Verb-Initial Constructions in Modern Hebrew
Nurit Melnik, 2002
Ph.D. thesis, University of California at Berkeley
INTRODUCTION
The goal of this study is to provide an integrative analysis of verb-initial (V1( constructions in Modern Hebrew (MH). These constructions pose a challenge to a syntactic analysis, since the NP dependents which can be considered subjects in the constructions do not exhibit the properties associated with prototypical subjects in the language. In addition, most of the V1 constructions considered here have an SV(O) counterpart whose propositional content is identical to that of V1. The converse statement is not valid, as not all sentences with SV(O) word order can be `transformed' into V1. This poses two interesting questions:
· Why would a language employ two different ways of expressing the same thing?
· Why is it not the case that all SV(O) sentences have a V1 counterpart?
These questions, as well as the question of the grammatical function of the postverbal NP in V1, are considered and ultimately answered in this study. The answer is presented in the form of a constructional HPSG grammar-fragment.
The structure of this dissertation is as follows. Chapter 2 is dedicated to a definition of the type of constructions under discussion, as well as a presentation of the data. This is preceded by a review of the data resources used for this study as well as a short introductory section on Modern Hebrew syntax. Chapter 3 presents the theory of syntax assumed here - Head-Driven Phrase Structure Grammar (HPSG). The purpose of this chapter to introduce to readers who are not familiar with HPSG the concepts and principles necessary to follow the analysis.
Chapter 4 contains the syntactic analysis of V1 constructions. It begins with a review of Shlonsky's (1987, 1997) transformational analysis of a subsection of the data. An alternative analysis is then developed by first considering V1 constructions and the grammatical functions of their `subjects' from a typological perspective. The conclusion is that only agreement-triggering arguments exhibit the necessary properties to be considered syntactic subjects. The proposal is then formulated in the HPSG framework. In the process, the Possessive Dative Construction, which plays a prominent role in V1 constructions, is considered and an HPSG-based analysis is presented.
Chapter 5 addresses the two questions raised above. More specifically, it considers the licensing conditions of V1 in MH. Once we identify the licensing conditions of V1 we are in a position to answer the second question, that is what distinguishes between those SV(O) sentences that have a V1 counterpart from those which do not. As it turns out, the answer to the second question includes the answer to the first. V1 constructions are employed by MH as an information packaging devices which encodes thetic expressions in distinction from `unmarked' categorical expressions. Thus, although SV(O) sentences and their V1 counterparts contain the same propositional content, they are pragmatically divergent. V1 constructions are licensed when they express a thetic judgment.
Chapter 6 completes the analysis by integrating the syntactic component with the information packaging one. This is achieved via a multi-inheritance type hierarchy in which types are cross-classified according to the two components. A preliminary step is the presentation of a proposal regarding the representation of information packaging, and more specifically judgment types in HPSG.
Chapter 7 presents the conclusions of this study.