1. 网站地图
  2. 设为首页
  3. 关于我们


英语语言文学 现代汉语“比”字结构的句法研究

发布时间:2020-02-06 12:59
Abstract
Comparative sentences are frequently used across languages. After multitudes of research findings have been made in the study of comparative constructions in western languages, the Chinese comparative construction marked with the character hi is becoming a research hotspot at home and abroad. Generally, bi comparative construction follows the pattern of comparative subject + bi + comparative standard + predicate of comparison. Carried out under the framework of the Generative Grammar and based on previous studies, this thesis makes an attempt at exploring bi comparative constructions and thus raises two questions: ( i) What syntactic category does the marker bi belong to? (ii) What is the syntax of comparative result after bil Is it an elided clause or just a DP itself?
Concerning the above questions, through syntactic analysis, the thesis comes to two conclusions. Firstly, different from previous studies in which marker bi is regarded as a preposition, a verb, a conjunction, or a complementizer, we make the proposal that bi is a light verb embedded in HOLD, a type of eventuality predicates expressing state. The combination of HOLD and bi serves as the predicate of a sentence. We will provide evidence for this proposal. Secondly, there are two types of comparative standards, clausal and phrasal standards parallel with each other instead of being in a relationship of derivation. Based on Larson’s (1988b) analysis of the double object construction, we believe that in phrasal 6i-comparatives, the comparative standard and the comparative result serve as the direct and the indirect object of marker bi respectively. This can be proved by the adjacency principle. After the comparative standard receives accusative Case and gets DP-moved, the S-structure is formed. Regarding clausal ^/-comparatives,we reckon there’s only one CP base-generated in [Spec, VP]. The subject of this CP is co-referenced with the comparative subject, and this can be proved by the control theory.
Key words: Mandarin Chinese, Comparative Construction, bi, light verb
英语语言文学 现代汉语“比”字结构的句法研究
List of Abbreviations
ACD antecedent-contained deletion
AgrOP object agreement phrase
AP adjective phrase
ASP aspect marker
BCC bi comparative construction
CL classifier
CP complementizer phrase
DegP degree phrase
DO direct object
DP determiner phrase
D-structure deep structure
EPP Extended Projection Principle
IO indirect object
IP inflection phrase
MC Mandarin Chinese
MLC Minimal Link Constraint
Neg negation
NP noun phrase
PP preposition phrase
PRO/pro pronoun
S sentence
Spec specifier
S-structure surface structure
Sj subject
St standard
TP tense phrase
VP verb phrase
vP voice phrase
 
 
 

Chapter One Introduction
Language, an important tool used for human communication, has a large array of ways to convey different relations between objects or people. Among these relations, the expression of differences and similarities, i.e., making comparisons, is a common occurrence. Refer to the following examples:
(1) Tom works as hard as Mary.
(2) Jack did better in Math than Jerry.
(3) Among all the candidates, Nick performed the best.
Among the above examples, (1) is used to express the same level of working strength on Tom and Mary, (2) and (3), though different in level of performance, have something in common. That is, they both express differences among individuals. To be more precise, when making comparisons in English, we use (2) to refer to a comparative sense and (3) a superlative sense. Yet, ifs fair and reasonable for the author, a native Chinese speaker, to study comparative expressions in Mandarin Chinese. The following corresponding examples are given.
(4) Lijun he Wangxia pao de yi-yang kuai.
Lijun and Wangxia run DE same fast
“Lijun runs as fast as Wangxia.”
(5) Ta tiao de bi wo gao.
he jump DE bi I high
‘‘He jumps higher than me.”
(6)  Liming kao de zui gao.
Liming score DE most high “Liming scores the best.”

Different in meaning as they are, examples (4), (5) and (6) are pretty similar in form to examples (1), (2) and (3) respectively. In the following part, we will present research background, research questions, significance of study, as well as thesis organization.
1.1 Research Background
Within the framework of the Generative Grammar, heated discussions have been made on comparative constructions. According to Ma (1935), comparisons in Chinese can be classified into three types, namely, equal, unequal and superlative comparisons. They were used to point out levels of difference between entities. To exemplify his classification, (4) is a case for equal comparison, (5) for unequal comparison, and (6) for superlative comparison. As is conspicuously shown in (5), the general pattern of bi construction goes like this: comparative subject + bi + comparative standard + predicate. Following this pattern, it’s not complicated to make a comparison. Nevertheless, ifs not that simple to encode comparative constructions from linguistic perspectives. Considering this, ifs necessary to show characteristics concerning this pattern before a detailed discussion is made.
According to Tsao (1989) and Lin (2009), categories allowed to serve as the comparative items range from DPs, including subject, object and temporal DPs, to VPs, PPs, locative phrases and clauses. Respective examples are given as follows:
(7) Subject DP
Zhangsan bi Lisi pang.
Zhangsan M Lisi fat
‘‘Zhangsan is fatter than Lisi.”
(8) Object DP
Zhangsan li-ke bi wen-ke xue de hao.
Zhangsan science M arts learn DE well
“Zhangsan is better at science than at arts.”

(9) Temporal DP
Wo jin-nian bi qu-nian gao.
I this year M last year tall
 
 
 
“I’m taller this year than last year.”
(10) VP
Shuo de bi chang de hao ting.
Say DE M sing DE good sound
uHe/She has a glib tongue.”
(11) PP
Wo dui han-yu bi dui ying-yu gan xing-qu.
I to Chinese M to English feel interest
“I feel more interested in Chinese than in English.”
(12) Locative Phrase
Ta cong jia-li bi cong Xi-an dao La-sa jin.
He from home M from Xi-an to La-sa near
“He travels faster from home to Lasa than from Xi’an.”
(13) Clause
Ta shuo ying-yu bi wo da kuai-ban hai kuai.
He speak English M I perform clapper talk even fast
“It’s faster for him to speak English than for me to perform clapper talk.”
Whafs exemplified above not only shows acceptable categories of compared items, but also exhibits one hidden mechanism, as is noticed by Tsao (1989). That is, when at least two items are compared, they must follow the order of subject-temporal-locative. For instance, the combination of (9) and (12) leads to the following examples:
(14) Ta qu-nian cong jia-li bi wo jin-nian cong Xi-an dao La-sa kuai.
He last year from home M I this year from Xi-an to La-sa fast
“It’s faster for him to travel from home to Lasa last year than for me to travel
3
from Xi’an this year (to Lasa).’’
(15) *Ta cong jia-li qu-nian bi wo cong Xi-an jin-nian dao La-sa kuai.
Intended meaning: ditto.
(14) is grammatical while (15) is unacceptable in that the order of the comparative items in the latter violates Tsao (1989) and Lin’s (2009) proposal. That is, temporal phrases qunianljinnian (last year/this year) must precede locative phrases congjiali/congxi'anifrom home/from xi^n), but not vice versa.
In addition, based on Xiang (2005), in bi-comparatives, an embedded standard is not allowed, which contrasts strikingly with the English counterpart. Refer to the following example:
(16) *Zhangsan bi Lisi ren-wei Wangwu kai-xin.
Zhangsan M Lisi think Wangwu happy
“Zhangsan is happier than Lisi thought Wangwu was.”
(Xiang 2005: 79)
Tsao (1989) also notices that the subject after the marker bi can be replaced with a PRO if identical to the comparative subject at the beginning of the sentence. Refer to example (17):
(17)  W〇i jin-tian bi PROi zuo-tian kun-juan.
I today M yesterday sleepy
“I’m sleepier today than I was yesterday.”
The Chinese character bi is the central part in comparative constructions in
Mandarin Chinese. How bi is defined regarding its syntactic status brings about
different syntactic results. Based on distinct aspects, parts of speech concerning bi vary. They range mainly from such classes as preposition, verb, co-verb, to
conjunction and complementizer. Yet, in terms of its word class, the most widely
4

accepted one in academia is preposition. According to this theory, bi lost its features in verb uses and has gradually been grammaticalized to a total preposition. This lexical entry can be found in The Contemporary Chinese Dictionary. In addition, foreign researchers like Liu (1996) and Lin (2009) accept it as well.
Still, Xiang (2005) and Erlewine (2007) consider comparative marker bi as a verb. Disagreeing with Liu (1996) in treating bi as a preposition, Xiang (2005) sees bi as a verb which must c-command the comparative standard and the comparative result. The c-command relation will no longer exist once bi + comparative standard as a whole is treated as an adjunct. However, this conclusion bears no solid ground for two reasons. One is that no justification has been given as to why the measure phrase is a must in ^/-comparatives. The other is that in Xiang (2005), verbal bi like bi sai, bi shi (compete) are confused with marker bi. Erlewine (2007) doesn^ provide enough evidence in terms of verbal bi. It is argued in Erlewine (2007) that the negation marker bu always appears before marker bi instead of the comparative standard; moreover, in A-not-A questions, what is overlapped is bi rather than the predicate corresponding to the comparative standard. Thus, bi is studied as a verb. Besides, the following three cases are analyzed in his paper, namely, syntactic features of ge (each, assignment operator) in comparative constructions, syntactic distributions of ziji (reflexive), coupled with passive voices of bi sentences. Thus he holds that bi + comparative standard is not adjunct to predicate, and moreover, the comparative subject is not in parallel structure with it. He goes on to argue for the possibility of putting bi sentences under the framework of vP-shell. In his opinion, marker bi is a verb, which, together with the comparative standard, serves as the predicate locating after the comparative subject.
Hong (1991) and Hsing (2002) look on bi as a conjunction and a complementizer respectively. Their conclusions are drawn out of comparisons with the English comparative marker than. Yet, than is not the same as bi in Mandarin Chinese.
Yet, due to limited space and time, we will not discuss all the different kinds of

/)z-comparatives listed above. Instead, this thesis is mainly focused on ditransitive phrasal ^/-comparatives where measure phrases are covered and clausal ^/-comparatives where only one dimension is involved.
1.2 Research Questions
Based on the previous research, this thesis aims to explore bi comparative construction (CC for short) in Mandarin Chinese under the framework of the Generative Grammar. We will focus on two points in bi CC—comparative marker bi, and comparative items. In short, the thesis attempts to find out the answers to the following two questions:
1) What is the syntactic category of bP.
2) In terms of comparative standards, how are they derived? Are phrasal comparatives derived from clausal ones by deletion or any other means?
The first question deals with the word class of marker bi. This directly determines where bi will be placed in tree diagrams. For instance, if bi is treated as a preposition, then it can possibly be placed in the adjunct position of VP. As for the second question, it has to be replied after the first one has been worked out. What syntactic status bi is assumed to have determines where the comparative standard and the comparative result are base-generated and thus provides justifications for their possible movements.
1.3 Significance of the Research
As to theory and practice, this thesis has its own values that go as follows:
Firstly, the present research can promote a better understanding of CC in Mandarin Chinese. By studying its part of speech, we can apprehend the syntactic role

of bi, and meanwhile, how comparatives are transformed. In this way, this special construction in Chinese can be better comprehended.
Secondly, the study of hi CC can facilitate both future teaching and learning of the Chinese language. It helps foreign learners to gain a clear idea of Mandarin bi constructions. Moreover, problems which cannot be solved by semantic or typological means at first can find their syntactic ways out.
Thirdly, it is hoped that this research can be of some help to further study of Mandarin Chinese by employing formal linguistic theory.
1.4 Organization of the Thesis
Six chapters constitute the thesis. They go as follows:
Chapter One gives a brief introduction to the present study, including research background, questions and significance of the research, and thesis organization.
Chapter Two is a literature review. It first provides a general overview of previous studies on bi CC, including various hypotheses of bi, and the derivation of comparative items. It is then followed by the author’s criticisms.
Chapter Three presents the theoretical framework of the current study. As the Generative Grammar is the overall framework of the whole research, a brief overview of it will be covered, followed by some concrete concepts like c-command, binding theory and movement constraints. Concerning them, detailed theory will be annotated first before we answer why they are employed in particular.
In Chapter Four and Chapter Five, we propose the role of bi as a light verb. The proposal on the syntactic role of bi will help to locate it in tree diagrams. After that, the author will argue that comparative items, clausal and phrasal, are parallel with each other. In so arguing, detailed evidence will be provided. The analysis will be carried out in accordance with the theoretical framework mentioned.
Chapter Six is a summary of the major findings and limitations of the research. Suggestions for further study will be involved as well.
Chapter Two Literature Review
After the thesis introduction is given in the first chapter, we will offer some basics on CC? unequal CC in particular. Following this will be a review on both previous scholars9 definitions of the syntactic role of marker bi and three mainstream analyses of the syntax of BCC. This part serves as a preparation for more detailed analyses in Chapter 4 and 5.
2.1 Introduction to CC
2.1.1 Definition and Classification of CC
According to Stassen (1985), the comparative construction generally refers to the construction that has the semantic function of assigning a graded (non-identical) position on a predicative scale to two (possibly complex) entities. Scholars have made substantial discoveries on the development of CC throughout history. Among them, some important constructions like bi sentence draw particular attention from academia.
As is mentioned at the beginning, Ma (1898) first classifies CC into equal, unequal and superlative comparisons. He then defines them respectively, which greatly influences later related studies. Besides, Lv (1942) goes even further as to classify CC into 11 subclasses—bini, jinsi, buji, shengguo, youzui, deshi, buru, yibian. Meticulously classified as they are, the scope for comparative sentences seems too general. Liu (2011) claims that concepts like leitong, deshi, buru andyibian should all be treated as similar semantic categories to bijiao and thus need to be eliminated from CC. For the following ones, bini belongs to the category of either equal comparison or figures of speech. Gaoxia, shengguo, and buji all stand in line with unequal comparison, youzui left with superlative comparison. Other scholars also try sorting out CC. They are Ding (1961) and Chao (1968). Yet, as they each have different understandings for CC, their criteria for defining CC differ.
2.1.2 Unequal CC
Unequal CC includes bi, yu and ru comparatives. Ma (1898) first studies the unequal construction. He points out that unequal sentences in ancient Chinese were mainly marked with yu and ru, with yu taking the lead. He also mentions the deletion ofyu. Yet, unequal CC with bi is hardly mentioned for lacking linguistic data.
Tai (1987) first introduces two types of word order in unequal CC: Type A: A + Adj + Prep + B; Type B: A + Prep + B + Adj. For the first type, comparative words can be yu, ru, si. As for Type B, the marker is bi. Based on his study, unequal CC markers like ru and si appeared in Tang Dynasty. Moreover, bi sentences in this period can already be treated as unequal CC. His viewpoint sheds light on later researches. As is mentioned at the beginning, Huang’s (1989) research on the origin and development of marker bi is closely tied to Tai^ (1987) discoveries. Following Huang (1989), others who have done similar researches include Shi (1993) and Li & Shi (1998).
In addition to bi, another focus in the study of unequal CC lies in comparatives marked by ru or si. Zhang (1991) considers the formula X + A+ ru/si + 7 as starting from Song Dynasty. He also attributes the replacement ofyu with ru/si to the deletion of yu. However, Wei (2007) disagrees. He owes the disappearance of yu to polysemy and to its limitations in expression. These two factors leave it hard for yu to achieve the full function of making comparisons in Mandarin Chinese. Also, the transition of the same formula X + W + ru + Y from equal CC to the unequal one lies in specific lexical item and tone of ru. Bi finally takes the place of ru/si because of structural ambiguity.
Judging from the above, more research findings have been made in unequal CC in ancient Chinese, compared with equal CC. Yet, in terms of all-round researches on unequal CC of Mandarin Chinese, we still have a long way to go. More research on this topic remains to be done.

In this section, we are going through previous studies concerning the categorical status of comparative marker bi. Many scholars have done research on this topic. Among them, Liu (1996), Chao (1968), and Lin (2009) hold that bi is a preposition. Xiang (2005) and Erlewine (2007) propose that bi is a verb. Hong (1991) treats bi as s conjunction. Xu (1993) suggests a functional analysis of bi. Hsing (2002) reckons bi as a complementizer . As prepositional and verbal analyses are the mainstream views, we will take a view of the two accounts in the following part.
2.2.1 The Prepositional Account
Liu (1996) argues that bi is a preposition, which, when combined with the comparative standard, forms a PP. He argues against the proposal that is a coordinator by employing the adverb yizhi (always) to falsify it.
(1) Guojing yi-zhi bi Huangrong xing-fu.
Guojing always M Huangrong happy.
“Guojing was always happier than Huangrong.”
(2) * Guojing bi Huangrong yi-zhi xing-fu.
Guojing M Huangrong always happy.
Intended meaning: ditto.
(Liu 1996: 219)
The fact that (1) is grammatical while (2) is not proves that bi can not be treated as a typical coordinator which can either follow or precede adverbs. He thus argues for the prepositional analysis of bi. Liu (1996) employs the coordination test to support his prepositional analysis. Refer to example (3):

uZhangsan is taller than Lisi or Wangwu.
(Liu 1996: 221)
By means of coordination test, Liu (1996) shows the evidence that bi + standard alone is a constituent. However, no further evidence is given in support of his prepositional analysis.
2.2.2 The Verbal Account
Xiang (2005) argues against Liu^ (1996) prepositional analysis as mentioned above. By proving that bi + standard is a constituent, Liu (1996) draws the conclusion that bi is a preposition which takes the comparative standard as its complement. Yet, Xiang (2005) argues that Liu^ (1996) proposal which is based on simple comparison between bi and than in English is superficial. Xiang (2005) asserts that bi is more like a verb rather than a preposition. Examples are given as follows:
(4) a. Ni bi guo na Hang ben shu de jia-qian ma?
You compare Asp that two CL book DE price Q
“Have you compared the prices of the two books?” b. Ta lao na wo he Zhangsan bi.
He often take me and Zhangsan compare “He often compares me with Zhangsan.”
Xiang (2005) demonstrates that compare is the meaning bi shows. Besides, in (4a), the perfective marker guo can be added to bi, which is in line with the fact that aspect markers are generally attached to typical verbs. Yet once we agree that there are two kinds of bi in Mandarin Chinese, the verb bi is mistaken for the comparative marker bi. A detailed discussion can be referred to in Chapter 4.
li

Ifs necessary to make a comparison between English and Chinese phrasal and clausal comparatives to avoid insisting that the same holds true in bi CC. That is, the phrasal comparative may be derived from a clausal one. Refer to the following examples:
(5) a. Tom earns more than [ Jerry does earn ].
b. Tom earns more than [ Mary thinks [ Jerry does ]].
However, it turns out incorrect in Mandarin Chinese. As is mentioned in the first chapter, Xiang (2005) and Lin (2009) suggest that in bi CCs, an embedded standard is not allowed. Refer to the following data:
(6) *Zhangsan bi Lisi renwei Wangwu gaoxing.
Zhangsan M Lisi think Wangwu happy
Intended meaning: Zhangsan is happier than Lisi thinks Wangwu is.
Therefore, in this way, why phrasal comparatives in Chinese are not underlying clausal comparatives can be explained. In addition, sub-comparison is illegal in bi CC in Mandarin Chinese as well, as exemplified in (7):
(7) *Zhe-zhang zhuozi bi na-zhang zhuozi chang kuan.
This-CL desk M that-CL desk long wide
Intended meaning: This desk is wider than that table is long.
Then, we may question what the underlying structures of the Mandarin phrasal and clausal comparatives are? That is, what is the syntax of the comparative standard following marker bil Is it a DP or a CP? In the following part, we will only examine

two major approaches, namely, clausal analyses (Fu 1978 & Liu 1996) and phrasal analyses (Lin 2009).
2.3.1 The Reduction Analysis
Fu (1978) advances the underlying semantic structure of bi comparative construction in the deep structure with a Z)z-clausal standard, as illustrated in (8):
Fu (1978) considers the abstract bijiao (to compare) as a predicate that takes two parallel prepositions as its arguments. The two comparative items, Si and S2 are modified by variables x and which denote the degrees of Si and S2 respectively, and
x>y. Fu (1978) further claims that the comparative marker bi is a complementizer-like
element, which does not exist in the underlying structure but is added by process of transformation. Taking Zhangsan bi Lisi gao (Zhangsan is taller than Lisi) as an example, and the transformation process is shown in (10).

 
 
Si S2 bijiao  
 
 

 


 
 
Lisi gae-
The transformational process in (10) is as follows: first, NPi has undergone Subject Raising and Chomsky-adjunction   to S〇,the comparative marker bi is adjoined to NP2 mdy is deleted, the original node So is replaced by So5, resulting in (10b); second, the variable degree marker x has undergone VP lowering and substitutes the VP of So9, the adjective in S2 has been truncated by Identity Deletion Rule, S2 is replaced by the left NP,and original So’ node is replaced by a new Si node, leading to (10c); finally, bijiao-dQlQtion brings out the surface structure in (lOd).
2.3.2 The Antecedent-contained Deletion Analysis
Liu (1996) argues that BCC in Mandarin Chinese has an adjunction structure

with antecedent-contained deletion. He claims that the comparative marker bi is a preposition taking a CP in which there is an empty /'and the ^/-phrase functions as a preverbal adjunction, as exemplified in (11):
(11)
IP
 
P CP
 
I △
bi degree clause
Liu (1996) then argues for the clausal analysis of Mandarin ^/-comparatives. First, he offers the following examples in (12) and the underlying structure is shown in (13).
(12) Guojing jintian bi Huangrong zuotian gaoxing.
Guojing today M Huangrong yesterday happy
today is happier than 喂rcwg was yesterday.”
(13) Guojing j intian [pp bi [cp Huangrong zuotian gaoxing]] gaoxing.
Guojing j intian [r [pp bi [ Huangrong zuotian[ r ]]] gaoxing].
today is happier than 喂 was yesterday.”
(Liu 1996: 221)
Liu (1996) argues that the deletion in the comparative standard is a case of /'-ellipsis. Without a local / node which checks the nominative Case with the
comparative subject, the subject would not pass the Case filter. He solves this by letting bi exceptionally Case mark the standard across the clausal boundary. And this also explains why we don’t yield embedded clausal standard as shown above: is too far away to check the Case.
Liu (1996) also recognizes that the descriptive complement structure of /)z-comparatives cannot be explained without the /'-ellipsis analysis, consider the example.
(14) Zhangsan qima bi Lisi qiniu qi-de kuai.
Zhangsan ride horse M Lisi ride bull ride-DE fast Zhangsan rides a horse faster than Lisi does a bull.
(15) *Zhangsan qima bi Lisi gan yang qi-de hao Zhangsan ride horse M Lisi drive sheep ride-DE good Zhangsan rides a horse better than Lisi drives the sheep.
By comparing the above examples, Liu (1996) suggests that the meaning of the elided site must be identical to its antecedent, and the violation of it will lead to the ungrammaticality of the sentence in (15). Similarly, the underlying structure of (14) is demonstrated below:
(16) Zhangsan qima [pp bi [cp Lisi qiniu [ — ]]] qi-de kuai.
Zhangsan rides a horse faster than Lisi does a bull.
Although Liu’s (1996) analysis can explain why we can’t yield embedded clausal standards of Z?z-comparatives in Mandarin Chinese, as Xiang (2005) and Erlewine (2007) point out, the clausal analysis put forward by Liu (1996) is problematic. First, Liu (1996) has difficulties explaining why in English a comparative with sub-deletion like (17) where the main clause and than clause are both full clauses is allowed, yet its Chinese counterpart is ungrammatical.
(17) a. This desk is wider than that table is long.
b. *Zhe-zhang zhuozi bi na-zhang zhuozi chang kuan.
To solve the problems existing in his clausal analysis of Z?z-comparatives, Liu (2011) suggests a hybrid analysis for the Z?z-comparatives with the following hypothesis: treat the comparatives with only one comparative standard constituent as phrasal comparatives and those with more than at least two standards as the clausal ^/-comparatives. We will not give more detailed description of his new idea here.
2.3.3 The Phrasal Analysis
Lin (2009) argues that superiority comparatives in Mandarin Chinese are all phrasal comparatives that could be directly interpreted. He admits that ^/-comparatives are truly phrasal at the underlying level, rather than reduced constructions derived from clausal comparatives.
Along the line developed by Kennedy (2007), Lin (2009, pp. 139-40) makes the following hypotheses in his paper:
1. The Chinese ^/-comparative is a phrasal in place of a clausal comparative. Chinese only has individual comparison,while it doesn’t have degree comparison.
2. The comparative marker bi is a dyadic degree operator that might take one or more than one argument, either individuals, times, locations, instruments, or even prepositions, as long as they are arguments of the predicate of comparison.
Assumption Two echoes with the background mentioned in Chapter 1. That is, comparative items can be DPs,PPs,and others like locative phrases and temporal phrases. Whafs special in Lin^ (2009) hypothesis is that they have to be arguments of predicate of comparison, which means they have to be assigned Cases.

‘‘Zhangsan was happier yesterday at school than Lisi is at home today.”
(19) [s Zhangsan [ap [np zuotian][Ap [pp zai xuexiao][Ap [Deg bh [Degp [np
Lisi] [Deg’ [Deg ti] [Deg [Npjintian] [Deg’[Deg ti][pp at jiali][Apkaixin]]]]]]]]]]•
Lin’s (2009) analysis of the sentence in (18) is shown in (19) where the degree word bi takes three arguments: the individual argument Lisi, time argument jin-tian (today), and the location argument zai jia-li (at home). And these arguments must be flanked by constituents of the same type. The whole Deg-shell headed by the degree word bi occurs as an adjunct degree phrase adjoined to the predicate of comparison.

Chapter Three Theoretical Framework
Since the syntactic research is done within the framework of the Generative Grammar, ifs necessary to provide an overview of both the theory itself and key concepts including c-command, binding theory and movement constraints as employed in this thesis.
3.1 An Overview of the Generative Grammar
The Generative Grammar, also known as the Transformational-generative Grammar (TG Grammar for short), first appeared in 1957 with the publishing of Syntactic Structures by Noam Chomsky. Over the past 60 years, the Generative Grammar has undergone constant innovation and development. As a direct result of its development, alternate theories like Lexical-Functional Grammar and Head-Driven Phrase Structure Grammar have also branched off this research program. As the originator of the Generative Grammar, Noam Chomsky was greatly influenced by philosophy and mathematical logic, which accounts well for why the Generative Grammar is characterized by formal abstraction and metonymical concretization. These features distinguish it from Cognitive Linguistics and System-functional Grammar.
Since the proposal of formal syntactic theory, syntactic constituents have become the focus of the Generative Grammar. Yet, when TG Grammar and Semantics had differences and contradictions, what was studied in the former expanded. In the course of its development, many different names have emerged, ranging from Transformational Grammar (TG), Standard Theory (ST), Government and Binding Theory (GB), to Principles and Parameters, and the Minimalist Program (MP). The blanket name, Generative Grammar, has been used to this day.
Syntax is often defined as the study of how sentences are structured. This definition is also the underlying thesis of the Generative Grammar. In generating a
series of sentences subconsciously in our minds, grammatical rules are needed, hence
19

the name generative grammar. These rules cannot be seen, but they do model the behavior of the mind. For example, we say She left me yesterday, but not She yesterday left me. This is an example of how adjacency principle works. In this sense, syntax becomes a scientific method in sentence structures.
3.2 Some Key Concepts
This thesis draws upon theories in the Generative Grammar before the period of Minimalism. Some key concepts like c-command, Binding Theory and Movement Constraints will be applied.
3.2.1 C-command
C-command is a structural relation on which binding is based. Carnie (2006) defined formal C-command in the following way:
C-command (formal): Node A c-commands node B if every node dominating A also dominates B, and neither A nor B dominate each other.
 
 
In this tree diagram, A c-commands all those nodes circled, for both node A and node B are dominated by the same node O yet they don’t dominate each other. Similarly, node N c-commands all nodes dominated by node O (O itself included).
C-command is an important and basic relation in this thesis. First, in Chapter 4, c-command must be applied in arguing against the proposal that bi + comparative standard is an adjunct. Second, the analysis of clausal ^/'-comparatives in Chapter 5 will be made by working out how c-command takes effect in the D-structures of BCC.
3.2.2 Binding Theory
Based on c-command, binding concerns a kind of coindexation (Carnie, 2006). It is coindexation that happens when one of the two NPs c-commands the other. It requires both coindexation and c-command. There are three binding principles. Principle A takes the anaphor, i.e., reflexives and reciprocals, as its research object. Principle B and C concerns pronouns and R-expressions respectively.
Binding Principle A (final): One copy of an anaphor in a chain must be bound within the smallest CP or DP containing it and a potential antecedent.
Binding Principle B (final): A pronoun must be free within the smallest CP or DP containing it but not containing a potential antecedent. If no such CP or DP can be found, the pronoun must be free within the root CP.
Binding Principle C: R-expressions must be free.
(Carnie 2006:412-415)
The above principles can be exemplified through the following examples.
(1) Zhangsanj ting-shuo Lisik bi zi-ji*j/k qu-nian kao de gao.
Zhangsan hear Lisi M self last year score DE high
 
^Zhangsan heard that Lisi scored higher than he did last year/5
 
 
(2) Zhangsanj ting-shuo Lisik bi taj/*k kao de gao.
Zhangsan hear Lisi M him score DE high
 
 
 
‘‘Z/za卿a” heard that ZzW scored higher than him.”
(3) Zhangsanj ting-shuo Lisik bi Wangwu*j/*k kao de gao. Zhangsan hear Lisi M Wangwu score DE high

In (1), we can treat bi + standard + predicate of comparison either as a CP or a DP for the purpose of discussion. Nevertheless, as no governor is available in this CP or DP, the smallest CP or DP serving as the governing category must be the embedded clause after ting-shuo {that being its complementizer in English interpretation). Thus, ziji can only be co-indexed with Lisi. Similarly, (2) is special in that ta (him) can be co-indexed with Zhangsan but doesn^ have to be. The reason is as there9s no possible smallest CP or DP required, ta (him) must be free within the embedded clause. Thus, crossing this embedded clause, the pronoun ta can refer to someone else in the context, or refer to Zhangsan, in this sense. As for (3), Wangwu is an R-expression, thus it should be co-indexed with neither Zhangsan nor Lisi.
3.2.3 Movement Constraints
In terms of movement constraints, there are quite a few. Yet, for the sake of present research topic, we only introduce subjacency condition here.
Subjacency Condition: Movement cannot cross more than one bounding node in a single step, where bounding nodes are IP and DP.
(Ouhalla 1999: 262)
(4) The news that [[the debate on pay dp] was futile ip] is pertinent.
(Crystal 2008: 342)
In (4), ifs legal to move PP on pay to the right of futile, for it crosses only one bounding node—[dp the debate]. Nevertheless, ifs otherwise ungrammatical if we move [pp on pay] to the right of pertinent. This is because in this case, PP has crossed both bounding node DP and the lower IP, which violates subjacency condition.
Subjacency condition helps when we analyze both English and Chinese comparative constructions. Examples are given as follows:
(5) Tom is smarter than they think Mary is.
(6) *Tom is smarter than the set mind that they think Mary is.
(7) Ta dui wu-li bi dui li-shi gan xing-qu de shi-shi,wo-men zhi-dao.
He to physics M to history feel interest DE fact we know “The fact that he’s more interested in Physics than in history is knows to us.”
(8) * Ta bi dui li-shi gan xing-qu de shi-shi, wo-men zhi-dao, dui wu-li.
Intended meaning: ditto.
(5) and (7) are grammatical. Yet, in the case of (6)? the invisible smart has crossed not only the smallest CP, but the DP set mind. Thus, the sentence is ungrammatical. The condition holds true in terms of (8).
The above theoretical framework serves as preparation for our later discussion on the formation of CC in Mandarin Chinese.

Chapter Four The Syntax of
4.1 The Syntactic Properties of Bi
Not only is controversy commonplace in English comparative marker, it exists in Mandarin Chinese as well. In studying comparative constructions in Mandarin Chinese, most light is shed on comparative marker bi, including the word class and the relation between bi and items that follow it. Presently, there are at least three word class definitions on marker bi, namely, conjunction, verb and preposition, which lead to three distinct syntactic analyses. In this chapter, we will discuss the syntactic properties of bi and probe its syntactic nature. We assume that the comparative marker bi in unequal constructions is not a preposition, a verb, a conjunction, or any other functional element, but a light verb. We will provide evidence against the previous word class definitions of marker bi, and then we will raise our own hypothesis.
4.1.1 Bi—Non-prepositional
From a diachronic perspective, prepositions in Chinese are derived from verbs. That is, the presently so-called preposition once served as a main verb. Yet, under the influence of specific mechanism, the original meanings in these main verbs were gradually delexicalized. As a result, their usages changed, and they were later grammaticalized to prepositions. Zhu (1982) suggests that Chinese prepositions, more often than not, bear verbal features. Although prepositions aren’t predicates, the construction preposition + object + predicate closely resembles that of serial verb construction.
In Chen^ (2002) opinion, some prepositions in Mandarin Chinese have completely lost their features when first used as verbs and were thus changed into typical prepositions, such as cong (from), duiyu (for), zi (since). Other prepositions like zai (in), dao (till), bi (than) also share verbal attributes. Words like these roll verbs and prepositions into one.
Based on the study of Mandarin Chinese, the above are generally classified into prepositions. Yet, compared with typical prepositions, comparative marker bi has many distinctive features in syntactic function. The following three features to be analyzed show that bi doesn^ necessarily have those typical features of a preposition.
First, bi + comparative standard as a whole cannot be put before subject. This resembles the case of ba/bei + NP, exemplified below. On the contrary, typical prepositions can be placed before subjects. For instance, prepositions like cong (from), duiyu (for), combined with complements, can be put either in the front or to the end of sentences.
(l) a. Zhangsan bi ta piao-liang.
Zhangsan M she beautiful
“Zhangsan is more beautiful than her.”
*Bi ta Zhangsan piao-liang.
b. Zhangsan ba ta zou le yi dun.
Zhangsan Prep-Ba she beat Asp-LE one time
‘‘Zhangsan beat her.”        
一 *Ba ta Zhangsan zou le yi dun.    
 
 
 
c. Zhangsan cong Shanghai dai hui-lai Hang he li-pin.
Zhangsan from Shanghai bring back two box gift
“Zhangsan took home two boxes of gifts from Shanghai.”
— Cong Shanghai Zhangsan dai hui-lai liang he li-pin.
—Zhangsan dai huilai liang he lipin, cong Shanghai.
d. Zhangsan dui zhe jian shi mei-you fa-biao ren-he yi-jian.
Zhangsan on this CL matter Neg make any comment
‘‘Zhangsan didn’t make any comment on this matter.”
—  Dui zhe jian shi Zhangsan mei-you fa-biao ren-he yi-jian. —Zhangsan meiyou fabiao renhe yijian, dui zhe jian shi.

Second, negation form bu must always precede marker bi. In many languages across the world, the comparative standard is negated directly. However, in terms of Mandarin Z?/-comparatives, bu can only appear before bi. In discussing Chinese prepositions, negation bu can appear not only before prepositions to negate PP as a whole or XP serving as complement to prepositions, but also before verbs to negate VP as a whole or XP serving as complement to verbs. Given below are relevant examples.
(2) a. Zhangsan bu bi Xiaoli cong-ming.
Zhangsan Neg M Xiaoli smart
“Zhangsan is less smarter than Xiaoli.”
—*Zhangsan bi Xiaoli bu congming.
b. Zhangsan bu bi ta tiao de gao.
Zhangsan Neg M he jump DE high
“Zhangsan jumps less higher than he is.”
— *Zhangsa bi ta bu tiao de gao.
c. Zhangsan bu dui zhe ge wen-ti fa-biao yi-jian.
Zhangsan Neg on this CL issue make comment
“Zhangsan didn’t make any comment on this issue.”
—Zhangsan dui zhe ge wen-ti bu fa-biao yi-jian.
d. Zhangsan bu cong Shanghai qu Meiguo le.
Zhangsan Neg from Shanghai go America Asp
“Zhangsan didn’t travel from Shanghai to America.”
—Zhangsan cong Shanghai bu qu Meiguo le.
Ifs obvious that all the above sentences are grammatical when negation bu
precedes bi or prepositions. Yet, ifs another story when bu precedes the predicate of comparison, as is shown in control groups in (2a-b). This is quite different from the grammatical control groups in (2c-d). That is to say, ifs unacceptable to put negation
bu before the predicate of comparison in syntactic structure. However,
26
counter-examples are everywhere as shown in (3).
(3) a. Zhangsan bi Lisi you9 bu congming duoshao.
Zhangsan M Lisi EMPH Neg smart much
“Zhangsan is not that smarter than Lisi.”
b. Zhangsan bi ta you bu gaoxing duoshao.
Zhangsan M he EMPH Neg happy much
“Zhangsan is not that happier than him.”
Based on the above counter-examples, those sentences that were previously considered ungrammatical become acceptable after the insertion of you. As is mentioned above, ifs not allowed to put bu immediately before the comparative subject. Yet, negation focuses on the comparative subject rather than the comparative standard. This is similar to many other languages and can be tested by various means.
According to speakers9 intention, what is truly negated by bu is the comparative subject. Lefs say, bu bi is different in meaning from meiyou, for the former does not necessarily highlight which one is stronger, whether it is the comparative subject or the comparative standard, but merely focuses on differences. However, the basic meaning of the latter is X is about the same as Y. Lv et al (1999) also believe what is expressed in bu bi is that there^ a difference between A and B, with only two possible cases expressed. 7〜fT generally presupposes Z fT. For example, TVa 双
bu bi zhe ge difang liangkuai (That place is no cooler than this place) presupposes Zhe ge difang hen bu liangkuai (Ifs quite hot here in this place). In this case, ifs likely to get the proposition <X-W>? namely, Na ge difang liangkuai (Ifs cool in that place.). However, ifs also possible that speakers agree in an implicit way that listeners hold <X-W> to be true. Under this circumstance, the sentence elicited by bu bi can be employed to refute it in order to hold Y. fT to be true. This shows in using bu bi, speakers always seem to retort ideas and intentions boasted by listeners. With this, the   conclusion can be drawn: when speakers are using bu hi in sentences, they are not negating the comparative standard, but rather, underscoring the fact that the comparative standard doesn’t have the same features as the comparative subject. Negation bu focuses on the comparative subject.
We can even prove the specific negative focal point in bu hi CC based on the relation between negation and focal point. As is pointed out by Lv (1985), there is, more often than not, a negative focal point in negative sentences. The very focal point is usually a part in the end of a sentence where stress falls. Yet, when a comparative stress lies in the head, the negative focal point will move to where the former stress is. Lv^ (1999) opinion is approved of by Xu and Li (1993). That is, the choice of a negative focal point is dependent on the choice of the focal point independent of negation itself. Liu (2012) held that in unequal CC marked with bi, the comparative subject is topical. If it holds true, the part that enjoys topicalization cannot be the focal point in a sentence. Therefore, the negative focal point can only be the comparative standard after the subject, and that is where the natural stress falls.
In addition, as is mentioned in the previous chapter, what follows bi varies from NP, AP/VP, IP or even PP. Different from English prepositions which can only head NP, Chinese prepositions are generally considered to be able to head not only NP, but VP, IP as well. Given below are relevant examples:
(4) a. Xiaozhang cong mei-guo huilai le. (NP)
Xiaozhang from America return Asp
“Xiaozhang has returned from America.”
b.  Zicong li-kai xue-xiao, wo zai mei shang guo xue. (VP)
Since leave school I again not go Asp study
“Since leaving school, I haven’t had any further study.”
c.  Zi ta shang le da-xue, wo-men bu zai lian-xi le. (IP)
Since he go Asp university we not again contact Asp
“We haven’t contacted each since he went to university.”
d. Dui-yu ni lai zuo ban-zhang, wo mei-you shen-me yijian.(IP)
28
About you to be monitor, I not have any objection “For you to be monitor, I’m cool with that.”
Yet, what follows bi is so encompassing as to include all categories in Chinese. Compared with typical prepositions, unequal comparison marker bi has a significant design feature and that is, bi can be followed by PP. Examples are given below:
(5) a. Zhangsan bi Lisi gao. (NP)
Zhangsan M Lisi tall “Zhangsan is taller than Lisi.”
b. Xiaowang dui Zhangsan lai zuo ban-zhang bi dui Lisi lai zuo
Xiaozhang for Zhangsan to be monitor M for Lisi to be
Ban-zhang geng bu-man. (VP)
monitor more unsatisfied
''Compared with Lisi, Zhangsan is the one with whom Xiaowang is more unsatisfied for his being appointed minitor.”
c. Ni pi-ping wo bi wo ling-dao pi-ping wo hai yan-li. (IP)
You criticize me M my head criticize me even stern
“You criticized me more sternly than my head did.”
d. Jiao-ao bi xu-xin wei-hai geng da. (AP)
pride M modesty danger more big “Pride is more dangerous than modesty.”
e. Qu Dalian cong Beijing bi Shanghai geng jin. (PP) go Dalian from Beijing M Shanghai more near “It’s nearer to travel to Dalian from Beijing than from Shanghai.”
What we can say with certainty is that the comparative marker bi does not have due syntactic features shared by general prepositions. However, you may refute our idea by questioning like this: Since dangling prepositions are not allowed in Chinese
and given the fact that ^/-comparatives without the comparative standard are
29

ungrammatical, why can’t W be a preposition? Yet, dangling categories with no objects include not only prepositions. Based on Yuan (1996), verbs like relative verbs cannot go without objects. We might as well compare the following sentences. In them, a dangling part after bi or shu-yu (belong to) is never allowed
(6) a. Xiaowang bi Xiaozhang geng gao-xing.
Xiaowang M Xiaozhang more happy “Xiaowang is happier than Xiaozhang.”
Xiaozhang, Xiaowang bi ta geng gao-xing.
*Xiaozhang, Xiaowang bi geng gao-xing.
b. Gege shi didi ku le qilai.
Elder brother make younger bro cry Asp happen ‘The elder brother made his younger brother cry.”
Didi, gege shi ta ku le qilai.
^ *Didi, gege shi ku le qilai.
c. Tongxian shu-yu Beijing.
Tongxian belong to Beijing “Tongxian County belongs to Beijing.”
Beijing, Tongxian shu-yu na-er.
^ *Beijing? Tongxian shu-yu.
The above examples show to us whether dangling object is allowed can^ be the sole means to distinguish between verbs and prepositions.
4.1.2 Bi—Non-verbal
Comparative marker bi lacks not only prepositional features, but also verbal features. Chinese academia classified bi into verb bi\ and preposition bh. We have
proved bh to be non-prepositional before. Now, by means of three syntactic tests, we are going to discuss differences between bh and verb bi\.
Test 1: Stand-alone Test. Typical verbs can be main verbs in sentences. They can stand alone as answers to questions. We might as well refer to the following cases:
(7) a. Wo he Xiaozhang bii saipao.
I and Xiaozhang bii running “Xiaozhang and I had a running race.”
b. …Ni gen Xiaoming jin-tian bii shen-gao le?
You and Xiaoming today bii height Asp
“Did you compare your height with Xiaoming today?”
…Bii le.
Bii Asp “I did.”
c. Zhangsan b'n Lisi gao.
Zhangsan b\2 Lisi tall “Zhangsa is taller than Lisi.”
—^Zhangsan b\2 Lisi.
Zhangsan b'n Lisi *Zhangsan than Lisi.
d.  …Xiaozhang zai xue-xiao W2 zai jia-li hai gao-xing?
Xiaozhang at school bi2 at home even happy uWas Xiaozhang happier at school than at home?^
…*Shi-de, bi2 le.
Yes W2 Asp
The above examples show that bi\ can serve as main verbs, as is shown in (7a). Shown in (7b), they can stand alone as a grammatical sentence. Thus, bii can be verbs. Yet, bi2 is quite the opposite. Comparative marker bh can neither be main verbs,

shown in (7c)? nor stand alone as answers to questions. Therefore, bii might not be a verb, at least not a typical verb.
Test 2: Addition of aspect markers Zhe (progressive), Le, Guo (perfective). Typical verbs can precede aspect markers. Refer to the following cases:
(8) a. Guang zhe yi xiang, Zhangsan he Lisi bii le yi xiao-shi.
Only this one item Zhangsan and Lisi bii Asp one hour
“It took Zhangsan and Lisi one hour to compete simply on this task.”
b. Ta-men lia bii zhe bii zhe jiu da le qi-lai.
They two bii Asp bii Asp then fight Asp happen
“They two fought when they were competing against each.”
c. An cheng-xv, yao bii guo shengao hou cai bi tizhong. By order need bii Asp height after when bi weight Weight comparison is programmed after height comparison.
d.  *Xiaowang W2 le Xiaozhang gao.
Xiaowang bh Asp Xiaozhang tall.
“Xiaowang is taller than Xiaozhang.”
e. *Qu Xinjiang bh guo qu Heilongjiang geng haowan.
Go Xinjiang bh Asp go Heilongjiang more exciting “Xinjiang is a more exciting place to travel to than Heilongjiang is.”
Based on above examples, what can be easily found is that aspect markers Zhe (progressive), Le, Guo (perfective) can be attached to bi\3 which means bii is a verb. Yet, ifs not so with bii. That is, comparative marker bi also lacks general features shared by typical verbs.
In short, two syntactic tests vividly demonstrate that marker 6/2 doesn’t distribute in the same way as general verbs. This denotes that bii is not a typical verb.
Test 3: Insertion of adverbs with strong active senses, namely, adjuncts carrying a strong active sense. These adjuncts include gu-yi (deliberately), xiao-xin (carefully),
si-yi (recklessly). In terms of typical verbs in Mandarin Chinese, we can insert these
32

adverbs before them without changing the original meanings as a whole. Take a look at the following instances: (all adjuncts inserted have been underlined)
(9) a. Lisi da sui le huaping.
"Lisi broke the vase.^
Lisi guyi da dui le huaping.
“Lisi deliberately broke the vase.”
b. Lisi xia le che.
“ZW got off the bus.”
—Lisi xiaoxin de xia le che.
"Lisi got off the bus carefully.”
c. Lisi cai hua.
“Lisi picked flowers.”
Lisi si-vi de cai hua.
“Lisi picked flowers recklessly.”
d. Lisi bi ta hao-qi-xin zhong.
“Lisi is more curious than him.”
*Lisi guvi bi ta hao-qi-xin qiang.
^Lisi is deliberately more curious than him.
e. Lisi shu-xue xue de bi ta hao.
“Lisi learns Math better than him.”
-^*Lisi shu-xue xue de xiaoxin bi ta hao.
*Lisi learns Math carefully better than him.
f. Lisi bi Zhangsan ai da lan-qiu.
"Lisi takes a greater interest in playing basketball than Zhangsan^ —^Lisi siyi de bi Zhangsan ai da lan-qiu.
^Lisi takes recklessly a greater interest in basketball than Zhangsan.
At first glance, we might find that for all the sentences above, (9a)-(9f), in terms of theta role, Lisi serves as the agent. Yet, this being true, why (9d), (9e) and (9f) are ungrammatical? The reason is that compared with (9d), (9e) and (9f), (9a), (9b), and (9c) are more agentive. Thus, verbs like da-sui (break), xia-che (get off), and cai (pick) are better at expressing their verbal senses. Based on it, we are able to explain the counter-examples then. The last three control groups are ungrammatical in that they are not that agentive themselves. Grabbing this justification, we can further conclude that their umain verbs^, bi in this case, is not verbal enough itself. This test again confirms our proposal that marker bi is non-verbal.
4.1.3 Bi—Non-con junctional
First, if bi is treated as a conjunction, both comparative items are parallel with each other in terms of word class and meaning. Moreover, both comparative items, when separated, can form a logical relation with the predicate. Besides, as for typical conjunctions like he, yu (and), we can alter the relative positions of DPs connected by them without changing original meanings. Take the sentences below as an example:
(10)  Xiao-yu he Xiao-sun qu le Shanghai.
Xiaoyu and Xiao-sun go Asp Shanghai
uXiaoyu and Xiaozhang went to Shanghai.55
Xiao-sun he Xiao-yu qu le Shanghai.
(11) Zai Shanghai bi zai Shenyang ya-li da.
in Shanghai M in Shenyang stress big
“It’s more stressful to live in Shanghai than in Shenyang.”
*Zai Shenyang bi zai Shanghai ya-li da.
This analysis does highlight the function of bi in coordinating sentences. However, when positions of comparative items get exchanged, the original meaning changes as well. This semantic change means bi as a conjunction does not hold.

(12)  Zhangsan ge-zi bi Lisi gao san-fen-zhi-yi.
Zhangsan height M Lisi tall one third “Lisi is 3/4 as tall as Zhangsan.”
In sentence (12), the measure phrase san-fen-zhi-yi (one third), can only modify Lisi rather than Zhangsan. Obviously, this goes against normal conjunctional functions, leaving bi as a conjunction inaccurate.
Second, in terms of typical Mandarin conjunctions, adverbial modifiers are not allowed. This is because once modifiers are applied, the original meanings cease to exist. Refer to the following examples: (adverbs before conjunctions have been underlined)
(13) *Wo kai-xin de gen ta chang zhe ge.
I happily DE with her sing Asp songs
“I sang songs happily with her.”
—*Ta kai-xin de gen wo chang zhe ge.
‘‘She sang songs happily with me.”
(14) Ta qi-ma bi wo da wu sui.
he at least M me senior five year
“He is at least five years older than me.”
We have to pay special attention to (13). (13) is grammatical and acceptable in Mandarin Chinese, yet it is illegal here, for the control group shows quite the opposite meaning. That is, the insertion of the adverbial modifier kai-xin-de (happily) directly indicates who is real happy in the act of singing. According to the first test used, the insertion of it will not change the state of happiness when we change the order of the wo (I) and ta (he) in it. But it does. Gen (with) under this context is a preposition rather than a conjunction. Based on above, (14) should have been ungrammatical in
that no adverbial modifier is allowed to closely precede the conjunction bi. Yet, the
35

fact that (14) is both grammatically and semantically correct shows that marker bi is not a typical conjunction.In this case, bi is more of a preposition than a conjunction. However, as is illustrated previously, bi is not a typical preposition.
4.1.4 Bi—Non-functional
We will argue in this part against the proposal that bi is a function word. Refer to
(15):
(15) a. Xiaowang da Xiaozhang wu sui. Xiaowang old Xiaozhang five year b. Xiaowang bi Xiaozhang da wu sui. Xiaowang M Xiaozhang old five year
“Xiaowang is 5 years older than Xiaozhang.”
Xiang (2005) lists three NPs in (15): the comparative subject Xiaowang, the comparative standard Xiaozhang, and the measure phrase wu sui (5 years). Based on Larson^ (1988b) VP-shell analysis, Xiang suggests that there should be a DegP shell, which is similar to double object structure, in (15). The comparative predicate da (old) is a three-place predicate, with Xiaowang serving as its external argument, the standard Xiaozhang and the measure phrase wu sui (5 years) as its internal ones. In order for three arguments to get their due positions, an empty degree morpheme, denoting older than and phonologically implicit, combines with two internal arguments in lower DegP. They then get moved to AP internal position based on head-movement. To meet EPP, the comparative standard Xiaozhang is moved to Spec position. In this way,a structure similar to small clause’s thus arises. Lastly, to make external argument Xiaowang stand out, the head combination da wu sui moves to the head position of higher DegP through head-movement. In so analyzing, (15a) can be obtained. (15a) and (15b) are generated in similar ways, for they are close to each other in semantic respect. In (15b), marker bi, serving as functional head,
combines immediately with AP and they are projected into higher DegP.
Still, there are limitations in Xiang's DegP shell approach. The first one concerns semantic question as to AP position. In both analyses, AP forms a small clause in which Xiaozhang serves as the subject. Semantically, this AP is interpreted as JT/aoz/zawg (Xiaozhang is 5 years older), which is contrary to the intended
meaning. The second question deals with treating marker bi as a three-place predicate like gei (give) and fang (put). The fact is that bi is not the same as the latter. It can be easily worked out once we leave out measure phrase, wu sui in this case. This measure phrase is unnecessary in that omission of it will not affect the grammaticality of the whole sentence. However, ifs quite the opposite in terms of typical three-place predicates.
4.1.5 Bi—Non-adjunctive
The structure of bi sentences is generally analyzed as comparative subject + bi + comparative standard + predicate of comparison. Bi + comparative standard is adjunct in the sentence (known as adverbial in Chinese) and is added in relation to parameter. In this sense, Zhangsan bi Lisi gao (Zhangsan is taller than Lisi) shares the same basic meaning with Zhangsan gao (Zhangsan is taller), with only different comparative standards.
(16) a. Zhangsan bi Lisi gao.
Zhangsan M Lisi tall “Zhangsan is taller than Lisi.”
b. Zhangsan hen gao.
Zhangsan very tall “Zhangsan is very tall.”
c. Zhangsan geng gao.
Zhangsan more tall.
“Zhangsan is taller.”
Based on Chu (1999), comparison can be classified into two types—general comparison and single comparison. Zhangsan Gao (Zhangsan is tall) should bear the same meaning with Zhangsan hen gao (Zhangsan is very tall). They are general comparative sentences. By contrast, Zhangsan hi Lisi gao (Zhangsan is taller than Lisi) is thought to mean the same as Z/zan炉an gao (Zhangsan is taller). These are two
single comparative sentences. In this sense, the above (16a&b) sentences are not identical. Then arises the question: is bi + comparative standard still an adjunct? This will be discussed later in this thesis.
Xu (2001) maintains that traditional SVV sentences are the overlap of two IPs, for instance:
(17) a. Ta gezi gao.
He height tall “He is tall.”
^[ip[spec Ta][r[ip[sPec gezi][r[AP gao]]]]].
b. Zhe ke shu yezi da.
This CL tree leaf big “The tree has big leaves.”
—[IP[Spec Zhe ke shu][r[ip[spec yezi][r[AP da]]]]].
We hypothesize that bi + comparative standard in Zhangsan bi Lisi gezi gao (Zhangsan is taller than Lisi) is PP serving as an adjunct. The truth is that the whole PP can only be added under secondary IP instead of the main IP, analyzed below:
(18) a. [ipZhangsan [ip gezi gao]]. [ip Zhangsan [pp bi Lisi][ip gezi gao]].
b. [ip Zhangsan [ip gezi gao]]. —►[ip Zhangsan [ip[pp bi Lisi ] gezi gao]].
c. [ip Zhangsan [ip gezi gao]]. —[ip Zhangsan [ip gezi [pp bi Lisi] gao]].
The comparative subject can c-command the comparative standard, thus the whole comparative sentence should obey Binding Condition B. That is, when the comparative standard is replaced by a pronoun, the latter can be co-indexed with the comparative subject. Refer to the following cases:
(19) a. [ip Zhangsan^ [ip bi tavy gezi gao]]. b. [ipZhangsarb [ip gezi bi ta*i/j gao]].
Based on (19), Zhangsan can not co-index with the comparative standard ta, which means the original hypothesis bears no ground at all. According to He (2011), with regard to English and Chinese pronouns, the governing domain is the same. Both include simple sentence, matrix clause, and pivotal structure plus small clause. As Zhangsan is in different domain from ta, the whole comparative sentence is not derived from the combination of typical SVV structure Zhangsan gezi gao {Zhangsan is tall) and adjunct bi + comparative standard. But rather, bi sentences like this have a lower IP in the matrix clause. Obviously, bi + comparative standard is not an adjunctive PP.
4.2 Our Proposal
Based on the above analyses, the unequal comparative marker bi contains neither syntactic features typical verbs are supposed to have, nor design features typical prepositions have. Then, how to define the syntax of marker bil
We tend to treat bi as a light verb.
The theory of light verbs has gained its development since it was first proposed by Chomsky (1995). Ifs considered in Chomsky (1995) that light verbs should be semantically empty. They serve mainly to convey grammatical meanings. That is, light verbs refer to meanings simple in connotation yet shared by many verbs. Whafs more important is that light verbs have to perform their syntactic functions. That is, they must undergo syntactic operations. In terms of grammatical performance, light

verbs are more of a type of functional category, similar to auxiliary verbs which take VPs as their complements. In terms of phonation, based on Shen et al (2001), ifs more likely that light verbs play the role of affixes like bound morphemes whose existence must be dependent on roots, or have no phonological content at all and thus can never stand alone as a word.
In Huang (2008), he pays special attention to light verbs in Mandarin Chinese. Ifs considered that Mandarin Chinese is an analytic language which has lost its many syntactic features. One of the design features of highly analytic languages is that each semantic unit has to be conveyed by independent lexical item. Based on Huang (2008), light verbs are no longer empty category or affix, but eventuality predicates that have their own semantic content. Huang’s (2008) idea is similar to that of Lin’s (2001).
In treating bi as a light verb, we have the following justifications.
First, based on Jespersen (1942), such verbs as have, take and give should be called light verbs, for these verbs, compared with constituents following them, are not as important. It is considered in Jespersen (1942) that light verbs are light for semantic reasons. For instance, the otherwise verbal take and give both have their semantic contents. One means to carry or move something from one place to another, and the other to hand something to someone. However, in phrases like take a walk and give someone a kiss, take and give are less important in meaning than a walk and a kiss are. That is, their original lexical meanings fail to be obvious. Moreover, both phrases can be replaced simply by verbs walk and kiss without changing their original meanings. Thus, take and give are treated as light verbs. Since being ulight,5 is one feature of light verbs, we can employ it as a test to see if bi is qualified for a light verb. Refer to the following example:
(20) a. Mary bi Tom gao. b. Mary gao guo Tom.
“Shared interpretation: Mary is taller than Tom.”
The control group in (20) shows that the comparative marker bi has the same
40
quality of being ulighf, as take and give do. First, (20a) and (20b) have the same meaning. Second, even when we substitute gao (tall), the comparative result in (20a)? for marker bi, the acceptability doesn^ fade at all. Thus in the whole ^/-comparative, the AP comparative result shoulders more responsibility in meaning expression than bi does. Moreover, the meaning of bi in dictionary, i.e., the lexical meaning, is weak. Bi is used to compare differences in trait and degree. Bi can be reduced to a light verb.
Second, according to Dai (2017), light verbs assign Cases and take arguments just as typical verbs do. We have ruled out the prepositional, verbal, conjunctional, and adjunctive analyses. Among them, verbal analysis may be used as proof to argue against our proposal. This counter-example can be falsified. Refer to the following example:
(21) a. Mary bi Tom pao de kuai.
Mary M Tom run DE fast.
b.  Mary pao de bi Tom kuai.
“Shared interpretation: Mary runs faster than Tom.”
c. John sent Mary a letter.
d. John sent a letter to Tom.
Based on the above control group, the direct and indirect objects of a ditransitive verb receives their theta roles from the verb and are preceded by it. (21a) is an example that is in line with the fact. Yet, (21b), which is also well-formed in Mandarin Chinese and shares the same meaning with (2la), violates the fact should we treat bi as a verb. The problem appears: since the comparative result kuai (fast) is a must in this ^/-comparative, it serves as one argument of bi, then bi alone takes three arguments. In this case, bi has to assign nominative Case to the comparative subject Mary before it gets DP-moved to [Spec, TP]. However, ifs illegal in this example, for in the embedded CP pao de, there is also the nominative Case to be assigned and pao also requires one argument. In this case, bi can only be reduced to a light verb and
merges with the a higher pure verb to check the Case feature.
41

The question thus follows is: typical light verbs like send and give select for a pure verb, then which kind of pure verbs should bi select for?
Cao (1997) considers bi as a co-verb. Co-verb is generally considered as a special category in the study of Chinese grammar. It once served as main verbs, but has now lost its main features as a verb, though having not been grammaticalized to complete prepositions. This means marker bi is neither a verb in lexical sense, nor a typical preposition. Huang (1997) looks upon the position of light verbs as equivalent to co-verbs. He thus attributes the same features to light verbs and co-verbs, like bi and ba. Due to use limitations of co-verbs, we will classify marker bi as a light verb in this article. Followed next will be syntactic derivation of bi structures.
Based on Huang (1997), in terms of Lexical Relational Structure, eventuality exists in verbs. Verbs with eventuality is embedded in abstract pure verbs. According to him, for example, active verbs, including unergative and transitive ones, are complement of DO. Similarly, inchoative verbs are embedded in BECOME, while stative verbs are embedded in HOLD. Based on Carnie (2006), causative verbs are embedded in CAUSE. Take a look at the following examples:
(22) mo }hei malign—causative—[CAUSE [Amohei B]]
(23) re }ai love一stative一[HOLD [A reai B]]
(24) chi Jan eat一active一[DO [A chifan]]
Ifs essential to state clearly why we treat comparative marker bi in Mandarin Chinese as a light verb embedded in HOLD. Look at the examples:
(25) Tabi ni jinshen.
“He is more careful than you.”
(26) Zhangsan yinwei Lisi shuo huang bi Wangwu yinwei ta dajia geng fennu.
"Zhangsan is angry more because Lisi told lies than Wangwu is angry because Lisi fought with others.
In sentence (25), the comparative result is AP jinshen (careful), with ta (he) and ni (you) being the comparative subject and standard respectively. In sentence (26), reason clauses are compared. The truth is whatever kind of argument is compared, there’s always a degree difference. The feature of this degree difference consists in the fact that it usually will not change with the passage of time. Take sentence (25) as an example, and we will have no idea as to when ta (he) became more careful than ni (you), nor do we know when ni (you) will otherwise become more careful than ta (him). It is the state that he is more careful than you that counts. Thus, we consider bi as a light verb which is put inside HOLD.
4.3 Summary
In the previous sections of this chapter, a succession of syntactic tests have been employed to demonstrate that problems do exist, whether we classify bi as a preposition, a verb, a conjunction, or an adjunct. These classifications are imperfect, for at least there are some design features that typical verbs and conjunctions are expected to have yet bi lacks. Till now, no consensus has been achieved on this topic. With the help of both semantic and syntactic means, we propose that bi is a light verb embedded in HOLD.
Yet, to nail down the category of bi is merely the first step of our arduous work. Subsequent questions like what the theta grids for HOLD and bi are, how to distinguish between the direct object and the indirect object (DO and IO for short) of bi, how Case filter is achieved, and what the D-structures of clausal and phrasal W-comparatives are remain to be answered.

Chapter Five The Syntax oiBi Comparative Construction
In Chapter 4, a list of evidence has been given on the possible word class of marker bi. Possibilities of bi treated as a conjunction, a preposition, and a verb have been ruled out respectively. We then make the proposal that bi is a light verb which contains the sense of HOLD. Also, related evidence has been given in arguing for it. In this chapter, we will analyze the syntax of bi comparative construction. A different perspective from previous ones will be taken. That is, we assume that neither phrasal nor clausal analysis makes complete sense. Phrasal comparatives are not derived from clausal ones. Thus how to analyze the syntax of bi comparative construction will be the central point of this chapter.
5.1 The Syntax of Phrasal ^/-comparatives
Based on antecedent-contained deletion, phrasal Z?z-comparatives are derived from clausal ones by deleting the same component, whose antecedent can be traced in the comparative subject. In this way, the previously clausal comparative standard is shortened into a phrasal one. However, we tend to treat it in another way. Look at the following examples:
(1) Mary bi wo da wu sui.
Mary M me old five year
‘‘Mary is five years older than me.”
(2) John sent a letter to Mary.
(3) John sent Mary a letter.
(Larson 1988b: 352)
In treating bi as a light verb, we make a similar analysis to Larson5s (1988b) double object construction. Larson (1988b) puts forward the analysis of VP-shell for
double object construction and argues that (3) is derived from (2) under the operation
44
of the dative shift. The following structures are given to exemplify Larson^ (1988b) analysis.
(4) a. D-structure: IP

a letter V pp
 
b. S-structure:
VPi 
 
八A
NP a letter
When Mary, the indirect object, gets moved forward, the verb sent loses its
inherent Case to so that the preposition 如 is deleted. The direct object a /以纪r is
dethematized as an adjunct and adjoined to the V in VP2. Finally, the verb send raises
45
to the head position of VPi, assigning Case rightward to the indirect object Mary. In D-structure, the indirect object Mary is Caseless and the VP subject position is non-thematic. Then the indirect object undergoes DP movement to the VP subject position in (4b)? leading to the S-structure form for the VP in John sent Mary a letter in S-structure.
In light of Larson’s (1988b) analysis of the double object construction,we hypothesize that the phrasal 6z-comparatives are similar to it. Yet one question appears. That is, what are the theta grids for HOLD and bil The answer to this question needs to be straightened out first. Before figuring it out, refer first to the following examples:
(5) a. Mary bi Tom gao.
“Mary is taller than Tom.”
b. Mary bi Tom gao wu gongfen.
“Mary is five centimeters taller than Tom.”
c. *Mary bi Tom.
d. *Mary bi gao.
e. *Bi Tom gao.”
Sentences in the above control group show the argument structure of marker bi. The contrast between (5 a-b) and (5 c-e) demonstrates that at least three arguments are needed in ^/-comparatives, for lacking any single one can result in the ungrammaticality in (5 c-e). A further notice of (5a) and (5b) proves that the measure phrase wu gongfen (five centimeters) only serves to supplement its degree of difference but is not in the least an indispensable part of the matrix clause (5a). This being true, the possibility of the measure phrase serving as an argument can be ruled out. Then, to account for how the [dp wu gongfen] fits in the structure, we propose that

it can be the adjunct of the adjective gao (high) rather than its complement, because leaving it out does not influence the acceptability of the AP. By analyzing so, we contend that [ap gao wu gongfen] is in the following diagram:
 
wu gongfen
gao
Another question that follows naturally is: which is closer to hi, the comparative standard or the comparative result? Namely what is the DO of bil This question matters as it concerns the movement and Case assignment of the DO. As we make a similar analysis of ^/-comparatives to that of Larson^ (1988b) , one plausible answer that can be required with ease is: we can first add a preposition in front of either of them, and then exchange their original order to see if the newly-formed sentence still makes sense. There is only one possible case. That is the construction of bi + PP-comparative result + comparative standard. However, this construction can be eliminated very soon for two reasons. First, hardly can such a preposition preceding the comparative result be found. Second, in Mandarin Chinese, the comparative standard is seldom seen to be preceded by the comparative result. Yet, there’s still one way out. We can resort to the Universal Grammar between the lines of which lies the adjacency parameter positing that in SVO languages like English and Chinese, nothing can be inserted between direct object and the verb in D-structure. Assuming this, we obtain the following results after the test has been performed on them.
(6) a. Mary bi Tom hai gao.
Mary M Tom even tall .Mary is even taller than Tom.

b. *Mary bi hai Tom gao. “Mary is even taller than Tom.”
The fact that (6a) is well-formed yet (6b) is ungrammatical shows the adjacency parameter is valid in telling apart the DO and the IO. The insertion of hai (even) before AP gao (tall) doesn^ affect the acceptability of the sentence. However, when hai (even) is put before DP Tom, the sentence becomes ungrammatical. The sharp contrast thus proves that Tom is closer to the marker bi, while gao (tall) is less so than previously thought is. By analyzing so, we can acknowledge that after the application of the X-bar theory, the comparative standard is right the sister of bi in the D-structure. That is, the former serves as the complement of the head bi. It is by moving the comparative standard upward to a higher position that ultimately leaves it precede the comparative result in S-structure.

Till now, we have drawn two conclusions. The first one is that at least three arguments are needed in HOLD and bi. The second one is that contrary to ditransitive verbs like give and tell in whose cases DPs that are given or told are IO, it is the comparative standard that serves as DO in ^/-comparatives. However, as is mentioned above, based on Case filter, DO must be assigned the accusative Case.
(Carnie 2006:419)
Based on the above diagram, we can propose that there’s a fourth argument in theta grids. That is, the HOLD selects for an AgrOP (standing for object agreement phrase) to achieve a Case position. Thus finally, we have the theta grids for the HOLD and bi.

HOLD
Agent
DP
AgrOP
i j
 
 
hi
Tlkeme
DP
Proposition
AP/CP
k i
 
 
After the theta grids have been worked out, we can draw syntactic trees for phrasal ^/-comparatives and make justifications for them. Yet, as is referred to before, Xiang (2005) also treats bi as a light verb. To illustrate our differences in analysis, we will contrast his argument against ours. Take Mary bi wo da wu sui in (l) for example, based on Xiang, we obtain the following syntactic structure in (7):
(7) Syntactic structure for (1):
TP
 
^UErUl
 
Xiang (2005) suggests that the first DP is the comparative standard while the second one indicates the value of different degrees. The phrasal 6/-comparative can be regarded as the syntactic behavior of the degree of predicative adjectives. Similar to Larson^ (1988b) analysis, the degree predicative takes two internal arguments: the comparative standard in the Specifier position of AP, and the comparative result in the complement position of AP. However, Xianganalysis fails to explain how the comparative standard is Case assigned, nor does it account for how theme is assigned in the diagram. Moreover, even if we put aside the question whether DP2 wusui (five years) is complement of the AP or not, one misleading part in his analysis is that the comparative standard^ final movement to [Spec, VP] can leave us the false impression that wo da wusui (Pm 5 years older), which is quite contrary to the original intended meaning that Mary da (wo) wusui (Mary is 5 years older).
Considering the weak points in Xiang?s argument, we now exemplify the same example with the tree diagram shown below:

(8) Syntactic structure for (1):
 
  
 
No trace is left in the tree diagram14 and instead, movement lines are cleared marked to better show how these items are base-generated after Lexicon selection and how head movement and DP movement are achieved. In the D-structure of the expanded VPs, the subject DP Mary is in the specifier of vP, AgrOP being the complement of HOLD. AP lies in the specifier of VP and the comparative standard DP is the only complement of hi. With the VP-internal subject hypothesis, the external argument for HOLD is base-generated in [Spec, vP] and then gets DP-moved to [Spec, TP] for two reasons. First, as is required by EPP, all sentences must have a subject; second, the subject DP has to check its nominative Case feature with the nearest T node. Bi then is head-moved to HOLD through AgrO because of MLC. After being theme assigned in its base position, the comparative standard gets DP-moved to [Spec,  

AgrOP] to check its accusative Case feature with AgrO. This is driven by both the theta criterion and Case filter. The comparative result AP is still in [Spec, VP] after all the above movements have been completed. This diagram solves the misleading part in Xiang9s analysis in that it sticks to the truth that it is Mary that is da wusui (five years older) in D-structure. Additionally, it echoes with our previous argument that the comparative standard must be moved to a higher position than the result to achieve the S-structure.
In this section, the very many tasks completed prove that the phrasal /)z-comparatives are not derived from their clausal counterparts. But rather, they all have their own base-generative positions. Driven by different motivations, they get DP-moved or head-moved to realize the S-structure. Our last task is to argue for the syntax of clausal ^/-comparatives.
5.2 The Syntax of Clausal ^/-comparatives
As is asserted at the very beginning of the thesis, the phrasal and clausal ^/-comparatives should be distinguished from each other and can not be derived in the same process. After having an idea of the syntax of phrasal comparatives, we are attempting to show how clausal Z?z-comparatives are derived. Refer back to the theta grid for bi, and we will find that the second argument can either select for an AP or a CP. This CP is where we are working at in this section. We make the proposal that the crucial mechanism behind clausal Z?/-comparatives is the control theory. Look at the following example:
(9) Mary bi Tom tiao de yuan.
Mary M Tom jump DE far
“Mary jumps farther than Tom.”
Based on previous studies, this sentence is cut from Mary tiao de bi Tom tiao de yuan. The shared part, VP tiao (jump), need only appear once in S-structure. Moreover, the adjectival (farther) is the comparative result. We don’t agree with those supportive of deletion approach. Whafs more is that we reckon that tiao de yuan (jump farther) as a whole instead of a single yuan (farther) serves as the comparative result. The following tree diagram shows its structure:
(10) Syntactic structure for (9):
TP
 
 
This diagram resembles the previous one drawn when we analyze phrasal /)z-comparatives in that both the DP movement and Head movement happen again. As the motivations driving these movements are the same as previous analyses, we will leave them out here. Two differences lie between our analysis and previous studies. One consists in how to deal with the comparative result. Contrary to previous analyses, we assume what is base-generated in [Spec, VP] is a CP rather than a VP that contains only a single verb tiao (jump). This whole CP is the previous embedded VP tiao (jump)and the comparative result yuan (farther) combined. Thus the CP alone is considered the comparative result that is the IO of hi. The other difference is less obvious, though. It concerns the insertion of null PRO/pro. We will answer why this insertion is necessary before telling them apart in this context.
The necessity of PRO/pro lies in the theta grid of the verb tiao (jump):
tiao
Agent
DP
k
Tiao (jump) is a one-place predicate. It has an external agent role to assign. Thus, there must be a DP to receive the theta role. However, in this single CP, no DP is available, thus we insert a null PRO/pro in it. It can be noticed that the little pro rather than the big PRO is employed. Ifs not chosen at random but for a reason. Based on Carnie (2006), PRO can only appear in Caseless position, namely, the specifier position of non-fmite clauses. As the Chinese language lacks morphological representations, we can^ tell whether tiao de yuan 〇ump farther) is finite or non-fmite by morphological means. However, one test is to see if a subject is obligatory in an embedded clause. Refer to the following examples:
(11) Wo zhi-dao *(ta) tiao de yuan.
‘‘I know that he/she jumps far.”
The example shows that in the complement clause, a subject is obligatory. Thus, we conclude that the CP is finite. Finiteness naturally calls for the little pro.
Having set the stage, we now focus on how control takes effect. In its D-structure, the subject DP Mary c-commands the pro in the whole lower CP. Mary can be co-indexed with the pro. Therefore it is the subject of the matrix clause that determines who the pro refers to . In this sense, we obtain Mary tiao de yuan (Mary jumps farther) which is in line with the intended meaning of this ^/-comparative. Otherwise, co-indexation between the pro and Tom will result in Tom tiao de yuan (Tom jumps farther) that runs counter against the original meaning. Whafs worse is that neither Tom nor the pro c-commands one another. In the final analysis, this phrasal 6/-comparative is derived under the mechanism of subject control.
5.3 Summary
This chapter is based on Chapter 4 where the category of bi has been classified as a light verb. We have asserted in the beginning that different syntactic operations lie underneath phrasal and clausal ^/-comparatives.
For the former, based on theta grids, four arguments are required in order to elicit the D-structure of them. The main mechanism behind their derivation is movement, with DP movement and head movement included. The DP movement here is operated on both the comparative subject and the comparative standard, driven by EPP and Case reasons. The head movement is operated mainly on bi that moves from the head of the lower VP to that of AgrOP in the midway and of higher vP in the end. This movement is driven by feature checking and MLC.
For the latter, similar movements are also needed. Yet, they differ from their phrasal counterparts in that subject control is required. That is, the comparative subjective determines the concrete agent role of the propositional comparative result. To tell it apart from object control, how co-indexation and c-command takes into effect needs to be clarified first.

Chapter Six Conclusion
Having opined that bi is a light verb and elucidated phrasal and clausal /)z-comparatives, major findings and limitations will be summarized in this chapter.
6.1 Major Findings
The comparative construction has been widely used in Chinese. Yet, different ways of expressing comparison can be found in ancient and Mandarin Chinese. We focus merely on CC with marker bi in Mandarin Chinese. In studying it, two central questions are at issue. One concerns the word class of the marker bi. The other deals with the derivation of ^/-comparatives.
As for the first question, it is considered in the present thesis that marker bi can be treated as a light verb which has undergone syntactic operations. In analyzing so, the fact in modern use can be explained. That is, bi has been grammaticalized, which means it has progressively lost its role as a typical verb, but more gained its functional role. For the second question, we make a similar analysis to that of Larson’s (1988b) double object construction. We propose that phrasal and clausal ^/-comparatives should be distinguished from one another. In terms of the former, we conclude that the comparative subject has undergone DP movement because of EPP and Case filter. As for the comparative standard and the comparative result, they are similar to direct and indirect object in VP-shells. The comparative standard gets DP-moved to [Spec, AgrOP] for Case filter. Marker bi is moved to the head of vP and merges with HOLD through AgrO in the midway for Case checking reasons. As for the latter, we hypothesize that subject control comes into play in distinguishing the agent of the CP that serves as the comparative result. Answers offered to the two crucial questions in this thesis enable us to have a basic idea of how ^/-comparatives work.
6.2 Limitations and Suggestions for Future Research
Although we have told phrasal and clausal ^/-comparatives apart and argued for
56

the movement motivations, still, we have to admit that at least the following six problems remain obviously to be figured out.
The first problem is related to how light verbs are defined. In previous studies, consensus has never been reached as to whether light verbs are phonologically null, whether they have semantic content, or whether they need to get merged. This inevitably adds to the difficulty in defining and selecting the design features of them. Therefore, weak points are sure to be found in our treating bi as a light verb.
The second problem involves the analysis of phrasal ^/-comparatives. In our analysis of the derivation of Mary bi wo da wusui, ifs surprising to note that the sentence in its D-structure, namely, Mary da wusui bi wo, also makes sense and is totally acceptable. Thus the question appears whether any movement is operated on the latter case. If the answer is positive, how is it achieved?
The third problem deals with Case. We have touched upon accusative Case assigned to the direct object after bi. Then, what Case should be assigned to the comparative result, the indirect object based on our analysis?
The last three problems belong to the analysis of clausal Z?/-comparatives. One is what if the result CP follows the comparative subject? That is, is movement operated in those clausal ^/-comparatives where the comparative result precedes marker bil Another is since morphological marks cannot be found in Mandarin Chinese, can there be more convincing ways to tell whether an embedded CP is finite or not? The last one is how to analyze those ^/-comparatives where more topics are compared, especially in terms of Case thought?
Given the above limitations, this thesis is in no way a complete one. More efforts should be put into clausal Z?z-comparatives particularly in future research. We expect more corroborating evidences to be offered to indicate where the comparative results are base-generated and how they get moved, for what we have presented is nothing but our hypotheses. Thus, this thesis calls badly for further examination.
References
[1] Carnie, A. (2006). Syntax: A generative introduction (2nd ed.). Oxford: Blackwell.
[2] Chao, C, P. (2005). On Phrasal and Clausal Comparatives in Mandarin Chinese
(Unpublished M.A. Thesis). National Chiao Tung University, Hsingchu, Taiwan.
[3] Chao, Y. R. (1968). A grammar of spoken Chinese. Berkeley: University of California Press.
[4] Chen, C.L.[陈昌来],(2002),《介词与介引功能》。合肥:安徽教育出版社。
[5] Chomsky, N. (1995). The minimalist program. Cambridge, MA: MIT Press.
[6]  Chu,Z. X.[储泽祥],(1999),通比性的“很”字结构,《世界汉语教学》,(1): 36-44。
[7] Crystal, D. (2008). A dictionary of linguistics and phonetics (6th ed.). Oxford:
Blackwell.
[8] Dai, T.T•[戴婷婷],(2017),浅析英语中的轻动词结构,《广东蚕业》,51(3)^75-78。
[9] Ding, S.S. [丁声树],(1961),《现代汉语语法讲话》。北京:商务印书馆。
[10] Erlewine, M. (2007). A new syntax and semantics for the Mandarin hi comparative. M.A. Thesis. Chicago: University of Chicago.
[11] Fu, Y. C. (1978). Comparative structure in English and Mandarin Chinese. Ph.D. Dissertation. Michigan: University of Michigan.
[12] He,Y.J.[何元建],(2011),《现代汉语生成语法》。北京:北京大学出版社。
[13] Hong, W. M. (1991). Comparative Structure in Mandarin Chinese (Unpublished M.A. thesis). National Tsing Hua University, Hsingchu, Taiwan.
[14] Hsing, J. C. (2002). On comparative sentences with bi-marker in Mandarin Chinese. M.A. Thesis. Taiwan: National Tsinghua University.
[15]  Huang, C. T•[黄正德],(2008),从“他的老师当得好”谈起,《语言科学》,(3): 225-241。
[16] Huang, C.T., Li, Y.H., & Li, Y.F. (2009). The Syntax of Chinese. Cambridge: Cambridge University Press.
[17] Huang, X.H.[黄晓惠],(1992),现代汉语差比格式的来源及演变,《中国语文》,
⑶:213-224。
[18] Jespersen, O. (1942). A modern English grammar on historical principles (Part
VI: Morphology). London: George Allen & Unwin Ltd.
[19] Kennedy, C. (2007). Vagueness and grammar: the semantics of relative and absolute gradable adjectives. Linguistics and Philosophy (1): 1-45.
[20] Larson, R. (1988b). On the double object construction. Linguistic Inquiry (19): 335-91.
[21]  Li, N.,&Shi. Y.Z•[李讷,石毓智],(1998),汉语比较句嬗变的动因,《世界汉语 研究》,(3): 16-28。
[22] Lin, J. W. (2009). Chinese comparatives and their implicational parameters. Natural Language Semantics (1): 1-27.
[23] Lin, T. H. (2001). Light verb syntax and the theory of phrase structure. Ph.D. Dissertation. Irvine: University of California.
[24] Liu, C. S. (1991). Anaphor in Mandarin Chinese and binding at the interface. Ph.D. Dissertation. Irvine: University of California.
[25] Liu, C. S. (1996). A note on Chinese comparatives. Studies in the Linguistic Sciences (2): 217-235.
[26] Liu, D.Q.[刘丹青],(2012),汉语差比句和话题结构的同构性:显赫范畴的扩张力 一例,《语言研究》,(4): 1-11。
[27] Liu, J.G.[刘建国],(2011),《“朱子语类”比较句研究》。北京:北京大学硕士
学位论文。
[28] Liu, Y.[刘焱],(2004),《现代汉语比较范畴的语义认知基础》。上海:学林出版社。
[29] Lv,S.X.[吕叔湘],(1942),《中国文法要略》。北京:商务印书馆。
[30] Lv,S.X•[吕叔湘],(1985),疑问•否定•肯定,《中国语文》,(4): 241-250。
[31] Lv, S.X.[吕叔湘],(1999),《现代汉语八百词》。北京商务印书馆。
[32] Ma,J.Z•[马建忠],(1935),《马氏文通》。上海:商务印书馆。
[33] Ouhalla, J. (1999). Introducing transformational grammar: From principles and parameters to minimalism (2nd ed.). Hodder Education.
[34] Shen,J.X.[沈家煊],(2001),跟副词“还”有关的两个句式,《中国语文》,(6):
483-493〇
[35] Shi,P.X•[史佩信],(1993),比字句渊源,《中国语文》,(6): 456-461。
[36] Stassen, L. (1985). Comparison and Universal Grammar. Oxford: Basil Blackwell.
[37] Tsao, F. F. (1989). Comparison in Chinese: A topic-comment approach. Journals of Chinese Studies, new series (1):151-189.
[38] Xiang, M. (2005). Some topics in comparative constructions. Ph.D. Dissertation. East Lansing: Michigan State University.
[39] Xu, J., & Li, Y. Z.[徐杰,李英哲],(1993),焦点和两个非线性语法范畴:“否定”、
“疑问”,《中国语文》,(2): 81-91。
[40]  Yuan,Y.L.[袁毓林],(1996),话题化及相关的语法过程,《中国语文》,(4): 241-254。
[41] Zhang, X.C.[章新传],(1991), “朱子语类”的“比”字句及其汉语史价值,《上
饶师专学报》,(4): 36-42。
[42] Zhu,D.X•[朱德熙],(1982),《语法讲义》。北京:商务印书馆。
【本文地址:https://www.xueshulunwenwang.com/jingjilei/shangwuyingyu/1472.html

上一篇:英语语言文学论文写作 张爱玲作品中的异国形象研究

下一篇:外国语言学及应用语言学方向 中学英语网络自主学习满意度探究及其教学启示 ——以“求解答”和“昂立求解答”平台为例

相关标签: