Xiang (2003)

Arguing against the proposal in Liu (1996), which suggests that the bi-constituent be

analyzed as an elided CP containing an I’ gap, Xiang (2003) proposes the phrasal analysis of

bi-comparatives in Mandarin Chinese. The post-bi constituent is analyzed as a simple DP, as

shown below in (6). Moreover, Xiang (2003) adopts the direct analysis in Heim (1985) and

Kennedy (1997) to derive the semantics of phrasal comparatives in Chinese.

(6) Zhangsan [PP bi [DP Lisi]] gaoxing

Zhangsan BI Lisi happy

‘Zhangsan is happier than Lisi.’

Significantly, three problems encountered by the clausal analysis are presented in Xiang

(2003) to claim that the phrasal analysis is more adequate than the clausal one in terms of the

analysis of phrasal comparatives in Chinese. To begin with, following Chomsky’s (1977)

analysis, Xiang (2003) argues that wh-movement is involved in comparative constructions³

which are claimed to contain degree variables by researchers in support of the clausal analysis.

Since long-distance wh-movement is generally allowed in Chinese, it is expected that it is also

allowed in comparatives under the clausal analysis. However, it turns out that

bi-comparatives disallow long-distance wh-movement. This shows that the clausal analysis makes

wrong prediction in terms of long-distance wh-movement in bi-comparatives. This problem is

illustrated with the contrast between the English example in (7) and the Chinese example in

(8).

(7) John is taller than [CP Opi [CP Max thought [CP Bill is ti]]].

(8) *Zhangsan bi Wangwu renwei Lisi zuotian gaoxing

Zhangsan BI Wangwu think Lisi yesterday happy

Intended: ‘Zhangsan is happier than Wangwu thought Lisi was yesterday.’

3 The wh-movement involved in comparative constructions is the movement of the degree operater (Op), as shown in (7).

Under the clausal analysis, the two sentences in (7) and (8) above are analyzed as comparatives

with an embedded standard. As can be seen, both of the two standards of comparison Bill and

Lisi occur in an embedded clause. Chinese is different from English in that it doesn’t allow

embedded standards, but English clausal comparatives do. This contrast cannot be accounted

for by the clausal analysis, which predicts that (8) is grammatical. However, it is not a problem

for the phrasal analysis. By adopting the phrasal analysis, it is claimed that the standard of

comparison in bi-comparatives is actually a DP rather than a CP; therefore, embedded

standards, which are certainly CPs, are disallowed in bi-comparatives.

The second problem faced by the clausal analysis is that, contrary to English, Chinese

bi-comparatives lack subdeletion (subcomparative), as demonstrated in (9) and (10a) below.

(9) The table is longer than [CP the door is wide].

(10) a. *Zhuozi bi [CP men kuan] chang

table BI door wide long

Intended: ‘The table is longer than the door is wide.’

b. Zhangsan bi [CP Lisi gao⁴] gao

Zhangsan BI Lisi tall tall

‘Zhangsan is taller than Lisi is.’

In the English subcomparative construction (9), than takes a full CP as its complement;

however, (10a), which also takes a full CP men kuan ‘the door is wide’, is ungrammatical.

Then, compare (10a) with (10b), where bi takes an elided CP. The fact that (10b) but not (10a)

is grammatical seems to suggest that the comparative morpheme bi in Chinese can only take

elided CPs. Examples in (9) and (10) show the problem of subcomparatives for the clausal

analysis. First, it is in conflict with the prediction of the clausal analysis, which claims that the

post-bi constituent is a clause rather than a simple phrase. Furthermore, as pointed out by Xiang,

generally speaking, a full CP is an unmarked construction, and an elided CP is marked. It is

hard to explain why bi in Chinese tends to take a marked construction as its complement instead

of an unmarked one. According to Xiang (2003), this problem is circumvented in the phrasal

analysis, which can directly rule out (10) by arguing that what bi can take is actually a DP.

4 Striking-through indicates deletion of the identical elements on the surface.

The behavior of the Chinese distributor dou in bi-comparative constructions is the third

problem of the clausal analysis. Following Lin (1998), Xiang (2003) makes two claims: first

of all, there is a leftness constraint of dou, which regulates that the licenser of dou must be at

the left side of dou. Moreover, in most cases, if the subject is quantified by a strong quantifier

like mei-ge ‘every’, the presence of dou is obligatory. With these two constraints, Xiang (2003)

finds that when dou appears in a bi-comparative construction, the clausal analysis will make

incorrect predictions, as shown in (11) and (12).

(11) Mei-ge-nanhaizi dou bi [CP mei-ge-nuhaizi gao] gao

every-CL-boy all BI every-CL-girl tall tall

‘Every boy is taller than every girl.’

(12) * Mei-ge-nanhaizi dou bi [CP mei-ge-nuhaizi dou gao] gao

every-CL-boy all BI every-CL-girl all tall tall

Intended: ‘Every boy is taller than every girl.’

The clausal analysis will predict (11) to be ungrammatical and (12) to be grammatical.

However, the facts are the opposite. Under the clausal analysis, mei-ge-nuhaizi ‘every girl’

serves as the subject of the CP complement; therefore, the occurrence of dou after it should be

legitimate. Nevertheless, (12) shows that dou cannot appear after mei-ge-nuhaizi ‘every girl’

in the post-bi constituent. On the other hand, according to Xiang (2003), the examples in (11)

and (12) can be captured with the phrasal analysis. Under the phrasal analysis, the post-bi

constituent mei-ge-nuhaizi ‘every girl’ in (11) and (12) is only a DP rather than the subject of

the CP complement. Thus, the only dou in (11) as well as the first dou in (12) are licensed by

the matrix subject mei-ge-nanhaizi ‘every boy’, and the second dou in (12) is unnecessary

because mei-ge-nuhaizi ‘every girl’ is not a subject at all.

With the three problems faced by the clausal analysis, Xiang (2003) proposes that we

should adopt the phrasal analysis for the phrasal comparatives in Chinese instead of the clausal

analysis claimed in Liu (1996). Besides, Xiang (2003) provides an obligatory A-to-V

movement analysis to account for the contrast in (13) in Chinese comparatives, as shown below.

(13) a. Zhangsan bi Lisi gao liang-mi Zhangsan BI Lisi tall two-meter ‘Zhangsan is two meters taller than Lisi.’

b. *Zhangsan bi Lisi liang-mi gao Zhangsan BI Lisi two-meter tall

As can be seen, the measure phrase liang-mi ‘two-meter’ cannot precede the predicate gao

‘tall’, or it will lead to ungrammaticality in (13b). Xiang (2003) provides an account for this

phenomenon by arguing that the movement of the predicate from A to V is obligatory in

Chinese comparatives, as illustrated in (14).

(14)

Since the V head is obligatory occupied by the moved predicate from A⁵, it is impossible for

the measure phrase to precede the predicate.

5 Xiang (2003) argues that the movement is obligatory in order to satisfy the locality constraint and to derive the correct interpretation of the sentence.