Peripheries in Kashmiri and Hindi-Urdu: The syntax of discourse-driven movement (Language Faculty and Beyond)

Peripheries in Kashmiri and Hindi-Urdu The syntax of discourse-driven movem ent Emily Manetta John Benjamins Publishin...

Author: Emily Manetta

61 downloads 720 Views 9MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Peripheries in Kashmiri and Hindi-Urdu The syntax of discourse-driven movem ent

Emily Manetta

John Benjamins Publishing Company

Peripheries in Kashmiri and Hindi-Urdu

Language Faculty and Beyond Internal and External Variation in Linguistics Language Faculty and Beyond (LFAB) focuses on research that contribuJ:es to a deeper understanding of the properties of languages as a result of the Language Faculty and its interface with other domains of the mind/brain. While the series will pay particular attention to the traditional tension between descriptive and explanatory adequacy, the series will also address issues such as the level of linguistic design, through new lines of inquiry often referred to as 'physiological linguistics' or 'biolinguistics'. LFAB aims to publish studies from the point of view of internal and external factors which bear on the nature of micro- and macro-variation as, for example, understood in the minimalist approach to language. For an overview of all books published in this series, please see http://benjamins.com/catalog!lfab

Editors Kleanthes K. Grohmann

Pierre Pica

University of Cyprus

CNRS,Paris

Advisory Board Paola Benincit University of Padova, Italy

Anders Holmberg

Cedric Boeckx

Lyle Jenkins

ICREA/Universitat Autonoma de Barcelona

Biolinguistics Institute. Cambridge, USA

Guglielmo Cinque University of Venice, Italy

Richard K. Larson Stony Brook University. USA

Noam Chomsky

Andrew Ira Nevins

Massachusetts Institute of Technology, Cambridge. USA

Alain Rouveret

University of Newcastle. UK

University College London, UK

Stephen Crain

University of Paris Vll, France

Macquarie University, Sydney, Australia

Esther Torrego University of Massachusetts, Boston USA

Marcel den Dillin CUNY Graduate Center, New York, USA Naama Friedmann Tel Aviv University, Israel

Anna Papafragou University of Delaware, Newark, USA Akira Watanabe University of Tokyo, Japan

Volume4

Peripheries in Kashmiri and Hindi-Urdu. The syntax of discourse-driven movement by Emily Manetta

Peripheries in Kashmiri and Hindi-Urdu The syntax of discourse-driven movement

Emily Manetta The University of Vermont

John Benjamins Publishing Company Amsterdam I Philadelphia

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences - Permanence of Paper for Printed Library Materials, ANSI z39.48-1984.

Library of Congress Cataloging-in-Publication Data Manetta. Emily. Peripheries in Kashmiri and Hindi-Urdn: the syntax of discourse-driven movement I Emily Manetta. p. em. (Language Fa.c:uhy and Beyond, ISSN 1877-6531; v. 4) Originally presented as the author's thesis (doctoral- University of California, Santa Cruz) under the title: Peripheries in Kashmiri and Hindi-Urdu. Includes bibliographical references and index. 1. Grammar Comparative and general--Syntax. 2. Kashmiri language--Syntax. 3· Hindi language--Syntax. 4· Urdu language--Syntax. I. Title. P291.M29

2011

4914--dC22 ISBN 978

90 272 o821 7 (Hb ; alk. paper)

ISBN 978

90 272 8699 4 (Eb)

2011003713

© 2011- John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher.

John Benjamins Publishing Co.· P.O. Box 36224 • 1020 ME Amsterdam· The Netherlands John Benjamins North America· P.O. Box 27519 ·Philadelphia PA 19118-0519 • usA

Table of contents

Acknowledgements

IX

CHAPTER 1

Introduction: Peripheries in Kashmiri and Hindi-Urdu Empirical overview 1 1.1.1 The left periphery of Kashmiri 1 1.1.2 Long-distance wh-dependencies in Kashmiri 2 1.1.3 The wh-expletive construction in Kashmiri and Hindi-Urdu 1.1.4 Sluicing in Hindi-Urdu 4 1.2 Theoretical context 4 1.2.1 Minimalist assumptions 5 1.2.2 Phases 6 1.2.3 A and A-bar movement 8 1.3 Organization of the book n

1

1.1

3

CHAPTER 2

Feature stacking: The Kashmiri periphery 2.1 Kashmiri: A brief introduction 13 2.1.1 Kashmiri data 15 2.1.2 Syntax 15 2.1.3 Previous work on Kashmiri 19 2.2 The Kashmiri left periphery 20 2.3 The cartographic approach to the left periphery of Kashmiri 24 2.4 New opportunities 26 2.4.1 The specifier-head relation 27 2.4.2 The cartographic project and the phase 29 2.4.3 Order of projections: encoding variation 32 2.5 Feature stacking 34 2.5.1 Features and the lexicon 34 2.5.2 More on feature stacking 37 2.5.3 Regularity and idiosyncracy 39 2.5.4 An additional empirical question: The Kashmiri element ki 2.5.5 Theoretical advantages 43

13

40

vx


CHAPTER

3

Full and partial wh-movement in Kashmiri Introduction 47 Kashmiri question formation and the structure of the clause 49 3.2.1 The Kashmiri question 49 3.2.2 Assumptions about the structure of the Kashmiri clause 50 3·3 Analyzing full and partial wh-movement in Kashmiri 53 3.3.1 A new account of A-bar movement 53 3.3.2 Restrictions on wh-expletives 6o 3·3·3 Previous approaches to wh-expletive constructions: Indirect and direct dependency 64 3·3·4 Interpreting wh-expletive constructions 69 3·4 Additional empirical investigations 71 3.4.1 A Kashmiri issue: Factive predicates 71 3.4.2 A crosslinguistic issue: Multiple wh-expletives 73 3·5 Conclusion 83 Appendix: C heads in the lexicon of Kashmiri 84

47

3.1 3.2

CHAPTER

4

Wh-expletives in Hindi-Urdu 4.1 Introduction 88 4.2 Wh-dependencies in Hindi-Urdu and Kashmiri 90 4·3 The position ofwh-material in Hindi-Urdu 94 4.3.1 Focused constituents 94 4.3.2 Adverbs 96 4·4 A-bar movement in Hindi-Urdu: Extending an account of Kashmiri 99 4·4·1 Kashmiri wh-dependencies 99 4.4.2 Extending the proposed account to Hindi-Urdu 103 4·5 Wh-dependencies in Hindi-Urdu: The vP phase 108 4.5.1 Wh-movement in Tagalog: A case for [Q]-bearing v 108 4.5.2 An account of Hindi-Urdu wh-dependencies 111 4·5·3 Comparison with other accounts 120 4·5·4 Conclusion 126 CHAPTER

5

Sluicing in Hindi-Urdu Sluicing in Hindi- Urdu and Kashmiri 127 Accounts without movement to Spec, CP 129 A new account: Movement to spec, CP 137

5.1 5.2 5·3

87

127

Table of contents vu CHAPTER

6

Conclusions A theory of the periphery 143 Wh-expletives and the role of expletives in the grammar Phases and their edges 146 Displacement and formal features 147 New research opportunities 147 Summary 148

6.1 6.2 6.3 6.4 6.5 6.6

143 144

References

151

Index

159

Acknowledgements

Work on the research program that eventually became this book spanned nearly a decade, and I am deeply grateful to the inimitable Linguistics faculty at the University of California Santa Cruz, in particular Jim McCloskey, Judith Aissen, Sandra Chung, and Jorge Hankamer. I am especially indebted to Jim for his incredible insight, patience, interest, and attention through many versions of the analyses presented here. To my peers at UCSC I also owe thanks: Pete Alrenga, Vera Lee-Schoenteld, Anne Sturgeon, Irena Folic Richter, Chris Potts, Line Mikkelson, Ascander Dost, Dylan Herrick, Nathan Sanders, Anya Lunden, Lynsey Wolter, Kyle Rawlins, Ruth Kramer, and many others. The talks, papers, drafts, and thoughts that eventually became this book benefited tremendously from the insightful comments of audiences and individuals over the years. I particularly thank Rajesh Bhatt, for his lighting-fast responses to questions, and for his comments, suggestions, and discussions at various stages. I am also grateful to Rakesh Bhatt, Anoop Mahajan, Ben Breuning, Adam Albright, Pranav Anand, Kashi Wali, Theresa Biberauer, Maziar Toosarvandani, Alice Davison, Chris Kennedy, Tanmoy Bhattacharya, Peter Svenonius, Heidi Harley, Ayesha Kidwai and Jason Merchant. I also owe an enormous debt to my native-speaker informants and their families, who have been patient, helpful. gracious and generous. Thanks goes to Vijay Chowdhury and her family, Fran Kaul, Rakesh Bhatt, Saleem Ali, Uzma Rizvi, Ali Afzal, Adnan Afzal, Chandra and Dev Gupta and their family, Subhan Ali Hunzai and his family, Tabinda Khan, and Khurram Khan. A number of others have also provided important support, including my colleagues in the Program in Linguistics at the University of Vermont, Julie Roberts, Maeve Eberhart, Guillermo Rodriguez, and Jennifer Dickinson. For their careful reading of my work, I am also grateful to the interdisciplinary writing group at UVM, including Danilyn Rutheford, Ben Eastman, Kabir Tambar, Kelda Jamison, Andrea Voyer, and Vicki Brennan. Finally, I thank my family for their unflagging support and enthusiasm for every stage of this book. Most of all I am grateful to Jonah and Zaki, who mean everything to me.

CHAPTER 1

Introduction Peripheries in Kashmiri and Hindi-Urdu

This book investigates the structure and organization of the periphery through an exploration of the A-bar systems of Kashmiri and Hindi-Urdu. Traditionally the periphery has been understood to be the position or positions at the left edge of the clause at which constituents with certain discourse status are found (topic, focus) and where clausal type and force is expressed. Generative research on the left periphery has been guided by three important questions. First, how is the periphery itself structured to accommodate the often rigid ordering and co-occurrence restrictions of the wide range of elements found there (and to what degree is this fixed by universal principles)? Second, what are the mechanisms that drive displacement to the periphery? Third, how does the periphery mediate instances of long-distance (or apparently unbounded) displacement? This book presents a micro-comparative analysis of A-bar positions, wh-dependencies, and wh-expletive constructions in Kashmiri and Hindi-Urdu in an effort to shed new light on these questions.

Empirical overview

1.1

1.1.1

The left periphery ofKashmiri

I will first present the core empirical landscape with which this book will be concerned. Unlike many more familiar Indic languages, Kashmiri is a verb-second language that features a rich left periphery. In the immediately pre-verbal position we find focused phrases. (1)

bi-ti chu-s masTar lSG-FOC be.PRS.M-lSG teacher 'I too am a teacher'

In case the pre-verbal constituent is an interrogative focus, a topic can precede it, literally throwing the verb into third position.

2

Peripheries in Kashmiri and Hindi- Urdu

(2)

Raj-an kamis haa-v nav kitaab? Raj-ERG who.DAT show-PST-FSG new book 'As for Raj, to whom did he show his new book?'

(Wall & Koul: 12)

In embedded clauses, which retain verb-second order, the particle ki can also optionally precede the constituents on the left periphery. (3)

miiraayi cha pataa (ki) mohn-an kamis Mira-DAT AUX.PRS.FSG know that Mohan-ERG who.DAT di-ts kitaab. give.PST-FSG book. 'Mira knows who Mohan gave a book to'.

Any productive investigation of wh-movement in Kashmiri must first address the organization of the periphery. Much recent work on highly structured peripheries such as that found in Kashmiri has relied on the so-called "cartographic" account (Rizzi 1997), featuring a hierarchy of distinct functional projections. In this book I explore ways to maintain the empirical advances of the cartographic approach while offering an analysis of the left periphery that is more compatible with the notion of the phase (Chomsky 2000, 2001, 2004). This work sets the stage for an exploration of the role of the periphery in wh-movement, and in particular in mediating apparently unbounded wh-dependencies. 1.1.2

Long-distance wh-dependencies in Kashmiri

Wh-movement to the left periphery of the clause is obligatory in Kashmiri. (4)

(5)

Raj-an kamis haa-v nav kitaab? Raj-ERG who.DAT show.PsT-FSG new book 'To whom did Raj show his new book?'

(Wall & Koul: 12)

*Rajan heav kamis nev kitaab

Kashmiri permits long-distance wh-movement, in which a wh-phrase generated in an embedded clause can be displaced into the matrix clause. (6)

Tse kam' chu-y baasaan ki Mohn-as 2sG.DAT who.ERG AUX-DAT think that Mohan-DAT di-ts kitaab? give-PST.FSG book 'Who do you think gave Mohan the book?'

Kashmiri also features the so-called wh-expletive construction (also termed scope-marking construction). A wh-expletive construction is one in which the full wh-phrase remains in an embedded clause while a minimal wh-element occupies the position at which the full wh-phrase is interpreted.

Chapter 1. Introduction

(7)

Tse k:'aa chu-y baasaan ki Mohn-as kam.' 2SG.DAT EXPL AUX-DAT think that Mohan-DAT who.ERG di-ts kitaab? gave-PST.FSG book 'Who do you think gave Mohan the book?'

Although a great deal of work has been done on wh-expletives and wh-expletive constructions in the related language Hindi-Urdu (Dayal1994, 1996, Mahajan 2000, Lahiri 2002), relatively little has been said about this construction in Kashmiri. I claim here that in order to build a complete account of the periphery; we must analyze the meaningless wh-element as a true expletive in the A-bar movement system, on par with the better-understood expletives of the A-movement system. This approach to wh-expletive constructions will be shown in the chapters below to reveal an underlying similarity of design between the A and A-bar movement systems, and indicates that the two can be understood as driven by the same basic set of mechanisms. 1.1.3

The wh-expletive construction in Kashmiri and Hindi-Urdu

Hindi-Urdu, like Kashmiri, features two ways to form a long-distance whdependency. In the first, in (8), the full wh-phrase originating in the lower clause, kis-ko 'who: is displaced into the higher clause, forming a root question. In the second, in (9), a wh-expletive kyaa. appears in the higher clause, while the full whphrase remains embedded. (9) also has a matrix question reading. (8)

Sita-ne kis-ko soc-aa ki Ravii-ne dekh-aa? [Hindi- Urdu] Sita-ERG who-Ace think-MsG.PRF that Ravi-ERG see-PRF 'Who did Sita think that Ravi saw?'

(9)

Sita-ne kyaa soc-aa ki Ravii-ne kis-ko Sita-ERG EXPL think-MSG.PRF that Ravi-ERG who-Ace dekh-aa? see-PRF 'Who did Sita think that Ravi saw?'

[Hindi- Urdu]

Superficially, (8)-(9) appear strikingly similar to (6)-(7), the corresponding constructions in Kashmiri. And yet, Hindi-Urdu is typically analyzed as a verb final, wh-in-situ. language, while Kashmiri is a verb-second language with full wh-movement This puzzle motivates the investigation of A-barmovement presented here, which provides a detailed comparison of the long distance wh-dependencies of Hindi-Urdu with those of Kashmiri. Though Hindi-Urdu features both long-distance wh -displacement and wh -expletive constructions, wh -material is not found at the clause edge. In this book I will argue that Hindi- Urdu and Kashmiri feature the same clausal topology, but occurring lower in the clause in Hindi-Urdu (at the vP layer) and higher in the clause in Kashmiri (at the CP layer). This lends

3

4


support to the claim that the specifier position of vP may play the same role often attributed to the specifier position of CP in wh-movement: it can be a position for wh-expletives and partially moved wh-phrases. In effect, this account locates wh-material in Hindi-Urdu in a second, clause-internal periphery. 1.1.4 Sluicing in Hindi-Urdu As most influential accounts of sluicing derive the ellipsis operation via movement of the wh-remnant (Merchant 2001) it may seem surprising that Hindi-Urdu (traditionally considered wh-in-situ) exhibits sluicing at all. However, we do see what appear to be English-style sluices in both Kashmiri and Hindi-Urdu. Hindi-Urdu: (10)

Aisha-ne ek ciiz khariid-ii lekin mujhe nahiiN pataa (ki) kyaa Aisha-ERG a thing buy-FsG.PRF but lsG.DAT not know (that) what 'Aisha bought something, but I don't know what'. (Mahajan 2005)

Kashmiri: (11)

kaNsi kh-yav batl magar me chu-ni someone.ERG eat-MSG.PST food but lSG.DAT AUX.PRS-NOT pataa :bm know who.ERG 'Someone ate the food, but I don't know who'.

Sluicing provides an important probe for the nature of wh-movement and the structure of the clause. The account of sluicing presented here claims that Hindi- Urdu sluicing is an exceptional instance of wh -movement to Spec, CP in Hindi- Urdu, driven by a C head possessing the ellipsis feature E (Merchant 2001; Toosarvandani 2007). This represents a significant break from previous approaches (including Manetta 2006), and is based on the analysis of wh-displacement and the periphery developed in this book.

1.2

Theoretical context

At the level of linguistic theory, the book aims to make contributions in two linked areas. The first centers on the nature and structure of the phase, and in particular on the crucial role of the phase-defining heads in determining crosslinguistic variation. In this section I will review previous and current work on the phase in order to set the stage for an understanding of phase as periphery. The second contribution is concerned with the system of movement for case and agreement (the A-movement system) and the system of movement to non-argument positions (the A-bar movement system). Much work over the past fifty years has drawn a distinction between these two classes of syntactic processes.


Recent work in the Minimalist Program holds out the promise of unification of the A and A-bar systems, though many (including Chomsky) are skeptical that such a unification is possible. In this book I will explore symmetries between these two systems, particularly with respect to the behavior of expletives. On this basis, I will argue for a unification of the mechanisms underlying the A and A-bar systems. Here I will review some current thinking on this topic, and preview why a study of wh-expletives can give us new insights into the inner workings of A-bar movement.

1.2.1

Minimalist assumptions

1his section serves to describe the set of assumptions in which most of the theoretical developments in this book are grounded. I will outline the basic underpinnings of what has come to be called the Minimalist Program. Components of this theory will be explored in some detail in subsequent chapters, so this introduction will ground the more sophisticated work to follow as well as to establish a common set of terminologies. Each language possesses a lexicon L. The lexical items in L have sets of features, interpretable and uninterpretable. The derivation of a sentence begins with a selection of lexical items from L. called the numeration. The lexical items in the numeration can then be composed into a syntactic object via the operation Merge, which combines two syntactic objects to form a new one. Uninterpretable features must be valued during the course of the syntactic derivation -they cannot be shipped to the interfaces of phonological form (PF) and logical form (LF). Uninterpretable features receive values via the operation Agree. In this operation, a head with uninterpretable features, here called the Probe, searches its c-command domain. When it encounters a lexical item with matching features, here called the Goal. Agree causes the features to mutually value one another. The complex operation Move, which is comprised of Agree followed by Merge, is also available in this context. A Probe can interact with a Goal and can cause that Goal to Merge into an additional specifier of the Probe (beyond those required by semantic selection). The feature which prompts the more complex operation Move is called the EPP. Through the operations Merge, Agree, and Move, all of the uninterpretable features on the lexical items in the derivation must be valued. A syntactic object with uninterpretable features remaining is illicit, and cannot be sent to the interpretive interfaces (this is, in essence, the principle of Full Interpretation (Chomsky 2004)). Of course, the operations described above are limited by a notion of locality, called the phase. Our understanding of the phase will be fleshed out in the immediately following section, as well as in subsequent chapters. However, it is important to point out that the numeration from which a specific sentence is

5

6


derived may be divided into sub-arrays oflexical items. The operations constructing each of these sub-sections of the whole derivation may be performed in parallel (Chomsky 2000, 2001, 2004). In this way, some sense oflocality is built-in, since a Probe can only potentially interact with Goals in a domain thus delimited. What is inherently minimal about this approach to syntactic derivations? We must first note that there is no room, in this theoretic view, to add any features to the derivation once it is in progress. The Inclusiveness Condition dictates that all features that participate in syntax, whether interpretable or uninterpretable, come into the derivation as a component of a lexical item from the syntax (Chomsky 2000). No additional features, such as indices or chains, may be introduced during the course of the derivation. There also can be no language specific operations in the syntax; the syntactic processes are limited to Merge, Agree, and the composite operation Move. We will see that this basic principle of derivation in this framework will play a large role in the analyses developed below, and we will explore what can be gained by limiting language-specific information to the lexicon. There are a number of more nuanced choices to be made concerning the theoretical framework described here. I will leave it to the remainder of the introduction and subsequent chapters to flesh out these important details. 1.2.2

Phases

Although introduced explicitly only in Chomsky (2000), the theory of phases is the latest instantiation of a very old idea in generative syntax- the notion of the cycle. This idea holds that derivation of a sentence proceeds in stages (cycles, phases) that are relatively independent and self-contained. Chomsky (2000) defines the phase as a 'propositionaf object: a verb phrase (vP) in which all theta-roles have been assigned, or a clause (CP) including tense and force. After all Merge, Move, and Agree operations have taken place in each phase, the output is sent to the interfaces of LF and PF. In some sense, the notion of phase is a way of encoding locality restrictions on operations, because Goals within a phase are inaccessible to Probes in subsequent phases. Goals on the edge of a phase, on the other hand, are accessible to higher Probes. The notion of phase edge will be crucial to the discussion of wh-movement in this book, and so I will define it here. According to Chomsky (2000, 2001, 2004), the phase edge consists of the phase-defining head, any specifier of that head, or anything adjoined to that head. This is fleshed out in the Phase Impenetrability Condition (PIC) as follows:

Phase Impenetrability Cottdition - "In a phase a with head H, the domain of H is not accessible to operations outside a, only H and its edge are accessible to such operations" where the edge includes specifiers and adjuncts to H. (Chomsky 2000: 108)


1his condition suggests that Goals on the phase edge are accessible to Probes in subsequent phases, and are therefore available to Agree with these heads or to undergo Move into higher structure. Chomsky (2004) adopts the principle that the interpretation and evaluation of each phase takes place at the level of the next higher phase. In this way, phase edges and the material they contain play a central role in knitting together the links of the A-bar chains in apparently unbounded dependencies. An important goal of this book is to develop an understanding of the structure of the periphery that is consistent with the characterization of phase and phase edge above. According to Chomsky (2000), building on Fox (2000) and Nissenbaum (2000), the phase-defining heads are C and v, due to their 'propositional' nature. The case for DP as a phase has also been made, however here we will be primarily concerned with the characteristics of the phase-defining heads C and v; and how the nature of these heads determines the properties of the A-bar systems of languages like Kashmiri and Hindi-Urdu. In recent work, the phase-defining heads themselves have taken on a role of even greater importance. Chomsky (2008) suggests that crosslinguistic variation may be in large part attributable to properties of the phase-defining heads themselves. That is, the featural makeup of the phase-defining heads in a language determines much of its syntactic character. 1his idea will come into play in two important ways. I will propose here that A-bar movement, much like A-movement, is controlled by sets of interpretable and uninterpretable features. Though the idea that such features are present on the C head is relatively familiar, I will show here that we should consider vas a location for wh-features as well (see Rackowski & Richards 2005). Further, Chapter 4 of this book will show that certain systematic contrasts brought out in a micro-comparison ofKashmiri and Hindi-Urdu can be understood given the assumption that certain structures are associated with the phase-defining head C in Kashmiri, but with the lower phase-defining head v in Hindi-Urdu. This portion of the book tests a specific prediction of this proposed equivalence between the phase-defining heads - if there are wh-expletives which appear at the edge of CP. there should also be wh-expletives which appear at the edge of vP. I argue that Hindi-Urdu is a language in which this prediction is borne out, and furthermore that we can understand the contrast between Hindi-Urdu and Kashmiri in terms of properties of the phase-defining vocabulary (C and v) in the two languages. We will also be concerned with the specific form features must take when located on the phase-defining functional heads. In particular, we will examine the complex left periphery of Kashmiri, and the way in which the language organizes the CP domain in particular. Insofar as we entertain the proposal that many language-specific properties can be attributed to the features present on these

7

8


phase-defining functional heads, we must recognize that the question of how these features are organized on the head itself becomes a crucial one. The analysis we will propose is an attempt to capture the insights of the cartographic investigations. which require a hierarchy of projections on the left edge of the clause, within the terms of a more spartan phrase structural system. Structuring features on a single C head, and thereby allowing the presence of multiple specifiers to a single head, provides an account of the complex left periphery of Kashmiri that is also more in line with current theoretical understanding. We will also show that this system can be extended to the v head, emphasizing again the equivalence of the phase-defining categories. Overall this book will then advance current research developments concerning the phase by taking seriously the hypothesis that inter-language variation is in large part determined by the featural properties of the phase-defining heads. Ultimately we are able to offer strong empirical support for this view, based on data involving long-distance wh-dependencies in Kashmiri and Hindi-Urdu.

1.2.3

A and A-bar movement

A primary focus of this book is the nature of A-bar movement, or movement to non-argument positions. From an empirical standpoint, we will examine not only wh-movement and wh-expletive constructions, but also other types of A-bar movement, including non-interrogative focus movement and topicali:zation. The position we will arrive at is that the A-bar system can be best understood in the current theoretical framework as driven by the same basic mechanisms as the A-system. These two systems have historically been understood as distinct by definition and as responding quite differently to a variety of tests. Movement in the A-system is typically movement of a DP argument for the purposes of licensing case assignment and/or agreement morphology; to a case-position such as the specifier of IPfl'P. On the other hand, a wider variety of phrases (DP, PP, etc.) can undergo A-bar movement, and these phrases need not be arguments. A-bar movement is not driven by a need for case, but instead by some (less clearly-defined) need to satisfy a wh-related or discourse-related property of the language. This movement takes place typically not to a case position, but instead to a non-argument position like the specifier of CP. It has at least the appearance of being unbounded, with dependencies spanning an unlimited number of clauses. Beyond these intrinsic differences lie a host of correlational properties. A number of tests have been devised in the large body of work on this subject that distinguish between these two kinds of movement. These tests include the triggering of weak crossover, the ability to strand quantifiers, the ability to reconstruct, and the licensing of parasitic gaps. The tests are categorical


enough that they can be applied to some unclassified form of movement such as scrambling (Mahajan 1990; Kidwai 2000) or QR (Hornstein 1995) to determine which category it falls into. The distinct properties of A and A-bar movement have led to divergent analyses for the dependencies formed by A and A-bar movement in the Government and Binding/Principles and Parameters framework. While I will not spend a great deal of time addressing these approaches here, I will briefly review this body of work based primarily on Chomsky (1981) and Chomsky (1986). Under this view, A and A-bar movement are characterized by the formation of a chain comprised of syntactic objects including the moved item and its traces. A chain formed by A-movement is a set of linked syntactic elements (al' ... , an) in which the head a1 is in an A-position. The head of an A-chain is in a position to which Case is assigned, satisfying the case filter requiring each NP to be associated with a Case position. The tail of an A-chain (an) must be in a position to which a theta-role is assigned. In an A-bar chain, the head a1 is in an A-bar (non-argument) position. The lowest position in a DP chain must be Case-marked, and can either be theta-marked or associated (via an A-chain) with a theta-marked position. A-bar movement operations obey a set of locality conditions, characterized under Subjacency (Chomsky 1981). This approach to A and A-bar movement served to highlight the divergent purposes and endpoints for each type of displacement, as well as the distinct locality constraints that seemed to govern them. Both the inherent and correlational properties of the two types of movements found explanation in these sets of conditions. However, as the Minimalist framework developed, so too did the hypothesis that the feature-checking mechanisms that drive A-movement could potentially underlie all movement operations (Chomsky 1995). Chomsky (2000, 2004) briefly considers the notion that wh-movement specifically could be understood as motivated by sets of features on the head C. He establishes that A-bar movement in this case would be point by point analogous to A-movement, and suggests how successive cyclicity, the wh-island constraint, and wh-in-situ. effects might be derived. At this point, however, Chomsky (2000) did not extend these processes to A-bar displacement such as topicalization, claiming that this was not feature-driven movement. Though I will not adopt precisely the account sketched in Chomsky (2000), the proposal I will make in Chapter 3 will build on these basic ideas. Even at this point in the development of the Minimalist framework, it was already clear that feature-checking mechanisms presented an opportunity to approach A and A-bar movement in a unified way (see also Kuroda 1988). One of the claims I defend here is that the phenomenon of wh-expletives provides us with a uniquely valuable probe as we investigate these questions. By "wh-expletives~ I mean minimal wh-words, with no independent interpretation

9

10


of their own, which appear in the position at which a more deeply embedded full wh-phrase is interpreted. Wh-expletives, also sometimes called scope-markers, have been addressed by a body of previous research which I will discuss in detail in the chapters to follow (McDaniel1989, Mahajan 1990, Dayal1996, among others). Wh-expletive constructions in German, Romani, Hungarian, and Kashmiri are exemplified below, with the wh-expletive itself in bold and glossed "EXPL': (12)

(13)

(14)

(15)

was glaubst du, wenn sie gekommen ist EXPL think 2SG, when 3FSG come AUX.PST 'When do you think she came?' so misline savo film o Demiri dikh-la EXPL think which fihn DEF Demir see-PST 'Which fihn do you think Demir saw?' mit mond-tak hogy kit hiv-ott fel Mari EXPL say-PST that who.ACC call-PST up Mari 'Who did they say that Mary had called up?'

[German] (McDaniel1989) [Romani] (McDaniel1989) [Hungarian] (Horvath 1997)

Tse k'aa chu-y baasaan ki Mohn-as learn' you.DAT EXPL AUX-DAT think that Mohan-DAT who.ERG di-ts kitaab? gave-PST.FSG book 'Who do you think gave Mohan the book?

(Wali & Koul: 18)

Here I will develop an account that analyzes wh-expletives as playing the same role in the A-bar system as DP-expletives play in the A-system. In my view, this comparison serves to reveal a fundamental symmetry of design between the two systems themselves. Let us first discuss our understanding of DP-expletives in this framework. Chomsky (2000, 2004) claims that the EPP on T can be satisfied by Merge of some nominal from within the command-domain ofT. with which it agrees, or alternatively by Merge of an expletive, such as English there. Since by definition, an expletive has no interpretable features of its own, it cannot value the set of features on the head with which it merges ("Expl cannot delete the probe ofnondefective T" (Chomsky 2000: 125)). Therefore these features remain active following the merge of the expletive, and must enter into an agreement relation with an element in their domain. The account ofwh-expletives developed here offers a treatment which closely parallels this approach to DP-expletives. In this approach, wh-expletives have only uninterpretable features. A wh-expletive undergoes Merge into the specifier of the criteria! head for wh-movement (in the case of Kashmiri, C). This merge satisfies the EPP requirement on C, and permits C to interact with some other


un-raised wh-phrase in its domain. This analysis accounts for the fact that it is the wh-expletive, not the full wh-phrase, which occupies the highest wh-position in the clause (in these cases, the specifier of matrix CP). This approach to DP-expletives and wh-expletives suggests that there are some heads in the functional vocabulary of a language that require additional material in their specifiers. This purely syntactic requirement can be satisfied by the associate of the head, or by an expletive of some kind. The fact that both the A and A-bar systems have expletives that are amenable to this analysis seems to reveal a deeper symmetry between the two, as will be detailed in the chapters to follow.

1.3

Organization of the book

Previous work on Kashmiri has focused predominantly on the system of agreement and argument realization of the language (with the so-called 'A-system'). However, this book will turn to a lesser-studied aspect of Kashmiri, the intricacies of the A-bar system. Chapter 2 will first briefly review the basic syntax of Kashmiri and the recent research needed to inform our work here. It then turns to a detailed investigation of A-bar movement to the left edge in Kashmiri, which includes focus movement, wh-movement, and topicalization. From an empirical point of view, this chapter will attempt to map in a detailed way the left periphery of the clause in Kashmiri, and develop a theory of what principles organize it. The goal is to establish an understanding which meshes well with comparative work on these questions and the complex empirical picture, and which is also well-integrated with current theory. The third chapter is centered on long-distance wh-dependencies in Kashmiri. To the best of my knowledge, this phenomenon has not yet been a primary topic of investigation in work on the language. Building on the analysis of the periphery developed in Chapter 2, I will propose an account of wh-movement and of wh-expletive constructions in Kashmiri which suggests that the A and A-bar systems can be understood as governed by the same set of basic mechanisms. At the empirical level, this chapter presents a wide range of data dealing with Kashmiri question formation, including three-clause wh-dependencies, which have been revealing for other languages with wh -expletive constructions (see McDaniel1989). At the theoretical level this chapter addresses the mechanisms which drive wh- movement to and through the clausal periphery; and the way in which the properties of the periphery mediate this movement. Chapter 4 provides a detailed micro-comparison of the syntax oflong-distance wh-dependencies in Kashmiri and in the related and more widely-researched language, Hindi- Urdu. The goal of this section is to provide an account of the very

11

11


different facts in Hindi-Urdu along the lines of the account provided for Kashmiri in Chapter 3. Crucial to this chapter is a series of systematic contrasts between the two languages, which are afforded an explanation in terms of the characteristics of the phase-defining heads C and v. From an empirical perspective this chapter offers a side-by-side comparison of a number of constructions in the two languages. The analysis presented here questions long-held assumptions about Hindi- Urdu, ultimately suggesting that it is not in fact wh-in-situ, but instead a language with wh-movement to a lower periphery at the edge of the verbal domain. The fifth chapter explores an account of the ellipsis construction sluicing in Hindi-Urdu. This is a challenging task since Hindi-Urdu is one of a number of languages traditionally understood as wh-in-situ that exhibit a sluicing construction which seems to be fed by wh-displacement. In Chapter 5 I propose a new analysis of sluicing in Hindi-Urdu that synthesizes the account of wh-movement in Chapter 4 with the understanding of the organization of the periphery built in Chapter 2. This analysis of sluicing posits that it is the properties of the C head that drive the move-and-delete operation, reinforcing our understanding of the periphery not only as a locus of variation, but as a point of crosslinguistic lexical commonalities. Chapter 6 concludes the book by assimilating the new understanding of the periphery afforded by this investigation. It also turns to how these findings address some of the larger questions posed by the current research program. In particular, to what degree can we attribute inter-linguistic variation at the periphery, whether in its fine structure or in the characteristics of long-distance dependencies, to specific properties of the phase defining heads?

CHAPTER2

Feature stacking The Kashmiri periphery

Crosslinguistically, a wide range of elements tends to appear at the left edge of the clause; among these are wh-phrases. topic phrases, focused phrases. and complementizers. Accounts of this subsystem typically rely on a hierarchy of distinct functional projections that appear in an order fixed by universal principles. Each of these projections hosts a single type of element (say, topic or focus) (Rizzi 1997, 2001, Beninca. 2001). This approach, sometimes called "cartographic~ has been a source of considerable empirical discovery, describing a wide range ofleft-edge phenomena. Kashmiri exhibits a relatively rich left periphery in both main and subordinate clauses. The region includes the second position verb, topic, focus, complementizer, and wh-phrases. all of which display rigid ordering and co-occurrence restrictions. For this reason, Kashmiri provides an empirical context for an investigation of how the periphery is organized. In this chapter I will explore some theoretical and empirical ramifications of this so-called cartographic approach to the left periphery, with an emphasis on how the cartographic view interacts with current theoretical developments. In particular, this exploration, and the account of the Kashmiri left edge that we will develop here, will provide a necessary basis for the work on A -bar movement to be considered in coming chapters. The first section of the chapter will present the basic facts of the Kashmiri clause edge. The second section discusses the cartographic approach to such phenomena, and presents a cartographic account of the Kashmiri left periphery. I then turn to a number of theoretical developments that have emerged since the introduction of the cartographic approach. These developments, I claim, provide an improved theoretical context in which to understand the left periphery while maintaining the empirical advances of the cartographic effort. Finally, I conclude the chapter with a look at some possible extensions.

2.1

Kashmiri: A brief introduction

Kashmiri is a lesser-studied language, addressed in a relatively small body of both descriptive work and formal linguistic research. For this reason, I will take this

14


opportunity to establish the context for current work on Kashmiri, by describing the syntax of the language and the body of research on the language upon which the present piece of work builds. Kashmiri, or kaashur as it is called by native speakers, is an Indic 1 language spoken in the greater region known as Kashmir, which includes the Indian state of Jammu and Kashmir, as well as other territory administered by Pakistan but under dispute between the two governments. There are between 3 and 4 million speakers of this language, primarily in India, but also in significant numbers in Pakistan, the United Kingdom, and the United States (Wall & Koul1997). A number of dialects of the Kashmiri language have been identified, although the terms used to refer to these dialects tend to vary. The novel data in this book comes from speakers of the yamraaz variety, spoken in and around Srinagar, which is regarded as standard (Bhatt 1999). Among those studying Kashmiri over the past one hundred years, a dispute arose concerning the family affiliation of the language. Some of the earliest thorough work in the language (Grierson 1919; Chatterji 1963) contended that there existed a separate branch of Indic called Dardic, and that Kashmiri should be grouped as a Dardic language alongside languages like Pashai and Shina. More recently, however, researchers (Fussman 1972; Koul & Schmidt 1983; Zakharyin 1984; Wali & Koul1997) have cautioned that the Dardic designation is a geographic one, not a linguistic one, and there is now a consensus that Kashmiri dearly belongs to the hill language family of the Indic group. Kashmiri has been described in only a few grammars, most incomplete or relatively unavailable. It is relevant to note that, although Kashmiri has a revered literary tradition originating in the 14th century, it has never adopted a standardized writing system. For this reason descriptions of the Kashmiri language have been written in Devanagri, Arabic, or Roman scripts, using widely varying transcription systems. Edgeworth (1841) and Leech (1844) provided skeleton grammars and vocabularies of the language. while Kaula (1898) published a Paninian-style grammar in Sanskrit. Grierson included Kashmiri in the Linguistic SU1-vey of India. published in 1919, following a smaller phrase book and dictionary published in 1911. Kachru ( 1969) contributed a more recent reference grammar for the language, but this was unfortunately not published. The definitive modern grammar is that of Wall and Koul (1997), to which I will frequently refer.

t. I will use the term 'Indic' in reference to this language family over the perhaps more widely used 'Indo-Aryan' (e.g. in the title of Masica 1991).

Chapter 2. Feature stacking 2.1.1

Kashmiri data

The Kashmiri data in this work comes from several sources. First, I have incorporated many observations from the only published modern grammar of the language (Wali & Koul1997), and have indicated this with citation. Second, I have conducted linguistic fieldwork, interviewing native-speaker informants speaking the Srinagar variety. The following are my primary informants: Initials

Gender

Age

Languages spoken (in order of acquisition)

PK

M

56

Kashmiri, Hindi-Urdu, English

JC BC

F

51


F

80+

Kashmiri only

vc

M

59


Hgure 1.

All were born in Kashmir and left the region during cycles of significant violence and civil disruption. I have worked with these informants, and with other Kashmiri-speaking individuals and families, since the Spring of 2004. In each case, I have pursued a combination of techniques, including elicitation of narrative, elicitation/translation of specific example types, and requests for grammaticality judgments. Interviews were typically conducted in English, with the use of Hindi-Urdu where required. Interviews with BC were conducted through a translator. In this book, Kashmiri data attributed to any of these informants is marked with the speaker's initials and the date of the recording. A third source of Kashmiri data in this book is edited stories and poems, and a limited selection of naturally occurring data online, including discussion group postings and news items. These examples are cited with the source and date obtained where applicable. 2.1.2

Syntax

In this section I will present a brief overview of the basic features of Kashmiri syntax, which differ significantly from the better-known Indic languages such as Hindi- Urdu. The various chapters of this book deal in detail with a range of specific issues. To situate those discussions and to make them easier to follow, I ofter here a snapshot of the basic syntactic processes. This will serve to facilitate comprehension of examples in future chapters, and permit the reader to focus on the important constructions at hand. Kashmiri has a rich system of nominal case declensions, including nominative/absolutive, dative, ablative, ergative, and genitive. Nominative/absolutive

15

16


is unmarked, and is the case given to subjects of all intransitive clauses, and the subjects of all transitive clauses except for those in the perfective aspect In the perfective aspect it is the direct object of the transitive clause that is unmarked. Case is marked with suffixes that vary according to number and gender. These are displayed in Figure 2. Case

Masculine

Feminine

Singular

Plural

Singular

Plural

Nom/Abs

0

0

0

0

Ergative Dative Ablative

an is/as ill

av an av

av an av

Figure 2.

Case stacking (dual case marking) is possible in Kashmiri, in particular with the genitive and another oblique case (Wali & Koul 1997). This occurs, for instance, when an object of comparison is also the complement of a postposition or comparative postpositional phrase. (1)

... farid-ni-s mukaabal-as manz ... Farid-GEN-DAT comparison-OAT in '... in comparison with Farid .. .'

(Wall & Koul: 157)

Both animate and inanimate nouns in Kashmiri are gendered (either Masculine or Feminine), and the gender of many nouns can be discerned by the morphological form of the word. Feminine nouns typically have endings such as -en', -In,- a:n, Ir, and -baay, as shown in Figure 3. There are, of course, many exceptions (Wali & Koul). Masculine

Feminine

marld 'man'

zanaan 'woman' gaglr'raf

ooluv 'potato' maastar 'teach' ph'ok 'shoulder'

maastar-baay 'teacher' ~r'window'

Figure 3.

Suffixation and stem vowel changes typically mark the plural form of nouns, as depicted in Figure 4 below.

Chapter 2. Feature stacking

Noun

Singular

Plural

'window'

~r

daari

'shoulder'

ph'ok

phek'

'potato'

ooluv

'rat'

gaglr

oolav gagri

Figure 4.

Kashmiri finite verbs and auxiliaries display inflection for number, person, and gender, such as in (2). (2)

a.

b.

bi ch-u-s skuul gatsh-aan lSG AUX-M-lSG school go-PRT 'I go to school.'

(Wall & Koul1997: 152)

mohn-an chal' palav. Mahan-ERG wash.PST.MPL clothes. 'Mohan washed the clothes.'

(Wali & Koul: 153)

haa manshi, k.'aazl chu-kh vuth-aan sekhi lavar 0 man, why AUX-2SG twist-PRT sand rope '0 man, why do you twist a rope of sand?' (Lal Ded, 14th century) The types of inflection that appear on the verb are of two broad categories. Core agreement is obligatory, and is controlled by the argument in nominative/absolutive case. This agreement encodes number and person features, and gender features in all non-future tenses. (3)

(4)

bi go-s lSG-NOM go-PAST.M.lSG 'I went.' Aslam-an vuch-u-kh tsl. Aslam-ERG see-MSG-2Ps 2sG.ABS.MSG 'Aslam saw you.'

(Bhatt 1999)

(Wall & Koul1997:248)

The second type of agreement involves sets of pronominal suffixes. These clitics may be doubled by their associated arguments, though not in all cases. First and third person ergative arguments need not be marked unless subjects are null. Second-person arguments must be marked, whether the pronoun is null or overt. A person hierarchy conditions the appearance of suffixes marking nominative objects in the nonperfective aspects and dative arguments. Examples of verbs with pronominal suffixes are supplied below (Wali & Koul1997).

17

18


(5)

a.

b.

tse vichi-th-as bl 2SG-ERG saw-2SG-1SG me-ABS 'You saw me.'

(Wall & Koul: 253)

bl chu-s-an-ay su tse havaall kar-aan lSG.NOM AUX-1SG-3SG-2SG 3SG.ACC 2SG.DAT hand over do-PRP 'I am handing him over to you.' (Wall & Koul: 253)

These inflectional elements on the verb are strictly ordered. Core gender-number agreement must precede all pronominal suffixes, and the pronominal suffixes follow a hierarchical order determined by the grammatical function of the argument; specifically. suffixes referring to the subject precede those referring to the object, which in turn precede those referring to the indirect object. Like Hindi-Urdu, Kashmiri is a split-ergative language. That is, the system of case-marking is nominative-accusative in all nonperfective aspects, and ergative-absolutive in the perfective aspect. In the perfective aspect the subjects of intransitive predicates and the direct objects of transitive predicates bear the same (unmarked) case. Kashmiri is generally claimed to be a language with underlying verb-final word order (SOV) in which tensed clauses surface as verb-second. Non-finite clauses, as well as relative clauses, are verb-final. (6)

su laRkl [yus dill ch-u roozaan] ch-u COR boy REL Delhi AUX.PRS-MSG Jive-PRP be.PRS-MSG m'oon booy POSS.MSG brother 'The boy who lives in Delhi is my brother'

(7)

(Wall & Koul1997: 54)

salim chu yatshaan [me baagaas manz vuch-un] Selim AUX.PRS-MSG want-PRP [lsG-DAT garden in see-INF.NEUTER] Selim wants to see me in the garden. (Wall & Koul1997: 46)

All other tensed clauses, including complement clauses, exhibit verb-second. We will examine this property of Kashmiri in more detail in the sections to follow. In general, Kashmiri is a head final language. In NPs, specifiers, genitives, and complements precede the head (8). Adpositions follow their complements (9). (8)

(9)

TuurisTan-hund makaan Tourists-GEN house 'Tourist's house' a.

Tern an zanaan [maal-is khaatr] 3SG-ERG bring.PST wife father-OBL for 'He brought (his) wife for (the sake of his) father.'

(Bhatt 1999)

(Bhatt 1999)


b.

yath forum-as m.anz this forum-OBL in '.. .in this forum'

(3/27/06, GreaterKashmir Forum)2

Complement clauses in Kashmiri appear uniformly to the right of the matrix clause, unlike in some other Indic languages (for Bengali, see Bayer 1996). And unlike in some Germanic languages, verb-second order prevails. Finite complement clauses in Kashmiri are optionally introduced by the element ki, which does not "count" in determining second position. (10)

laRk-as ch-a khabar [ki swa yii-na) boy-DAT AUX.PRS-3SG knowledge that 3SG.F come.FUT-NEG (Bhatt 1999: 74) 'The boy knows that she will not come.'

The internal structure of complement clauses, and in particular the left edge of these clauses, will be crucial to the discussions that follow. The other properties of Kashmiri syntax that will be relevant to our discussion principally concern movement to A-bar positions. Detailed discussion of these facts will be provided in the chapters below.

2.1.3

Previous work on Kashmiri

Just as there are few grammars of Kashmiri, there has been comparatively little formal syntactic research on the language. This is particularly surprising considering the significant attention given to other Indic languages such as Hindi- Urdu, Marathi, and Bengali. Most of the previous work on Kashmiri has been concentrated in three areas: the complex case system, the system of agreement and cliticization, and the verb-second phenomenon. Overall, it is the word order and the syntactic and morphological intricacies of the A-system of Kashmiri that have to this point interested linguists. Many of the early formal observations of Kashmiri, particularly concerning word order, were made by Peter Hook (see Hook 1976, Hook 1984, Hook & Manaster-Ramer 1985). Bhatt (1999) attributes to Hook first mention of verbsecond word order in Kashmiri. Other general work on word order in Kashmiri appeared in the volume edited by Hook and Koul (1994), such as Subbarao (1984). A number of researchers have undertaken shorter explorations of the Kashmiri system of case and agreement. Bhatt (1993a, 1993b) has investigated a range of case-related issues in Kashmiri, particularly with respect to ergative-nominative

2. http ://greaterkashmir. com/forum/topic.asp ?which page= 1&TO PIC_ID=84&REPLY_ ID=661

19

20


structures. Kachru, Kachru, and Bhatia (1976) made comparative observations about subjecthood in a range of languages including Kashmiri. Linked research on Kashmiri agreement and cliticization includes work by WaH and Koul (1992), Subbarao (2001), and Subbarao and Mushi (2000). Hook and Koul (2004) represents a fairly recent contribution to this line of research, comparing case and agreement in Kashmiri with that in Shina, Poguli, and Gujarati. Larger scale work on Kashmiri beyond the grammars mentioned above is comparatively rare. Raina (1991), in her dissertation, investigated word order and argument structure of the language. Her central claim, disputed by subsequent researchers (see Bhatt 1999), was that Kashmiri is a nonconfigurationallanguage in which the subject and the object mutually c-command one another. Most recently Bhatt (1999) has published a significant study of the verbsecond phenomenon in Kashmiri. featuring detailed comparisons with Germanic verb-second. This work builds on earlier research by Bhatt and Yoon (1992), in which they proposed a functional Mood projection on the left edge of the clause. The second-position verb ordering was attributed to the fact that the verb appeared as a reflex of Mood marking in the head of MoodP. This proposal was extended to the range of Germanic verb-second as well in Bhatt (1999). Because Kashmiri and a few related varieties are unique among the Indic languages in featuring verb-second, this aspect of the language has rightfully attracted research attention. The second chapter of this book will build in part on the work of Bhatt (1999), analyzing the syntax of verb-second structures, but attempting also to understand the phenomenon in the context of a broader theory of movement to the left edge.

2.2

The Kashmiri left periphery

Kashmiri is unusual among the Indic languages in exhibiting the verb-second (V2) property, more familiar from Germanic languages. To the left of the verb, a number of constituent types may be found at the clause edge. The finite verb appears as the second constituent of a finite declarative clause. Any of the arguments (or other constituents) may appear first. (lla) exhibits the unmarked order, and (llb-e) are also grammatical (all from Wall & Koul1997:89). (11)

a. aslam-an di-ts mohn-as kitaab raam-ini kh;ltrl raath aslam-ERG give.PST-FSG Mohan-DAT book Ram-DAT for yesterday 'Aslam gave Mohan a book for Ram yesterday.' b. mohn-as di-ts aslam-an kitaab raam-ini k~trl raath Mohan-DAT give.PST-FSG Aslam-ERG book Ram-DAT for yesterday 'Aslam gave Mohan a book for Ram yesterday.'


( 12)

c.

kitaab di-ts aslam-an mohn-as raam-ini khatrl raath yesterday book give.PST-PSG Aslam-ERG Mohan-DAT Ram-DAT for 'Aslam gave Mohan a book for Ram yesterday.'

d.

raam-ini khatrl di-ts aslam-an mohn-as kitaab raath Ram-DAT for give.PST-FSG Aslam -ERG Mohan-DAT book yesterday 'Aslam gave Mohan a book for Ram yesterday.'

e.

raath di-ts aslam-an mohn-as kitaab raam-ini khatrl yesterday gave Aslam-ERG Mohan-DAT book Ram-ERG for 'Aslam gave Mohan a book for Ram yesterday.'

bshura zabaana chi akha arya zabaana. Aryan language Kashmiri language be.PRS.3s an 'The Kashmiri language is an Aryan language.'

(2/20/06)3

We can further probe the position of the verb by examining the position of sentential negation and the distinction between auxiliaries and main verbs in Kashmiri. Sentential negation follows the second position verb, attaching as a suffix. (13)

raath khyav-na lark-av batl yesterday eatMSG-NEG boy-ERG.PL food 'The boys did not eat the food yesterday'.

(Bhatt 1999: 104)

In a sentence with a tensed auxiliary, it is the auxiliary that occupies second position, and not the main verb. This is frequently taken as evidence that the verb is underlyingly in final position (Bhatt 1999). It is also the auxiliary to which negation attaches, as in (14d). (14)

a.

laRk ch-u dohay skuul gatsh-aan boy AUX-3MS daily school go-IMPFV 'The boy goes to school every day.'

(Bhatt 1999)

b. *laRk dohay skuul gatsh-aan ch-u c.

50 lacha lukha ch-i yeh boolaana. 50 (100000) people AUX-3MPL this speak 'Five million people speak it' (2/20/06, Kashmiri Wikipedia)4

d.

bl chu-s-nl azkal garl gatshaan 1SG AUX-1MS-NEG nowadays home go-IMPFV 'I don't go home nowadays.'

(Bhatt 1999)

3· More commonly accepted terms for this language familyinclude'IndiC: used here, and 'IndoAryan'. Example from: http:///ks.wikipedia.org/wiki/K%C 5%8F%C5%9Bura_zab%C4%81 na. 4

http://!ks.wikipedia.org/wiki/K%CS%8F%C5%9Bura_zab%C4 %81 na.

21

:u Peripheries in Kashmiri and Hindi- Urdu Let us now turn to the constituents that precede the second-position verb. The nonsubject pre-verbal constituents in (llb-e) are generally interpreted as focused. For instance, the focus-particle -ti can only appear suffixed to a constituent in this position (Bhatt 1999). (15)

bl ti goo-s gari vakht-as peth 1SG FOC go.PST-1SG home time-OAT on 'I too went home on time.'

(Bhatt 1999)

Note that the suffixation of -ti to huun 'dog' in (16) is grammatical only if huun is found in the pre-verbal position, as in (16), not when it follows the auxiliary, as in (17). (16)

behna broNh panin jaay goD huun-ti ch-u dog-Foe AUX-3MSG seat before self's place first saaf kar-aan clean do-IMPFV

(Bhatt 1999)

'Even the dog cleans his place before sitting.' (17) ?*panin jaay ch-u huun-ti behna broNh goD saaf kar-aan self's place AUX-3MSG dog-Foe seat before first clean do-IMPFV Intended: 'Even the dog cleans his place before sitting.' In constituent questions, the focused interrogative phrase must appear immediately before the verb, as in (18). Other positions for the interrogative constituent are strongly dispreferred.

(18)

a.

learn' haa-v shill-as nav kitaab raath who.ERG show.PST-FSG Sheila-OAT new book yesterday 'Who showed a new book to Sheila yesterday?' (Wall & Koul: 12)

b.

kam-is ch-i vaariyaah paasl? who-OAT be-3MPL lot money 'Who has a lot of money?'

(Wall & Koul: 14)

(19) *?shill-as haa-v learn' kitaab raath Sheila-OAT show.PsT-FSG who.ERG book yesterday Intended: 'Who showed a book to Sheila yesterday?' (judgment: PK 9/21/04) In one important case, an additional constituent can precede the verb, which will thus no longer be "second~ though it is not in its base position. This additional pre-wh constituent in (20) may occur just when the wh-word is present, and it is interpreted as a Topic (Bhatt 1999). (20)

a.

raj-an learn-is haa-v nav kitaab? Raj-ERG whom-OAT show-PsT-FSG new book 'As for Raj, to whom did he show his new book?' (Wall & Koul: 12)


b.

mohn-an k'aa kor panlni gari Mohan-ERG what do.PST.MSG his-at house 'As for Mohan, what did he do at his house?'

(Wali & Koul)

It is ungrammatical to have more than one topic (as in ( 21 a)), to have the wh-phrase precede the topic (21b), or to have a topic precede a non-interrogative focus (21c) Qudgments all JC 9/8/05).5 (21)

a. *rajan nav kitaab kam-is haa-v Raj-ERG new book whom-DAT show-PST.FSG Intended: 'As for Raj, as for the new book. to whom did he show it?' b. *kam' tse chu-y baasan ki mohn-as who-ERG 2sG.DAT AUX-DAT think that Mohan-DAT di-ts kitaab give-PST.FSG book Intended: 'As for you, who do you think gave Mohan the book?' c. *gari bl goo-s vakht-as peth home lsGwent.PST-lsG time-DAT on Intended: 'As for home, I went there on time.'

Subordinate clauses are identical to matrix clauses in their word order, except that they are optionally preceded by the particle ki 'that'. This particle is not counted in determining verb-second position. These facts are exemplified by the balded material in the sentences in (22)-(23). (22)

mohn-an kitaab. miiraayi ch-a pataa ki kam-is di-ts AUX-3SG know that who-DAT give-PST.FSG Mohan-ERG book Mira 'Mira knows who Mohan gave a book to.' (Wall 2002)

(23)

di-ts kitaab. miiraayi cha pataa ki mohn-an kam-is Mira Aux-3sG know that Mahan-ERG who-DAT give-PST.FSG book. (JC 9/8/05) 'Mira knows who Mohan gave a book to.'

5· Bhatt (1999) provides other evidence that, with the exception of subjects and temporal adverbs, the immediately preverbal constituent is a focus and not a topic. For instance, universally quantified nominals, which have been argued to be incompatible with topicalization (May 1977), appear in this preverbal position. (i)

sooruikeNh khyav ramesh-an everything eat.PST Ramesh-ERG 'Ramesh ate everything'

Fwther, kaNh 'someone: which is inherently unfocused, is not a good initial constituent. (ii) ?*kaNh oosu-yi tse tshaanD-aan someone AUX.PST-FSG 2SG.DAT looking-PRP 'Someone was looking for you.'

23

24


In summary, the left periphery of the finite clause in Kashmiri can take two essential forms. The first (in (24a)), is when a single focused constituent, whether interrogative or non-interrogative, precedes the verb. The second (in (24b)), is when a topic phrase precedes a wh-phrase which precedes the verb. (24)

a. b.

[Focused (wh or non-wh) XP] [verb] [... ] [Topic XP] [Focused wh-XP] [verb] [... ]

In the case of subordinate clauses, either order can be preceded by the element ki.

2.3

The cartographic approach to the left periphery ofKashmiri

The left periphery of the Kashmiri clause is a relatively complex one, and any coherent account ofwh-movement and wh-expletive structures in the language is going to require an understanding of how the periphery is structured. Rizzi (1997) initiated a research program in which the left periphery, or 'C-domain', is regarded not as a single functional projection, but rather as an articulated hierarchy of distinct projections. The program has yielded rich empirical results (see the volumes edited by Belletti (2002) and Rizzi (2004)) and has been influentiaL In this view, the left periphery of the clause is comprised of a sequence of functional projections whose hierarchical order is fixed universally. Each of these heads hosts a unique element in its single specifier. The expansion of the CP layer into this sequence conceptually echoes the expansion of the IP layer into a series offunctional projections (Pollock 1989). In its original conception in Rizzi (1997), this theory posits at least the following projections: (25)

ForceP

~ Forceo ~ Top°

FocP

~ Foc TopP ~ TopO FinP 0

~ This hierarchy divides into two types of projections. Force and Finiteness projections, on the edges of this structure, are required. They are present at every clause edge for all languages. The Force projection contains information that determines


the force of the clause to follow (i.e. interrogative, exclamative, imperative, and so on). The Finiteness head contains information about whether the clause will be finite or non -finite. Each of these heads may (or may not) host morphological material. The other projections are in some sense optional. Topic and Focus projections appear in the structure "when needed~ or when a constituent with topic or focus features in the main clause needs to enter into a specifier-head relation with the relevant functional head. Note here that the Topic head can be recursive, allowing for multiple topics in a single clause edge, while the Focus head cannot. Rizzi (1997) suggests that there cannot be more than one focus in a given clause because if there were, an interpretive paradox would arise. While a lower focus must have a focused or 'new' interpretation, it must also simultaneously be interpreted as given or 'old' as part of the presupposition of a higher focal head. In the system introduced above, all movements to the left periphery are driven by the need to satisfy some criterion. That is, constituents with a topic or focus feature must ultimately be in the specifier-head relation with a head bearing those same features. It is this feature that motivates both the presence of the relevant optional projection in the structure, and the movement itself. Importantly, each functional head is the realization of precisely one syntactically relevant feature (Rizzi 2004). Let us examine how this system might account for the left periphery of a Kashmiri clause. In a simple declarative clause as in (llb), repeated here as (26), we find a focused constituent on the left edge, followed immediately by the second position verb. (26)

Mohn-as di-ts aslam-an kitaab raarn-ini khatrl raath Mohan-DAT give-PST.PSG Aslam-ERG book Rarn-DAT for yesterday ~slam gave Mohan a book for Ram yesterday.'

In order to form this sentence, a focus projection must appear on the left edge, sandwiched between ForceP and FinP. A focus feature present on the Foe head attracts the focused constituent in the clause, prompting a move to Spec, FocP. According to Rizzi's approach to Germanic verb-second, when the Focus head is projected it also attracts the finite verb. We will assume this is also the case in Kashmiri. (27)

ForceP

~ Force° FocP

moha.n~ 'Mohan' Foco dits 'gave'

FinP

~ Fin

IP

15

:16


Note that in this Kashmiri sentence there is no audible material in the Force or Finiteness projections - that is, there seems to be no morpheme that corresponds in particular to the interpretations designated for these heads. In the case of a more complex interrogative clause such as (20), repeated as (28) below, the preverbal position is occupied by a focused wh-word, mutually exclusive of any other focused constituent. (28)

Raj-an kamis haa-v nav kitaab? Raj-ERG who.DAT show.PsT-FSG new book ~s for Raj, to whom did he show his new book?'

Preceding the wh-element is a constituent interpreted as a topic. This topic can only be present when a wh-word occupies the focus position; it is otherwise ungrammatical, as shown in (21c). Under the cartographic approach, in this sentence a Focus projection must again appear between ForceP and FinP. A single Topic projection (recursive topics are not possible in Kashmiri as seen in (21a)), must appear between ForceP and FocP. It is also not possible for this single TopP to appear between FocP and FinP, as the topic must precede the second position verb in the head of FocP (as seen in the ungrammatical sentence in (21b)). The focus and Q feature in the Foe head interact with the focus and ensure that it moves to the left periphery. The topic feature on the Top head motivates movement to that specifier as well. Again, the presence of the Foe head attracts the second position verb. The resulting structure is below. (29)

ForceP

~ Force X 0

.

ra;~ ~ 'Raj' Topo

FocP

ka~ 'who' Foe FinP

haav ~ 'showed'~ FinO IP

Again, there is no audible linguistic material in this sentence that would appear in the head or specifier of either the ForceP or FinP, nor in the head ofTopP.

2.4

New opportunities

The cartographic approach to this point has been the most successful analysis of languages such as Kashmiri or Italian, which exhibit an articulated left periphery.

Chapter 2. Feature stacking 27

The hierarchy of functional projections not only permits a wide range of material to be incorporated, but also prescribes the relative order in which this material can appear. To a significant degree, the cartographic approach fixes the order and number of these projections universally, helping us to describe many varied peripheries. I won't attempt a broad critique of the cartographic approach here (see, for instance, Newmeyer 2004, Lahne 2007). Instead I attempt to show that since the initial proposals were made, there have been several theoretical developments that let us look at these sorts of facts in a new way. When constructing a more complete account of the periphery of Kashmiri, a close examination of this approach may present us with new explanatory opportunities, and allow us to expand our understanding of periphery as the edge of the derivational unit. 2.4.1

The specifier-head relation

The first of these developments involves the specifier-head relation. In the cartographic view of the left periphery, it is the formation of specifier-head relation, satisfying criteria on the peripheral heads, which causes such a range of projections to appear on the left edge. For each projection, there is a single specifier in a unique relation to its head. If we are committed to the notion that there is a single specifier for each projection, we are likewise committed to the position that there must be a unique projection for each constituent that undergoes A-bar movement to the left edge. That is, whether or not we have evidence for a head in that position, we must postulate that one exists in order to provide room for a specifier. However, recent work has suggested that the restriction that there be just one specifier per head is neither theoretically nor empirically justified (Chomsky 1995, 2000, 2008, Ura 2000). Abandoning this restriction, we could permit multiple specifiers to be hosted by a single head. This shift in theoretical perspective is in harmony with two kinds of empirical observations. The first is that while evidence for a sequence of phrasal constituents on the left periphery is overwhelming, evidence for a sequence of distinct head positions among these phrasal constituents is delicate at best. The second observation has to do with the positioning of audible linguistic material in the heads of the left periphery. As described above, Kashmiri is a "verb-second" language, in the sense that the finite verb must follow at least one major clausal constituent in declarative sentences. The crudeness of the term "verb-second" becomes obvious when we examine interrogative clauses. in which the verb is actually in third position, preceded by the topic and a theoretically unlimited number of wh-phrases. In the cartographic view, this seems to indicate that the second-position verb is located in the Focus head. However, looking at

:18


the hierarchy of projections in (215), there are at least four heads to which the verb could potentially move: Force0, Topic 0, Focus0, and Fin°. It would seem, given this structure, that it would be possible for the verb to raise further to Topic 0, in which case it should directly follow the topic (and precede wh-material) in linear order. Possibly it could move to an even higher head, such as Force 0, in which case it could precede the topic. Yet both of these alternative orders are degraded to ungrammatical Qudgments JC 9/8/05). (30)

a. *raj-an haa-v kam-is nav kitaab? Raj-ERG show.PST-FSG who-DAT new book Intended: 'As for Raj, to whom did he show the new book?' b. *haav raj-an kam-is nav kitaab? show.PsT-FSG Raj-ERG who-DAT new book Intended: 'As for Raj, to whom did he show his new book?'

In order to rule out the ungrammatical constructions in (30), we will have to require that head movement raises the Kashmiri verb as far as the Focus head, and no further. That is, an analysis such as that in (18) multiplies analytical possibilities, in the sense that it provides multiple possible landing sites for raising of the finite verb and provides no principled basis for choosing among them. It seems that the concept of multiple specifiers may be forced upon us for Kashmiri in any case, given that Kashmiri is a language in which multiple whphrases can all appear displaced to the periphery. (31)

(32)

kam-is kam' k'aa d'ut. Who-DAT WhO.ERG what give.PST.MSG 'Who gave what to whom?'

(Wali & Koul: 26)

Me ch-u ru pataa ki kam-is kam' 1SG.ERG AUX.PRS-MSG NEG know that who-DAT who-ERG k'aa d'ut. What give.PST.MSG 'I don't know who gave what to whom.

(Wali & Koul: 26)

We might assume that each of these wh-phrases in (31)-(32) is in the specifier of a Focus projection. Since this form of the cartographic account stipulates only a single Focus projection, the dearest path is to assume that the Focus projection has multiple specifiers in Kashmiri to host multiple interrogative foci (but crucially not non-interrogative foci). Any analysis placing each of these wh-phrases in the specifiers of separate heads risks creating even more potential landing sites for the verb, but in a sentence in Kashmiri with multiple fronted wh-phrases, the verb can only appear following the last wh-phrase.


(33) *k.ati kus oos kam-is meellni gatsh-aan going-IMPERF where who.NOM AUX.PST.3SG who-DAT meet Intended: 'Who was going to meet whom where?' An account in which there is a single C head with multiple specifiers narrows the range of possible analyses (assuming that heads may only move to head-positions) to one, and leads us to expect what is in fact the case - namely that the finite verb will appear to the right of all fronted phrasal constituents in the C-domain. 2.4.2

The cartographic project and the phase

The clause edge not only functions as a position for the placement of constituents with certain discourse-related functions, but also has historically been viewed as a transition point between one clause and another, particularly for successive-cyclic movement. A relatively recent theoretical development concerning the nature of this transition point is the concept of the 'phase: as defined in Chomsky (2000, 2001, 2004). Phases are self-contained subparts of a derivation, each beginning with a numeration and ending with transfer of the objects created to the interfaces. CP and vP have been identified in the literature as the minimal phases, with other functional projections such as DP claimed to have phase status as well (Svenonius 2003). The clause edge as addressed by the cartographic project is also identified as the edge of the phase, a region with a special status. Constituents on the edge of the phase do not transfer to the interfaces along with the phase itself, but instead remain accessible to probes in the next higher phase (Chomsky 2004). This is the process that makes successive cyclic wh-movement possible, for instance. To determine the edge of the phase, we must know what the phase-defining head is. This process is outlined in the Phase Impenetrability Condition (PIC) as follows: Phase Impenetrability Condition- "In a phase a: with head H, the domain of H is not accessible to operations outside a;, only H andits edge are accessible to such operations" where the edge includes specifiers and adjuncts to H. (Chomsky 2000: 108)

The map in (25) is a theory of CP. CP is also the category whose status as a phase is best established. To the extent, then, that we want to maintain results and analyses that depend on the notion of the phase, theories of the CP-domain must provide us with a reasonable way of defining phasehood. The first task in correlating the phase and the cartographic hierarchy of the left periphery as in (25) is to identify the phase-defining head. However, this becomes a challenge once the CP is split into a hierarchy of projections. It is unclear which

30


of these projections becomes the phase-defining head, and more importantly what material is then by definition considered to be on the phase edge. This question can be addressed empirically when we examine a wh-movement construction such as the one in (34). (34)

Tse k-am' chu-y baasaan [ki Mohn-as di-ts kitaab?] 2sG.DAT who.ERG AUX-DAT think that Mohan-DAT give-PST.FSG book 'Who do you think gave Mohan the book?'

Though we will turn to address constructions such as that in (34) in greater detail in Chapter 3, at this juncture it serves to illustrate a very specific concern. The balded wh-word bm' 'who' originated in the lower clause in (34) as the indirect object of the verb dits 'gave'. Given our understanding of wh-movement in the current framework, kam' must have moved to the edge of the bracketed subordinate clause at some point According to the cartographic view, at this point kam' would occupy the specifier position of the Focus phrase on the left periphery of this lower clause. (35)

ForceP

~ Forceo

FocP

ke~FinP

'who' FocO dits 'gave'

~ . ~ Fin IP

In this position, kam' must be on the phase edge, so that it is able to interact with probes in the higher clause and ultimately move to its final position in the matrix focus projection. From this we could conclude that the phase-defining head in the split-CP is Focus, and so any material in the Focus head, or in its specifier, is on the phase edge. However, this conclusion will prove too simplistic, even for simple interrogative clause in which a Topic is present, such as that in (36). (36)

Raj-an kam-is haa-v nav kitaab? Raj-ERG who-DAT show.PST-FSG new book 'To whom did Raj show his new book?'

In Kashmiri, the TopicP dominates FocusP. If Focus0 is the phase-defining head, then the probe in the Topic head would be outside of the phase, and would be unable to probe any material inside the domain of the Focus head. In particular,


the phrase with topic features in this derivation, rajan 'Raj: would be unable to interact with the topic features on the Topic head, being inaccessible to it. (37)

ForceP

~ Forceo TopP ~ Topo FocP [top] ~ hmis ~ 'to whom' Foe FinP haav ~ 'showed'· ~ FinO IP

/"'-. raja.n ... [top] The derivation would crash since the topic features cannot be satisfied, and so designating Focus as the phase-defining head will certainly not achieve the desired result. Of course, if the phase-defining head is instead Topic0, a difierent problem arises, in that any wh-material in the immediately lower FocusP will be inaccessible to any probe in a higher clause. That is because such material will no longer be on the phase edge, being lower than the specifier ofTopicP. This would make it impossible to successfully derive the wh-question in (36). Another option logically available would be to view the entire left periphery as the phase edge. This is a natural move, given that this hierarchy of projections was intended to represent an expansion of the single CP. If the single C is phasedefining in the current theoretical view, then so too must be the array of projections created when this CP is split. This then would require that we re-vamp our notion of phase such that we can designate an array of heads as the phase-defining unit, and all of the linguistic material in that array as being on the phase edge. We might simultaneously ask whether other phase boundaries, such as the vP, actually represent an expanded array of functional projections, all of which are phase-defining. It may be possible to work this proposal out in some reasonable way, but doing so depends on a definition of formal relatedness among the various distinct heads of (25) which for the moment, at least, remains vague. None of these elaborations is necessary if, instead of (25), we posit a single (phase-defining) head of category

31

32


C allowing multiple specifiers - all of which will be on the phase-edge according to the definition of the PIC above. 2.4.3

Order of projections: encoding variation

The cartographic project also raises an important theoretical question concerning the way in which parametric variation is handled by the grammar. In particular, what is the source of the language-to-language variation in the order and number of the constituents on the left periphery? While some languages, like Kashmiri or Italian, make very elaborate use of the left periphery, other languages, such as Irish and English, make markedly less use of such resources. This section explores where this kind of variation might be regulated and encoded. Within the cartographic view, let us first tum to the way in which the order of projections that appear on the left periphery of a given language is determined. While it is true that certain patterns emerge consistently on the left periphery, there is also significant crosslinguistic variation. It is useful here to compare Kashmiri with other languages that exhibit verb-second order in subordinate clauses, such as the Germanic languages Yiddish and Icelandic. In the case of indirect questions in all three of these languages, there are a number of constituents on the left-periphery, including the topic, the wh -word, and the verb. In indirect questions in Yiddish, the order of the constituents is wh-topic-verb, but never *topic-wh-verb (Diesing 1990; Bhatt 1999). (38)

Ikh veys vos bay mir tut zikh. 1sG know what by me does REFL. 'I know what goes on with me.' b. *Ikh veys bay mir vos tuto zikh. a.

On the other hand, in Kashmiri indirect questions, the order of the constituents is the reverse: topic-wh-verb, but never *wh-topic-verb (Bhatt 1999). (39)

a.

me ch-i pataa ki batl kam' khyav 1SG.DAT AUX.PST-MSG know that rice who.ERG eat.PST. 'I know (that), as for rice, who ate it' (Bhatt 1999: 166)

b. *me ch-i pataa ki kam' bati khyav 1SG.DAT AUX.PST-MSG know that who.ERG rice eat.PST Intended: I know (that), as for rice, who ate it (Bhatt 1999: 166) We would need to address how this intra-language variation is encoded, and how it is expressed in the syntax. Within Kashmiri, there are also some particular restrictions. The hierarchy of projections presupposed by (25) suggests that the Topic projection has the potential to iterate (though the Focus projection cannot). However, only one


topic is permitted per clause in Kashmiri, so the order *topic-focus-topic or *topic-wh-topic is unavailable, as exemplified in (21) above. Again, we would need to ask what it is about the syntax ofKashmiri in particular that restricts the number of topics. More generally, under the assumptions of the cartographic approach, how could we determine which projections may appear in the left periphery of a given language, and in what order? Cinque (1999) is largely agnostic about what mechanisms determine these patterns. Rizzi (1997) suggests that when an element bearing the relevant features (say topic) appears within the sentence, the optional projection associated with that feature (in this case, TopP) will project on the left periphery. Of course, this alone does not dictate in what order those phrases may project, nor does it limit how many could potentially appear. It seems that we would also need some device with essentially the effect of traditional phrase structure rules to restrict the inventory of left-periphery projections for a given language, and to indicate their relative order. In the case ofKashmiri, these mechanisms must permit all and only the hierarchies in (40) (where the arrow indicates immediate containment). (40)

a.

ForceP--? TopP--? FocP (wh only)--? FinP

b. ForceP --? FocP --? FinP For other languages, very different patterns must be guaranteed, particularly with respect to the order and number of Topic and Focus projections. In this sense, the larger cartographic hierarchy represented in (25) is something of a template, indicating an upper bound on what arrays of structures languages might employ on the left periphery. The broader theory-internal question that opens at this point is how we wish to account for parametric variation in the grammar. To the degree that we are committed to an approach like the cartographic one, we must also be committed to the existence of phrase structure rules or similar language-specific mechanisms. On the other hand, there is a line of research that adopts as a premise that all parametric variation resides in the functional lexicon. That is, language-specific characteristics are located in the learnable functional vocabulary of a language. It would be desirable, then, to develop a theory of the variation just sketched which would be compatible with this program. I will suggest below that such a theory can be readily constructed given the view that the CP-domain is shaped by a single head. 6

Rizzi (2004) argues that the tension between the cartographic approach and focus of minimalist analyses on certain core categories (C, T, v, etc.) is only apparent However, if we take the notion of a reduced vocabulary of functional structures seriously, Section 2.4 suggests 6.

33

34


In the section that follows I develop a theory that tries simultaneously to maintain the empirical gains of cartographic work, and to take advantage of some of the theoretical opportunities reviewed here. The hope is that this proposal will provide good descriptions of languages like Kashmiri but also interact seamlessly with current theoretical assumptions that we may want to adopt for independent reasons.

2.5

Feature stacking

My starting point will be the idea that the attributes of the left periphery can be accounted for by way of a single functional head whose features have some internal organization- specifically in that they are ordered, or form a stack (see Bobaljik & 1hrainsson 1998, Muller 2010). Constituents that interact with this head can potentially undergo Move, creating multiple specifiers of this single head. Most importantly, the order of features in the stack is intended to mirror (as well as to capture) the patterns being uncovered in cartographic work. 2.5.1

Features and the lexicon

Features are linguistic properties that are made available by UG. A given language makes a one-time selection from these features and organizes them to form a lexicon (Chomsky 2000). I propose that features are grouped into bundles. Each bundle is a unit that will eventually be valued in a single Probe-Goal interaction in a derivation. Each syntactic head in the lexicon is comprised of a stack of feature bundles. This stack is simply an ordered list of one or more bundles of features. The composition of the feature bundles and the order in which they appear in a stack on a head is language-specific. In fact, the selection and organization of features into lexical items is, in this view, a principal locus of grammatical variation. Of particular interest in this chapter, the unique characteristics of the left periphery from language to language are attributed here not to phrase structure rules but instead to the featural composition of the clause-peripheral head. What are universal across languages are the mechanisms (Merge, Agree) by which these features interact, are valued, and are transferred in phases to the interfaces.

that the two approaches make substantially different empirical predictions. I claim here that we can retain the essential advances of the cartographic project with a smaller inventory of functional heads. Moreover, this is not a pursuit of simplicity purely for simplicity's sake, but instead an effort to account for the significant intra-language variation we see on the periphery while limiting this variation to the learnable lexicon of a language.


Let us now be more precise about what feature stacking is and how it might function. When a head His introduced into the derivation as in (41), the features of feature bundle F1 must be accessed in the derivation first, followed by those in F2 to Fn sequentially. (41)

HP

~ H [Ftl [Fzl [F3] So in a stack consisting of feature bundles Fl, F2, and F3, all features in F1 (a, b, c) will be valued before all features in F2 (d, e), which will be valued before all features in F3 (g)-1 (42)

If a feature bundle is made of interpretable and unintepretable features, it will interact with an available goal vial the Agree operation. If a feature bundle in addition contains the EPP property, the relevant Goal will undergo Move (Agree + Internal Merge) to successive specifiers of the head H. In principle, it makes no difference whether we assume that those specifiers attach successively further outward from the head (in the case that the feature bundles on the stack are accessed topdown) or "tuck-in" (Richards 2001) and are successively more proximate to the head (in the case that the feature bundles on the stack are accessed bottom-up). It only matters that the order of the feature stack will mirror the appropriate surface order among the specifiers. Here I will assume for illustrative purposes that the feature stack is accessed top-down (following Miiller 2010), and specifiers attach successively further outward, but with respect to the question we are interested in here, the choice between these alternatives is largely arbitrary. Below, I will follow convention in calling the single left-peripheral functional head C, and the phrase that it projects, CP. Let us turn to a more specific case, or the left periphery of a Kashmiri constituent question as in (43). (43)

Raj-an bmis haa-v nav kitaab? Raj-ERG who.DAT show.PST-FSG new book 'To whom did Raj show his new book?'

7· For another implementation of ordered Merge of multiple specifiers, see Lahne (2007); for a recent approach to ordered feature stacks on lexical items (heads) see Milller (2010).

35

36


Under the feature stacking view, the left periphery of this clause is comprised of a single CP projection. In (43), the C head must bear sets of features related to interrogative focus movement, the raising of the topic, and the raising of the second-position verb. The features controlling wh-movement and wh-expletive constructions in Kashmiri, and in fact the syntax of A-bar movement more generally, will be the primary concern of Chapters 3-4. Let us at this point choose a basic set of features to illustrate the proposal. Let us say that the features controlling interrogative focus movement are [Focus] and [Q], the feature controlling topicalization is simply [Topic], and that uninterpretable [Tense] is the feature triggering verb movement (Biberauer & Roberts 2005). These features must be organized into a sequence of sets (bundles), which is in turn associated with or constitutes the C head in the lexicon ofKashmiri. One bundle motivates wh-movement, a second topicalization, and a third verb-second.

(44)

CP

c [iQ, uFoc, EPPh [uTop, EPPh [uTenseb haw 'showed'

The bundle containing the uninterpretable [Focus], the interpretable [Q] feature and the EPP will be valued first, interacting with the wh-word kamis 'who' Due to the presence of the EPP in this bundle, kamis will Move into Spec, CP. The second bundle offeatures on the C head (which includes the uninterpretable [Topic] feature and another EPP) is thereby rendered accessible and triggers movement of a Topic-DP to another specifier of CP. The third feature, uninterpretable Tense, motivates the head raising of the verb into the C head, resulting in verb-second (or in this case, literallythird). 8 In this way, the entire left periphery of an interrogative clause in Kashmiri is contained within a single CP.

8. Lahne (2007) notes that this type of account more accurately reflects V2 patterns crosslinguistically, in that left peripheral material typically appears to the left of the fronted verb.


As far as Kashmiri is concerned, the C head in (44) is one of the C heads available in the lexicon of the language. The full range of possibilities is expressed in (45). 9

(45)

a.

C b. [(iQ), uFoc, EPP] [uTense]

c [iQ, uFoc, EPP] [uTop, EPP] [uTense]

(45a) represents a C head in a clause in which a sole focused constituent precedes the second-position verb, whether that focus is interrogative or not. The top bundle on the C in (45a) is the set offeatures attracting that focused constituent, and the second bundle, consisting only of uninterpretable Tense, attracts the verb for head movement. (45b) represents a C head in a clause like (43), in which an interrogative focus precedes the second position verb and a topic precedes the interrogative focus. These are the only manifestations of the C head in the Kashmiri lexicon. 10 Note that these are not at all dissimilar to the basic observations made in (24) or (40) about what combinations of constituents are typically found at the left edge. What is different about the statement in (45) is where it locates this variation - in the lexicon. 2.5.2

More on feature stacking

Feature bundling and stacking clearly represent an increase in technological complexity over alternative possible conceptions of the internal make-up of functional heads. We are required to view the features on a head not simply as a set, but as a list of sets. It is not clear, however, that this is an unwarranted increase. It may not go beyond the kind of complexity already observed in studies of lexical structure (Anderson 1977, 1982, Zwicky 1990). If it is in the lexicon that we locate

9· It is dear that the similarities between (45a) and (45b) are not accidental (in both, Focus is paired with EPP, and both heads bear uTense). A more articulated structuring of features could capture these co-occurrences- perhaps something like a feature geometric account (see Cowper 2005). While I will not elaborate on this further here, we may very well want to encode the notion that the appearance of some feature bundles is dependent on the presence of others. 10. In addition, there is the C heading relative clauses, which are unusual in Kashmiri in that they are verb-final (Wali & Koul 1997). I do not address these clauses in detail here, though see the conclusion to this book (Chapter 6), as well as Bhatt and Munshi (2009). For multiple wh -questions (addressed in Chapters 3-4 of this book), the C head would merely have multiple instances of the first bundle in the stack in (45a-b).

37

38


parametric variation, we will need to have systems for distributing and organizing features on lexical items. 11 It is already clear that certain features display a tendency to bundle together, and that certain features tend to bundle with the EPP property. For instance, the feature responsible for assigning nominative case to some accessible argument frequently also requires internal Merge (raising) of that argument, a property we have couched in terms of an EPP feature. On the other hand, the feature responsible for assigning accusative case to some argument often does not require internal Merge of that argument. The process of feature bundling within the lexicon is an acknowledgment of such observed tendencies. Recent proposals for more articulated feature structuring meet or exceed in complexity the stack or ordered list discussed here. Chomsky (2008) recognizes "multiple probes" within C, and suggests that perhaps only one functional head may be necessary to account for the left peripheral region. Cowper (2005) introduces a feature geometric account of the inflectional node that requires that entailment relations hold among sets of features. The feature tree produced by these entailment relations is language-specific, and is a property of the inflectional head constructed in the lexicon of a given language (see Footnote 4). The proposal here similarly asserts that when the lexicon of a given language is constructed, the appropriate features are organized onto the C head, but this view only requires that sets of features be ordered in a list. Greater structure and organization of features may also be needed on another phase-defining head, transitive v. Constituents purported to be located in the specifier of v include externally merged subjects, shifted objects, and wh-phrases (Rackowski & Richards 2005). These issues will be explored further in Chapter 4. If, in fact, each of these constituents must occupy the single specifier of some functional

u. As a reviewer points out, one criticism of the cartographic approach is that it requires phrase structure rules to enforce appropriate ordering of the hierarchy of functional projections. On the one hand, the feature stacking proposal escapes this criticism in moving away from a range of functional projections. On the other hand, we must still establish how the language-specific ordering of multiple constituents on the left edge of the clause is determined One approach (Georgi & Milller 2010, Milller 2010) has been to claim that although the stack of "'structure-building" features are ordered by the 9-grids of predicates, the stack of "probe" features is unordered (or the order is immaterial) and the linearization of the resulting multiple specifiers is determined by language-specific principles following the syntactic derivation. Another approach is to assume that there are language-specific principles of the ordering of features in the stacks on phase-defining functional heads that apply in the lexicoiL In either case, the burden for determining the language-specific linear order of multiple specifiers (which must be lodged somewhere) rests outside the narrow syntax, distinguishing these accounts from the cartographic approach.


head, and must occur in a fixed order, we will need to expand the vP phase in a manner similar to the split-CP. On the other hand, feature stacking could provide a means to attract any number of constituents to the vP phase edge while maintaining a single functional projection. 2.5-3

Regularity and idiosyncracy

The feature stacking approach suggests that parametric variation on the left periphery is located in the lexicon (more specifically, in properties of the closed class lexical items), and that it is the types of C head a language possesses that determine what constituents appear at the left edge in that language. At the same time, we must also address what principles or mechanisms account for the crosslinguistic regularities we see on the left periphery. In the cartographic approach, to the extent that we observe regularities in the order of constituents on the left edge, it is the hierarchy of projections in (25) that encodes these tendencies and regularities. However, in the feature-stacking account no such hierarchy exists. Instead, this information must be encoded in the operation that selects features from UG and creates lexical items in each language. For instance, this operation must possess a restriction that states that a C head cannot possess more than one feature bundle containing (non-interrogative) Focus features. This would be the analogue of the phrase structure restriction concerning focus that Rizzi ( 1997) places on the hierarchy in (25), and might be grounded in similar considerations. At first glance, the lexicon-forming operation may not seem to be the appropriate location for the encoding of this kind of crosslinguistic regularity; however it is likely that a number of similar crosslinguistic lexical patterns and tendencies already need to be encoded in such a wayP For instance, the observation that the introduction of an external argument is accompanied by the assignment of accusative case to the complement of a verb, usually called "Burzio's generalization" (Burzio 1986), has in recent years been instantiated by a single functional head in the syntax, transitive v. This head has something like (at least) the following clustering of properties: [externalS-role, case= uacc]. What determines that this particular information appears on a particular head in any number of languages? In the present account, this would be attributed to the way in which

u. Newmeyer (2004) presents a separate set of arguments that claim that some of the ordering restrictions on left peripheral constituents fall out from other independentlyneeded principles and need not be derived via a hierarchy of functional projections. While I won't explore these arguments in detail here, I want to point out that it may be the case that not all ordering restrictions on constituents in the left periphery necessarily result from variation in the lexicon either.

40


the lexicon-forming operation constructs this head, encoding this regularity in the process of lexicon building. That is, we would claim that there is a process that constructs the lexicon, and that this process tends to construct functional elements like the C-heads investigated here, as well as the transitive v described above. Though I will not pursue it in great detail here, the same hypothesis could be extended to help us understand the prevalence of other functional verbal heads, such as the applicative v [externalS-role (inherently oblique), no case=acc feature] (Pylkkiinen 2008). In other words, the observation that certain caseassigning and selectional properties appear associated with specific heads can be seen as a lexical encoding of a crosslinguistic tendency. Those lexical items that do not conform to these tendencies, though they might occur, should be expected to be both rare and marked. In sum, while the feature-stacking proposal does suggest that the featural content of a lexical item must have more internal structure than is implied by a simple set, the increase in technological complexity does not seem unwarranted. Other recent accounts of functional heads require even more structure, and it is clear that at each phase boundary multiple constituents of different types will need to be hosted. Further, we commit ourselves in this view to a lexicon-formation operation on which we must place certain universal constraints. However, this does not seem like an increase in theoretical machinery given that we already require a way to express strong crosslingustic tendencies in lexical structure. If we have an interest in locating the source of parametric variation in the lexicon, then this account simply represents a fine- tuning of our notion of the structuring of lexical objects. 2.5.4

An additional empirical question: The Kashmiri element ki

So far we have provided an account for the left periphery of matrix clauses in Kashmiri that captures the observations of the cartographic work in a single C head. In this section we turn to an additional empirical challenge for the proposal made here: Kashmiri subordinate clausesP Subordinate clauses in Kashmiri differ from matrix clauses only in that all of the above mentioned constituents can optionally be preceded by the element ki.

Lahne (2007) discusses what might be viewed as an additional empirical challenge for a proposal with a single functional head: how should morphological markers such as those attached to focused material be accommodated? Under the cartographic view, these markers occupy distinct left -peripheral heads. However Lahne argues that many such markers can be better analyzed not as heads, but as attached to the displaced constituent.

13.


(46)

bi oosu-s yi zaanaan ki seliim ga-v raath lSG AUX.PST-lMSG this know.IMPFV ki. Selim go.PST-3MS yesterday raj-as sit Raj-DAT with 'I knew that Selim went with Raj yesterday.'

(Wall & Koul: 48)

Notice that if we are committed to hosting the verb in verb-second clauses in a single C head, and the focus and topic constituents in its specifier, the element ki must be in some location at which it can precede all of these elements. One possibility under the Rizzi (1997) cartographic approach is to assume thatki is a Force particle. If this were the correct analysis, many of the difficulties that we identified above for the definition of phase-hood would re-emerge. Fortunately, however, there are a number of reasons to believe that the element ki plays no particular role in determining the force of the clause it precedes. First, ki is optional and is never required in a subordinating construction. In fact, there are instances in which ki must not appear, such as when a clause is preposed. (47)

raath raajas sit yi oosus Selim go.PST-3Ms yesterday Raj-DAT with this AUX.PST-lMSG

(*ki) selim gav

ki. bi

zaanaan

lSG know.IMPFV 'Selim went with Raj yesterday; this I knew.'

(Wali & Koul: 48)

Ifwe can assume that the first clause in (47) is in fact a preposed subordinate clause, we can observe that ki cannot appear when the clause it precedes is preposed. If

this is the case, it suggests that ki is not selected by verbs like zanaan 'know: If ki were selected by this verb, it should appear regardless of the ultimate location of the subordinate clause. Note that the facts in (32)-(33) are almost the mirrorimage of those which hold of English that (considered a typical Force head). 14 ( 48)

a.

I know that Selim went with Raj yesterday. That Selim went with Raj yesterday I know. c. *Selim went with Raj yesterday I know.

b.

Further clarification of the role of ki comes from embedded questions. This particle ki can appear preceding an interrogative complement, such as that of the verb prutS 'ask'.

14· Sobin (2002) asks if even the English that can always be understood as inhabiting the Force head, since in Middle English that can appear to the right of wh- material.

41

42


(49)

tam prutS me [ki mohn-an oos-aa bulvmts miiraa] 3sG.ERG ask.PST lsG ki Mohan-ERG be-Q invite-PERF Mira 'He asked me [if/whether Mohan invited Mira]' (Davison 2003)

The fact that ki can appear not only in declarative contexts, but also preceding an embedded question, suggests that it cannot be a marker of force. There are two imaginable approaches to its distribution. The first, and more conservative, would not be consistent with the single-C view of the left periphery of the clause that has been proposed here. This approach would locate ki in a position in the syntax, such as the head of some specifierless phrase (we could call it SubP). Of course, a number of questions arise, including whether this category has other members, and why the head does not seem to have semantic content. The head would, furthermore, have to be itself transparent to selection, in the sense that a governing predicate would have to ignore it and instead target some lower head (the locus of clause-typing information in our analysis). These are not insurmountable problems, but they would require additional stipulation. Further, an approach of this kind requiring multiple projections on the left edge would force us to revisit earlier questions about how phase-hood can be defined. A second, somewhat more radical approach, which is in line with the singleheaded view of the left periphery proposed here, would be to claim that ki is not present in the syntactic derivation, but is instead a morphological marker of the Force phase edge - one that is inserted following spell-out. That is, the reason ki does not seem to be selected by the verb, is transparent to selection, and appears to have no semantic content is that it is not actually present during the syntactic derivation. To describe this approach more formally; ki would be optionally inserted in each CP phase by the morphological component in the position between the V head and the material forming the edge of the immediately lower phase selected by V. 15 In this way, ki serves as an audible marker of the boundary between one CP phase and another. 16 This approach offers a way of understanding the facts in (47), repeated here. (47)

(..ki) selim gav raath raajas sit yi oosus Selim go.PST-3Ms yesterday Raj-DAT with this AUX.PST-lMSG ki

bl zaanaan lSG know.PRP 'Selim went with Raj yesterday; this I knew.'

(Wali & Koul: 48)

15. The optionality of insertion of a phase-edge marker could be language-specific. As a reviewer points out, in many environments in Hindi-Urdu the presence of ki appears to be obligatory.

For an alternative way of implementing this view using an OT-based account within the Distributed Morphology framework, see Lahne (2007).

16.


The ungrammaticality of ki in the structure in (33), and the contrast with English shown in (34), would be hard to understand in a view in which ki, like that, is a functional head high in the C-domain. In the proposal developed here, however, there is no similarity implied between ki and that. This places us in a better position to understand the contrast between (33) and (34). In the morphological account, ki would not be inserted by the morphological component in the position in (33) (or any other sentence-initial position for that matter) because it is not located between a V head and the material on the edge of some lower phaseP This approach to ki (and possibly other morphemes like it) deserves exploration in greater detail, but that is beyond the scope of this chapter. For now I will tentatively adopt this proposal. We will refer to this view in the discussion of Hindi-Urdu in Chapter 4. 2.5.5

Theoretical advantages

The feature stacking analysis presented here permits us to not only maintain the empirical ground covered by the cartographic view, but also to align the account of the left periphery with recent theoretical developments. This subsection returns to some of the developments mentioned in above and considers the feature-stacking approach in a cross-linguistic light. The feature-stacking account clearly takes advantage of the notion that a single projection can have more than one specifier. If the restriction that each head may have only one specifier can be abandoned, then we are free to assume that anumber of A-bar-moved constituents may be associated with a single head. This view

A reviewer has pointed out that the proposal that ki is inserted between the V head and the material forming the edge of the immediately lower phase it selects might be challenged by sentences in Hindi-Urdu in which a DP from the matrix clause appears at the right edge of the matrix clause, preceding the embedded CP.

17.

(i)

Tina-ne kah-aa Mona-se ki [ .... ] Tina-ERG say-PRF.MSG Mona-INSTR that .. . 'Tina said to Mona that'

In (i) ki appears to be non-adjacent to the verb. However, whether this is an instance of nonadjacency at the relevant point in the derivation depends on one's understanding of postverbal constituents in Hindi-Urdu. Under Bhatt and Dayal's (2007) remnant-VP approach the embedded CP in (i) is still adjacent to a trace of the verb. In Manetta's (to appear) approach, the embedded CP is right-aligned in the phonological component, and so there is still adjacency between V and the embedded CP at the point at which ki is inserted In an approach to postverbal constituents in Hindi-Urdu in which there was no adjacency between the V and embedded CP in (i), we could change the formulation to require ki to demarcate the left edge of the phase from material not in that phase.

44


not only eliminates a restriction on the grammar, but has a nice empirical result for Kashmiri as well. As mentioned above, in Kashmiri verb movement to any position that is not immediately following the focused wh or non-wh constituents is ungrammatical. Recall that in the cartographic approach there were a number of other heads available into which the verb could potentially raise, and so we required additional stipulation to ensure that the verb raised only as far as the Focus head. In the feature stacking view, there is only a single CP, and therefore only one candidate target position to which the verb could raise, C. The mechanism that ensures that the verb always immediately follows the focused element in Kashmiri is in fact the order of feature bundles in the stack. In Kashmiri, the first bundle to be valued on C must be the Focus bundle. In this way, the focusedDP or wh-phrase will be the constituent most proximate to the raised verb in the C head. This characteristic of Kashmiri is captured here as a feature of the Kashmiri lexicon. A second theoretical development discussed with reference to the cartographic analysis in Section 3 has to do with the concept of the phase. The phase provides a specific way of understanding the closed unit of the clause, and of particular relevance here, the transitional nature of the clause edge. In the cartographic approach, we determined that the definition of the phase edge would need to be tailored to include the hierarchy of left edge projections. On the other hand, in the case of the feature-stacking account presented here, establishing the phase-defining head and the phase edge is less problematic. The single C head is widely claimed to be phase-defining (Chomsky 2000, 2001, 2004). All of the specifiers of CP are unambiguously at the phase edge. This achieves the expected results in the case of successive cyclic wh-movement and topicalization in Kashmiri, without the need to alter the definitions of phase or phase-edge to accommodate arrays of projections. In this way, the feature-stacking approach achieves the expected empirical results while at the same time incorporating smoothly a useful theoretical development. This will become particularly important as we turn to an account of successivecyclic wh-movement and wh-expletive constructions in Chapters 3-4 and sluicing in Chapter 5. In this chapter we have developed an empirically advantageous way of accounting for the syntax of the clause-edge in Kashmiri using a single C head. However, the question remains, is it the case that the left periphery of every language is comprised of a single C head? Could some languages require an exploded C-domain, with many distinct heads? The answers to these questions could follow one of two possible paths. The first would claim that languages vary parametrically as to whether a single head or an exploded array of projections appears on the left edge. Such a view has been proposed for the inflectional layer by Bobaljik and Thrainsson (1998). In this


view, some languages would accommodate left-edge material in a single head with multiple specifiers. All the necessary features would be bundled and stacked on that head. Other languages might require that each feature bundle occupy a distinct head, and that this hierarchy of heads host left-edge material. A second, bolder, and potentially more interesting route, is to claim that only a single functional head, C, is available on the left periphery of any language. Though the order and composition of feature bundles on this head may vary, there is no language that displays an array of distinct functional heads appearing on the left edge (in fact, these distinct heads are not available to any lexicon). Importantly, this is the view that articulates most clearly with our understanding of the phase, as this chapter has demonstrated. A challenge to this bolder view are cases in which there appear to be multiply filled overt C heads on the left edge of complement clause, such as in Scandinavian languages permitting a complementizer and second position verb on the left edge (Iatridou & Kroch 1992) or in instances of multiple complementizers separated by adjuncts as in English (McCloskey 2005). Because in these cases it appears to be the same type of head that is duplicated at the left edge, they have been analyzed as instances of recursive CPs, in which a C takes a CP complement. In the cartographic literature, CP-recursion proposals have typically been understood in terms of the articulated C-domain. However, in the framework of the present discussion, the older approach now seems more appropriate. We might then ask how phasehood could be defined in these recursive structures. The most natural assumption would be one in which both CPs are phases. As far as I know, the evidence is at least consistent with this view. It does not seem that these cases in and of themselves would necessitate a parametric view of expansion on the left edge of the clause. 18 The feature-stacking approach is an attempt to restrict variation of this sort to the learnable functional lexicon of a language. That is, the order, number, and nature of elements found on the left edge of the clause is dictated by the presence and structuring of features on the functional head C. All the information the grammar requires to attract the appropriate constituents to the left edge is contained within the lexicon, in this view. In the syntax, the left periphery is comprised of a single CP, which selects a TP of the appropriate kind. As this selection is universal, there is no need for any phrase structure rules to determine the order of projections in a clause. That is, the attributes of the left-periphery in a

18. The same issues may also arise in the v-domain, in proposals in which a v takes a vP complement (e.g. Pylkannen 2008, Folli & Harley 2004, Harley 2008).

45

46


specific language reduce to the presence of certain features in the lexicon, which increasingly looks like the primary locus of parametric variation in the syntax. The feature stacking approach may have ramifications beyond accounting for the left periphery. Chomsky (2008) has suggested that properties of a language not only reduce to the properties of the functional heads, but in tact specifically to properties of the phase-defining heads. Recent research indicates that the phasedefining head C and the phase-defining head v may have a number of characteristics in common. In the case of successive-cyclic movement, constituents must move to the edge of each phase in order to be accessible to a higher phase. Work by Rackowski and Richards (2005) proposes that vP is the position at which the wh-criterion is satisfied in Tagalog. In Chapter 4, I will argue that we can even find wh-expletives at the vP phase edge (in Hindi-Urdu), just as we find them in the CP domain. If these investigations are on the right track, we might expect to find a similar constellation of constituents appearing at the vP-phase edge as at the CP phase edge. The feature-stacking approach may then help us to account not just for the left periphery, but also for the range of constituents appearing at the edge of vP. In this way, the technology introduced in this chapter can serve to clarify the source of parallelism and variation in these domains.

CHAPTER3

Full and partial wh-movement in Kashmiri

3.1

Introduction

Kashmiri exhibits both full and partial wh-movement as question formation strategies in sentences with multiple clauses. (1)

tse kam.' chu-y baasaan ki mohn-as di-ts kitaab 2SG.DAT who.ERG AUX-DAT think that Mohan-DAT give-PST.FSG book 'Who do you think gave Mohan the book?' (PK 9/21/04)

(2)

tse k'aa chu-y baasaan ki mohn-as kam.' 2SG.DAT EXPL AUX-DAT think that Mohan-DAT who.ERG di-ts kitaab give-PST.FSG book 'Who do you think gave Mohan the book?' (Wall & Koul1997: 18)

The question word in the subordinate clause receives a matrix scope interpretation in both (1) and (2). In (1) this interpretation is achieved by movement of the question word kam' 'who' into the matrix clause. In (2) this is achieved via the presence of a minimal question word k'aa. ('what') in the scope position in the matrix clause, while the contentful question word kam' remains in the subordinate clause. The clausal periphery in Kashmiri plays a crucial role in mediating long-distance wh-dependencies. As discussed in Chapter 2, it is the periphery that must host displaced wh-phrases, wh-expletives, and partially moved wh-phrases. Insofar as we understand wh-movement to proceed successive-cyclically, it is through the periphery that wh-material must pass when displaced into a higher clause. And it is the specific properties of the peripheral. or phase-defining, head that determine how these processes proceed. This chapter investigates the minimal question word k'aa 'what' in Kashmiri, analyzing it as a wh-expletive comparable to the nominal expletives in the A-movement system (following in the spirit of recent approaches by Simpson (2000) and Fanselow and Mahajan (2000)). If an A-movement system expletive is a semantically null element that serves a specific syntactic function, we must ask what the corresponding function of an expletive might be in the A-bar movement system. This leads us into a deeper exploration

48


of the structure of the periphery, and the way in which the periphery serves as a potential site for expletives. The issues ultimately at stake here are large - in particular the question of whether the syntax of the left periphery (the A-bar movement system) is governed by the same organizing principles as clause-internal syntax (the A-movement system). If the proposals developed here are on the right track, then the two systems emerge as being completely parallel - in a way that does not emerge so clearly from the more closely studied languages. Previous approaches to these constructions generally fall into two classes. So-called 'direct dependency analyses (McDaniel1989, Rizzi 1992, Mahajan 1990, inter alia) contend that a direct syntactic connection is formed between the meaningless question word and the contentful question word, mediated by chains and conditions on chain formation. The second class of approaches, called 'indirect dependency accounts (Kiss 1987, Dayal1994, 1996, Lahiri 2002), denies that any such syntactic connection exists. Instead they claim that the meaningless question word is coindexed with or replaced by the clause containing the contentful question word at the level of Logical Form (LF), and this is how the correct interpretation is achieved. The account proposed here falls squarely into neither of these classes. It does build on important properties of each, maintaining that the meaningless question word is base-generated in a clause-internal (not clause-peripheral) position as in the indirect dependency approaches, but also that the role played by the meaningless question word itself is entirely a syntactic one, as in direct dependency analyses. However, the essential point of contrast with these two camps is the claim developed in this chapter that there is no connection at all between the meaningless question word and the contentful question word in a lower clause, whether syntactic or at some level of interpretation. Instead, the meaningless question word in the A-bar movement system will be analyzed as a wh-expletive, satisfying the property of a peripheral head that requires phonologically overt material to appear in its specifier. Section 3.2 of this chapter details important features of Kashmiri A-bar syntax, particularly with respect to the way in which questions are formed. This section relies on the account of the left periphery of the Kashmiri clause proposed in Chapter 2 above, and establishes a working view of the internal phrase structure of the Kashmiri clause. Section 3.3 provides an account of wh -movement and wh-expletive constructions in Kashmiri and compares the reach of this account with previous approaches to wh-expletive phenomena. In Section 3.4, I discuss how this account can be extended to two more empirical puzzles, one specific to Kashmiri and one crosslinguistic in nature. Section 3.5 concludes the chapter.

Chapter 3. Full and partial wh- movement in Kashmiri 49

Kashmiri question formation and the structure of the clause

3.2 3.2.1

The Kashmiri question

Recall that Kashmiri is unusual among the Indic languages in exhibiting the verb-second (V2) property; more familiar from Germanic and the older Romance languages. The unmarked word order of a Kashmiri tensed root clause is: (3)

subject - finite verb - indirect object- direct object

The finite verb must be the second constituent in the clause, but any of the arguments (or other constituents) may appear first. The order of the postverbal elements is also fairly free, though the subject must immediately follow the second-position verb if the sentence is not subject-initial. In constituent questions, the question word must appear before the verb in addition to some other constituent. Only if the question word is the subject may it naturally appear alone in sentence-initial position. (4)

a.

b.

Raj-an kamis haa-v nav kitaab? Raj-ERG who.DAT show.PST-FSG new book 'To whom did Raj show his new book?'

(Wali & Koul: 12)

shill-as haa-v kam' nav kitaab raath who.ERG show.PsT-FSG Sheila-DAT new book yesterday 'Who showed a new book to Sheila yesterday?' (Wali & Koul: 12)

Kashmiri has two strategies for forming constituent questions with more than one wh-phrase. In the first, all wh-phrases are moved to the preverbal position. In the second, only one wh-phrase is moved, and the remaining wh-phrases are found in-situ within the clause. (5)

(6)

kam' di-ts kamis kitaab? who.ERG who.DAT give-PST.FSG book 'Who gave the book to whom?'

(Wali & Koul1997:26)

kam' di-ts kam-is kitaab who.ERG give-PST.FSG who-DAT book 'Who gave the book to whom?'

(Wali & Koul1997:27)

Subordinate clauses are identical to matrix clauses in their word order, except that they are optionally preceded by the particle ki. This particle is not counted in determining verb-second position, and will be considered here to be inserted as the marker of the CP phase boundary; as discussed in Chapter 2. A typical indirect question is in (7).

50


(7)

miiraayi ch-a pataa ki kam-is di-ts mohn-an kitaab. AUX-3SG know that who-DAT gi.ve-PST.FSG Mahan-ERG book Mira 'Mira knows who Mohan gave a book to: (Wall 2002)

Multiple wh-phrases can appear in indirect questions as well, just as they can in matrix clauses. (8)

Me ch-u-nl pataa kus learn-is oos 1SG AUX-1SG-NEG knowwho.NOM who-DAT AUX.PST.3SG mee.lmi gatsh-aan going-IMPERF meet

'I don't know who was going to meet whom:

(Wa.li & Kou.l1997: 27)

The focus of this chapter shall be constructions that permit matrix scope interpretations of question words originating in subordinate clauses. Bridge verbs permit just such a construction, in which an invariant wh-expletive k'aa, appears in the pre-verbal position in the matrix clause. I will call this the whexpletive or partial movement construction throughout. The specific behavior of non-bridge verbs with respect to these constructions will be explored further below. (9)

tse k'aa chu-y baasaan ki mohn-as kam' 2SG.DAT EXPL AUX-DAT think that Mohan-DAT who.ERG di-ts kitaab gi.ve-PST.FSG book 'Who do you think gave Mohan the book?'

(Wall & Koul: 18)

Direct questions formed by full wh-extraction from the clausal complements of these non-factive verbs are also possible, as in (10). (10)

a.

ditsmits' baasaan ki raaj-an aasi tse k'aazi chu-y 2SG.DAT why AUX-DAT think that Raj-ERG AUX.FUT gave-PERF mohn-as kitaab Mohan-DAT book 'Why do you think that Raj gave the book to Mohan?' (Wall & Koul: 19)

b.

3.2.2

Aslaam k'aa ch-i yatshaan ki su gotsh anun that 3MSG should bring Aslaam what AUX-MSG want 'What does Aslaam want that he should bring?' (Wall 2002)

Assumptions about the structure of the Kashmiri clause

The facts of Kashmiri could be construed to support an approach in which the unmarked word order in Kashmiri [subject-verb-object] is derived by

Chapter 3. Full and partial wh- movement in Kashmiri

movement of the verb to T and the subject to the specifier of TP. Bhatt (1999) reports that the preverbal position is generally the focus position, and that elements in this position always bear focal stress, with the important exception of subjects. 1 This would indicate that subjects may not necessarily be in the specifier of some head containing focus features, but can instead be in the specifier of a head unmarked with respect to focus. Putting these observations together, we are led to assume that in subject-initial declarative clauses, the finite verb raises only as far as T. Unfocused subjects then appear in Spec, TP. In the case of structures with focus and topic constituents, a C bearing these features attracts the finite verbs to raise further. This split view of verb-second is similar to that offered in Zwart (1997) (and earlier in Travis 1991). According to Zwart, in Dutch (as in other Germanic languages) subject-initial main clauses do not involve movement to CP. as has often been assumed. Zwart also offers evidence against generalized V-to-C movement in Dutch, demonstrating that there is no clear motivation for assuming that the verb has always raised to C. From these considerations, Zwart concludes that the subject and verb in subject-initial main clauses in Dutch are not located in CP, but instead in AgrS, the highest head of the inflectional layer. He asserts that the only way a subject can move into the specifier of CP is if it is attracted by some feature of CP (i.e. a wh-feature) beyond that which involves the normal interaction with T. Only initial constituents that are wh-words or non-subjects are analyzed by Zwart as raising to a CP or higher phrase. It seems that this approach can also be applied to the verb -second facts of Kashmiri, as we have seen above. 2 The proposal I have developed extensively in Chapter 2 and the suggestions I have made here for verb-second in Kashmiri differ significantly from the account offered in Bhatt ( 1999). Bhatt claims that verb -second crosslinguistically is the result of movement of the verb and some sentential constituent to a functional projection MoodP (MP). MP is a component of an articulated CP - a phrase that he asserts is universal across languages whether or not they possess explicit morphological mood markers (Kashmiri does not). Bhatt directly addresses Zwart's (1997) claims concerning Dutch V2, arguing against the approach primarily on theory-internal grounds. It seems that most or all of his concerns disappear when the proposal is updated along the lines of the Minimalist Program as it is presented in Chomsky (2000). The empirical argument he makes against Zwart's proposal involves

1. Bhatt also mentions that temporal adverbs do not appear stressed in this position. Why this is so is beyond the scope of this discussion. 2.

Fora new argwnent for the split view ofV2 for Danish see Mikkelsen (2009).

51

52


subordinate clauses and complementizers in Swedish, and I will not address these issues here. The only mention of Kashmiri in Bhatt's discussion of Zwart's proposal is that under the view that only in subject-initial sentences is the verb not in CP, it would be difficult to explain that sentence-initial temporal adverbs in Kashmiri do not receive sentential (i.e. focal) stress. If there are other reasons why such adverbs may not receive sentential focus intonation, we can move forward in explaining why subjects behave differently from all other sentence-initial constituents- because (as in many languages) they may remain in Spec, TP. These issues do not bear on my central goals in this chapter, and so I will proceed on the assumption that something like Zwart's approach is correct for Kashmiri. This chapter and this book are primarily concerned with the left periphery of Kashmiri, and in particular the CP layer. However it is necessary at this point to establish what our basic assumptions will be about the internal structure of the Kashmiri clause. Bhatt (1999) claims that the placement of the verb in Kashmiri indicates that lexical projections in Kashmiri are head-final, and functional projections are head-initial. Overall, Kashmiri exhibits the properties of a head-final language. For instance, adpositions appear following their complement. (11)

... baag-as manz ... garden-we in '... in the garden

(Wall & Koul: 45)

The verb appears in clause final position in non-finite and relative clauses, and when the tensed auxiliary appears in second position the main verb is still clause final (as in ( 13)). (12)

[yi bl khyvaan chus] su chu-y-aa tse khosh karaan REL lSG eat-PRP AUX DEM AUX-DAT-Q 2SG.DAT like do 'Do you like what 1 eat?' (Bhatt 1999)

(13)

lark-as kitaab divaan bl chu-s 1SG AUX.PRS-1SG boy-DAT book give-PRP 'I give a book to the boy:

(Bhatt 1999)

In general, the inflected verb appears in second position in all tensed clauses, rna trix and subordinate, as discussed extensively in Chapter 2. This means that the functional projection at the edge of the clause (CP) must have its head on the left to arrive at the grammatical verb-second word ordering (otherwise, raising to C would be string-vacuous). Following Bhatt and based on these facts, I will assume the following structure in (14) for basic Kashmiri clauses. The external argument, complement to the verb, and the verb in (14) are in the positions into which they first merge in the structure. In the course of the


(14)

CP

~ C TP

~ T vP Sub~

VP

v

~

Object

V Verb

derivation of a simple subject-initial declarative, the external argument would raise to Spec, TP, and the verb to T. If any other constituent is initial, it will merge into Spec, CP and the verb will raise to C. In addition to the above, I will assume, following Bhatt (1999) and Bayer's (1996) claims for Bengali and Hindi-Urdu, that subordinate clauses in Kashmiri are complements of the verb that exceptionally appear on the right. This will be discussed further for the case of Hindi-Urdu in Chapter4.

3·3

Analyzing full and partial wh-movement in Kashmiri

With an understanding of Kashmiri phrase structure in place, based on the account presented in Chapter 2 and above in this chapter, we can now take a closer look at wh-movement in Kashmiri, in particular full and partial movement from subordinate clauses. In Section 3.3.1 I will present in detail a new approach to A-bar movement that accounts for the facts in Kashmiri. In 3.3.2, I contrast this account with previous approaches to partial movement constructions, indicating theoretical and empirical differences. Finally, Section 3.3.3 suggests how the interpretation of partial movement constructions might proceed under the syntactic account proposed here. 3.3.1

A new account of A-bar movement

Recall the basic data under discussion. Kashmiri exhibits both full and partial wh-movement (wh-expletive constructions) as question formation strategies in sentences with multiple clauses.

53

54


(1)

tse kam' chu-y baasaan ki mohn-as di-ts kitaab 2SG.DAT who. ERG AUX-DAT think that Mohan-DAT give-PST.PSG book 'Who do you think gave Mohan the book?' (PK 9/21/04)

(2)

k:'aa chu-y baasaan ki mohn-as kam.' tse 2sG.DAT EXPL AUX-DAT think that Mohan-DAT who.ERG di-ts kitaab give-PST.PSG book 'Who do you think gave Mohan the book?'

(Wall & Koul: 18)

The current theoretical framework provides a feature-based understanding of A-movement. In this chapter I will consider approaching A-bar movement in the same way. Following the spirit of the recent work of Simpson (2000) and Fanselow and Mahajan (2000), I argue that the analysis of these two systems can be unified, using Kashmiri as a test case for this hypothesis. I will show that both full extraction and partial wh-movement of question words in Kashmiri can be analyzed using a system of interpretable and uninterpretable features in a manner similar to the approach to the A-system. The distinction between full extraction from subordinate clauses and partial wh-movement in Kashmiri can be analyzed as the distinction between the operations Move and static Agree to satisfy uninterpretable features. In the feature-based approach to A-movement, a nominal enters into an agreement relation with a higher accessible head. This Agree operation is simply a mutual exchange of information between a head and a nominal bearing the relevant features -an exchange of information which takes place in a particular structural configuration as follows:

(15)

H ..... a .... p a.

H commands a, which in turn commands p

b.

a is 'closer to' H than p

'Closer to' is defined in terms of asymmetric c-command. That is, a is 'closer to' H than~ iffH commands a, a commands~. and~ does not command a (a commands ~ iff~ is contained within the sister of a). (Adger 2003). In this configuration, it will be possible for the features of the head (the Probe) and those of the nominal (the Goal) to mutually value one another. If the required relation is not established, features remain unvalued, and the derivation will not result in a well formed syntactic object Once all of the features of an element are valued, the element is inactive, and its participation in head-nominal interactions will be limited. Within the A-system, a nominal may Agree with a higher head, and may also raise if the probing head has the EPP property. This composite operation is called


Move, and will be defined here as Agree+ pied-piping+ re-Merge. 3 If the EPP is not present, the uninterpretable features of the Probe and Goal may of course be valued by static Agree over a distance, as described above (Chomsky 2000). I also assume (following Chomsky 2004) that all and only uninterpretable features have unvalued instances when they enter the derivation. The operations described above are constrained by locality considerations. Agree (and hence the composite operation built upon it, Move) can only take place within a phase. Phases are self-contained subparts of a derivation, each beginning with a numeration and ending with transfer of the objects created to the interfaces. A Probe can only interact with Goals within its own phase or on the edge of the previous phase. Elements that are not within the current phase (that are contained in previously constructed phases) are not available. For the purposes of this chapter, I will focus on the CP phase only, though Chapter 4 will introduce the relevance of the v-phase for wh-movement (Rackowski & Richards 2005). Chapter 2 has set the stage for our understanding of the rich left periphery of the clause as a single C-head with multiple specifiers, and this understanding will become crucial here. Our theory of the CP domain provides us with a reasonable way of defining phasehood. We must elaborate the way in which CP phases mediate wh-movement, and in particular wh-expletive constructions. The claim that the specifier of CP is an expletive-hosting position has much to tell us about the nature of the C head, and the phase-defining heads more broadly. In this chapter I will claim that the A-bar system ofKashmiri functions identicallyto the A system in the following respects. Heads and wh-phrases will possess interpretable and uninterpretable features. If a higher probe possesses the EPP in addition to other features, an accessible wh-phrase will undergo Move. Alternatively, if the wh-expletive k'aa is in the numeration, the merging of k'aa can satisfy the EPP on the Probe, much like an expletive in the A-system. If this occurs, the uninterpretable features on the A-bar probe will be valued by interacting with an accessible wh-phrase via static Agree over some distance. There are three features at work controlling movement and agreement in the A-bar system: the EPP (common to the A and A-bar systems), the [Q] feature, and the [wh] feature. The [wh] feature is interpretable on wh-phrases and uninterpretable on all heads, activating probes that interact with wh-phrases. The interpretable feature on wh-phrases is its "wh -hood"; that which triggers the interpretation of the wh-phrase as a Reinhart (1998)-style choice function variable. The feature [Q] is uninterpretable on the wh-phrase but present and interpretable

3· The first merge is when element is merged into the 'workspace' from the numeratioiL

55

56


on the highest head in an A-bar movement sequence. This interpretable [Q] feature marks the position at which a wh-phrase will be interpreted. Like the category-defining features, it allows the hosting head to enter into selectional relations, and is interpreted as an unselective binder of(choice) function variables (Reinhart 1998). The role of these interpretable features will be further explored in Section 3.3. In a sentence comprised of a single clause in Kashmiri, a wh-phrase must always move to the specifier of the C-head containing the second-position verb. Wh-in situ is impossible except in instances of multiple wh-phrases (at least one wh-phrase must raise). This means that the C-head which contains the second position verb will necessarily possess an uninterpretable [wh] feature and the EPP in all interrogative clauses, both matrix and subordinate. It is the uninterpretable [wh] feature that makes the Can active Probe, and it is the EPP which requires that the wh-phrase in its domain not only Agree but also Move into its specifier. As in any account of wh-movement, we must encode the observation that the EPP in wh-questions Kashmiri must be satisfied by wh-material. One way of doing this is to assume thatthe EPP is the property of a head that it have an additional specifier beyond its selectional requirements (Chomsky 2000, 2001, Hiraiwa 2001), and that the EPP is designated such that it can only be satisfied by certain kinds of moved goals. The EPP in this case is EPP Q' meaning it can only be satisfied by wh-material, since all wh-elements bear a interrogative [Q] feature (see Ura (2000) for the analogous EPP for DP arguments). I will refer to it simply as EPP throughout. 4 Within an interrogative sentence comprised of a single clause, as in (16), the C-head will also possess the interpretable [Q] feature, as in (17). (16)

Raj-an kamis haa-v nav kitaab? Raj-ERG who.DAT show-PST.FSG new book 'To whom did Raj show his new book?'

4· An alternative view of the EP P is that it is not a property of a head, but instead a property of a feature of a head, or a "subfeature" (Pesetsky & Torrego 200 1). In the account presented here it would not be possible to claim that the EPP is a subfeature of a single specific feature in the wh-system since multiple features drive wh-dependencies. However, Pesetsky and Torrego (2007) propose a radically revised system of feature sharing in which valuation and interpretability of features are independent, meaning that each feature has four guises (combinations of valued/unvalued and interpretable/uninterpretable). The increase in complexity of the feature itself might permit a single-featured account of wh -dependencies, which in turn would allow us to assume the EPP to be a subfeature of a single feature. As it is unclear whether this alternative is empirically distinguishable from the view of the EPP adopted here, I leave this question to future research.


(17)

c [iQ] [uwh]

EPP The presence of the interpretable [Q] feature signals the position at which the binder of the wh-phrase variable will be introduced. This feature will value the uninterpretable [Q] feature on the wh-phrase, and will both arrest the movement of the wh-phrase and allow the sentence to be a well-formed syntactic object (with no unvalued uninterpretable features). In addition, the scope of the wh-phrase is determined at the position of the interpretable [Q] feature, in the sense that this is the position from which the wh-indefinite is unselectively bound. In this single clause, the scope of the wh-phrase is determined at the position at which the whphrase itself is ultimately located, however we will see that this need not always be the case. The goal of this section is ultimately to describe extraction and partial whmovement in Kashmiri subordinate clauses, so let us now consider a clause that is embedded. In this scenario, the wh-phrase may not remain in the lower clause, but instead must raise all the way to matrix scope position, as in (1), repeated here. (1)

tse learn' chu-y baasaan ki mohn-as 2sG.DAT who.ERG AUX-DAT think that Mohan-DAT di-ts kitaab give-PST.FSG book 'Who do you think gave Mohan the book?'

(PK 9/21/04)

In a sentence like (1), the subordinate C-head will lack an interpretable [Q] feature (the embedding verb basaan does not select a question). A wh-phrase that has raised to the specifier of this CP will still have uninterpretable features that require valuing because it bears a [Q] feature which is not interpretable and which has no matching feature on the embedded C head. The wh-phrase in the specifier of the subordinate CP will be in the same phase as the matrix C-head. The matrix C-head will have an uninterpretable [wh] feature, the EPP, and the interpretable [Q] feature, just as in (17). As a Probe, it will find the wh-phrase in the specifier of the subordinate CP and will enter into an Agree relation with this wh-phrase and attract it to its specifier (it will undergo Move). The wh-phrase will raise to the specifier of the matrix CP, and the result will be full extraction. The particle ki is inserted post-syntactically. as discussed in Chapter 2 and indicated by an arrow. This process and the features involved are diagrammed in (18). (18) represents the extraction of the wh-phrase and its passage through the specifier of the embedded CP into the matrix CP. At this point all uninterpretable

57

58


(18)

CP

ts~ ka;;; ~

'yo:'

'who'

t 1

[

C u;Qh]

EPP chu-y 'aux'

TP

~

+ 1

ki

CP

Moha~ ~•

;i::

C [u.wh]

EPP :: dits I-----------------------------':1: 'gave'

TP

~ 1

L-----------------~

features on the wh-phrase are valued and it is rendered inactive, or frozen. Again, it so happens that in this derivation the position to which the wh-phrase is ultimately displaced is also its scope position. An obvious question arises at this point. If static Agree (that is, Agree without Move) is one of the operations available to the derivation, why can't the uninterpretable feature of the wh-phrase be valued by the matrix C-head while it remains in the specifier of subordinate CP? In fact, it can. However, the EPP on the matrix C-head must be satisfied by some [wh]-bearing element. In this particular case, no wh-expletive happens to be in the numeration, and so there is no alternative way for the EPP on the matrix C-head to be satisfied; the derivation would therefore crash. Let us now turn to the other strategy by which wh -phrases originating in subordinate clauses take matrix scope in Kashmiri the partial wh-movement or whexpletive construction as in (2). (2)

tse k'aa chu-y baasaan ki mohn-as kam' 2SG.DAT EXPL AUX-DAT think that Mohan-DAT who.ERG di-ts kitaab give-PST.FSG book 'Who do you think gave Mohan the book?'

(Wall & Koul: 18)

In a partial movement construction, the subordinate C-head will once again have only an uninterpreatable [wh] feature and the [EPP]. The numeration happens to contain a wh-expletive k'aa, which can be merged to satisfy the EPP on the matrix C Probe. This expletive differs from a full wh-phrase in its feature content.


As an expletive, it consists entirely of uninterpretable features, and contributes to the syntactic computation only an (uninterpretable) interrogative feature (written uQ here). This understanding of the wh-expletive has two consequences: (i) the expletive may appear only in questions, (ii) since it lacks a wh-teature altogether, it cannot render inactive the probe in whose specifier it is merged. As a result, the matrix C-head will still need to value its uninterpretable [wh] feature, and will therefore pro be its domain. It finds the wh-phrase in the specifier of the subordinate CP, and will enter into static Agree with it. In this way, all uninterpretable features are valued. This theoretical view takes wh-expletives to be the A-bar counterparts to expletives of the A-system. That is, they do not contribute any interpretable features to the derivation, but have only uninterpretable features. This means that they can only serve to satisfy the EPP and permit a head's features to be valued statically by some other accessible element (Simpson 2000). This process is shown in (19). (19)

CP

ts~

'yo~'

k';;

'expl'

~

C

[ ~oh] EPP chu-y 'aux'

TP

~

~ ki

CP

Moha~ n-aL~ ~ 'who' C t luwh]

: LEPP :

!

TP

~

dits 'gave'

The analysis presented here accounts for both full wh-movement in Kashmiri as in (1) and partial wh-movement as in (2), and provides an understanding of how C-heads and wh-phrases interact in the course of forming long-distance whdependencies. I have proposed that the two distinct strategies for construing matrix scope for embedded wh-phrases are driven by identical mechanisms- that is, the features involved in the derivation of each strategy are precisely the same. The wh-movement and wh-expletive constructions, in this view, differ only in their numerations. If a wh-expletive happens to be present in the numeration, it will be merged into the matrix Spec, CP, allowing the features of the lower wh-phrase to be valued via static agreement over a distance. If the expletive is not present, the

59

6o


wh-phrase will itself raise to value the features on the matrix C-head. Either way, it will be the wh-phrase from the embedded clause that will value these features. A number of questions concerning the properties and distribution of wh -expletives still remain to be answered, however, and these are addressed in the following sections. 3.3.2

Restrictions on wh -expletives

Consider the ungrammaticality of the Kashmiri clause in (20a), or its equivalent in German, (20b). (20)

a. *raj-an k'aa haa-v kam-is nav kitaab? Raj-ERG EXPL show-PST.FSG who-DAT new book Intended: 'To whom did Raj show his new book?' b. *Was glaubst du was? EXPL believe 2SG What? Intended: -what do you believe?'

In such cases, a wh-expletive is in the same clause as the full wh-phrase whose position of interpretation it is meant to indicate. It~ as we have explained above, an expletive can be merged into Spec, CP to satisfy the EPP and the uninterpretable [wh] feature on the C-head can be valued via static Agree with a wh-phrase in its phase, there should be no problem with (20a-b ). It seems that wh-expletives, unlike DP-expletives, are constrained by a kind of anti-locality. Compare the ungrammaticality of the wh-expletive constructions in (20) with the English DPexpletive there, in (21), which can appear in the same clause as its DP associate. (21)

There are three unicorns in the garden.

Simpson (2000), along with Horvath (1997) and Fanselow and Mahajan (2000), claims that the anti-locality property of wh-expletives can actually be reduced to a question of case. If the wh-expletive in fact needs case just like any other wh-phrase, it must actually be merged into a case position in a sentence like (20a). However, this is not possible, because the full wh-phrase has occupied the relevant position and received this case. 5 The wh-expletive's need for case will block instances like (20a), in which the case is instead being assigned to the full wh-phrase bmis 'whO, but permit (2) in which there is no competitor for case in the matrix clause.

5· Note that the need for DP case would not alter the status of k' aa or was as a wh -expletive. It is still defective in the sense that it cannot value the uninterpretable features on the C-head to whose specifier it raises. This will force the uninterpretable [wh] feature on the C-head to seek another wh-phrase with which to interact. In addition, case features are uninterpretable, and so that part of the understanding of wh -expletives also remains constant.


(2)

tse k:'aa chu-y baasaan ki mohn-as kam.' 2SG.DAT EXPL AUX-DAT think that Mohan-DAT who.ERG di-ts kitaab give-PST.FSG book 'Who do you think gave Mohan the book?'

(Wall & Koul: 18)

Further empirical support for this view comes from Kashmiri and the related language Hindi-Urdu. In addition to k'aa, Kashmiri has a pleonastic element yi that can be optionally inserted into a matrix clause. (22)

b1 oos-us yi zaanaan ki seliim ga-v lSG AUX.PST-lMSG EXPL know.PRP ki Selim go.PST-3MS raath raj -as sit yesterday Raj-DAT with 'I knew that Selim went with Raj yesterday:

(Wall & Koul: 48)

A similar construction exists in Hindi-Urdu, as shown in (23). (23)

Sita yeh jaanti hai ki Raj kis-se Sita EXPL know-HAB.F AUX.PRS that Raj who-INST [Hindi- Urdu]

baat kare-gaa talk AUX.FUT-MSG 'Sita knows who Raj will talk

to:

In both Hindi-Urdu and Kashmiri the expletive object (yi or yeh) cannot co exist with a wh-expletive, suggesting that they occupy the same case position in the clause. (24)

[Hindi- Urdu] *Sita-ne yeh kyaa soc-aa ki ravii-ne Sita-ERG EXPL (WH)EXPL thought-PRF.M that ravi-ERG kis-k.o dekh-aa? who-Ace saw-PRF.M Intended: 'Who did Sita think that Ravi saw?'

Now let us turn to the position at which the wh-expletive is first introduced in the clause. Simpson (2000) claims that wh-expletives are base generated in the specifier of the agreement projection AgrO, and when other DPs are present which need to check case in this position, it will not be possible to generate the whexpletive. This would rule out examples like (20) and (24), in which another wh-DP or the expletive object yeh must occupy this position. Updating this view to reflect the framework of this chapter, we need to ensure that the wh-expletive is base generated in a position at which it can have accusative case valued; that is, within the c-command domain of the accusative case licenser transitive v. Further,

61

62


we know that the wh-expletive is not a semantic argument of any predicate (having no interpretable features) and cannot be introduced by semantic selection. For this reason, the wh-expletive must be merged into the specifier of a head that has the EPP property. The EPP is a quasi-selectional feature that causes the head to seek an additional specifier beyond those it needs to satisfy its core selectional requirements. Aspect is a functional head in the c-command domain of v on which it would be reasonable to posit the EPP property (Travis 1991, 2010). AspectP is the projection claimed to introduce aspect morphology, such as the perfective suffix -mut on the Kashmiri past verbal stem (Travis 1991; Bhatt 1999). The wh-expletive introduced into the specifier of AspP has three features: an uninterpretable D feature, an uninterpretable case feature (accusative), and the uninterpretable Q feature. Note that it has no interpretable features at all. When the transitive vis introduced, it will interact with the wh-expletive and the expletive's uninterpretable accusative case feature and [D] feature will be valued. By the time the C-head is introduced, the wh-expletive has only an unvalued uninterpretable [Q] feature remaining, and will interact with the C-head in the way described above. In this view, examples like (20) and (24) are impossible because only one goal can interact with the v head and have its uninterpretable case feature valued. If there is more than one potential goal, such as an additional wh-phrase or a clausal expletive, the uninterpretable case feature on one or the other will go unvalued, and the derivation will fail to converge. Observations about Hungarian provided by Horvath (1997) offer additional empirical evidence for this proposal. Hungarian is relevant because it has a partial wh-movement construction whose properties closely parallel those of Kashmiri (and Hindi-Urdu). The wh-expletive in the matrix clause actually exhibits case morphology appropriate to its role in relation to the matrix verb, and the whphrase in the subordinate clause is assigned a separate case, determined by properties of the embedded clause, as expected. (25)

Mit mondtal hogy kinek vett Janos szinhazjegyet? EXPL.ACC said that who.DAT bought Janos theater ticketACC 'Who did you say Janos bought a theatre ticket for?'

Horvath originally intended this data to argue against the so-called 'direct dependencY, or chain-based approaches to partial wh-movement, claiming that the chain itself could not be assigned different cases at its head and tail. Note that under the analysis outlined here, (25) would not pose such a problem, because the wh-phrase and wh-expletive are not syntactically connected, and can therefore be assigned difierent cases as necessary.


The ramifications of the claim that wh-expletives (at least in some languages) need case go beyond explaining the ungrammaticality of the examples in (20), as Simpson points out. Chomsky (1995, 2000) asserts that the operation Merge, in which an element is simply merged into the structure, is more economical than Move, the composite operation which involves Agree, pied-piping, andre-Merge combined. A sentence like (27) below is barred because in the derivation of the clause to be a man in the g(mlen, there are two choices if the numeration contains an expletive. The first is to Merge the expletive there (later raising it further) as in (26). The second is to Move the DP a man and postpone Merging the expletive, as in (27). (27) is ungrammatical because the less economical operation was chosen at that point in the derivation. (26)

There seems [t to be a man in the garden].

(27) *There seems [[a man] to bet in the garden]]. Returning to wh-expletives, if wh-expletives are simply merged into the specifier of CP, this will always be more economical than moving a full wh-phrase. According to Simpson (2000), the ungrammaticality of (28a-b) demonstrates that whexpletives are not simply merged into a wh-position. (28)

a. *Was glaubst du [was Hans wen gesehen hat] EXPL believe 2sG EXPL Hans whom seen has Intended: 'Whom do you believe Hans saw?'

[German]

b. *Was glaubst du [t Hans wen gesehen hat] EXPL believe 2sG Hans whom seen has Intended: 'Whom do you believe Hans saw?'

[German]

If Merging of a wh-expletive was more economical than Moving the wh-phrase into the lower Spec, CP, we would expect (28a-b) to be possible. However, if we accept that was as a wh-expletive must be assigned case, it would not be possible to generate it in the subordinate clauses in (28) in which all available cases have been assigned to full wh-phrases. This is an extension of the account of the ungrammaticality of the sentences in (20). Simply stipulating anti-locality as a requirement ("an expletive and any full wh-phrase cannot be clausemates) cannot be adequate, since it is possible for a full wh-phrase to appear in the same clause as the wh-expletive if the full whphrase is a subject (nor would this requirement prevent ungrammatical sentences as in (28)). In this case, the matrix wh-expletive indicates the position of interpretation of the wh-phrase in the lower clause, and has nothing to do with the wh-phrase in the matrix clause. This construction is exhibited in German by (29) (originally from Fanselow and Mahajan (1996)) and in Hindi-Urdu by (30).

63

64


(29)

(30)

Was glaubt wer wen sie liebt6 EXPL believe who whom 3FSG loves 'Who believes she loves whom?'

[German]

kis-ne kyaa soc-aa ki aap-ne kyaa paRh-aa? who-ERG EXPL thought-PRF.M that 2PL-ERG what read-PRF.M 'Who thinks you read what?' [Hindi- Urdu]

As long as the wh-expletive is being assigned a DP case separate from that of the full wh-phrase in the same clause the construction is grammaticaF 3·3·3

Previous approaches to wh-e>..1'letive constructions: Indirect and direct dependency

Indirect dependency approaches to wh-expletive constructions generally follow the account of Hindi-Urdu first proposed by Dayal (1994, 1996). Although we will address concerns particular to Hindi-Urdu in Chapter 4 (especially the revision to Dayal's approach proposed by Lahiri (2002)), it is essential to review the indirect dependency account here as it is frequently cited as the leading account for whexpletive constructions in a range oflanguages (as recently as Rackowski & Richards

6. There has been some disagreement about the grammaticality of (29). While Dayal (1994) claims that this sentence is ungrammatical, both Simpson (2000) and Beck and Berman (2000) indicate that this sentence is grammatical or nearly so for most speakers. 7. Sandy Chung has pointed out to me that we must also ask whether it is possible for a wh -expletive to co -occur with and indicate the position of interpretation of an interrogative subject or adjunct. This is ungrammatical in Kashmiri: (i) *k'aa von kami ki Sita ch-i sehat -as manz e.xpl say.PsT who ki Sita AUX.PRs-3sG health-we in Intended: 'Who said that Sita is healthy?' (ii) *tse k'aa chu-y kati pataa ki Sita chi 2SG.DAT EXPL AUX-DAT how know that Sita AUX.PRs-3sG

sehat -as manz health-we in Intended: 'How do you know that Sita is healthy?' The account in Chapter 4 correctly predicts that such sentences will be ungrammatical because of the properties of v. In the case of (i), v probes its domain and finds no interrogative material (besides the expletive with no interpretable features). Its [wh]-feature therefore cannot be valued In (ii), assuming that both the wh -expletive and full wh -phrase originate in the domain ofv, if the wh-expletive raises through the specifier ofvP, the full wh-phrase kati 'how' will not be able to have its [uQ] feature valued (it will be too far from C, and v can have no [Q] in this derivation or the wh-expletive would have been frozen in its specifier).


2005, Bruening 2006). The core claim of the indirect dependency proposal is that all apparently subordinate clauses in Hindi-Urdu are in fact adjunction structures, and that what we have called a wh-expletive in this chapter is in fact a scope-marker that is coindexed with the adjoined CP. The process of semantic interpretation of the coindexed components proposed by Dayal allows the scope of the wh-phrase in the adjoined CP to be interpreted at the position of the scope-marker. Dayal refers to this as 'indirect dependency' to contrast it with chain-based, 'direct dependency' approaches which rely on syntactic connections between the element kyaa (construed now as an expletive) and the full wh-phrase (these will be discussed in greater detail below). The indirect dependency approach is an attempt to address the following puzzle: wh-in-situ phrases in a single Hindi-Urdu clause take scope over the entire clause. (31)

tum kis-k.o pasand kar-te ho 2PL who-ACC like do-HAB.PL AUX.PRS 'Whom do you like?'

[Hindi- Urdu]

According to Dayal, this indicates that the wh-phrase moves to Spec, CP at the level of Logical Form (LF). In a two-clause question in which the wh-in-situ phrase is in the lower clause, a matrix scope interpretation is not available. (32)

tum jaan-te ho [ki us-ne kyaa kiy-aa] 2PL know-HAB.PL AUX.PRS that he-ERG what do-PRF.M 'You know what he did: :J! 'What do you know that he did?'

[Hindi- Urdu]

According to Dayal, this fact demonstrates that the embedded wh-phrase cannot move to the Spec, CP of the matrix clause at LF. If this is the case, then there must be some reason why this movement is blocked. Dayal claims that finite subordinate clauses are adjoined to the matrix clause CP or IP. Support for this claim is derived both from the scope facts in (32), and from constituent order in Hindi-Urdu. In a standard Hindi-Urdu clause, complements precede verbs in an unmarked sentence. (33)

Hamid-ne pani piy-aa Hamid-ERG water drank-PRF 'Hamid drank water:

[Hindi- Urdu I

However, finite complement clauses appear exclusively to the right of the verb. (34)

a.

Vo jaan-tii hai ki anu aay-ii 3SG know-HAB.F AUX that Anu come-PRF.F 'She knows that Anu came:

b. *Vo ki anu aayii jaantii hai

[Hindi- Urdu] [Hindi- Urdu I

65

66

Peripheries in Kashmiri and Hindi-Urdu It~

as Dayal claims, all seemingly finite complement clauses in Hindi-Urdu are in fact adjoined, these clauses could be considered strong islands. Following the definition of barrier proposed in Cinque (1990), this would be because an adjoined clause is in a position not directly selected by the verb and is in the non -canonical direction. It is therefore a barrier for both binding and government. Further, if we follow Dayal's proposal that subjacency is operative at LF, it will be impossible for any in situ wh-phrase to escape the finite complement clause. For this reason, it cannot take matrix scope under any circumstances. It then becomes a puzzle why the scope-marking structure in (35) permits a matrix interpretation of the wh-phrase. The phrase should never be able to escape the lower clause at LF. (35)

Sita-ne kyaa soc-aa ki Ravii-ne kis-ko dekh-aa? Sita-ERG EXPL thought-PRF.M that Ravi-ERG who-Ace saw-PRF.M 'Who did Sita think that Ravi saw?' [Hindi- Urdu]

In these structures, Dayal claims that the complement position of the main verb is occupied by a scope-marker (kyaa), which we have called a wh-expletive to this point. The scope-marker and the adjoined CP are coindexed. Since HindiUrdu is an SOV language, in this approach the pre-verbal position that the whexpletive occupies appears also to be the canonical object position. According to Dayal in Hindi-Urdu all in situ wh-material must raise to the edge of its CP at LF, creating two local wh-dependencies in a two-clause partial movement construction. The first is due to the LF movement of the full wh-phrase to the specifier of the adjoined CP, and the second to the LF movement of the scope-marker kyaa to the specifier of the matrix CP. Since the scope-marker and the adjoined CP are related by coindexation, the net result gives the effect of a single longdistance dependency. The LF structure this process would produce for (36a) is represented in (36b ). (36)

a.

sita-ne kyaa socaa ki ravi-ne kis-ko dekh-aa sita-ERG EXPL thought-PRF.M that Ravi-ERG who-Ace saw-PRF.M 'Who does Sita think Ravi saw?'

b.

[cP kyaa1 [[1p sita-ne t 1 socaa] [CP ki kis-ko 2 ravi-ne t2 dekhaa] 1]]

t

I

t

I

This view achieves the effect of interpreting the full wh-phrase on the left periphery of the matrix clause because that is where the entire adjoined CP will be interpreted. Dayal extends this approach to German question formation, proposing a nearly identical LF for German wh-expletive constructions. The indirect dependency approach raises a number of questions, both theoretical and empirical. Perhaps the most unusual claim of this approach is that


what is normally viewed as a complement CP is in fact adjoined to the matrix CP. If this is the right view, we should expect that interactions which depend on command relations into the rightmost clause will work in very different ways than in languages in which that clause is a complement to V (as is presumably the case in English). For instance, it should not be possible for quantifiers in the matrix clause to bind variables in the adjoined CP, since under this view these quantifiers would not c-command the variables. However, in Hindi-Urdu wh-expletive constructions like (37) it appears that the quantifier har aadmii 'each man' in the matrix clause in fact can bind the pronoun us-ne 'he' in the second CP (Mahajan, 2000). (37)

Har aadmii-nei kyaa soc-aa ki us-nei kis-ko dekh-aa each man-ERG EXPL think-PRF.M that he-ERG who-ACC SaW-PRF.M 'Who did every mani think that hei saw?' [Hindi- Urdu]

Under the indirect dependency approach, the bound variable reading in (37) is unexpected. Similarly, in (38) a matrix clause complement binds a pronoun in the CP that follows the verb bolaa 'told'. This bound variable reading would be surprising if the CP is adjoined higher than the object, since the object could not c-command the pronoun inside CP. (38)

aap-ne har aadmii-sei kyaa bol-aa ki voi fon-par you-ERG each man-INST EXPL tell-PRRM that he phone-on he kis-se bol sak-taa who-INSTR call can-HAB.M AUX.PRS [Hindi- Urdu] 'Who did you tell each man that he could telephone?'

Along these same lines, Beck and Berman (2000) point out that if German clauses containing expletives was andes cause their complements to be adjoined, then the binding in (39) should be impossible. (39)

daB keine Studentini es bedauert, daB siei die Vorlesung geschwanzt hat that no student it regrets that she the lecture skipped has '... that no student regrets it that she has skipped the lecture: [German)

Again, since the embedded clause would hypothetically be adjoined to the first CP, the matrix subject should not c-command any material within it, and this binding would be ruled out. 8 On a theoretical note, the origin of the coindexation of the wh-expletive (and the free variable it is translated as) with the embedded wh-clause is unclear. There

8. Chapter 4 offers a detailed critique of the most recent version of the adjunction approach for Hindi-Urdu, as proposed by Lahiri (2002), including additional evidence such as negative polarity licensing.

68


appears to be no motivation beyond achieving the correct interpretation of the adjoined CP since, as Beck and Berman point out, the indexation is neither referential nor anaphoric. Further, if we are to explore the viability of the Strong Minimalist thesis for language, the Inclusiveness Condition requires that no new features be introduced in the course of a derivation (Chomsky 2000: 113). Indices and coindexation become unavailable in this view. Of course, this condition would also be problematic for incarnations of the direct dependency approach that rely on syntactic chains established between wh-items to arrive at the appropriate connections within the structure. We will return to the indirect dependency approach below addressing its similarities and differences with the account that will be offered here. The account proposed in this chapter follows the indirect dependency approach in assuming that the wh-expletive is base generated in a position inside the clause in which it is assigned case. This claim was supported by evidence of various kinds in Section 3.3.2. However, it departs from the indirect dependency view in that here the role of the wh-expletive is taken to be syntactic only, and the wh-expletive does not play a role in the semantic interpretation of the question (for more on this point, see Section 3.3.4). It may seem, then, that the account presented here has more in common with direct dependency approaches, but this is probably misleading. The representative direct dependency approach to partial wh-movement and wh-expletive constructions is McDaniel's (1989) account of German and the Indic language Romani. In this account, a syntactic wh-chain with a specific set of well-formedness conditions connects a wh-phrase and its trace with a wh-expletive. McDaniel claims that the unique properties of wh-expletive constructions can be explained if Subjacency is considered to be a condition on representations (not just movement). This type of analysis seems especially geared toward wh -expletive constructions of three clauses or more, as discussed in 3.4.2 below. The direct dependency approach attempts to codify the notion that full and partial wh-movement are difierent manifestations of the same phenomenon. Under an account of this kind, both full movement and wh -expletive constructions result in chains with similar (though not identical) well-formedness conditions. This is a view shared by the approach proposed here, though it is encoded in a very different way. In this case each head in a series of clauses in a matrix question must have its features valued and EPP satisfied. This can be accomplished either via movement, or via wh-expletive insertion, depending on the numeration; the end result is essentially the same. In the indirect dependency approach, on the other hand, full wh-movement constructions are a syntactic phenomenon, while wh-expletive constructions are not. That is, the crucial role played by a whexple tive itselfis a semantic one. These two question formation strategies are viewed as totally distinct. The claim, on the other hand, that both full wh-movement and wh-expletive constructions are manifestations of the workings of the same set of


mechanisms is something that the direct dependency account and the account presented here share. In the approach in 3.3.1, a connection is made in the syntax between the position at which the wh-phrase will be interpreted and the wh-phrase itself. It is the features of the full wh-phrase in a sentence like (2) which value the uninterpretable features of the head in the matrix position where the wh-phrase will be interpreted. Unlike the direct dependency view, however, there is no link made between the whexpletive and the full wh-phrase. The connection that is forged makes use of neither chain nor index and is effected in the course of normal syntactic processes, requiring no additional mechanisms beyond featural satisfaction and associated movement This feature-based approach permits an important simplification over chainbased direct dependency approaches. McDaniel's account requires a stipulation contained in the well-formedness condition on chains to ensure that wh -expletives appear only above the full wh-phrases whose scope they indicate in syntactic structure. This is stated in (40): (40)

For any scope-marker a~, 1

Peripheries in Kashmiri and Hindi-Urdu: The syntax of discourse-driven movement (Language Faculty and Beyond)

Verb Movement and the Syntax of Kashmiri

Peripheries in Kashmiri and Hindi-Urdu: The syntax of discourse-driven movement

Correlatives Cross-Linguistically (Language Faculty and Beyond)

The Faculty of Language

Epistemology And The Logical Syntax of Language

Logical syntax of language

Movement and Silence (Oxford Studies in Comparative Syntax)

Carnap's Logical Syntax of Language

The language of propositions and events: issues in the syntax and the semantics of nominalization

Syntax and Semantics of Prepositions (Text, Speech and Language Technology)

On Pied-Piping: Wh-Movement and Beyond

Heian Japan, Centers And Peripheries

Spontaneous Spoken Language: Syntax and Discourse

Optimality-Theoretic Syntax (Language, Speech, and Communication)

Tourism in Peripheries

Beyond Coherence: The syntax of opacity in German

The Language of Propositions and Events: Issues in the Syntax and the Semantics of Nominalization (Studies in Linguistics and Philosophy)

Understanding Syntax (Understanding Language)

The Acquisition of Syntax in Romance Languages (Language Acquisition and Language Disorders)

Beyond the Language Classroom

Reconnecting language: morphology and syntax in functional perspectives

The Architecture of the Language Faculty (Linguistic Inquiry Monographs)

The Architecture of the Language Faculty (Linguistic Inquiry Monographs)

Constraints on Displacement: A phase-based approach (Language Faculty and Beyond)

Language universals and linguistic typology: syntax and morphology

Lexicon-Syntax Interface in Second Language Acquisition

Language Universals and Linguistic Typology: Syntax and Morphology

The Grammar of Q: Q-Particles, Wh-Movement, and Pied-Piping (Oxford Studies in Comparative Syntax)

Faculty Diversity: Problems and Solutions

After Rhetoric: The Study of Discourse Beyond Language and Culture

Peripheries in Kashmiri and Hindi-Urdu: The syntax of discourse-driven movement (Language Faculty and Beyond)

Verb Movement and the Syntax of Kashmiri