X-Bar Theory (222)

X-Bar Theory

Linguistics 222/322

One of the goals of any theoretical framework is for it be minimal. That is, the less structure the better, as long as the goals of theory are met. In the following discussion on the basic structure of the sentence--the projections of a category--we start with the standard X-bar theory. It is less minimal than the second one we discuss. It is less minimal than the third one briefly discussed last.

In the standard X-bar theory it is assumed that there are three levels: X, X-bar, and XP. The head is always the first (X) level, the phrase is the XP level, and X-bar is an intermediate level. The complement of a head always forms an X-bar level: X-bar --> X plus Comp of X. XP --> Specifier of X plus X. The basic P-rules expanding XP to X-bar and then to X are:

(1) XP --> ({D, NP}) X-bar
(2) X-bar --> X (PP).

(3)

Either a determiner or a possessive NP may function as the specifier of N:

(4) The glass of water

(5) John's glass of water.

The modifier of X is normally adjoined to N-bar:

(6)

The P-rules for NP, for example, including prenominal adjectives are:

(7)     NP --> (D) N-bar (non-iterative)

(8)     N-bar --> N-bar {PP, CP} (iterative)

(9)     N-bar --> AP N-bar (iterative)

(10)     N-bar --> N (PP).

The middle two rules are adjunction rules. All adjunction rules are by nature optional without exception. The two middle rules do not appear to be ordered with respect to each other. It is virtually impossible to determine whether the adjective or adjunct PP has scope over the other. Note that the complement rule (4) is an obligatory rule. The PP complement is optional.

In a somewhat more minimal theory of x-bar, there is no intermediate node. The is the phrasal level and the head level. A modifier is adjoined to XP, either to the right or to the left depending on the parameters of word order of a given language.

(11)

In the NP a determiner and a possessive NP are adjoined to XP like any other modifier: The determiner must take scope over numerals and adjectives; hence it applies first.

In the two-level theory, the following P-rules apply including prenominal AP:

(12)     NP --> D NP (non-iterative)

(13)     NP --> NP PP (iterative)

(14)     NP --> AP NP (iterative)

(15)     NP --> N (PP).

Again, there seems to be no ordering preference for rules (2) and (three).

(16)

In the most minimal form of x-bar theory, all branching is binary. No other kinds of domination is allowed. XP cannot dominate X alone because XP does not branch. In the simplest form, X is both a head and a maximal (phrasal) projection. In the first example, X takes no complement and it is not modified:

(17) X (head, phrase)

The extension of X plus a complement determines the level XP distinct from X:

(18)

If a head without a complement is modified, the modifier is adjoined to X:

(19)

If a head with a complement is modified, the modifier is adjoined to XP:

(20)

Note that modifiers are adjoined the phrasal level. If X has no complement, then it is both a minimal projection (head) and a maximal projection (=XP). For more examples of actual sentences, click on determiner below.

The first of these three theories is the oldest and the most entrenched. There is a trend, however, toward the second theory. We will adopt this theory, since the latter theory is still working is finding some acceptance in the linguistic community. The third, the most optimal theory, is yet to find a noticeable audience in the linguistic community.

The concept of the specifier appears to be the weak link in the three-level theory. It is easy to define configurationally:

(21) Definition of Specifier of X

The specifier of X is the node immediately dominated by the maximal projection of X.

There is nothing wrong with the definiton, per se, but it appears to have no function. Let's first look at one claim for the specifier of N. Ouhalla notes that D, the alleged specifiier of N, is non-iterative in the sense that the rule expanding NP as DP + NP is non-iterative. There is but one determiner that modifies a noun. The adjective is iterative in the sense that the rule expanding NP as AP + NP is iterative. It is true that there is but one determiner that modifies a noun. Now consider the following NP:

(22) the first five eager soldiers in each platoon.

We have considered five a modifier with the property of cardinality (a cardinal number). Cardinal numerals are not iterative:

(23) *the first five 32 eager soldiers in each platoon.

(24) *John ate 4 9 oranges.

Note that some numerals appear to be iterative:

(25) We're going to take the six twenty-eight train to Chatanooga.

Here six twenty-eight is a compound name referring to the traint that leaves at 6:28. 6:28 is not a cardinal numeral in its referrence to train, but it is a set of two cardinal numerals in reference to time, each modifying hour and minute, respectively: 6 hours and 28 minutes after 12 (midnight or noon). The two numerals occur in conjoined structures, which is legitimate.

If there is but one cardinal numeral that can modify a noun (excluding conjunction) then by the same logic applied by determiners, cardinal numerals should be specifiers. Now consider:

(26)      all the first five men in the brigade

(27)     both the (two) women

(28)     any three stories

(29)     ?most five men

(30)     each essay

(31)     some (*five) birds.

All and both are what we will call group-quantifiers. Some linguists call them pre-determiners. They may modify definite or indefinite nouns. The group-quantifiers tend to marks groups of things large or small. For example in:

(32) all the books,

the reference is all the books of a given set. In

(33) all the books in the universe,

the set is all the boks that exist in the universe. Both refers to two in a set. Each refers to single entities in a set. The word any is more interesting. It marks the feautre [-Specific], as opposed to [+Specific]. Given a group of books, any means any one fo them without specific reference to any of them:

(34) John can read any book on the table.

The speaker is not saying that he can read a specific book on the table. It holds true for the entire set, though no one book is specified. The following sentence:

(35) , John can read a book on the table.,

can be interpreted to mean that there is a book on the table that John can read. This is the specific reading.

The remaining quantifiers can only modify indefinite nouns. Most seems to be able to modify a noun modified by a cardinal numeral where the numerals indicates an indefinite group. Some cannot do this where some refers to an indefinite quantity.:

(36) *Mary bought some five eggs for dinner.

Perhaps some is an indefinite numerals rather than a quantifier--this woudl explain why *some five birds is ungrammatical. Numerals specify exact amounts, quantifiers refer to relative quantity. Both occur in (1) and (2) above. Only one quantifier may modify a noun:

(37)     *all any three stones

(38)     *both some birds

(39)     *any both three stories

(40)     *most all books

(41)     *any all lies

(42)     *all all men.

No doubt that these are constrained by semantics, but iteration just simply is not possible. These too would have to be considered specifiers.

Now let us consider adjectives that are not involved with numerals or quantifiers of any sort. First consider:

(43) *the yellow black shirt.

Neither a shirt nor any object be modified by a two colour adjectives that are not conjoined. The following phrase:

(44) the yellow and black shirt,

implies that part of the shirt is yellow and part is black. The previous sentence does not imply. Some colors may occur in compounds spelled with a hyphen:

(45) the blue-green sweater.

Blue-green refers to a shade that occurs somewhere between blue and green but seeming to have characteristics of both colors. Now we may claim that only one colour adjective may modify a noun except for conjunction. Therefore, colour adjectives should be specifiers, too. The same kind of argumentation holds for size:

(46)     *A tall short boy

(47)     *Some fat skinny women

(48)     *many long short streets.

(49)     *The fresh moldy pear.

Size adjectives that refer to same the same physical attribute are non-iterative. These must be specifiers, too. This kind of argumentation applies to most if not all adjective classes.

Go to determiner

Go to adjunction

Go to x-bar theory (L322)

To return to 222 course outline Click here.