Between Syntax and Semantics

C.T. James Huang

• Publisher: Routledge
• Year: 2010
• ISBN:
  9781135217570
• Edition: 1
• Topic: Languages & Linguistics
• Subtopic: Linguistics
• Index: Languages & Linguistics

This indispensable volume contains articles that represent the best of Huang's work on the syntax-semantics interface over the last two decades. It includes three general topics: (a) questions, indefinites and quantification, (b) anaphora, (c) lexical structure and the syntax of events.

Part I
Questions, Indefinites, and Quantification

1 Move WH in a Language without Wh-Movement*

1 INTRODUCTION

1.1

The study of constraints in grammar has been one of the most persistent topics in generative studies. Among the most important results of this enterprise is a set of locality conditions, including most notably Ross’s island constraints and Chomsky’s Subjacency Condition. These locality conditions are usually defined over certain structural configurations and a rule of movement, deletion, or at least some kind of dependency between two structural positions.

One type of inquiry that arises from studies of these conditions concerns their status in languages where there is no structural configuration or overt dependency between two structural positions meeting the definition of the various islands and constraints. For example, many languages do not have a wh-word fronting rule in syntax to form a wh-question, and some do not move any element for the purpose of subdividing a sentence into focus and presupposition (i.e., of forming a cleft sentence). Since questions and sentences with focus are universal sentence types, it is natural to ask what the proposed structural conditions have to say about these languages. It would seem that since these languages do not exhibit the defining configurations for the conditions to be applicable, they would be exempt from these conditions. If it turns out, however, that these types of sentences do show properties as if they form islands and obey island conditions, then it might seem that the conditions as formulated in structural terms must be rejected or revised. As a point of logic, however, this fact may lead us instead to ask if it is a mistake to regard these sentence types as exhibiting absolutely no structural dependencies in every possible sense. That is, although they do not show overtly the required structural properties, couldn’t they be seen as actually involving them in some abstract sense?

In this chapter, I will try to demonstrate that taking this latter position, instead of concluding the invalidity of the structurally based conditions, will lead us to the prediction of certain very interesting facts and provide a simple explanation of these facts. In particular, I will discuss certain properties of two major sentence types in Chinese, questions and cleft sentences, and show that, although they obviously do not involve any movement in Syntax, it is desirable to assume that they undergo movement in the LF component of grammar. Thus, for example, while it is clear that a wh-question like (1) is generated in Syntax with the wh-word never anywhere but in its base position, it will be argued that in the LF interpretive component the wh-word is moved to a position c-commanding the sentence, leaving a trace interpreted as a variable bound to it, as in (2):

(1) ni xihuan shei?
you like who
‘Who do you like?’

(2) [shei_i [ni xihuan e_i]]
who you like

The assumption that abstract movement rules of this sort exist in UG is an important feature of some recent works within the Extended Standard Theory, and is supported by a fairly wide range of facts observed in English and several other languages.¹ It will not be unreasonable to assume the existence of such abstract devices in Chinese simply as a consequence of UG, but it will be worthwhile to ask if there are some positive language-specific motivations for making such an assumption. The main purpose of this chapter is to show that there are indeed strong language-specific motivations.

1.2

The assumption that the LF representation of (l) is of the form (2) has as its immediate consequence the existence of:

(3) a. a quantifier
b. an empty category
c. a movement process

and the properties associated with each of (3a–c). The main body of this chapter will be concerned with case (c), the existence of movement and its properties. Case (b) will be briefly dealt with in the Appendix. As regards case (a), a brief discussion in this section will give some support for the treatment of wh-words as a kind of quantifier and provide some initial motivation for the rest of the chapter. Consider the following sentences:

(4) [Zhangsan wen wo [shei mai-le shu]]
Zhangsan ask me who bought books
‘Zhangsan asked me who bought books.’

(5) [Zhangsan xiangxin [shei mai-le shu]]
Zhangsan believe who bought books
‘Who does Zhangsan believe bought books?’

(6) [Zhangsan zhidao [shei mai-le shu]]
Zhangsan know who bought books
a. ‘Who does Zhangsan know bought books?’
b. ‘Zhangsan knows who bought books.’

The only surface difference among these sentences is in the choice of the matrix verb. In (4), wen ‘ask’ belongs to a class of verbs that require an interrogative complement. In (5), xiangxin ‘believe’ does not permit an interrogative complement. In (6), zhidao ‘know’ may optionally take an interrogative complement. As the translation shows, this single difference in the choice of the verb is responsible for the fact that (4) must be interpreted as a statement taking an indirect question, (5) must be interpreted as a direct question embedding no indirect questions, and (6) may be interpreted as either. It makes good sense to ask how the very different meanings of the virtually identical (4) and (5), as well as the ambiguity of (6), may be represented in an optimal theory. One natural approach to this question is to look at an indirect question like (4) as one in which the question word has scope over the embedded sentence, and a direct question like (5) as one in which the question word takes scope over the entire sentence, while in (6) the question word may take either scope. This amounts to postulating (7)–(9) as the logical forms for (4)–(6), respectively:

(7) [Zhangsan wen wo [shei_x [ x mai-le shu]]]
Zhangsan ask me who bought book
‘Zhangsan asked me for which x, x bought books.’

(8) [shei_x [Zhangsan xiangxin [ x mai-le shu]]]
who Zhangsan believe bought book
‘For which x, Zhangsan believes x bought books?’

(9) a. [Zhangsan zhidao [shei_x [ x mail-le shu]]]
Zhangsan know who bought book
‘Zhangsan knows for which x, x bought books.’

b. [shei_x[Zhangsan zhidao [ x mai-le shu]]]
who Zhangsan know bought book
‘For which x, Zhangsan knows x bought books?’

It is a well-known property of quantifiers that they exhibit scope phenomena. Since the difference between (4) and (5) and that between the two readings of (6) are naturally seen as difference in scope of a wh-word, treating a wh-word as a quantifier, as is shown in (7)–(9), is not unreasonable. Obviously, this is not the only possible way to account for these facts in Chinese, but note that these are not facts peculiar to Chinese; they are paralleled by corresponding sentences in other languages, including those in which a wh-word is actually moved to a quantifier position in surface form. Given that facts in UG should be treated in a uniform way for all languages and that there is already a natural representation given by the surface form of overt wh-moved languages, it is entirely natural to postulate the abstract representations like (7)–(9) for a language without overt movement.

Another construction for which a case can be made for an abstract movement rule in LF is the cleft construction in Chinese. The formation of a cleft sentence in this language clearly does not involve the dislocation of any constituent in Syntax. In surface structure a cleft sentence differs from a non-cleft in that in a cleft there is a focus marker (the copula shi) immediately preceding the focused constituent.² (In the glosses following, FM = Focus Marker.)

(10) a. shi wo mingtian yao mai neiben shu
FM I tomorrow want buy tha book
‘It is I that want to buy that book tomorrow.’

b. wo shi mingtian yao mai neiben shu
I FM tomorrow want buy that book
‘It is tomorrow that I want to buy the book.’

c. wo mingtian shi yao mai neiben shu
I tomorrow FM want buy that book
‘I do want to buy that book tomorrow.’

Since a cleft sentence has the universal semantic property of dichotomizing a sentence into focus and presupposition, it is natural to provide a unified representation of this dichotomy in LF. One reasonable assumption is that the focused material is also regarded as a quasi quantifier binding a variable in the presupposition, so that (10a) can be represented as (11):

(11) [(shi wo)_x [x mingtian yao mai neiben shu]]
FM I tomorrow want buy that book

By convention we may then interpret the operator and the open sentence in (11) as representing its focus and presupposition, respectively.^3...

Table of contents

• COVER PAGE
• TITLE PAGE
• COPYRIGHT PAGE
• PERMISSIONS
• INTRODUCTION
• PART I: QUESTIONS, INDEFINITES, AND QUANTIFICATION
• PART II: ANAPHORA AND BINDING
• PART III: LEXICAL STRUCTURE AND EVENTS
• NOTES
• BIBLIOGRAPHY