the N-bar "students of chemistry", and
finally the PP adjunct "from Paris" modifies (hence has semantic scope
over) the N-bar "very nice students of Chemistry".

Clearly, therefore, N-bar constituents are not only syntactically

active but also phonologically and semantically active. Thus, given their
linguistic activity, we conclude that they are linguistically real and that the
X-bar theory claiming their existence is adequate since it is empirically
motivated and justified.

2. Arguments for V-bar intermediate constituents in the Verb

Phrase

So far, we have presented syntactic, phonological and syntactic

arguments in support for the existence of intermediate N-bar constituents
in the English Noun Phrase, thus confirming the claim of X-bar theory that
the structure of phrasal (and clausal) categories happens to be binary and
more articulate than traditionally believed. In this section, we shall apply
the same set of empirical arguments confirming the existence of
intermediate V-bar constituents in the English Verb Phrase, thus showing
that the Noun Phrase are similarly structured not only because they
involve the same syntactic functions but also because they have the same
internal syntactic design, consisting of hierarchically related intermediate
constituents.

To see how these arguments apply to intermediate constituents in

the Verb Phrase in English, let us consider the VP in the following example
and its corresponding X-bar structure:

6) a. Mary will always read newspapers at home in the morning.

Downloaded by El anzaoui Abdelilah (abdelilahelanzaoui@gmail.com)

 lOMoARcPSD|32882462

b. VP

V'

AdvP V'

V' PP

Always V' PP

V NP at home in the morning

Read newspapers

Notice that the X-bar syntactic structure in (6b) above contains four
intermediate V-bar constituents, namely the lowest V-bar (or V')
constituent "read newspapers", the V-bar immediately dominating it,
namely "read newspapers at home", the immediately higher V-bar
constituent "read newspapers at home in the morning", and finally the
topmost V-bar constituent "always read newspapers at home in the
morning".

Each of these four intermediate V-bar constituents exist because, the

N-bar constituents we dealt with above, they are syntactically active,
phonologically active and semantically active, which means there is
syntactic, phonological and semantic evidence proving their existence.
Syntactically speaking, each of these V-bar constituents may undergo
coordination and may be replaced by a proform (namely do so). The
examples in (7) below may be used to illustrate the coordination
argument (or test):

7) a. Mary will always [read newspapers] and [listen to music] at home in

the morning.

b. Mary will always [read newspapers at home] and l[isten to music in

the garden] in the morning.

c. Mary will always [read newspapers at home in the morning] and

[listen to music in garden in the afternoon].

d. Mary will [always read newspapers at home in the morning] and

[never watch TV in her bedroom at night].

Downloaded by El anzaoui Abdelilah (abdelilahelanzaoui@gmail.com)

 lOMoARcPSD|32882462

In sentence (7a), the lowest V-bar, i.e. "read newspapers" has been
conjoined with a identical V-bar constituent. In sentence (7b), the V-bar
"read newspapers at home" has also been conjoined with an identical
constituent. The same also applies to the V-bar constituent "read
newspapers at home in the morning" in sentence (7c) and to the V-bar
constituent " always read newspapers at home in the morning", which
have also been conjoined with an identical V-bar constituent. Since only
full constituents may undergo coordination, we can safely conclude that
the V-bar intermediate constituents are real constituents of VP structure.

Concerning the Pro-form replacement argument, it also strongly

supports the existence of the so-called intermediate V-bar constituents, as
shown by the following examples:

8) a- Mary will always [read newspapers] at home in the morning, and

her husband will always do so in the library in the afternoon.

b- Mary will always [read newspapers at home] in the morning, but her
husband will always do so in the afternoon.

c- Mary will always [read newspapers at home in the morning] but her
husband will never do so.

d- Mary will [always read newspapers at home in the morning] but her
husband will not do so.

Notice that in each of the four sentences in (8) above, the pro-form "do so"
has been rightly used to replace/refer back to a specific V-bar constituent, thus
clearly confirming the existence of the so-called V-bar constituents as instances
of X-bar constituents, like N-bar constituents.

Finally, the same V-bar constituents can be confirmed by both

phonological and semantic evidence, given the fact that they are phonologically
active and semantically active, beside being syntactically active, exactly like N-
bar constituents. At the phonological level, each of these intermediate
constituents is pronounced as a unit, as you can verify for yourselves. And at the
semantic level, each one of the same intermediate V-bar constituents may be
involved in semantic relations, given the fact that it can be modified by another
constituent that is higher than it in the hierarchical syntactic structure. For
example, in the example given above, the lowest V-bar "read newspapers" is
semantically modified by the adjunct "at home", the V-bar constituent "read
newspapers at home" is semantically modified by the adjunct "in the morning",
and the V-bar constituent "read newspapers at home in the morning" is
semantically modified by the highest adjunct "always". Concerning the highest

Downloaded by El anzaoui Abdelilah (abdelilahelanzaoui@gmail.com)

 lOMoARcPSD|32882462

V-bar in this VP, we can say that it is semantically related to the next higher
constituent, which can be the subject (Mary) in its underlying position (claimed
to be the specifier of VP), or possibly the auxiliary "will".

3. Intermediate constituents in other phrasal categories

So far, we have presented different types of linguistic arguments (syntactic,

phonological and semantic ones) for the existence of intermediate X-bar
constituents in the Noun Phrase and the Verb Phrase, i.e. constituents that are
larger than the head but smaller than the phrase (the latter also called maximal
projection, within the framework of X-bar Theory). Today, we shall extend these
arguments to intermediate constituents in other phrasal categories, namely in
the Adjective Phrase, the Adverb Phrase and the Prepositional Phrase.

3.1. Intermediate constituents in AdjP

As we have already seen before, the Adjective Phrase in English, like all other
phrases, involve the same types of syntactic functions (Head, complement,
Adjunct and Specifier) which are merged in a step by step fashion, as a result of
the application of the (obligatory) complement metarule (even in the absence of
a complement), creating the lowest Adj-bar, immediately followed by the
(optional) Adjunct metarule (if there is an adjunct), producing another Adj-bar
(i.e. a copy of the lowest Adj-bar), finally followed by the (obligatory) Specifier
metarule (even in the absence of a specifier), creating the maximal projection
(AdjP). Thus, we have seen that an Adjective Phrase like the one put between
brackets in sentence (1a) below has the X-bar syntactic structure in (1b):

Downloaded by El anzaoui Abdelilah (abdelilahelanzaoui@gmail.com)

 lOMoARcPSD|32882462

1) a. Mary is [extremely fond of her parents].

b. AdjP

Adj'

AdvP Adj'

Adv Adj PP

Extremely proud of her parents

In this tree diagram, we notice that the Adj and its PP complement are
merged to form an intermediate Adj-bar (Adj') constituent, exactly like the
intermediate N-bar "students of Physics" in the Noun Phrase "These happy
students of Physics" and the intermediate V-bar constituent "read books"
in the verb phrase " always read books in the summer", as we have
already demonstrated in previous lectures.

Among the argument confirming the existence of our intermediate

Adj-bar constituent in Adjective Phrases like the one in sentence (1)
above, we can mention syntactic arguments, phonological arguments and
semantic arguments.

Syntactically speaking, our lowest Adj-bar constituent in (1b) above

exists as a full syntactic constituents because it is syntactically active with
respect to coordination and to pro-form replacement, as shown by the
following examples:

2) a. Mary is [extremely [fond of her parents]] and [proud of them].

b. Mary is [extremely [fond of her parents ]] but her brother is not very much so.

Notice that in (1a) above, our lowest Adj-bar constituent (consisting of the
Adjective and its PP complement) has been conjoined with a similar Adj-
bar constituent, namely "proud of them", thus clearly indicating that both
strings are full syntactic constituents, which are larger than the Adj head
and smaller than the higher Adj-bar, resulting from the application of the
Adjunct metarule, and maximal projection AdjP, resulting from the
application of the specifier metarule.

Downloaded by El anzaoui Abdelilah (abdelilahelanzaoui@gmail.com)

 lOMoARcPSD|32882462

In sentence (2b), the same intermediate Adj-bar constituent,

resulting from the application of the Complement metarule, is
syntactically active with respect to pro-form replacement, since it has
been replaced by a proform, namely 'so'. Since only full constituents may
be replaced by proforms, we conclude that "fond of her parents" in the
Adjective Phrase "extremely fond of her parents" is a full syntactic
constituent, referred to here as Adj-bar (or Adj').

Other arguments supporting the existence of the same intermediate

constituent are phonological and semantic in nature. Thus, at the
phonological level, this intermediate Adj-bar constituent, like all other
syntactic constituents in human language, are pronounced as phonological
units. As you can verify for yourselves, our Adj-bar 'fond of her parents" is
surely pronounced as a unit, including the adjective and its PP
complement, and excluding the Adjunct 'extremely', the latter being
pronounced as part of the highest Adj-bar to which it immediately
belongs. And the semantic level, the same Adj-bar constituents like "fond
of her parents", mentioned above, is semantically active with respect to
what we have already referred to as semantic scope, like all other full
syntactic constituents in English (and other human languages). Thus, in
the sentences in (2) above, the AdvP "extremely" clearly modifies and
therefore has semantic scope over our Adj-bar 'fond of her parents".
Again, since only full syntactic constituents may serve as domains of
semantic scope, we conclude that our Adj-bar is a full syntactic
constituents.

3.2. Intermediate constituents in the Adverb Phrase and the

Prepositional Phrase

Now, let us present arguments presented in support for the existence of

intermediate X-bar constituents may be extended to the intermediate Adv-
bar constituents and intermediate P-bar constituents, as shown by the
following examples:

3) a. He made up his mind [quite [independently of his parents] ]

b. I will put it [right [off the table]]

Notice that in sentence (3a), the string 'independently of his parents' is

treated as a syntactic unit, namely as the Adj-bar resulting from the
application of the Complement metarule. And in sentence (3b), the string
'on the top shelf' is also treated as a syntactic unit, namely P-bar resulting
from the application of the same complement metarule. Evidence in

Downloaded by El anzaoui Abdelilah (abdelilahelanzaoui@gmail.com)

 lOMoARcPSD|32882462

support of these intermediate constituents may involve the syntactic

operations of coordination, as shown by the following examples:

4) a. He made up his mind [AdjP [Adj' quite [[Adj' [Adj autonomously]] and [Adj' independently
of his parents]].

b. He will put it [PP right [P' [P' off the table] and [P' onto the floor]]].

Notice that in sentence (4a) above, the intermediate Adv-bar constituent

'independently of his parents' has been properly coordinated with a
similar intermediate Adv-bar constituent, namely 'autonomously' (which
behaves as an adverb and an adv-bar in this example), clearly indicating
that both coordinated strings behave as full constituents with respect to
coordination.

Concerning the pro-form substitution argument (or test), it applies

especially to the P-bar intermediate constituent, as indicated by the
following example:

5) He will not put his book right onto the top shelf, but his wife will put it right there.

In this sentence, the adverb 'there, which functions as a P-bar proform, has
been used to properly replace our P-bar intermediate constituent 'right
onto the top shelf', rightly showing that this string behaves as a full
constituent, with respect to this Proform substitution test.

Finally, as you can verify for yourselves, the Adv-bar and P-bar
intermediate constituents presented so far may be confirmed by both
phonological and semantic evidence, since each one of them may be
pronounced as a phonological unit and each one of them may serve as a
syntactic domain of semantic scope. For example, in sentence (3a) above,
the Adv-bar constituent is semantically modified (and hence serves as a
domain of semantic scope) for the AdvP 'quite' (notice that in this case
this degree AdvP does not modify the adverb 'independently' alone, but it
rather modifies the full Adv-bar 'independently of his mother'). The same
arguments apply to the P-bar constituent, like the one in sentence (3b)
above, where this intermediate constituent is pronounced as a unit and
also serves as a syntactic domain for the semantic scope of the AdvP 'right'
(again, notice here that this adverb phrase modifies (and hence has
semantic scope over) over the whole P-bar 'off the table').

In the next lecture, we shall deal with the syntactic structure of

clausal categories. More specifically, we shall argue that all clauses in
English (and other languages) essentially have the same syntactic

Downloaded by El anzaoui Abdelilah (abdelilahelanzaoui@gmail.com)

 lOMoARcPSD|32882462

structure as ordinary phrasal categories. In the other words, we shall

demonstrate that the same meta-rules we used to account the syntactic
structure of ordinary phrases may safely be used to account for the
syntactic structure of clausal categories (assuming that each of the two
major clausal categories S and S-bar are headed by the functional heads
T(ense) and C(omp), respectively).

Downloaded by El anzaoui Abdelilah (abdelilahelanzaoui@gmail.com)