Japanese Linguistics

1

Ling 301 Introduction to Japanese Linguistics HO7 Motohashi

5. Syntax

There are infinite number of sentences in a language.

Native speakers of a language subconsciously knows which sentence is grammatical or not.

Syntax deals with how the words are put together to create grammatical sentences.

5.1. Syntactic Structures

5.1.1 Syntactic constituency: which word is combined with which word to form a larger unit

(1) 赤 あか

い本 ほん

の表紙 ひょうし

(ambiguous)

a. [[赤 あか

い本 ほん

]の表紙 ひょうし

] ‘the cover of a red book’

b. [赤 あか

い [本 ほん

の表紙 ひょうし

]] ‘a red cover of a book’

(2) Principle of Compositionality

The meaning of a sentence is determined by the meaning of its component parts and the manner in which

they are arranged in syntactic structure.

(William O'Grady, John Archibald, and Francis Katamba (2011) Contemporary Linguistics, p. 214.)

5.1.2 Phrase structures

A common device to represent syntactic constituency: phrase structure trees or labeled bracketing

a) Phrase structure tree:

(3) NP

NP N

AP N 表紙 ひょうし

赤 あか

い 本 ほん

Note: AP is a short-hand drawing for: AP

赤 あか

い A

赤 あか

い

b) Labeled bracketing:

(4) [[ [AP赤 あか

い][N本 ほん

]][N表紙 ひょうし

]]

They represent:

i) syntactic constituency: which word is combined with which word to form a larger unit

ii) labeling of categories, lexical and phrasal (N, V, A; NP, VP, AP)

iii) right/left order: 赤 あか

い precedes 本 ほん

. 赤 あか

い本 ほん

is followed by 表紙 ひょうし

.

2

(5) うちで ‘at home’ (6) 新 あたら

しい本 ほん

を買 か

う ‘buy a new book’

PP VP

NP P NP V

N de AP N ka-u

uti A hon-o

atarasi-i

(7) 太郎 た ろ う

が学校 がっこう

で 新 あたら

しい本 ほん

を買 か

った。 ‘Taro bought a new book at school.’

S

NP VP

Taro-ga PP NP V

NP P AP N kat-ta

gakkoo de atarasi-i hon-o

5.1.3 Phrase Structure Rules: should have power to express an unlimited number of sentences and phrases; should be

restrictive enough to be able to characterize the set of sentences that does not belong to the language

(8) a. S → NP VP

b. NP → (NP)(AP) N

c. AP → (Deg) A

d. VP → (PP)(NP)(PP)(NP) V

e. PP → NP P

In (8a) S (the sentence) is rewritten as NP (noun phrase) and VP (verb phrase). NP and VP are ordered (i.e.,

NP precedes VP). So → (the arrow) defines the ‘is a’ relation (domination). In (8b) NP is rewritten as the

optional NP, optional AP (adjectival phrase), and the obligatory N (noun). The elements enclosed by the

parentheses are optional, otherwise they are obligatory (see (8a). NP, AP, and N are ordered. In (8c) AP is

3

rewritten as optional Deg (degree word) and A (adjective). They are ordered, too. In (8d) VP is rewritten as

optional PP (postpositional phrase), optional NP, optional PP, optional NP, and V (verb). They are all ordered.

IN (8e) PP is rewritten as NP and P (postposition). They are ordered.

(9) a. 学生 がくせい

が来 き

た。(Students came.)

b. 子供 こ ど も

が公園 こうえん

で遊 あそ

んだ。(Children played in the park.)

c. 太郎 た ろ う

が花子 は な こ

と東 京 とうきょう

へ行 い

った。（Taro went to Tokyo with Hanako.)

d. 太郎 た ろ う

が魚屋 さかなや

で古 ふる

い 魚 さかな

を買 か

った。(Taro bought old fish at the fishmonger.)

e. 次郎 じ ろ う

が正子 ま さ こ

に指輪 ゆ び わ

をあげた。(Jiro gave a ring to Masako.)

f. 太郎 た ろ う

が花子 は な こ

に宅急便 たっきゅうびん

で本 ほん

を送 おく

った。(Taro sent a book to Hanako by delivery service.)

To generate (explicitly characterize) (9a), apply (8a) S → NP VP. You always start with (8a). The symbol

S in (8a) is sometimes called the initial symbol. Then you apply (8b) NP → N (in (9a) the subject consists only of a

noun gakusei(-ga) ‘student’). Then you apply (8d) VP → V (in (9a) the predicate consists only of a verb ki(-ta)

‘came’). We will see how case particles and tense suffixes (functional categories) are introduced to a sentence later.

So you have the following phrase structure tree.

(9a)’ S

NP VP

N V

gakusei(-ga) ki-ta

To generate (9b), you apply (8a) (S → NP VP) and (8d) (VP → PP V). Then you apply (8e) (PP → NP P) and

(8b) (NP → N). Because NP’s appear twice in (9b) you apply (8b) after (8d). So (8b)-(8d) are unordered. You

apply the rule whenever you need to. Then you have the following phrase structure tree.

(9b)’ S

NP VP

N PP V

kodoo-ga NP P ason-da

N de

kooen

To generate (9c), apply (8d) VP → PP PP V and other relevant rules. You get the following:

(9c)’ S

NP VP

Taro-ga PP PP V

Hanako-to Tokyo-e it-ta

To generate (9d), apply (8d) VP → PP NP V and (8c) AP → A other relevant rules. Then you get the following:

4

(9d)’ S

NP VP

Taro-ga PP NP V

sakanaya-de AP N kat-ta

A sakana-o

huru-i

To generate (9e), apply (8d) VP → NP NP V and other relevant rules. Tsujimura takes 正子 ま さ こ

に as a dative NP.

Then you have the following phrase structure tree:

(9e)’ S

NP VP

Jiro-ga NP NP V

Masako-ni yubiwa-o age-ta

However, applying the three tests that distinguish postpositions and case particles, it seems that the ni-phrase in (9e)

seem to be a postpositional phrase:

(10) 次郎 じ ろ う

が三人 さんにん

の学生 がくせい

*（に）指輪 ゆ び わ

をあげた。‘Jiro gave a ring to three students.’

In (10) the postposition ni following gakusei ‘student’ is dropped and the sentence is ungrammatical.

(11) *次郎 じ ろ う

が学生 がくせい

に三人 さんにん

指輪 ゆ び わ

をあげた。‘Jiro gave a ring to three students.’

In (11) Numeral Quantifier sannin ‘three (people)’ is moved across the postpositional phrase gakusei-ni and the

sentence is ungrammatical.

So we revise (8d) to the following:

(8d)' VP → (PP) (PP) (PP) (NP) V

To generate (9f), apply (8d)' VP → PP PP NP V and other relevant rules. You get the following structure:

(9f)’ S

NP VP

Taro-ga PP PP NP V

Hanako-ni takkyuubin-de hon-o okut-ta

So far we have been able to generate simple sentences. How should we generate such complex sentences as (12a and

b)?

(12) a. 花子 は な こ

が太郎 た ろ う

が作 つく

った寿司 す し

を食 た

べた。’Hanako ate the sushi that Taro made.’

b. 太郎 た ろ う

が作 つく

った寿司 す し

が腐 くさ

った。’The sushi that Taro made got rotten.’

In (12a) Taro-ga tukut-ta ‘Taro made’ is added to the object sushi. In (12b) Taro-ga tukut-ta ‘Taro made’ is added to

the subject sushi. Therefore, we can add a clause to NP (subject and object) we can generate (12a and b). We

expand phrase structure rules to include (13):

(13) NP → (S) (NP) (AP) N (S): relative clause

That is, we add the optional S (sentence) to (8b). We call the clause inside a noun phrase a relative clause. To

expand the object noun phrase we apply (8a) again. So we apply (8a) recursively. Whenever you have a S inside a

5

phrase you go back to (8a) and apply the relevant rules all over again. So (12a) has the following underlying structure.

(12a)’ S

NP VP

Hanako-ga NP V

S N tabe-ta

NP VP sushi-o

Taro-ga NP V

tukut-ta

Notice that the unexpressed object of the verb tukut-ta ‘made’ is the same as the object sushi-o of the verb tabe-ta ‘ate’.

Similarly to generate (12b) you apply (16) and other relevant rules. You get the following structure:

(12b)’ S

NP VP

S N V

NP VP sushi-ga kusat-ta

Taro-ga NP V

tukut-ta

Japanese relative clauses:

i) noun modifying clause must appear to the left of the head noun within NP

ii) there is no relative pronoun in Japanese (Compare: Hanako ate the sushi (that/which) Taro made)

iii) there is an empty NP inside the relative clause which is co-indexed with the head noun

NP

S Ni

NP VP sushi-o Ni sushi and empty NPi are co-indexed.

Taro-ga NPi V

φ tukut-ta

Phrase structure rule (13) does not generate the following sentences:

(14) a. 太郎 た ろ う

が花子 は な こ

がおいしい寿司 す し

を作 つく

ったと言 い

った。

‘Taro said (that) Hanako made some delicious sushi.’

b. 太郎 た ろ う

が花子 は な こ

に先生 せんせい

が東 京 とうきょう

へ行 い

ったと言 い

った。

‘Taro said to Hanako (that) the teacher went to Tokyo.’

In (14a) there is an extra pair of the subject and the predicate in addition to the main clause Taro-ga … it-ta ‘Taro said

…’. Similarly, in (14b) the main clause Taro-ga Hanako-ni … it-ta ‘Taro said to Hanako …’ contains the subordinate

clause sensei-ga Tokyo-e it-ta ‘the teacher went to Tokyo.’ These subordinate clauses cannot be introduced by PS rule

6

(13) because they do not appear under a noun phrase. So suppose that the subordinate clause is the object clause of the

main (transitive) verb it-ta ‘said’. Then we expand (8d)' to (22).

(8d)' VP → (PP)(PP)(PP)(NP) V

(15) VP → (PP) (PP) (PP) (NP) V

(S)

(16) a. 太郎 た ろ う

が木曜日 もくようび

に図書館 としょかん

で花子 は な こ

に英語 え い ご

を話 はな

した。

'Taro spoke English to Hanako in the library on Thursday.'

b. 太郎 た ろ う

が木曜日 もくようび

に図書館 としょかん

で花子 は な こ

に先生 せんせい

が東 京 とうきょう

へ行 い

ったと話 はな

した。

'Taro said to Hanako in the library on Thursday that the teacher went to Tokyo.'

In (15) the optional S (a subordinate clause) is introduced inside a VP. That is,

(17) VP

S V

We call this S inside a VP as a complement clause. Notice that the complement clause in (16b) sensei-ga Tokyo-e it-ta

‘the teacher went to Tokyo.’ form a constituent with to. We call this particle as a complementizer (a morpheme that

introduces a complement clause). We need revise (8a) to include the complementizer.

(18) S’ → S COMP (S’ is ‘S prime or S bar’; COMP is a complementizer)

To generate (16b) we apply (15): VP → PP S V, (18): S' → S COMP, and other relevant rules and we get the

following VP.

(19) VP

PP S' V

Hanako-ni S COMP hanasi-ta

NP VP to

sensei-ga Tokyo-e it-ta

5.1.4 Head and Specifiers

A phrase consists of the head (lexical category) and the preceding grammatical category called a specifier.

Heads: define the category and core meaning

Specifiers: semantically specifiers help to make more precise the meaning of the head: syntactically specifiers

typically mark a phrase boundary:

(20) VP

NP V'

PP NP V

Specifier Complements Head

Subcategorize

7

Agree

(21) a. あの人 ひと

'that person'

b. *人 ひと

あの

In (21a) the head is a noun hito 'person' and its specifier is a demonstrative ano 'that' (the specifier preceding a noun will

be called a determiner from now on). The order of the determiner and the noun is fixed in Japanese: the determiner

must precede the noun. Otherwise the phrase will be ungrammatical (21b).

In (22a) the head is the verb mi-ru 'see' and its specifier is the auxiliary (y)oo 'will'. The auxiliary must follow the verb

in Japanese. Otherwise the phrase will be ungrammatical (22b).

(22) a. 見 み

よう'will see' (In Japanese the auxiliary (y)oo follows the imperfective form.)

b. *よう見 み

Notice that the auxiliary follows the verb. Later we will treat the order of Aux and V uniformly like Det and N:

(23) a. b.

NP VP

Det N' Aux V'

ano hito yoo mi

move

See Richard S. Kayne (1994) The Antisymmetry of Syntax. MIT Press.

In (24a) the head is the adjective kanasi-i 'sad' and its specifier is the degree word totemo 'very'. The degree word must

precede the verb. Otherwise the phrase will be ungrammatical (24b).

(24) a. とても悲 かな

しい 'very sad'

b. *悲 かな

しいとても

We call the phrase (21a) a noun phrase (NP) because its head is a noun. We call (22a) a verb phrase because its head

is a verb. We call (24a) an adjective (adjectival) phrase because its head is an adjective.

Notice that there is a mutual selectional relation between the head and its specifier. If you have a noun as the

head, its specifier must be a determiner. The wrong choice will result in an ungrammatical phrase. In (25b) the

auxiliary (y)oo 'will' follows a noun hito 'person'. The auxiliary cannot be the specifier of a noun therefore (25b) is

ungrammatical. Similarly, in (25c) the degree word totemo 'very' precede a noun hito 'person'. The degree word

cannot be the specifier of a noun therefore (25c) is also ungrammatical. (16) and (27) show the same point: there is a

mutual selectional relation between the head and its specifier. We will call this relation the specifier-head agreement

(Spec-head agreement).

(25) a. あの人 ひと

'that person'

b. *よう人 ひと

'*will person'

c. *とても人 ひと

'*very person'

(26) a. 見 み

よう'will see'

b. *とても見 み

（る）

c. *あの見 み

（る）

(27) a. とても悲 かな

しい 'very sad'

b. *あの悲 かな

しい

c. *悲 かな

しいよう'*will sad' (kanasi-i-yoo is acceptable if it means 'how sad (I am)!')

8

(28) a. NP b. VP c. AP

Det N' V' Aux Deg A'

ano hito mi yoo totemo kanasi-i

Later we will revise (28) and take the specifier to be a NP which is the 'subject' of NP, VP, and AP.

5.1.4.1 Head

The phrasal categories are built around the lexical categories of the same type:

NP → ... N, AP → ... A, VP → ... V, PP → ... P

i) The heads are always present (obligatory).

ii) The type of meaning associated with the head is always associated with the phrase.

If the head is a noun, the (noun) phrase denotes an entity. If the head is a verb, the (verb) phrase designates

an action/event/state. If the head is an adjective, the (adjectival) phrase designates a property or state.

5.1.5 Subcategorization: specifies the number and/or type (or category) of constituents (besides subject) with which an

individual verb appears

There is an important syntactic relation that holds between the obligatory head and the complements (the phrasal

categories that appear within the VP). The number and the type of complements subcategorize the head. The

prototypical subcategorization is the transitive verb: if a verb is subcategorized by a noun phrase (the object), the verb

is a transitive verb. In subcategorization we extend this relation not only to verbs but also to other lexical categories so

that a head is subcategorized, and furthermore, not only a noun phrase but also any phrasal categories subcategorize the

head. See Notes to HO7.

(29) a. 太郎 た ろ う

が死 し

んだ。'Taro died.'

b. 花子 は な こ

が寿司 す し

を食 た

べた。'Hanako ate sushi.'

c. 次郎 じ ろ う

が棚 たな

に本 ほん

を置 お

いた。'Jiro put a book on the shelf.'

In (29a) the VP consists only of the head V. So the verb sin-u 'die' is subcategorized by no complement. In (29b) the

VP consists of the noun phrase susi-o and the head V tabe-ta 'ate', so that the verb tabe-ru 'eat' is subcategorized by a

noun phrase (NP). In (29c) the VP consists of a postpositional phrase (PP) tana-ni 'on the shelf', a noun phrase (NP)

hon-o 'a book', and the head V oi-ta 'put', so that the verb ok-u 'put' is subcategorized by PP and NP. We represent

such information in the lexicon (the mental dictionary):

(30) a. sin 'die' [+ V] : [ ___ ]

(sin is the phonemic (underlying) representation of the verb sin-u. 'die' is the meaning of the verb sin-u.

[+V] represents the lexical category, that is sin is a verb. [ ___ ] represents the subcategorization frame:

no complement (nothing is enclosed by [ ]) subcategorizes the head sin.

b. tabe 'eat' [+V] : [ NP __ ]

(The subcategorization frame [ NP __ ] represents that NP preceded the head (indicated by the

underline), so that the verb tabe is subcategorized by (a single) NP.

c. ok 'put' [+V] : [ PP NP __ ]

(The subcategorization frame [ PP NP __ ] represents that PP and NP precede the head. That is, the

9

verb ok 'put' is subcategorized by PP and NP.)

Subcategorization explains why the following sentences are ungrammatical.

(31) a. *太郎 た ろ う

が寿司 す し

を死 し

んだ。'*Taro died a sushi.'

b. *花子 は な こ

が食 た

べた。'*Hanako ate.'

c. *次郎 じ ろ う

が本 ほん

を置 お

いた。'*Jiro put a book.'

In (31a) the VP consists of NP (susi-o) and V (sin-da), but the verb sin 'die' is subcategorized by no complement, so that

(31a) is marked ungrammatical (*). In (31b) the VP consists only of the verb (tabe-ta). However, the verb tabe 'eat'

is subcategorized by a NP so that (31b) is marked ungrammatical. Hanako-ga tabe-ta can be interpreted as a case of

Object Drop: in Japanese the object is freely dropped and the sentence is interpreted as 'Hanako ate it.' （花子 は な こ

がそれ

を食 た

べた。）Later we analyze this Object Drop as involving an invisible pronoun (invisible sore) and it is called null

anaphora. In (31c) the VP consists of a NP and V, but the verb ok 'put' is subcategorized by PP and NP, so that (31c)

is marked ungrammatical. However, similar to (31b), Jiro-ga hon-o oi-ta can be interpreted as a case involving null

anaphora: the invisible soko-ni 'there' is present in (31c).

5.1.6 Structural relations

Phrase structure trees consist of the nodes, branches, labels (categories).

A node refers to a labeled branching point in tree diagram. Branches are the lines that connect the nodes.

Labels designate the categories of the node (a constituent).

(32) ・← VP (・is a node; VP, S, and V are labels; and are branches.)

S → ・ ・← V

Pharse structure tress consist of the following relations.

(immediately) dominate, precede, sisters

If the two nodes are connected by a branch then the higher dominates the lower. That is, in (32), VP dominates

S and V. If there is no other node between the two connected nodes, the higher immediately dominates the

lower.

(33) VP

S' V

S COMP

In (33) VP dominates S', S, COMP, and V. VP immediately dominates S' and V; S' immediately dominates S

and COMP.

If a node A occurs anywhere to the left of another node B, A precedes B. In (33), S' precedes V, S precedes

COMP, and S and COMP precede V. If there is a domination relation between the two nodes, there is no

precedence relation between them. In (33) S does not precede S' or VP. S' does not precede VP.

If the two nodes are immediately dominated by a node (sometimes called the mother), they are sisters. In (33) S'

and V are sisters, and S and COMP are sisters.

See David Crystal (1985) A Dictionary of Linguistics and Phonetics. Blackwell.

10

5.2. X-Bar Theory

Because phrase structure rules have the following defects, they are revised by X-Bar Theory (or X-bar rules).

5.2.1. Phrase Structure Rules

a) The category of the phrasal level is determined by the category of the head. However, phrase structures rules

connot exclude such rules in (ii) that violate this principle:

i) NP → ... N, AP → ...A, VP → ... V, PP → ... P:

ii) *NP → ...A, *AP → ... V, PP → ... N, *NP → ...A, etc.

X-Bar Theory expresses this principle by using a variable X, whose value comprises N, A, V and P (subject to

language-variation).

NP → …N

AP → …A

VP → …V

PP → …P

__________

XP → ... X

b) There is an intermediate category between the phrasal and lexical categories (the existence of X’ category).

(34) a. John met a student of physics with long hair,

and Mary met one with short hair

In 34a) the pronoun one substitutes a student of physics and according to a test for constituency

('only constituents can be substituted'), a student of physics is a constituent. We call this category

X' (X-bar or X-prime) category. Because the head is a noun student, a student of physics is N'. So

we have the following categories now:

XP (phrasal category) → X' (intermediate category) → X (lexical category)

b. 太郎 た ろ う

は赤 あか

いイタリアの 車 くるま

を買 か

った。‘Taro bought a red Italian car.'

花子 は な こ

も青 あお

いのを 買 か

った。 'Hanako bought a blue one, too.’

In (34b) the pronoun no substitutes itaria-no kuruma 'an Italian car'. Therefore, according to the

test for constituency, itaria-no kuruma 'an Italian car' is a constituent. This constituent is larger

than a lexical category (kuruma 'car'is a noun) but smaller than a phrasal category (aka-i itaria-no

kuruma 'a red Italian car' is a noun phrase) In aka-i itaria-no kuruma 'a red Italian car', kuruma

'car' is the head, so that itaria-no kuruma 'an Italian car' is N' (aka-i itaria-no kuruma 'a red Italian

car' is N'').

11

(35) a. John will watch a movie today,

and Mary will do so tomorrow.

In (35a) the pronominal do so substitutes watch a movie. It is larger than a verb watch and smaller

than a verb phrase will watch a movie tomorrow. So watch a movie is V'.

b. 太郎 た ろ う

は今日 き ょ う

映画 え い が

を見 み

るだろう。'Taro will watch a movie today.'

花子 は な こ

も明日 あ し た

そうする だろう。'Hanako will do so tomorrow.'

In (35b) the pronominal soo suru 'do so' substitutes eiga-o miru 'watch a movie'. It is larger than a

verb miru 'watch' but smaller than a verb phrase kyoo eiga-o miru-daroo 'will watch a movie'. So

eiga-o miru 'watch a movie' is V'.

c) The internal structure of phrasal categories is the same.

A phrase consists of the three elements: specifier (optional), head (obligatory), and complements (optional)

The internal structure of phrases is basically the same: They consist of the three elements with basically the

same order:

(36) For Japanese, X” → … X’ (X'' is a phrasal category. X' is an intermediate catgory.)

(... is called the specifier position)

X’ → … X (X' is an intermediate category. X is a lexical category.)

( ... is called the complement position)

X” = phrasal category

Spec X’ = intermediate category

Comp Head = lexical category

(37) a. 太郎 た ろ う

（が）本 ほん

（を）読 よ

む (V") ‘Taro reads a book.’

b. 家 いえ

（が）海 うみ

に 近 ちか

い (A") ‘The house is near the sea.’

c. 太郎 た ろ う

（の）株 かぶ

（の）取 と

り引 ひ

き (N") ‘Taro’s transaction of stocks’

d. 東 京 とうきょう

から (P" with Spec absent) ‘from Tokyo’

In (37a) the NP Taro occupies the specifier position and the NP hon 'book' occupies the complement

position (notice that the NP hon subcategorizes the head V; yom-u 'read' is a transitive verb that selects a NP)

and the head is a V yom-u 'read'. Because the head is a V, Taro(-ga) hon(-o) yom(-u) is a VP. We will

see how functional categories such as case particles ga and o and non-past tense marker u are introduced into

a sentence later. In (37b) the NP ie 'house' occupies the specifier position and the PP umi-ni 'to the sea'

12

occupies the complement position (the adjective tika-i 'near' is subcategorized by a NP) and the head is an A

tika-i 'near'. Because the head is an adjective, ie(-ga) umi-ni tika-i is an adjectival phrase. In (37c) the NP

Taro occupies the specifier position and the NP kabu 'stock' occupies the complement position (the noun

tori-hiki 'transaction' is subcategorized by a NP). Because the head is a noun, Taro(-no) kabu(-no) tori-hiki

'Taro's transaction of stocks' is a NP. In (37c) there is no specifier. Postpositions are a defective category

in that there is no specifier position. The PP Tokyo-kara 'from Tokyo' occupies the complement position

(the postposition kara 'from' is subcategorized by a NP). Because the head is a postposition, Tokyo-kara is

a postpositional phrase.

(37a)' V" (37b)' A"

N" V' N" A'

Taro N" V ie P" A

hon yom umi-ni tika

(37c)' N" (37d)' P"

N" N' P'

Taro N" N N" P

kabu tori-hiki Tokyo kara

d) The maximal level of the categories is 2.

There are only X" (phrasal category), X' (intermediate category), and X (lexical category) (=X 2 , X

1 , and X

0 ).

e) The X’ level allows recursion.

If there are more than one complement, X' level recurse.

(38) 太郎 た ろ う

が花子 は な こ

に花 はな

をあげた。'Taro gave flowers to Hanako.'

V"

N" V'

Taro P" V'

Hanako-ni N" V

hana age

f) The positions of Spec and Comp are subject to language-variation (In English, Comp appears on the right of

the head X).

J: X" E: X"

Spec X' Spec X'

Comp X X Comp

Taro hon yom John read a book

5.2.2. Sentence = TP

Up to this point we have seen how phrases are formed. Phrase structure rule (8a) forms a sentence combining

an NP and a VP:

13

(8) a. S → NP VP

(11a) is usually called S Rule.

Furthermore, we expanded phrase structure rules to include subordinate clauses (relative clauses, (13), and complement

clauses, (15)) :

(13) NP → (S) (NP) (AP) N

(16) introduces a clause inside a NP (a relative clause). (22) introduces a clause inside a VP (a complement clause):

(15) VP → (PP) (PP) (PP) (NP) V

(S')

The clause introduced inside a VP accompanies a morpheme called a complementizer so that we needed (18):

(18) S’ → S COMP (S’ is ‘S prime or S bar’; COMP is a complementizer)

(18) is usually called S' Rule.

It is a long tradition that sentences are analyzed into the subject (NP) and the predicate (VP). However, the problem is

that we have two rule types: i) phrase structure rules like (8a) and (18) and ii) X' rules. (13) and (15) can be

incorporated into X' system once the problem with (8a) and (18) is solved. That is, if we can simplify the two rule

types into one we can achieve simplicity and the theory would be better (the simplicity measure: the simpler, the

better, if other things being equal). The direction that has been taken is to reformulate (8a) and (18) into X' rules. X'

rules make the following claims:

(39) X' rules

a) Phrases consist of the three elements: the optional specifier position, the obligatory head position, and the

optional complement position.

b) There are two relations in phrases: specifier-head agreement (Spec-head agreement) and

subcategorization (complements subcategorize the head).

To reformulate (11a) in terms of X' rules we first need to find the obligatory head of a sentence. Sentences have long

been thought as an exocentric construction (no head inside the structure). However, in English an inflectional

morpheme that is found in the following sentence has been proposed as the head of a sentence.

(40) John comes.

If we take the third person singular non-past suffix -s as the head of a sentence, the sentence would be a projection of

the inflectional morpheme. The third person singular non-past suffix -s is designated as I and then the sentence is I"

(IP). As for English, (8a) is changed to the following:

(41) English

a. I" → (Spec) I'

b. I' → I (Comp)

14

(42) I" = IP

N" I'

John I V"

-s come

move

agree subcategorize

(42) satisfies all the requirements for X' rules: IP consists of the three elements: optional Spec, obligatory head, and

optional Comp. Furthermore, (42) satisfies the two relations that must hold in a phrase: Spec-head agreement (the

subject John agrees with Inflection -s and the Comp V" subcategorizes I (Inflection -s only attaches to V").

It is well-known that Japanese does not have inflection such as the third person singular non-past -s. So the question

is: Does Japanese have Spec-head agreement?

Takezawa (1987, 73-74) proposes that Tense agrees with the subject in Japanese.

(43) a. ジョンはメアリーの横顔 よこがお

がとても 美 うつく

しいと思 おも

った。

'John thought that Mary's profile was very beautiful'

c.*ジョンはメアリーの横顔 よこがお

がとても 美 うつく

しく思 おも

った。

'John thought Mary's profile (to be) very beautiful.'

Takezawa, Koichi (1987) A Configurational Approach to Case-Marking in Japanese. Ph. D.

Dissertation, University of Washington.

In (43a) the subordinate clause has Mary-no yokogao-ga totemo utukusi-i 'Mary's profile is very beautiful.' and the

nominative case particle ga is licensed by the non-past suffix i that attaches to the adjective. However, in (43c) the

subordinate clause is Mary-no yokogao-ga totemo utukusi-ku 'Mary's profile (to be) very beautiful.' so that there is no

tense marker in the clause (utukusi-ku is a tenseless form called the continuative form). However, the subject Mary is

marked by ga and this nominative case particle is not licensed, therefore, (70c) is ungrammatical. Following

Takezawa (1987) we take Tense to be the head of Japanese sentences:

(44) Japanese

a. T" → (Spec) T' (T = non-past (r)u and past ta for verbs; non-past i and past katta for adjectives)

b. T' → (Comp) T

(45) T" (=TP)

NP T'

Mary-no yokogao A" T

totemo utukusi i

sucategorize

agree

We have made an important revision: we have assumed that X' rules only apply to lexical categories but from now on

we assume that X' rules apply to grammatical categories as well as lexical categories.

5.2.3. Sentence Embedding

Now that we got rid of (8a) our task is to get rid of (18):

15

(18) S’ → S COMP (S’ is ‘S prime or S bar’; COMP is a complementizer)

It is relatively obvious that the head of S' is COMP in (18). Taking C as the head of C", we have the following rules:

(46) Japanese

a. C" → (Spec) C'

b. C' → (Comp) C

(47) T"

N"i T'

John-wa V" T

N" V' ta

ti C" V

?Spec C' omot

T" C

Mary-no yokogao-ga utukusi-i to

subcategorize

agree?

It is not clear we have Spec-head agreement in (47). However, in the following sentence the interrogative

complementizer ka agrees with the interrogative phrase nani-o 'what'.

(48) 太郎 た ろ う

は花子 は な こ

が何 なに

を買 か

ったか尋 たず

ねた。'Taro asked what Hanako bought.'

(49) T" 1

N"i T'

Taro-wa V" T

N" V' ta

ti C" V

Spec C' tazune

T" 2 C

Hanako-ga nani-o kat-ta ka

subcategorize

agree

In (49) the interrogative phrase nani-o 'what' moves to the Spec position of C". We assume this covert movement of

the interrogative phrase in Japanese. Then the interrogative C ka agrees with the interrogative phrase nani-o in Spec

of C". If C has the (affirmative) complementizer to, the interrogative phrase nini-o in Spec will not be licensed:

(50) *太郎 た ろ う

は花子 は な こ

が何 なに

を買 か

ったと尋 たず

ねた。'Taro asked what Hanako bought.'

Furthermore, if there is no interrogative phrase in T" 2 in (49), the sentence will be ungrammatical:

(51) *太郎 た ろ う

は花子 は な こ

が買 か

ったか尋 たず

ねた。'Taro asked what Hanako bought.'

16

If we supply doo-ka 'whether/if' after the interrogative complementizer ka in (51), the sentence will be grammatical.

(52) 太郎 た ろ う

は花子 は な こ

が買 か

ったかどうか尋 たず

ねた。'Taro asked whether/if Hanako bought (it).'

In (52) doo-ka 'whether/if' agrees with the interrogative complementizer ka.

We subcategorize the complementizer C into the interrogative [+wh] and (affirmative) [-wh] complementizers.

(53) a. C b. C

[+wh] [-wh]

ka to

Summarizing so far, phrase structure rule (8a) (S Rule) is replaced by (27) in X' theory:

(8) a. S → NP VP

(44) Japanese

a. T" → (Spec) T' (T = non-past (r)u and past ta for verbs; non-past i and past katta for adjectives)

b. T' → (Comp) T

S' Rule (18) is replaced by (46) in X' theory:

(18) S’ → S COMP (S’ is ‘S prime or S bar’; COMP is a complementizer)

(46) Japanese

a. C" → (Spec) C'

b. C' → (Comp) C

Extending X' rules to grammatical/functional categories, we can get rid of S Rule and S' Rule (phrase structure rules)

and only X' rules are necessary, and we can achieve simplicity.