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 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

Peripheries in Kashmiri and Hindi-Urdu

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

Peripheries in Kashmiri and Hindi- Urdu

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.

Chapter 1. Introduction

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

Peripheries in Kashmiri and Hindi-Urdu

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)

Chapter 1. Introduction

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

Peripheries in Kashmiri and Hindi-Urdu

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

Chapter 1. Introduction

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

Peripheries in Kashmiri and Hindi- Urdu

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

Chapter 1. Introduction

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

Peripheries in Kashmiri and Hindi-Urdu

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

Peripheries in Kashmiri and Hindi-Urdu

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

Kashmiri, Hindi-Urdu, English

F

80+

Kashmiri only

vc

M

59

Kashmiri, Hindi-Urdu, English

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

Peripheries in Kashmiri and Hindi- Urdu

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

Peripheries in Kashmiri and Hindi- Urdu

(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)

Chapter 2. Feature stacking

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

Peripheries in Kashmiri and Hindi-Urdu

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.'

Chapter 2. Feature stacking

( 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)

Chapter 2. Feature stacking

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

Peripheries in Kashmiri and Hindi-Urdu

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

Chapter 2. Feature stacking

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

Peripheries in Kashmiri and Hindi-Urdu

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

Peripheries in Kashmiri and Hindi- Urdu

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.

Chapter 2. Feature stacking 29

(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

Peripheries in Kashmiri and Hindi- Urdu

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,

Chapter 2. Feature stacking

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

Peripheries in Kashmiri and Hindi-Urdu

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

Chapter 2. Feature stacking

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

Peripheries in Kashmiri and Hindi-Urdu

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.

Chapter 2. Feature stacking

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)

<{Fl=a, b, c}, {F2=d, e}, {F3=g}>

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

Peripheries in Kashmiri and Hindi-Urdu

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.

Chapter 2. Feature stacking

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

Peripheries in Kashmiri and Hindi-Urdu

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.

Chapter 2. Feature stacking 39

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

Peripheries in Kashmiri and Hindi-Urdu

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.

Chapter 2. Feature stacking

(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

Peripheries in Kashmiri and Hindi-Urdu

(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.

Chapter 2. Feature stacking 43

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

Peripheries in Kashmiri and Hindi-Urdu

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

Chapter 2. Feature stacking

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

Peripheries in Kashmiri and Hindi-Urdu

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

Peripheries in Kashmiri and Hindi-Urdu

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

Peripheries in Kashmiri and Hindi- Urdu

(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

Peripheries in Kashmiri and Hindi- Urdu

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

Chapter 3. Full and partial wh- movement in Kashmiri

(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

Peripheries in Kashmiri and Hindi-Urdu

(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

Chapter 3. Full and partial wh- movement in Kashmiri

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

Peripheries in Kashmiri and Hindi-Urdu

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.

Chapter 3. Full and partial wh- movement in Kashmiri

(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

Peripheries in Kashmiri and Hindi- Urdu

(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.

Chapter 3. Full and partial wh- movement in Kashmiri

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

Peripheries in Kashmiri and Hindi-Urdu

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.

Chapter 3. Full and partial wh- movement in Kashmiri

(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

Peripheries in Kashmiri and Hindi- Urdu

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.

Chapter 3. Full and partial wh- movement in Kashmiri

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

Peripheries in Kashmiri and Hindi-Urdu

(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).

Chapter 3. Full and partial wh- movement in Kashmiri

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

Chapter 3. Full and partial wh- movement in Kashmiri 67

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

Peripheries in Kashmiri and Hindi-Urdu

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

Chapter 3. Full and partial wh- movement in Kashmiri 69

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
In the account presented in Section 3.4, it follows automaticallythatwh-expletives will only occur in clauses higher than the one hosting the wh-phrase whose scope they mark. More precisely; the feature-based system requires that something with a wh-feature appear in the domain of the lowest interrogative C-head. This is because this head possesses an uninterpretable [wh] feature. Wh-expletives do not have a [w h] feature at all (by definition), so if a wh -expletive is the only wh-material to appear in the domain of the lowest interrogative head, the uninterpretable [wh] feature on that head will remain unvalued. In this case the derivation will crash. This result is achieved without additional stipulation.

3·3·4

Interpreting wh-expletive constructions

This section is intended to suggest an approach to the semantic interpretation of wh-expletive constructions that is compatible with the syntactic analysis presented above. The requirement imposed on such an approach is that the wh-expletive, which has no interpretable features at all, plays no role in the matrix interpretation of embedded wh-phrases. Only those elements of the A-bar system with interpretable features can be relevant for interpretation, since all uninterpretable features are deleted by the conclusion of the derivation. These elements are the matrix C-head, which hosts an interpretable [Q] feature, and the wh-phrase itself, which possesses an interpretable [wh] feature. Reinhart (1998) proposes a mechanism for bindingwh-in-situ constituents that is meant to be compatible with the general architecture of the Minimalist Program.

70

Peripheries in Kashmiri and Hindi-Urdu

In that vein, it seeks to eliminate the need for movement at LF to achieve wide scope, and instead proposes that these constituents must be interpretable in-situ. I will not attempt a full development of this analysis; I only want to sketch how a reasonable semantics can be built on the syntactic foundation developed here. Reinhart follows Karttunen (1977) in assuming that a wh-NP is simply an indefinite NP, and a question denotes the set of (true) propositions that are its answers. She claims that the question operator (at the position at which the scope of the question will be interpreted) introduces an existential quantification over a choice function variable. The wh -word, and in fact any weak determiner, acts as a (choice) function variable applying to the set denoted by the NP. This variable can be bound by an existential operator that may be arbitrarily far away. In this way, wide-scope interpretation can be assigned to in-situ wh-constituents.9 Note that on this view, no LF-raising of the wh-phrase is required. In fact, talk of the "scope" of the wh-phrase is somewhat misleading. This proposal dovetails well with the syntactic approach put forward above. In this view, the interpretable [Q] feature on the matrix (or highest relevant) C-head triggers as existential quantification over a choice function variable. The interpretable [wh] feature reflects the fact that the wh-indefinite is interpreted as a choice function variable that applies to a given set For the wh-expletive construction in (2), repeated here as (41 ), the relevant portion of the interpretation is sketched in (42). (41)

kam' baasaan ki mohn-as tse k'aa chu-y 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?'

(42)

(Wall & Koul: 18)

{PI (3f) (CH(f) &you think thatf(xl animate(x)) gave Mohan the book}1°

9· This requires an understanding of the semantic interpretation as having access to the entire derivation. The choice function variable introduced by the wh-indefinite in an embedded clause goes unbound until the iQ in the matrix clause introduces the existential binder, and this binder, crucially, may be separated from the wh-indefinite by an arbitrary number of phases (see, for instance, the case of three-clause questions in (50)-(51) in the text). As Chomsky (p.c. cited in de Villiers et al. 2007: 7) puts it: "Note that the total semantic and phonetic interpretation has to be global; both CI and SM (say, prosody) enter into full semantic interpretation, and in both cases the whole expression has to be surveyed. Yes-no question intonation, for example, is only determined at the very end of the computation, long past the point where the intonation appears." to. Reinhart assumes that pronominal wh-phrases such as who are determiners with empty nouns: [who [N'!(i)]].

Chapter 3. Full and partial wh- movement in Kashmiri

In this case, the [iQ] in the matrix clause provides existential quantification over a choice function, binding the function variable provided by the [iwh] of the wh-NP. 11 This question then denotes the set of true propositions P, each stating that for some choice function f that you think that the animate being selected by fgave Mohan the book. This binding can occur over an arbitrarily long distance, so it should not matter how deeply embedded the full wh-phrase is. Our syntactic proposals are thus compatible with a reasonable semantics one that was developed on entirely independent grounds. Another kind of question now arises: if the wh-expletives are irrelevant for effecting the matrix scope interpretation of the wh-phrase, why are they present? The analysis presented above makes clear that the role of the wh-expletive is purely syntactic in operational terms. It fulfills the syntactic requirement we have labeled with the name EPP - that certain heads require additional specifiers beyond those needed for selectional purposes. The appearance of these wh-expletives is not a semantic requirement, as the system of interpretation has no need for intermediate material to aftect the matrix scope reading. 12 The larger question - of why the design of natural language includes expletive elements - of course remains.

Additional empirical investigations

3·4

There are two important empirical questions concerning wh-expletive constructions that remain to be discussed. The first question, addressed in 3.4.1 is a Kashmiri-specific one, and its solution does not bear significantly on the analysis presented above. The second question, addressed in 3.4.2 is crosslinguistic in nature, and has ramifications for the wider study of wh -expletives across languages and dialects. 3·4·1

A Kashmiri issue: Factive predicates

Bhatt (1999) claims, and my informants agree (PK 9/21/04, JC, VC 9/8/05), that full wh-movement is not possible from clauses that follow non-bridge verbs in Kashmiri, but wh-expletive constructions are generally permitted (compared with the grammatical bridge verb versions of these sentences in (1)-(2)).

n.

The interpretable [Q] feature also seems to be what establishes the set of propositions.

u. This will be supported by certain varieties of wh-expletive languages that have three clause structures lacking wh-expletives in the intermediate clause, discussed in Section 4.2.

71

72

Peripheries in Kashmiri and Hindi- Urdu

(43)

a. *k'aa ch-a miiraa-yi khabar ki t mohn-an por what AUX-FSG Mira-DAT know that Mohan-ERG read Intended: 'What does Mira know that Mohan read?' b.

por k'aa cha miiraa-yi khabar ki k'aa mohn-an EXPL AUX-FSG Mira-DAT know that what Mohan-ERG read 'What does Mira know that Mohan read?'

Others (namely Wall & Koul1997) have claimed that wh-expletive constructions are also impossible in factive contexts, and that (43b) is ungrammatical. It is dear that there is further empirical work to be done here, but presuming for the moment that the judgments in (43) are correct for at least some varieties of Kashmri, we can explore several ways of accounting for them. Bhatt (1999) has proposed that subordinate clauses out of which extraction is impossible in Kashmiri are in fact adjoined to the matrix clause. The entire structure would be as in (44).

---------CP

(44)

CP

~

miiraayi cha khabar Mira AUX knows

CP

~

ki k'aa mohnan por that what Mohan read

In this view, wh-movement will not be possible between the two clauses because the first CP does not contain the second. There is, however, an approach to the structure in (44) in which k'aa could be viewed as marking the scope of the whphrase in the adjoined clause. This approach, along the lines of Dayal (1996), resembles structures proposed under the indirect dependency account, discussed in Section 3.3.2. Bhatt's proposal differs from complete indirect dependency, in that he proposes an adjoined structure like (44) only to account for contexts in which interclausal displacement of the wh-phrase is unavailable (not for all clausal complements). Suffice it to say here that under such an approach one could obtain the contrast in (43). On the other hand, binding facts in Kashmiri indicate that a hierarchical relationship obtains between the matrix and complement clauses of both factive and non-factive predicates (JC 6/05). (45)

a.

har insaan-asi ch-u basaan ki sui ch-u teez each man-DAT AUX.PRS-M thinks that 3SG AUX.PRS-M smart 'Each man thinks he is smart'

b.

te:z har insaan-asi ch-u pataa ki sui ch-u each man-DAT AUX.PRS-M knows that 3SG AUX.PRS-M smart 'Each man knows he is smart'

Chapter 3. Full and partial wh- movement in Kashmiri

In both sentences in (48), har insaanas binds the pronoun in the lower clause, indicating that it commands the pronoun. This lack of contrast is unexpected if the structure of the factive sentence was as in (44). An alternative solution to this puzzle could be related to the fact that factive verb complements are weak islands in other languages. (46)

a. How do you think he escaped the building? b. *How do you regret he escaped the building? (on matrix reading)

Assuming the wh-expletive construction to be an instance of the operation Agree (simplex), Agree may be able to obtain across weak island boundaries even though Move (complex: Agree+ 2nd Merge) is impossible across those same boundaries. In light of the facts in (46), this view seems more promising than an adjunction account, which will be shown to face additional empirical challenges in Chapter 4. Moreover, the controversial nature of the data (that is, whether any kind oflongdistance wh -dependency is possible across the complement of a factive verb) as it is currently reported in the literature indicates that further empirical clarity may be required. 3.4.2

A crosslinguistic issue: Multiple wh-expletives

The crosslinguistic issue that remains to be examined is centered on wh-questions consisting of more than two clauses. In German, Hindi-Urdu, and Kashmiri, an embedded wh-phrase can take scope across any number of clauses so long as this is mediated by a wh-expletive in every clause higher than the clause containing the wh-phrase. (47)

a.

Was glaubst du, was Jan meint, mit wem Ann gesprochen hat EXPL believe 2sG EXPL Jan think, with whom Ann talked has 'Who do you believe Jan thinks Ann talked with?' [German]

b. *Was glaubst du, Jan meint, mit wem Ann gesprochen hat? (48)

a.

Ram-ne kyaa soc-aa ki ravii-ne kyaa kah-aa ki konsa Ram-ERG EXPL think-PRRM that Ravi-ERG EXPL say-PRF.M that which aadmii aay-aa man come-PRF.M

[Hindi- Urdu] 'Which man did Ram think that Ravi said came?' b. *Ram-ne kyaa socaa ki ravii-ne kahaa ki konsa aadmia aayaa (49)

a.

Ram-an k'aa von ki tse k'aa chu-y baasaan ki Ram-ERG EXPL say.PST that 2SG.DAT EXPL AUX-DAT think that Mohn-as kam' di-ts kitaab Mohan-DAT who.ERG give-PST.FSG book

'Who did Ram say you think gave Mohan the book?' [Kashmiri] b. *Raman k'aa von k:i tse chu-y baasaan ki mohn-as kam' di-ts kitaab

73

74

Peripheries in Kashmiri and Hindi- Urdu

In (47a), (48a), and (49a) the wh-phrase has undergone partial movement to the left periphery of the lowest clause. A wh-expletive is present in both the intermediate and matrix clauses, and a matrix interpretation of the wh-phrase results. The ungrammatical examples, (47b), (48b), and (49b) are identical, except that they lack a wh-expletive in the intermediate clause. It seems that a wh-expletive must be present in every clause above the clause hosting the wh-phrase, up to and including the clause in which the wh-phrase is bound by the interrogative operator. The first question is how to account for the core pattern: the obligatory presence of an expletive in each clause above the clause containing the wh-phrase. The partial movement of the wh-phrase in examples like (49b) is explained just as it is in two-clause structures: in this way the EPP property of the lowest C-head is satisfied, and the wh-phrase becomes accessible to higher probes which will value its remaining features. If the wh-phrase remains in this low position however, the EPP on all higher probes must be satisfied by a wh-expletive. If an expletive is missing in any of these positions, the EPP will not be satisfied on the C-head in that clause. Upon insertion of an expletive, this C-head will still have features that need to be valued by Agree. It can do so by probing the lower phase edge and interacting with features found there. In what follows I will make this account more explicit Consider the three possible ways of forming a grammatical three-clause whquestion in Kashmiri in which the wh-phrase originating in the lowest clause takes scope over the entire sentence (Wali & Koul1997; Bhatt 1999, judgments confirmed by JC & VC 9/8/05). (SO)

Ram-an kam' von ki tse chu-y baasaan Ram-ERG who.ERG say.PST that 2sG.DAT AUX-DAT think ki mohn-as di-ts kitaab that Mohan-DAT give-PST.FSG book 'Who did Ram say you think gave Mohan the book?'

(51)

[Kashmiri]

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

(52)

[Kashmiri)

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

[Kashmiri)

Chapter 3. Full and partial wh- movement in Kashmiri

(50)-(52) illustrate the three different ways to ask the matrix question 'Who did Ram say you think Mohan gave the book to?' In (SO) we have full whmovement from the base position in the lower clause (indirect object of dits 'gave'), to the canonical wh-position in the matrix clause, preceding the secondposition verb. In (51), the wh-phrase has moved up only one clause, to the preverbal position in the intermediate clause, and a wh-expletive occupies the wh-position in the matrix clause. Below I will demonstrate that this case, not surprisingly, reduces to the analysis given to two-clause wh-expletive constructions. In (52), the wh-phrase has moved only to the left periphery of its own clause. A wh-expletive occupies the wh-position in both the intermediate and matrix clauses. In the case of full wh-movement to the matrix clause, as in (50), this approach would offer the relatively straightforward account outlined in the diagram in (53). (53)

[wh-phrase [Cl) ... ][t [C2] ... ][t [C3] ... t] [uQ] [iQ] [uwh] [uwh] [iwh] [uwh] [EPP] [EPP] [EPP]

The wh-phrase that originates inside the lowest clause is an active Goal. The C-head on the left periphery of this clause (C3) is an active Probe. The whphrase raises into its specifier, valuing its uninterpretable [wh] feature and satisfying the EPP. As it is on the edge of the lowest phase, it is visible for interactions in the next phase up. Furthermore, the wh-phrase still has unvalued features (since C3 is non-interrogative and cannot value its uninterpretable [Q] feature). Therefore, it can still function as an active Goal. The process is then repeated in the phase determined by C2. The wh-phrase at this point is still an active Goal because its uninterpretable [Q] feature has not been valued by C2 either. The process repeats once again in the phase determined by C1, but on this step, when the wh-phrase raises into the specifier of C1, its [Q] feature is valued. At this point, all uninterpretable features have been valued, and the derivation is licit. The sentence in (51) is a mixed case, in which a part functions just like the full wh-movement described above, and a part functions like a wh-expletive construction. This process is diagrammed in (54). (54)

[wh-expl [Cl] ... t][wh-phrase [C2] ... ][t (C3] ... t] [uQ] [iQ] [uQ] [uwh] [uwh] [uwh] [iwh] [EPP] [EPP] [EPP]

•

75

76

Peripheries in Kashmiri and Hindi-Urdu

In (54), the wh-phrase raises from the lowest to the intermediate clause in the same manner described above in (53). However in (54), a wh-expletive is present in the numeration and is introduced in a case position inside the matrix clause. Cl is an active Probe and finds the wh-expletive that is an active Goal. The wh-expletive moves into the specifier of Cl, and its single remaining uninterpretable feature is valued The EPP on Cl is satisfied, but is own uninterpretable [wh] feature has yet to be valued. Cl probes its domain once more and finds the full wh-phrase at the left edge of the lower phase. The wh-phrase is still an active Goal, because its [uQ] is still unvalued. Cl and the wh-phrase Agree with one another, mutually valuing one another's uninterpretable features and creating a licit derivation. The second part of this derivation is identical to the way in which a two-clause wh-question is analyzed. A third possible way to form a three-clause matrix question in Kashmiri is to permit the wh-phrase to remain in the lowest clause by employing wh-expletives in both the intermediate and matrix clauses, as in (52). The analysis that has been described here would give (52) the structure described by the schematic in (55). (55)

[wh-e:xpl [Cl] ... t] [wh-e:xpl [C2] t ][wh-phrase[C3] ... t] [uQ] [uQ] [uQ] [iQ] [u.Q] [uwh] [iwh) [EPP] [uwh] [uwh]

[EPP]

[EPP]

Let us walk through each step of (55) carefully. The initial step of (55) is identical to that in (53) and (54), in which the wh-phrase moves from its base position into the specifier of the lowest C-head. The EPP on C3 is satisfied, and its single uninterpretable feature is valued. The wh-phrase, on the other hand, remains active. In the intermediate clause, the C2 head probes its domain and finds the wh-expletive. The expletive moves into the specifier of C2 and satisfies the EPP while valuing its own uninterpretable [Q] feature. Though this wh-expletive is now inactive, the C2 head still has an unvalued uninterpretable [wh] feature. It probes its domain once again and interacts with the wh-phrase at the left edge of the phase below. The C2 head and the wh-phrase mutually value one another's features. The derivation continues in the matrix clause, in which the Cl head interacts with the wh-expletive originating in its own clause, and the process proceeds as above. The Cl head must then probe its domain again, as it is still active, and find the valued [wh] feature on the C2 head with which to value its own [wh] feature. At this point all uninterpretable features have been valued, and the derivation is licit. Notice that the wh-expletives in (55) do not interact with the wh-phrase in any way. In fact it is the C-heads in each clause that interact with the wh-phrase, not the wh-expletive. The role of the expletive is solely to satisfy the EPP- it has no other purpose. This approach permits an arbitrarily long series of C-heads to value one another via static Agree over a distance. In this view, there need be no direct

Chapter 3. Full and partial wh- movement in Kashmiri link between the expletives and the full wh-phrase. The wh-expletive does not possess interpretable features, so will ultimately be meaningless, and only the features of the full wh -phrase will be relevant in the interpretation of the question. 13 The analysis of long-distance wh-expletive dependencies presented in (55) captures an important intuition about the construction in (52). Wh-movement and wh-expletive constructions are in effect alternative strategies for forming a long-distance wh-dependency. In this approach to wh-expletive constructions, it is the features of the wh-phrase itself that cause features on the successive C-heads to be valued. In (55), it is the features of the wh-phrase that value the intermediate C-head, and those features which in turn value the features of the matrix C-head. Some information about the wh-phrase is affecting the series of heads, though it is not the wh-phrase itself which directly values the features on the matrix C. In this way, this account shares with direct dependency approaches the notion that wh-movement and wh-expletive constructions are two methods of creating the same long -distance wh -dependency. It diverges in that no explicit chain is formed. Instead, it is the features of the full wh-phrase that ultimately cause the features of the matrix C-head to be valued, whether it moves there itself, or whether a whexpletive fills the position at which the scope of the wh-phrase is interpreted. These proposals raise an important question about the way in which the feature content of the clause periphery is determined. In (55), the intermediate C-head, like the matrix C-head, has a [Q] feature. This causes the wh-expletive in its specifier to have all its uninterpretable features valued, and thus be frozen in place. The presence of the [Q] feature has an additional consequence, of valuing the uninterpretable [Q] feature on the wh-phrase in the specifier of the lowest C-head. In effect, this makes the lower two clauses of the structure in (55) analogous to any two-clause wh-expletive construction. The analysis in (55) ensures not only that the wh-expletive is frozen in place in the specifier of the intermediate C-head, but also that the wh-phrase in the lowest clause is frozen in the lowest C specifier. It may seem odd that the interrogative feature (though uninterpretable) is present on an intermediate C head. However, there is crosslinguistic evidence that intermediate heads in a wh -movement sequence may possess interrogative features of some kind. In particular, Henry (1995) observes that in Belfast English subjectauxiliary inversion takes place not only in the highest C head in a wh-movement

13. As discussed above in the text, the wh -phrase introduces a choice function which can be

existentially bound at an arbitrary distance. In this case the binder is introduced by the iQ on the matrix C-head, and the three-clause sentence will be interpreted as a matrix wh-question. This assumes, as noted in Footnote 4 above, a framework in which the interpretative component must at some point have access to the entire structure, not just a single phase or sub-part.

77

78

Peripheries in Kashmiri and Hindi-Urdu

sequence, but in intermediate heads as well. She takes this as an indication that at least in this context these intermediate heads have some interrogative status just like matrix C-heads (see also Rizzi 1996). In the present account, this notion is reflected in the [Q] feature that appears on some intermediate C-heads in Kashmiri. A central element in these proposals (evident in (55)) is the availability of three different flavors of C-heads - each with a slightly different set of interpretable and uninterpretable features. A legitimate concern arises at this point: what if it is not these exact combinations of features that appear on these heads in the numeration and in this just this order? For instance, what if an intermediate C-head in a three clause wh -expletive construction was not of the type containing a [uQ]? (Of course, the presence of a [iQ] at the top of an interrogative clause is going to be mandated by the selectional requirements of the governing predicate). One way of answering this question is to work through this and other scenarios for both wh-movement and wh-expletive strategies, an exercise conducted in the Appendix to this chapter. However the larger answer to this family of questions is that derivations and numerations with flaws such as these, should they arise, fail to converge. Some uninterpretable feature in each case will remain unvalued and will result in a crash. The system of features proposed here is such that heads that must trigger a certain semantic interpretation (for matrix questions, the matrix C-head) possess certain interpretable features. Uninterpretable features on those heads serve to enact movement of and agreement with various goals. Certain sets of features appearing on heads in a certain order limit the patterns that can appear. Other patterns will result in non-convergent derivations. As opposed to a list of maxims, principles, and conditions, in this framework the properties of the grammar are expressed via features and the systematic operations that result in their valuation. Though we will not explore this issue here, it should be noted that it is part of an important set ofbroader questions about the way in which the Minimalist Program accounts for ungrammaticality. Through (53)-(55), the feature-based approach to wh-expletive constructions in 3.3.1 has been extended to account for such constructions in sentences of three clauses (in fact to sentences of arbitrarily many clauses). Essential to this account are two operations that can function alone or in concert. The first is interclausal wh-movement, in which a wh-phrase moves from one clause to the next through a series of C-heads. These C-heads cannot value the uninterpretable feature on the wh-phrase, so the wh-phrase must continue until reaching the matrix clause. The second operation is wh-expletive clause internal movement, in which the C-head has the capacity to deactivate the wh-expletive by valuing its single uninterpretable feature. The wh-expletive only values the EPP on this type of C-head, and then remains fixed in its specifier. The C-head then acts as a probe and values its uninterpretable features via static Agree within its domain. An example like (SO), with

Chapter 3. Full and partial wh- movement in Kashmiri 79

long wh-movement, makes use only of the first operation described here. (51) and (to a lesser extent) (52), make use of both of these operations in combination. Given the different flavors of functional heads that can appear, and the different types ofwh-elements in languages like Kashmiri (meaningless wh-expletives and meaningful wh-phrases), it is predictable that there will be several ways to form a matrix question in a three-clause sentence. At this point, however, an unresolved issue arises concerning (49b), repeated here. ( 49) b.

*Raman k' aa von ki tse chu-y baasaan ki mohn-as bm' dits kitaab.

An alternative derivation of (49b) is conceivable in which the EPP of the intermediate clause is satisfied. If the wh-expletive that appears in the matrix clause originated in the intermediate clause and subsequently moved higher, then we would need to find an alternative explanation for the ungramrnaticality of (49b). Such a derivation would be one in which the wh-expletive in the intermediate clause raised into the specifier of the intermediate C-head, then raised further into the specifier of the matrix C-head. In this case the EPP would be satisfied in both positions, and (49b) should not be ungrammatical. This would be analogous to the behavior of DP-expletives. Consider the English DP-expletive there in (56). (56)

There seems to be a problem.

The standard account of (56) is that there originates in specifier of the lower TP [to be a problem], and then raises into the specifier of the higher TP. However, there are several reasons to believe that wh-expletives, unlike DP-expletives, cannot move from one clause to another. I will address these reasons below, and will ultimately conclude that the ungrammaticality of the (b) sentences above in (47)-(49) can be explained as a failure to satisfy the intermediate EPP. Part of the evidence for rejecting inter-clausal movement of wh-expletives comes from Hungarian. Recall above that the explanation for the so-called crosslinguistic anti -locality effect of wh -expletives was that wh -expletives originate in a case position inside their clause. Therefore sentences in which the wh-expletive appears in the same case-domain as a full wh-phrase will be correctly predicted to be impossible because of"case-competition" between the expletive and the whphrase. In Hungarian, the base-generated position of the wh-expletive is quite clear, since it bears a morphological indication of the case is assigned. Consider the three-clause question in (57). The verb in the intermediate clause, szamitasz, assigns allative case to the wh-expletive in its clause, resulting in the form mire. The matrix verb hitt assigns accusative case to its wh-expletive, resulting in the form mit.

so

Peripheries in Kashmiri and Hindi-Urdu

(57)

Mit hitt Janos, hogy mire szamitasz, hogy EXPL-ACC thought-INDEF Janos that EXPL-AL count-2SG that mit fognak mondani a gyerekek? [Hungarian] say-INF the kids-NOM what will 'What did John think that you expect that the kids will say?' (Simpson 2000)

Adding complexity to this question, Horvath (2000) and Simpson (2000) report that sentences of the type in (47b), (48b), and (49b), in which the intermediate clause lacks a wh-expletive, are possible in Hungarian, as well as in some varieties of German and possibly dialects of Kashmiri (these varieties will be discussed in detail below). If these sentences were formed via movement of the intermediate wh-expletive into the highest clause, we would expect that the sole wh-expletive in the sentence would be the allative form mire. This prediction is not borne out. (58)

Janos, hogy szamitasz, hogy mit Mit/*mire hitt EXPL-ACc/EXPL-AL thought-INDEF Janos that count-2SG that what fognak mondani a gyerek.ek? [Hungarian) say-INF the kids-NOM will 'What did John think that you expect that the kids will say?' (Simpson 2000)

Instead, it is the accusative form of the expletive mit that is grammatical, as is apparent in (58). If the expletive that ultimately appears only in the matrix clause originated in the intermediate clause, it would first have had to receive allative case from szamitasz, then accusative case from hitt. It seems clear that this wh-expletive was base-generated in a case position in the matrix clause. Simpson takes this to indicate that wh-expletives cannot move from one clause to another. We can also find support for this position in a variety of Kashmiri (let us call it Kashmiri B) which permits sentences like (49b). Let us compare the sentence with a wh-expletive in each clause, with the sentence that lacks a wh-expletive in the intermediate clause (presented here as (60), and marked grammatical). (59)

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

(60)

Ram-an k'aa von ki tse chu-y baasaan Ram-ERG EXPL say.PST that 2SG.DAT AUX-DAT think

[Kashmiri A]

Chapter 3. Full and partial wh- movement in Kashmiri ki mohn-as kam' di-ts kitaab that Mohan-DAT who.ERG give-PST.FSG book 'Who did Ram say you think Mohan gave the book to?' [Kashmiri B] [JC, VC (9/8/05), but PK (9/21/04) disagrees with this judgment] In all dialects of Kashmiri, the pronoun preceding the wh-expletive in (59), tse 'you', is interpreted as a topic. Any argument to the left of a wh-phrase in a Kashmiri clause is a topic and would not, for instance, be able to host the focusmarker -ti (Bhatt 1999). In the sentence in (60), on the other hand, tseis interpreted as focused material. It can grammatically host the focus-marker, and cannot be interpreted as a topic [JC, VC 9/8/05]. This distinction would be unexpected if the wh-expletive in the matrix clause in (60) had originated in the intermediate clause and then raised. If this were the derivation of (63), the wh-expletive would have valued the interrogative focus features on C (and no other focus features may simultaneously be present according to the Kashmiri lexicon). This means that no other focus elements could appear (Bhatt 1999). Tse would have to be located in the topic position above the first specifier position of C. Movement of the whexpletive to the matrix clause should not change the informational role of tse. For this reason (60) serves as further evidence that wh-expletives do not move out of the clause in which they originated. Under the feature-based analysis presented for wh-expletives in two-clause questions, it is expected that wh-expletives cannot move out of their clause. This is because the wh-expletive has a single uninterpretable [Q] feature. The feature is valued when the wh-expletive moves into the specifier of the C-head in the clause in which it originates because in intermediate clauses in which expletives are generated there is an uninterpretable [Q] feature on the C-head. Since the intermediate C-head in these constructions values all uninterpretable features on the wh-expletive, then the immobility ofwh-expletives is accounted for. The whexpletive becomes inactive and is 'frozen' in place, unable to undergo the operation Move again. 14 In this way, the system of features developed above to account for wh-expletive constructions predicts that wh-expletives cannot move from clause to clause.

14· Just as in the account of (52) found in (55) in the text above, the matrix C head will probe its domain and value its uwh feature with the valued wh feature on the intermediate C head Permitting a probe to interact with a goal whose features have already been valued requires that we assume that Agree can take place when the goal is inactive (though Move cannot). For further perspectives on this, see Pesetsky & Torrego 2007 and Bhatt 2005. This is discussed further in 4.4.1 the text below.

81

82

Peripheries in Kashmiri and Hindi-Urdu

I will for now conclude that wh-expletives may not undergo further raising after they raise to the specifier of a [Q]-bearing head and are thereby rendered inactive. Given this, we can assume that in the ungrammatical (b) sentences in ( 47)-(49) the intermediate wh -expletive is simply missing, and that each will fail due to the unsatisfied EPP in the first embedded clause. As is mentioned above, in Hungarian and some varieties of German and Kashmiri, structures like the (b) sentences In (47)-(49) are grammatical. In these sentences there is apparently no expletive in the intermediate position. Simpson attributes this pattern to phonological deletion of the intermediate expletive, and not expletive movement. However, (60) suggests that this explanation is not plausible. There is only one set of focus features possible on a C head in the lexicon of Kashmiri, and in the intermediate clause in (60) the pronoun tse 'you' has interacted with these features. This means that it is impossible for there to be another element, for instance a wh-expletive, in this position at all, regardless of whether or not it is pronounced. Although this question is not yet resolved, importantly the core pattern, in which the wh-expletive obligatorily appears in all intermediate positions in a multiclausal matrix question, does follow from the account presented here. Throughout this discussion, we have built an account of A-bar movement that is guided by the same general principles that underlie A-movement. The question then becomes why expletive movement is possible in the A-system, but not in the A-bar system. Consider once again (56), repeated here as (61). (61)

There seems to be a problem.

What is different about the DP-expletive the1-e that permits it to move, while wh-expletives may not? My tentative answer to this question here is: nothing. In fact, the DP-expletive and the wh-expletive are roughly equivalent in terms of feature content, each possessing only uninterpretable features. The difference lies instead in the types of heads that exist in the A and A-bar systems. In the A-system, the non-finite T head is often referred to as 'defective'. It has no ability to assign case to a DP or even to fully Agree with its phi-features. In the A-bar system, there is no analogous 'defective' head. Although there are different types of C-heads, all of them have uninterpretable [wh] features which are valued by some [wh] feature in their domain. So under this account there is no A-bar equivalent of the 'defective' T head. We can simultaneously make an empirical observation that there is also no visible equivalent of non-finiteness in the A-bar system - that is, there is no morphological realization of 'defectiveness' like the English word to. For this reason I will assume here that the fact that whexpletives cannot move from one clause to another does not represent a major

Chapter 3. Full and partial wh- movement in Kashmiri

split between the A and A-bar systems, nor does it represent a strong distinction between the two types of expletives. It is also relevant to note here that some researchers have concluded that A-expletives also do not undergo movement (Boskovic 2002), and that the reason for believing in A-expletive raising may only be as strong as believing that the raising T head bears the EPP. Although more research is certainly needed on this point, I will maintain that this supposed distinction does not represent a major difference between A and A-bar expletives themselves.

3·5

Conclusion

The investigation of A-system expletives and their properties has proved vital in the advancement of syntactic theory. This chapter claims that the A-bar system also features expletive elements, and that their investigation can provide similar insight into the nature of heads that mediate long-distance whdependencies. The account presented here claims that A-bar movement, like A-movement, is driven by a system of interpretable and uninterpretable features. The wh-expletive is a semantically empty element that satisfies the requirements of certain heads that phonologically overt material appear in their specifier. This permits that head to then interact over some distance with contentful wh-material. The resulting wh-expletive construction is one in which the meaningful wh-phrase appears at the edge of an embedded clause, while the wh-expletive appears in the position at which the scope of the whphrase will be interpreted. This analysis of long-distance wh-dependencies furthers the larger effort of this book: to understand the nature of the clausal periphery. In the view developed in this chapter, the functional head that defines the phase is that which mediates wh-dependencies and hosting wh-expletive element. The periphery is then the edge of a derivational unit, and its properties permit material to appear displaced from that unit, or to mimic the interpretive effects of that displacement through a static dependency. While the account presented in this chapter covers significant empirical ground in Kashmiri, it makes no use of devices particular to the construction under analysis here. Instead the mechanisms proposed apply to all of A-bar movement, and in fact reveal the symmetry of design between the A and A-bar systems. I now turn to test this approach on the long distance whdependencies of Hindi-Urdu, a related language that differs from Kashmiri in crucial ways.

83

84 Peripheries in Kashmiri and Hindi-Urdu

Appendix: C heads in the lexicon of Kashmiri Prom the diagrams in (56)-(58) above we see that there are three posslble interrogative feature bundles that can appear on a C-head in a three-clause matrix question in Kashmiri. The presence of the feature bundle that contains [IQ] is determined by the selecting predicate (or in matrix position). However, other predicates (such as the equivalents of think or say) could in princtple select a C-head with either of the two remaining bundles. These two bundles differ in that one contains [uQ] and one does not. The following schematics depict the three C-heads in a threeclause matrix wh -question in which the full wh-phrase originates in the lowest clause. They lllustrate that ifthe incorrect feature bundle is chosen on one of these two lower heads, the derivation will nearly always fall to converge. The two patterns already established above in (56)-(58) are shown in (i)-(ii). 15 (i)

Cl C2 C3 uwh uwh uwh iQ

(see (56)-(57))

EPP EPP

EPP (ii)

Cl C2 C3 uwh uwh uwh iQ

uQ

EPP

EPP

(see (58))

EPP

The feature bundles on the C-heads in (i) permit full wh-movement to the specifier of Cl, or partial wh-movement to the spectfier of C2 with a wh-expletive in the specifier of Cl, but do not permit partial wh-movement only to the specifier ofC3, because the [Q] feature on the full whphrase could not be valued (there is no other [Q] feature on an accessible Probe). The features in (ii) permit only partial wh-movement, to either the specifier ofC2 or C3, with wh-expletives in the specifiers ofhigher C-heads. The patterns in (iii) and (iv) are not discussed in 4.2. In (iii) the C3 head possesses uQ instead of the C2 head. (iii)

C1 C2 C3 uwh uwh uwh iQ EPP uQ

EPP

EPP

This combination of heads will not permit any derivation to converge. The wh-phrase will be frozen in the specifier of C3 because all of its features will have been valued However, a wh-expletive (having only the [uQ] feature) that originates in the clause beneath the C2 head (with no [Q] feature) will not be attracted by that head because they have no matching features. M. predicted, no derivation will converge given this selection. It is also posslble that both the C2 and C3 heads could have a [uQ] feature, as in (iv).

Clearly this list of features is not exhaustive; only the relevant features are displayed for the sake of illustration.

15.

Chapter 3. Full and partial wh- movement in Kashmiri (lv)

Cl uwh iQ EPP

C2 uwh uQ EPP

C3 uwh uQ EPP

In this case, the wh-phrase originating in the lowest clause will be frozen in the specifier of C3 because all of its features will have been valued So wh-movement to either the specifier of C1 or C2 will not be possible. However, the construction with a wh -expletive in the specifier of C1 and C2 is possible, as no features will remain unvalued This is the only unexpected converging derivation. It is worth considering whether this system ofheads would permit constructions in which there are only wh-expletives in every clause (no full wh-phrases). This would not be possible because the [uwh] feature on C3 head in (i)-(ii) would not be valued, since the expletive has no [wh] feature. This derivation would fall to converge. So too would a derivation in which full wh-phrases instead of wh -expletives occupied the specifiers of C 1, C2, and C3, though the intended interpretation was the same as a canonical wh-explettve construction. Assuming the heads were either like those in (i) or (ii) the [uQ] on the wh -phrase in the specifier of C3 would not be valued, because either the C2 head will not have its own [Q] feature (as in (i)), or the C2 Probe will be inactive before probing the wh-phrase in the specifier of C3, because it will have been valued through interaction with the wh-phrase in its own clause. In either case, this type of derivation also will faU to converge. The important thing to conclude from the exercise is that in this view the feature bundles that appear on the relevant heads permit all and only the family of grammatical constructions, whUe making no use of stipulations or devices particular to the construction under analysis here.

85

CHAPTER4

Wh-expletives in Hindi-Urdu*

In the preceding chapters of this book I have investigated the left periphery of the Kashmiri clause and have proposed an understanding of the structure and organization of the C domain. This approach successfully accounts for the rich range of elements that can be found on the periphery, as well as the way in which long-distance A-bar dependencies are mediated by the peripheral head. At the same time, the account proposed above allows us to reconcile the many roles of the clausal periphery with the theoretical notion of the phase. We can then understand the role that the derivational unit and the properties ofits edge play in A -bar displacement and A-bar dependencies. However, this newly synthesized view of the phase-defining head at the clause edge forces an important question: what about languages in which wh-dependencies and discourse-marked elements do not appear at the clausal periphery, but instead in some embedded position? Earlier treatments of this contrast suggested that the role of the clause edge in the overt syntax was parametrized and that in some languages, displacement of A-bar constituents occurred at a post-syntactic, and hence covert leveL In this chapter I will advance an alternative view - that there is an additional fixed position clause-internally for A-bar material, at the edge of the verbal domain. Investigating this clause internal periphery reveals that the properties previously ascribed only to the head C, that is, only to the clause edge, may in fact be properties of the edge of the dervivational unit. In particular, we should expect to find languages in which the edge of the vP phase serves the same function as the edge of the CP phase, as a position mediating A-bar movement and A-bar dependencies. Hindi- Urdu is a language closely related to Kashmiri that has been traditionally understood to be wh-in-situ. However, Hindi-Urdu exhibits long-distance wh-dependencies that closely resemble those in Kashmiri, including multi-clausal wh-expletive constructions. In what follows I will claim that a microcomparative

* Many components of this chapter appeared in an article in Linguistic Inquiry (41: 1) entitled "Wh-expletives in Hindi-Urdu: the vP Phase" (Copyright 2010 MIT Press). I am grateful to two anonymous LI reviewers whose comments have substantially enriched this work. Of course, all remaining errors are my own.

88

Peripheries in Kashmiri and Hindi-Urdu analysis of A-bar dependencies in Kashmiri and Hindi-Urdu have much to tell us about the nature and structure of the edges of derivational units, as well as the role of the properties of the periphery in determining crosslinguistc variation.

4.1

Introduction

In this section I will introduce wh-displacement and wh-expletive constructions in the Indic language Hindi-Urdu. I will consider these constructions in HindiUrdu alongside those in Kashmiri, and suggest that a unified account of the various strategies of forming long-distance wh-dependencies in the two languages can be constructed. There are two ways to form a long-distance wh-dependency in Hindi- Urdu. In the first, depicted in (1), the full wh-phrase that originated in the subordinate clause has been displaced into the matrix clause, and is found in the canonical preverbal wh-position. The semantic correlate of this configuration is interpretation as a root question: (1)

ki Sita-ne kis-ko soc-aa Sita-ERG who-Ace think-PRF.M that Ravli-ne Ravi-ERG

dekh-aa? see-PRF.M

[Hindi- Urdu]

'Who did Sita think that Ravi saw?' In the second way of forming a long-distance wh-dependency, depicted in (2), the full wh-phrase appears in the preverbal wh-position in the subordinate clause. In the matrix clause a minimal wh-word kyaa appears in the preverbal position. The result is also a matrix reading: (2)

Sita-ne kyaa soc-aa ki Sita-ERG EXPL think.-PRF.M that Ravli-ne kis-ko dekh-aa? Ravi-ERG who-Ace see-PRF.M

[Hindi- Urdu]

'Who did Sita think that Ravi saw?' Superficially, ( 1) -( 2) might appear very similar to long -distance wh -dependencies in Kashmiri, repeated in (3)-(4). However, keep in mind when considering these data that Hindi-Urdu is typically analyzed as a verb final, wh -in-situ language, and Kashmiri as a verb-second language with full wh-movement. 1

t. As discussed in Chapter 2, the exception to verb-second order in Kashmiri is when a wh -word appears immediately before the second-position verb. In this case a DP interpreted

Chapter 4. Wh-expletives in Hindi-Urdu 89

(3)

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 [Kashmiri] 'Who do you think gave Mohan the book?' (PK 9/21/04)

(4)

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 [Kashmiri] 'Who do you think gave Mohan the book?' (Wall & Koul: 18)

The earliest treatments of wh-movement assumed that such dependencies were of a potentially 'unbounded' span. Later in the course of research on wh-movement, it was argued that constituents were required to move through a sequence of intermediate positions, specifically through the CP layer of each containing clause. Recent research has questioned whether CP is the only intermediate landing site. The concept of the phase (Chomsky 2000) affords equal status to CP and vP, and therefore suggests that the edge of vP should be a forced stopping point for long-distance movement as well. An observation supporting the local view ofwh-movement is that expletive elements in certain languages can be found in exactly those positions proposed as intermediate stopping-off points, specifically in the specifier of CP. If this is an accurate understanding of wh-movement and its association with the appearance of expletives, then we should expect similar effects wherever intermediate stopping-off points are postulated- in particular, at the edge of vP. To develop an understanding of constructions of the type (1)-(2) in HindiUrdu, I will bring together three strands of work on wh-dependencies. The first strand of work is the argument pursued in Chapter 3 that wh-expletives, much like expletives in the A-system, are featurally deficient elements in that they possess only uninterpretable features, lacking an interpretation (Chomsky 2000). When these wh-expletives are merged into the specifier of a functional head, the wh-features on that head are then left free to Agree with an unraised wh-phrase in its domain. I will claim here that Hindi-Urdu wh-expletives, like those in Kashmiri, have precisely these properties and effects. The second strand of research relevant here is a specific view of the vP phrase. This strand argues, as mentioned above, that insofar as vP is a phase and v a phase-defining head, the specifier of vP may be as crucial an intermediate

as a contrastive topic may also appear preceding the wh-word, literally creating verb-third order (Wall & Koul1997). For further analysis of verb-second in Kashnllri see Bhatt (1999 ).

90

Peripheries in Kashmiri and Hindi- Urdu

stopping off point in the course of long-distance wh-movement as the specifier of CP (Chomsky 2008, Rackowski & Richards 2005, Bruening 2007). That is, the v head has features relevant to wh-movement, and any wh-phrase occurring within the vP phase must move first to the phase edge (Spec, vP) before interacting with any higher head. Finally, the third strand of research that will prove to be of interest also concerns the heads that are active in wh-movement. The account proposed in this chapter will lend support to the proposal that in some significant part, intralanguage variation can be attributed to the featural properties of the phase-defining heads (Chomsky 2008). That is, it is the organization of wh-related features on specifically the phase-defining functional heads C and v which determine many of the characteristics of questions crosslinguistically. The central claim of this chapter is that the sets of features driving whmovement and wh-expletive constructions in the two languages are quite similar. The contrasts between the two will be accounted for as a difference in the properties of the phase-defining heads in each language. I will argue that while the proposals introduced here incorporate many aspects of previous approaches, they provide a better overall understanding of the facts internal to Hindi-Urdu, allow a better understanding of the contrasts with Kashmiri, and are better integrated theoretically. I will also emphasize that the present account offers a solution to the longstanding puzzle ofwh-in-situ in Hindi-Urdu- why the wh-expletive construction exists at all.

4.2

Wh-dependencies in lllndi-Urdu and Kashmiri

Unmarked word order in Hindi-Urdu is verb final, and in a transitive sentence the subject typically precedes the object: (5)

Hamid-ne pani pi-ya. Hamid-ERG water drink-PRF.M 'Hamid drank water:

[Hindi- Urdu]

(6)

Bacci-ne mehmaan-ko phul pesh ki-ye. child-ERG guest-Ace flowers present do-PRF.PL 'The child presented flowers to the guest'

[Hindi-Urdu] (Schmitt 1999: 188)

In this section I will explicity contrast the properties of Kashmiri with those of Hindi-Urdu, with a view toward advancing our understanding the role of the periphery in intra-language variation. As discussed in preceding chapters, Kashmiri is a verb-second language, with a variety of phrase types potentially

Chapter 4. Wh-expletives in Hindi-Urdu

appearing before the verb. (7a) represents the unmarked word order, while (7b-d) are also grammatical (Wali & Koul, 1997:89). (7)

a.

Aslam-an dits Mohn-as kitaab Raam-ini aslam-ERG gave Mohan-DAT book Ram-DAT khatrl raath. for yesterday

[Kashmiri]

'Aslam gave Mohan a book for Ram yesterdaY: b.

Mohn-as dits Aslam-an kitaab Ram-ini khatrl raath.

[Kashmiri]

c.

Kitaab dits Aslam-an Mohn-as Ram-ini khatri raath.

[Kashmiri]

d.

Raath dits Aslam-an Mohn-as kitaab Raam-ini khatrl.

[Kashmiri]

Hindi-Urdu is often described as a wh-in-situ language. More accurately. in constituent questions, the unmarked position for the interrogative phrase is immediately before the sentence-final verb, regardless of the grammatical role that it bears (Bhatt 2003, Schmidt 1999, Kidwai 2000): (8)

Kitaab-ko kis-ne paRh-aa book-ACC WhO-ERG read-PRF.M 'Who read the book?'

[Hindi- Urdu]

(9)

a.

Hamid-ne kyaa paRh-aa? Hamid-ERG what read-PRF.M 'What did Hamid read?'

[Hindi- Urdu]

b.

haJ.."? Abhi kis-ko dekh-ta nowwho-ACC look-HAB.M AUX.PRS 'Who are you looking at now?'

[Hindi- Urdu] (3/31/06, Paklinks)2

Hindi-Urdu is a language that permits relatively free scrambling of constituents, so for examples like (5)-(6) and (8)-(9) a variety of other word orders are possible, bearing various interpretations. Though they have been extensively investigated (Mahajan 1990, 1994, Dayal1994, Kidwai 2000) these alternatives will be of less interest to us here than the unmarked order. In Section 4.4.2 I will address the role played by scrambling in the formation ofwh-questions in Hindi-Urdu. Under the analysis ofwh-movement presented in this chapter, the subject wh-phrase is located in Spec, vP in (8). In addition, the preferred word

2.

(http://www.paldinks.com/gs/archive/index.php/t-25313.htrnl)

91

92

Peripheries in Kashmiri and Hindi- Urdu

order in (8) is the result of subsequent movement of the object DP out of the vP. I assume here that this evacuation is the result of scrambling, and discuss it at length below. Kashmiri wh-phrases must immediately precede the second-position verb. A contrastive topic may also appear pre-verbally when a wh-phrase is present (Bhatt 1999): (10)

a.

b.

learn' haa-v Shill-as ruw kitaab raath? [Kashmiri] who.ERG show.PsT-FSG Sheila-DAT new book yesterday (Wall & Koul: 12) 'Who showed a new book to Sheila yesterday?' Raj-an learn-is haa-v nev kitaab? [Kashmiri] Raj-RG who-DAT show.PST-FSG new book (Wall & Koul: 12) 'To whom did Raj show his new book?'

Embedded clauses in Hindi- Urdu are optionally preceded by the subordinating particle ki. Finite embedded clauses appear obligatorily to the right of the verb, although all noun complements appear to the left (Dayal1996): (11)

a.

Vo jaan-tii hai [ki Anu aay-ii]. She know-HAB.F AUX.PRS that Anu come-PRF.F 'She knows that Anu came: aay-ii] jaan-tii he. b. *Vo [ki Anu

[Hindi- Urdu)

[Hindi- Urdu]

There has been much discussion of these clauses in the literature, particularly with respect to their status as complements. It has often been argued that finite clauses appearing to the right of the selecting verb in Hindi-Urdu do not originate in that position, but have somehow been extraposed (Davison 1988) or otherwise adjoined to CP/IP (Dayal 1996, Lahiri 2002). These claims will be addressed in detail in Section 4.5.3 below. It does seem, however, that these clauses behave as though they are c-commanded by material in the matrix clause. For instance, even when a wh-expletive is present in the matrix clause, a quantifier in the matrix clause can bind a pronoun in the embedded clause (Mahajan 2000). If the apparent CP complement were adjoined, this would not be expected since the wh-expletive kyaa. then would be occupying the sole complement position associated with the verb. Bayer (1996) discusses this fact for both Bangia and Hindi-Urdu, noting that while displacement of wh-phrases is possible from what he considers complement CPs, it is impossible for CPs that are clearly adjuncts. I will then maintain that finite clauses are complements in Hindi-Urdu, though they appear to the right of the verb. Additional arguments that support the view that these finite clauses are in fact complements are presented in 4.5.3 and in Manetta (to appear).

Chapter 4. Wh-expletives in Hindi-Urdu 93 Given the facts above, I will assume that the transitive light verb (v) and the VP projection are both head final, with the verb taking nominal and CP complements to the left as in (12): (12)

vP

Sub~ VP

v

~

CP

Verb

~ For sentences with finite clause complements in Hindi-Urdu, I will assume that in the process of post-syntactic linearization, the CP complement must be aligned to the right edge of the clause (Bhatt 2003: 16, Fox & Pesetsky 2005, Sabbagh 2007). 3 When a verb does take a finite complement clause in Hindi-Urdu it is possible for interrogative phrases originating in the embedded clause to take matrix scope using one of two strategies: the question phrase originating in the embedded clause can be displaced, appearing in the main clause, or can remain in its usual position in the embedded clause while the apparently meaningless question word kyaa appears in the main clause (as in (1) and (2)). Athough both Hindi-Urdu and Kashmiri are underlyinglyverb final (Kashmiri, for instance, features verb-final order in non-finite clauses (see Bhatt 1999)), Kashmiri has verb-second word order on the surface. In fact, Kashmiri seems to exhibit the typical profile of a verb-second language, familiar from the extensive literature on this language type among European languages. As addressed in detail in Chapter 2, I have analyzed the second position verb in Kashmiri as located in the C head (see also Bhatt 1999). 4 Wh-material (wh-expletives and wh-phrases)

3· This proposal forces us to wonder how the locality of this kind of linearization can be constrained. Tentatively, we could investigate whether the restrictions that Fox and Pesetsky (2005) and Sabbagh (2007) place on rightward movement maybe adequate (Order Preservation, Landing Sites). Though space constraints prevent us from exploring this further here, were these to apply to post-syntactic linearization, we could account for all of the grammatical positions of the complement, and we could prevent the ungrammatical orders (summary: the complement clause must be right peripheral in its clause, and cannot appear on the right edge of a finite clause that is higher than the finite clause in which it is embedded). I leave this effort to future research. 4 In Kashmiri all finite clauses surface as verb second, including embedded clauses preceeded by a complementizer. It is similar in this respect to Germanic V2 found in Yiddish and Icelandic (Diesing 1990; Bhatt 1999).

94

Peripheries in Kashmiri and Hindi- Urdu

in both Kashmiri and Hindi-Urdu appears in the preverbal position. According to the assumed approach, in Kashmiri this position is an A-bar position above the C head, specifically the specifier of CP. However, it is clear that the preverbal position ofwh-material in Hindi-Urdu must be significantly lower than CP. If the verb itself is within vP in Hindi-Urdu, then wh-material is in an A-bar position relatively close to the verb. At this point, I will introduce a hypothesis that will serve as a major theme for the remainder of this chapter. I will suggest that Hindi-Urdu wh-material is not in-situ, as has usually been claimed, but instead in a distinguished position in the overt syntax which is lower than Spec, CP. I will propose that this position is the specifier of vP. Among those who have argued that the vP phase is a possible stopping-off point for long-distance wh-movement are Rackowski and Richards (2005). I want to begin with this basic intuition: a wide range of facts about Hindi-Urdu and its wh-dependencies can be understood naturally if we take the specifier position of vP to be the position in which Hindi-Urdu whmaterial appears.

4·3

The position of wh-material in Hindi-Urdu

Preverbal position is the unmarked position for wh-material in Hindi-Urdu, irrespective of grammatical role or argument status of the wh-phrase, as shown in (8)-(9) above (Bhatt 2003, Schmidt 1999, Kidwai 2000). Since this is not consistent with unmarked declarative word order, unmarked interrogative order is evidently the result of obligatory displacement. In this section I will offer empirical evidence in support of the hypothesis that wh-material in Hindi-Urdu is displaced to a fixed position in the syntax, and this position is consistent with Spec, vP. 4.3.1

Focused constituents

The grammaticalized focus position in Hindi-Urdu is the position immediately preceding the verb in linear word order: (13)

MaiN-ne kamre-meN [in-hii tiin laRk-oN-ko] bhej-aa. [Hindi-Urdu] lsG-ERG room-to [these-Foe three boy-PL-Ace] sent-PRF.M 'I sent these three boys to the room: (Butt & King 1996)

(14)

Kitab-eN kal maiN lay-aa thaa. [Hindi-Urdu] book-PL yesterday lSG bring-PRF.M AUX.PST.M 'I brought the books yesterday (It is I who brought the books yesterday)' (Kidwai 2000)

Chapter 4. Wh-expletives in Hindi-Urdu

There is thus a fixed syntactic position for focused constituents, whether interrogative or non-interrogative. 5• 6 Previous work assumes a dedicated focus phrase dominating VP (Kidwai 1995). I suggest here that the word order pattern is consistent with that fixed position being the specitler of vP in Hindi-Urdu. Analogously, in Kashmiri the position for both interrogative and noninterrogative focus is the same. Both types of constituents appear in the preverbal position. The Kashmiri focus particle -ti can only appear suffixed to constituents in this immediately preverbal position (all Bhatt, 1999): (15)

Huun-ti ch-u behna broNh panin jaay goD dog-Foe AUX-PRS.M seat before self's place first saaf kar-aan. clean do- IMPRFV 'Even the dog cleans his place before sitting.'

[Kashmiri]

5· It is worth speculating whether the most deeply embedded XP in VP bears focus (similar to the Nuclear Stress Rule for Romance languages, though this precise requirement cannot necessarily hold in syllable-timed Hindi-Urdu (Kidwai 1999:240)). However, preverbal constituents in sentences with canonical linear order need not be focused, and focused phrases can appear outside the preverbal focus position provided they are marked with heavy (contrastive) stress and/or the suffix -hii. I assume here that after a focused constituent has valued its features in the specifier position of vP, it may subsequently scramble (also see Bhatt 2003). While there are alternative approaches to focus not making use of the specifier of a fixed functional projection in the syntax, the account presented here overcomes one of the primary objections to the syntactic view (Kidwai 1999). It provides a systematic understanding of the way in which languages exhibit similar positional requirements for focus at different projections (Le. CP in Kashmiri and vP in Hindi-Urdu). Further, the proposed analysis of wh -in-situ as an instance of the operation Agree could potentially be expanded to account for focus in languages without a positional focus requirement 6. Simspon and Bhattacharya (2003, fn 3) point out Mahajan's suggestion that the lack of normal question interpretation for wh-phrases in postverbal position in monodausal sentences (Mahajan 1997:209, fn 9) supports the view that these elements have moved to a fixed position in overt syntax. (i)

Savita-ne kitaab d-li kis-ko? Savita-ERG book give-PRF.F who-DAT 'Savita gave a book to WHO?'

The absence of matrix question interpretation falls out from the account presented in Section 5.2, since the wh-phrase in (i) is too distant to interact with the matrix C head (and, for that matter, will not satisfy the EPP on the v head).

95

96

Peripheries in Kashmiri and Hindi-Urdu

(16) ?*panin jaay ch-u huun-ti behna broNh goD self's place AUX-3MSG dog-Foe seat before first saaf karaan clean do-IMPFV

[Kashmiri]

Intended: 'Even the dog cleans his place before sitting.' (17) *Gary-ti kus ch-u-na kaam kar-aan. home-FOC who AUX-3MSG-NEG work do-IMPFV Intended: 'Who doesn't work even at home?'

[Kashmiri)

In the chapters above I developed an account in which both interrogative and non-interrogative focused constituents appear in Spec, CP in Kashmiri. It appears that focused constituents, whether interrogative or not, preferentially appear preverbally in Hindi-Urdu and Kashmiri. 7 The account presented here attributes the position of focused material in each language to the properties of the phase-defining heads C and v, respectively. In effect, what we see is the same clausal topology in Kashmiri and Hindi-Urdu, lower in Hindi-Urdu (on the v head) and higher in Kashmiri (on the C head). 4.3.2

Adverbs

Adverbs that are typically analyzed as adjoining to vP, such as 'always', can appear before the verb and the direct object or wh-phrase in an unmarked Hindi-Urdu sentence as in (18-19) (Schmidt 1999: 190). However, it is also possible for the adverb to follow the question word, as in (20).

7· A reviewer asks whether the postverbal position in Hindi-Urdu might also be a focus position, though certainly not as neutral as the preverbal positioiL Consider the question answer pair in (i): (i)

Q:

Raam kitaab kis-ko de-gaa Raam book who-DAT give.FUT.M 'Who will Ram give the book to?'

A:

Raam kitaab de-gaa Mona-ko Raam book give-FUT.M Mona-DAT 'Raam will give a book to Mona:

Garnbhir (1981) and Butt and King (1996) have argued that the postverbal position is not a typical position for new information, and that constituents in this position seem to be interpreted as given, either as a contrastive or continuing topics. Though the syntax of postverbal constituents has been under discussion in the literature in recent years (Bhatt & Dayal2007, Manetta to appear), relatively little wolk has been done on the discourse function of rightmost DPs. More systematic research is needed to determine whether this might also be some kind of emphatic focus position in Hindi-Urdu.

Chapter 4. Wh-expletives in Hindi-Urdu 97 (18)

Vo mujhe hamesha cai pil-aa-ta hai. 3SG 1SG.DAT always tea drink-CAUSE-PRF.M AUX.PRS 'He always has me drink tea~

[Hindi-Urdu]

(19)

Vo aap-ko hamesha kyaa pil-aa-ta hai? 3SG 2PL-ACC always what drink-CAUSE-PRF.M AUX 'What does he always have you drink?'

[Hindi-Urdu]

(20)

Vo aap-ko kyaa hamesha pil-aa-taa hai? 3SG 2PL-ACC what always drink-CAUSE-PRF.M AUX.PRS 'What does he always have you drink?'

(Schmidt 1999) [Hindi-Urdu]

One could argue that in (18) and (19), both the direct object and the interrogative pronoun could be in their base positions. However in (20) it appears that the whword cannot be in-situ. These observations fall into place naturally on our proposal that in both (19) and (20), the wh-word has moved to the specifier of vP. If the positioning requirement for this class of adverbs is that they attach to a complete vP, that requirement can be met at two derivational points on our assumptions: either (i) after both specifiers have been introduced, or (ii) after the external argument has been introduced but before the second specifier (the wh-element) has been merged. The first alternative yields (19), the second (20).8 It is unclear how we might account for (20) on the assumption that the whphrase is in-situ. However, some might object that (20) is simply evidence of scrambling of the wh-phrase out of its (in-situ) position as the complement to V, and into some higher position. In essence, this is not substantively different from what I propose here- thatthewh-phrase arrives at its surface position in (19) and (20) through movement out of the VP. However, there is another, related fact which indicates that the position to which kya 'what' has moved in (20) has some special status. Chandra (2005), citing Mahajan (1990), claims that when an adverb like jaldise 'quickly' appears following a non-agreeing direct object, the object requires focal stress: (21)

Sita kaam jaldise kar-tii thii. sita work quickly do-PRF.F AUX.PST.F 'Sita worked quickly.'

[Hindi- Urdu]

A comparison of (19) and (20) might be taken to suggest that wh-movement to Spec, vP is optional in Hindi-Urdu. However, recall that unmarked interrogative word order is not consistent with declarative word order except in the case of the direct object (which is preverbal in declarative structures). Since it would be both stipulative and undesirable to encode such a distinction for wh-movement, we can safely assume that all basic wh -questions in Hindi-Urdu are derived via displacement, and the overlap between interrogative and declarative word order in the case of (19) is incidentaL 8.

98

Peripheries in Kashmiri and Hindi-Urdu

This is in contrast to (22), in which kaam 'work' need not receive focal stress (see also Schmidt 1999: 190). (22)

sita jaldise kaam kar-tii thii. Sita quickly work do-PRF.F AUX.PST.F 'Sita worked quickly.'

[Hindi- Urdu)

Under the account presented here, kam 'work' in (21) is a focused phrase appearing in Spec, vP. The adverb can be adjoined before the object has shifted here for the purposes of focus. This suggests that when a wh-phrase or direct object appears to the immediate left of the vP-adjoined adverb it has been displaced to a specific A-bar position at which it checks focus-related features (interrogative or non-interrogative).9 I will take this evidence to support a correlation between the position of interrogative and non-interrogative focused constituents, and that that position is consistent with an A-bar position on the edge of vP (Spec, vP) in Hindi-Urdu (for similar argumentation concerning wh -expletives, see Malhotra & Chandra 2007). This data contrasts sharply with the facts in Kashmiri, in which these adverbs follow the second-position verb: (23)

Akhk-is slf ch-i hameeshi lad-aan. [Kashmiri] fight-IMPFV one another-OAT with AUX-3MPL always '(They) always fight with each other.' (Wali & Koul: 133)

(24)

Tam-is nishi as hameesh I nookar. 3sG- DAT near have.PRS always servants 'He always has servants.'

[Kashmiri] (Wali & Koul: 140)

In Kashmiri, the verb is clearly in a position higher than the edge of vP (analyzed here as C). Because wh-phrases and focused constituents always appear immediately before the second-position verb (in Spec, CP), they are never found immediately adjacent to vP-adjoined adverbs. No other permutations of these orders are possible in Kashmiri. We have now seen a set of systematic contrasts between Hindi- Urdu and Kashmiri, which allow us to better understand the position of wh-material in Hindi- Urdu. The facts in 4.3.1 and 4.3.2 would be surprising indeed if Hindi-Urdu were a true wh-in-situ language. Each data set suggests that there is a distinguished

9· As Chandra points out, this is not necessarily the case for agreeing or case-marked (hence specific) objects. In Chandra's account these objects shift for purposes of agreement/case and not focus. This does not affect the argument here since we are assessing only whether the position for interrogative focus and non-interrogative grammatical focus is the same, and is plausibly Spec, vP.

Chapter 4. Wh-expletives in Hindi-Urdu

position for interrogative and non-interrogative focus in Hindi-Urdu, and that this position is consistent with an A-bar position at the edge of vP (Spec, vP). The comparisons here highlight the difference in the relative positions of wh-material in Hindi-Urdu and Kashmiri, and suggest that the same clausal topology exists in both languages, with the difierence being that the specifier of CP in Kashmiri plays the role that is played by the specifier of vP in Hindi- Urdu.

44 44.1

A-bar movement in Hindi-Urdu: Extending an account of Kashmiri Kashmiri wh-dependencies

In the chapters above, I claimed that 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 Agree to satisfyuninterpretable features, just as in the A-movement system. Heads and wh-phrases possess sets of interpretable and uninterpretable features. If a wh-expletive happens to be in the numeration, it can merge to satisfy the EPP on a head. Much like a DP-expletive, the wh-expletive is defective in that it has no interpretable features of its own, so the head into whose specifier it has merged must value its uninterpretable features by interacting with an accessible wh-phrase via Agree over some distance. Reviewing the essential aspects of the account proposed for Kashmiri in Chapter 3, there are three features controlling movement and agreement in the interrogative A-bar system: the EPP (common to the A and A-bar movement systems), the [Q] feature, and the [wh] feature. Importantly, it is the interpretable [Q] feature that introduces an unselective binder of (choice) function variables (as in Reinhart 1998). The operations Agree and Move are limited to the phase in the A-bar system, just as they are in the A-system. The two phases considered here are the CP-phase and the vP phase. In the phase theory of movement (Chomsky 2000, building on Fox 2000 and Nissenbaum 2000) the v head defines a phase for the purposes of A-movement. That is, an argument within the vP may shift to the edge of the vP-phase (Spec, vP), and may subsequently interact with higher heads, such as T. Current work is exploring the role of the vP-phase in A-bar movement (Rack.owski & Richards 2005), and this strand of research will ultimately help to provide an understanding of Hindi-Urdu. From a comparative perspective, we will see that certain properties of Kashmiri and Hindi-Urdu can be attributed specifically to the properties of these two phase-defining heads. In what follows I will recast the account ofKashmiri presented in Chapter 3 to include the role played by the vP phase in wh-movement and wh-expletive constructions. This will provide

99

100

Peripheries in Kashmiri and Hindi-Urdu

us with the analytical tools to understand the contrasts between Kashmiri and Hindi- Urdu discussed here. No intermediate probes in Kashmiri, whether they are subordinate C heads, or v heads, will bear an interpretable [Q] feature, which appears only at the point referred to as the "scope position". Instead, intermediate probes will have an uninterpretable [wh] feature and an EPP property. This will cause them to be active probes that require the wh-material to appear in their respective specifiers. As discussed in Chapter 3, this EPP must be EPP [Q]' meaning that it can only be satisfied by wh-material, since all wh-elements bear an interrogative [Q] feature. Again, I will continue to refer to it as EPP throughout. In a clause that is embedded, the wh-phrase may raise into the root clause, as in (3), repeated here.1o (3)

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?' [Kashmiri]

In a sentence like (3), the subordinate C and v head will lack an interpretable [Q] feature. A wh-phrase that has raised to the specifier of the subordinate CP will still have uninterpretable features that require valuing. The matrix v will be much like the subordinate C, and will cause the wh-phrase to undergo Move into its specifier in the matrix clause. The matrix C head will have an uninterpretable [wh] feature, the EPP property, and the interpretable [Q] feature. As a probe, it will find the wh-phrase in the specifier of the matrix v and will enter into an Agree relation with this wh-phrase and attract it to its specifier. The wh-phrase will raise to the specifier of the matrix CP, and all features will be valued on the wh-phrase and all heads. The result will be full extraction. This process and the features involved are depicted in (25). Beneath each head are the features present on that head: (25)

[ep wh-XP [C ... ] [vP _ (uQ] [uwh] [iwh] [iQ] [EPP]

[v ... ] [ep _ [uwh] [EPP]

[C ... ] [vP _ [uwh] [EPP]

[v ]])]] [uwh] [EPP]

(25) represents the extraction of the wh-XP and its passage through the specifier positions of each phase-defining head until reaching the matrix CP.

10. Wall and Koul (1997) note that many speakers prefer the wh -expletive construction over long extraction, however the long extractions are accepted, and are malked by Wali and Koul from marginal (??) to grammatical (no marking) (pp. 19-20). The native-speaker informants I consulted judged (3) and sentences like it to be grammaticaL

Chapter 4. Wh-expletives in Hindi-Urdu

The derivation will not be licit if the uninterpretable feature of the wh-phrase is simply valued by the matrix C head via Agree while the full wh-phrase remains in the specifier of the subordinate CP. This is because, in the absence of a whexpletive in the numeration, the EPP property of higher probes (v and C) will not be satisfied. Let us now turn to a grammatical instance of partial wh-movement in Kashmiri as in (4), repeated here. (4)

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.FSG book 'Who do you think gave Mohan the book?' (Wali & Koul: 18)

[Kashmiri]

In a partial movement construction, the subordinate C head will have only uninterpretable features: an uninterpretable [wh] feature, an uninterpretable [Q] feature, and the [EPP] property. In the case of (4), the numeration happens to contain the wh-expletive k'aa. This expletive differs from a full wh-phrase in that it consists entirely of uninterpretable features, and contributes to the syntactic computation only an (uninterpretable) interrogative feature, written here as [uQ] (thus limiting the wh-expletive to constructions which will be interpreted as questions). The understanding of wh-expletives developed in the preceding chapter takes seriously the notion that they are the A-bar movement system counterparts to expletives of the A-movement system. That is, they do not contribute any interpretable features to the derivation. As a result of the uniterpretable interrogative feature present on the wh-expletive, it will move through the specifiers of the phase-defining heads until its feature is valued. Of course, though the EPP is satisfied on both the matrix v and C heads by the wh-expletive in the derivation schematized in (26), theirwh-features (uninterpretable) must still be valued via Agree. In this case the untinterpretable [wh] feature on v interacts with the interpretable [wh] feature on the full wh-phrase at the edge of the embedded CP. In turn, the uninterpretable [wh] feature on the matrix Cis valued by interacting with the (now valued) [wh] feature on v, which is available because it has not yet transitioned to the interface for deletion. 11 In this way; all uninterpretable features are valued, and wh-expletives in the A-bar system serve the same purpose as expletives in the

n. Permitting a Probe to interact with a Goal whose features have already been valued requires that we assume that Agree can take place when the Goal is inactive (though not Move). For further perspectives on this see Pesetsky and Torrego (2007) and Bhatt (2005).

101

102

Peripheries in Kashmiri and Hindi-Urdu

A-system - to satisfy the EPP and permit the head's features to be valued by some other accessible element (Simpson 2000). (26)

[ep wh-expl [C ... HvP _ [uQ] [uwh] [iQ]

[EPP]

[v.. _ . Hcp wh-XP [C ... HvP _ [uwh] [uQ] [uwh] [EPP] [iwh] [uQ] [EPP]

[ v ]]]]) [uwh] [EPP]

It may seem, in (25) and (26) above, that the vP phases play little role in longdistance A-bar movement in Kashmiri. For instance, we never see wh-material remain in the specifier of vP. This is attributed to two factors. First, the C head in Kashmiri always has the EPP (reflected in the tact that wh-material always precedes the second position verb). In a construction with no wh-expletive, the whphrase that has moved into matrix Spec, vP must then interact with the matrix C-head to have its uninterpretable [Q]-feature valued. This interaction will trigger an application of of Move, since C always has the EPP in Kashmiri. The second factor that prevents wh-material from appearing in Spec, vP in Kashmiri, argued in detail in Chapter 3, is that the wh-expletive originates in a position below v where it can be assigned case. This analysis follows Simpson's (2000) claim 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 wh-expletive. I have therefore claimed here that the wh -expletive is base generated in a position in which it can have accusative case valued; that is, within the c-command domain of the accusative case licenser transitive v. Since the wh-expletive cannot be introduced by semantic selection, we know that it must be merged into the specifier of a head that has the EPP property. Aspect is a functional head in the c-command domain ofv on which it would be reasonable to posit the EPP property. 12 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 v is introduced, it will interact with the wh-expletive and the expletive's uninterpretable accusative case feature and [D] feature will be valued. In this view, if there is more than one potential goal such as an additional object wh-phrase or a noninterrogative clausal expletive, the uninterpretable case feature on one or the other will go unvalued because only one goal can interact with the v head, and the derivation will fail to converge. I will suggest that the wh-expletive kyaa in Hindi-Urdu is base-generated in a similar manner in 4.5.2 below.

12.

For discussion of the Aspect phrase below vP in Hindi-Urdu, see Davison (2003).

Chapter 4. Wh-expletives in Hindi-Urdu 103

In the matrix clause of a two-clause wh-expletive construction in Kashmiri, the wh-expletive will first move into Spec, vP to value the EPP property on v, and must then move to Spec, CP for the reasons discussed above. If this is so, the partially-moved full wh-phrase on the edge of the embedded clause will never have the opportunity to move into Spec, vP (the wh-expletive has already valued the EPP property on v). As discussed in the preceding chapter, this derives the so-called anti-locality property of wh-expletives; a wh-expletive cannot appear in the same clause as the full wh-phrase whose position of interpretation it indicates (Simpson 2000, McDaniel1989)P Interestingly, the primary thrust of the analysis to follow is that in Hindi-Urdu it is in the intermediate vPs that we see partially-moved wh-phrases. 44.2

Extending the proposed account to Hindi-Urdu

Hindi-Urdu has traditionally been viewed as a wh-in-situ language. That is, whmaterial was not understood to occupy a distinguished position in the overt syntax. Previous approaches to Hindi-Urdu and similar languages (Mahajan 1990) proposed that wh-material be licensed and raise to a position of interpretation at the level of Logical Form (LF), via movement to the specifier of CP. This movement occurs covertly, following the operations in the syntax, and is therefore never visible in the surface representation. An alternative to this view presents itself in more recent work in the form of the operation Agree (Chomsky 2000, see also Bhatt 2005), permitting the features of a wh-phrase to be valued under local command in the narrow syntax, instead of forcing movement to Spec, CP. We can assume that C bears the interpretable [ Q] feature in a main clause question, but that this head need not necessarily possess the EPP property, so that the wh-word need not move to a C-peripheral position. Further, if wh-phrases have the semantics of indefinites unselectively bound by interrogative operators, there is no reason to think that they must raise, for interpretive reasons, to a "scope position': I will continue to use that term here (since it is convenient and well-established) but this is purely for exposition. Let us proceed stepwise through a derivation of a simple monoclausal whquestion as in (27), before turning to the long-distance wh-dependencies at

13. This analysis predicts that it is possible to derive the crosslinguistically familiar pattern in which an interrogative sentence comprised of three or more clauses features a wh -expletive in every clause between the clause hosting the wh -phrase and the clause in which the wh -phrase is interpreted. We find precisely this configuration in Kashmiri (see Chapter 3) and HindiUrdu (discussed further in Section 4.5.2).

104

Peripheries in Kashmiri and Hindi-Urdu

hand. The v head probes its domain and interacts with the wh-phrase through the operation Move, simultaneously valuing its uninterpretable [wh] feature and the EPP property. 14 The C head then probes its domain and values its uninterpretable [wh] feature in the local command configuration through the operation Agree. The uninterpretable [ Q] feature on the wh-phrase is simultaneously valued by its interpretable counterpart on C. This operation is represented by the diagram in (28). (27)

Hamid-k.o kis-ne maar-aa? Hamid-Acc who-ERG hit-PRF.M 'Who hit Hamid?'

(28)

[ep C ... hamid-ko [vP kis-ne [vP v [iQ] [uQ] [uwh] [uwh] [iwh] [EPP]

[Hindi- Urdu]

_ _ maar-aa]]

Note that in the derivation of simple, monoclausal wh-questions like (27) (and those in (29) and (31), discussed below) the preferred word order is the result of subsequent movement of the DP arguments that are not wh-phrases out of the vP, to a position that I have not strictly identified in (28). I assume here that this evacuation is the result of scrambling, extensively discussed elsewhere (Mahajan 1990, 1994, Dayal1994, Kidwai 2000). The precise characteristics of this intermediate scrambling operation (scrambling beyond the subject position) are somewhat mysterious and have resisted straightforward analysis. Investigations of this type of scrambling have demonstrated that it may exhibit characteristics of both A and A-bar movement, and researchers have suggested that it is movement to a "mixed" position (Webelhuth 1989), or is movement comprised of multiple steps (Mahajan 1990, 1994). However, what is clear is that this scrambling out of the vP is a necessary part of the formation of an unmarked wh-question. The question of why and how focal stress applies to a constituent in a particular position while other, defocalized constituents appear to shift elsewhere has been an important one crosslinguistically. One approach to the type of movement that vacates the vP in (27) has been to assume some principle requiring alignment of a prosodically prominent position and a focused constituent (for Romance see Cinque 1993, Zubizaretta 1998; for Bengali see Hayes & Lahiri 1991; for Japanese see Ishihara 2001). That is, there is prosodically-motivated movement forced by the need to resolve any mismatch between word-order derived by narrow syntax and the correspondence of focus and prosody. Kidwai (2000) rejects this type of approach

14.

The domain of a probe includes its specifier (Chomsky 2000: 123, Richards 2004).

Chapter 4. Wh-expletives in Hindi-Urdu

for Hind-Urdu, on the grounds that Hindi-Urdu lacks the Nuclear Stress Rule, and with the conceptual concern that this would mean that no causal connection exists between the vP-vacating movement process (scrambling) and the focusing effect that results. Instead, Kidwai introduces a syntactic account that attempts to directly link the evacuation of the vP with the immediately preverbal positioning of interrogative and non-interrogative foci, claiming that scrambling is nonoptional A-bar movement that serves to license or "activate" the preverbal focus position. In her account, it is the scrambling DP that activates a Focus head dominating vP. The Focus projection then hosts the focused constituent, which will appear in the preverbal position. The mechanisms that Kidwai employs to bring this activation about during the syntactic derivation are not available in the current framework, so we cannot adopt this account wholesale. Moreover, the account of wh-displacement and wh-dependencies presented here may fit better conceptually with an account of scrambling in which it is the properties of the interrogative element or interrogative clause that drive vP evacuation, not the scrambling which activates interrogative focus. What is important for us here is not to embark on a wholly new theory of scrambling, as this is outside the scope of this effort, but instead to determine whether there is room to build an account of the type of scrambling in (27) that is consistent with the approach to wh-dependencies proposed here. Notice that some account of this type of scrambling would be essential to derive the unmarked word order even in a wh-in-situ approach to Hindi-Urdu, so this issue is not one particular to the wh-movement approach. It seems to me that it is possible to find common ground between these two kinds of proposals. That is, there is way of understanding the vP-vacating scrambling as feature-driven movement (Miyagawa 1997, 2001, Ko 2007) while still maintaining that this scrambling is at some fundamental level prompted by requirements of the interrogative structure (not the reverse). In fact, it is the notion that both types of movement are driven by properties (features) of functional heads that might permit us to link the two. If this intermediate scrambling targets a functional head like Tense, we could assume that interrogative C heads require (or select for, perhaps in the sense of Muller 2010) a Tense head with features that drive scrambling. This would represent an encoding of the wh-verb-adjacency in the narrow syntax. It is important to establish how this is in principle different from a simple OT-style constraint favoring adjacency between the wh-phrase and the verb, which might appear to obviate the need for wh-movement to a criteria! position in the syntax at all. But crucially. feature-driven wh-displacement in the narrow syntax is not only doing the work of deriving unmarked word order in single interrogative clauses in Hindi-Urdu. It is also providing a coherent

105

106

Peripheries in Kashmiri and Hindi-Urdu

explanation for the ways in which long-distance wh-dependencies are formed in the language, and why those long distance dependencies share so much in common with those in an undisputed wh-movement language like Kashmiri. These puzzles require more than a simple stipulation that wh-material be verbadjacent whenever possible. For instance, such a stipulation has nothing to tell us about why multi-clause wh-dependencies in Hindi- Urdu require an expletive element in each clause between the clause containing the wh-phrase and its point of interpretation, as we will see below. In fact, it has nothing to tell us about whywh-expletives exist at all. In conclusion, while we have not built here a complete account of the broader phenomenon of intermediate leftward scrambling, it seems that there is room to find a formal way of associating a C head with interrogative features with heads that might be responsible for scrambling. Such an account would accomplish three important goals: (a) encoding the notion that for Hindi-Urdu this phenomenon cannot be a solely prosodic process, (b) capturing the intuition that itis the requirements of interrogative structures that drive this scrambling, not the reverse, and (c) maintaining the insights of the cross-linguistic approach to wh-dependencies pursued here. With that in mind, let us now briefly walk through the derivation of monoclausal wh-questions in which the questioned element is a subject (in (29)) or adjunct (in (31)), in order to set the stage for the analysis of long-distance whdependencies below. (29)

Hamid-ne kyaa ciiz dekh-ii Hamid-ERG what thing see-PRRF 'What thing did Hamid see?'

(30)

[ep C ... hamid-ne [vP kyaa ciiz [vP v [iQ] [uwh]

[uQ] [iwh]

_ _ dekh-ii]]

[u.wh] [EPP]

In the case of (30), the features on the v and C probes, as well as on the wh-phrase goal, are valued in precisely the same way as they are in (28). The primary difference here is that it is the subject hamid-ne 'Hamicf. originally merged into Spec, vP, must scramble outside the vP. (31)

Hamid-ne Azi.z-ko kab dekh-taa thaa Hamid-ERG Azi.z-Acc when see-HAB.MSG AUX.PST.M 'When did Hamid see Aziz?'

(32)

[ep C ... hamid-ne Aziz-ko [vP kab [vP v [iQ] [uwh]

[uQ] [iwh]

_ _ _ dekhtaa thaa]] [uwh] [EPP]

Chapter 4. Wh-expletives in Hindi-Urdu

Again, in (32) the valuing of features on interrogative probes and goals takes place in the same manner as described above for (28). 15 However in this case both the subject and object must scramble out of the vP in order to derive the unmarked order. The approach above highlights a primary source ofvariation that distinguishes Hindi-Urdu and Kashmiri. Interrogative C lacks the EPP property in HindiUrdu, but bears the EPP property in Kashmiri. Therefore Kashmiri will exhibit wh-movement to the left edge of the clause but Hindi- Urdu will not. However this apparent advance immediately creates a puzzle for the case of wh-phrases taking scope over more than one clause. Consider again the Hindi-Urdu wh-expletive construction in example (2), repeated here. (2)

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

In the case of (2), how is it that the features on the matrix C get valued? It is clear that the wh-phrase in the lower clause remains clause-internal, just as the whphrase does in (9), and does not move to a left-peripheral position in the clause. If this is the case, the features of the wh-phrase kis-ko are inaccessible to the matrix C Probe, because they are not contained within its phase, or even on the edge of the immediately preceding phase. Recall the Kashmiri example in which it is the interpretable [wh] feature on the wh-phrase in the embedded Spec, CP that values the

There is significant evidence that the class of wh-adverbs must be treated in a more nuanced way. Research in Chinese (Tsai 1994), Slavic languages (Rudin 1988, Boskovic 2000), English, German, and many others (Reinhart 1998, Stepanov 2001, Stepanov & Tsai 2008) indicate that different classes of wh -adverbs behave differently, and that this behavior is subject to a great deal of parametric variation Here, a reviewer asks how this approach to wh-movement and wh-dependencies can handle wh-adverbs adjoined higher than vP. Several possibilities emerge. Stepanov and Tsai (2008) argue that some adverbs typically assumed to adjoin as high as TP or CP in German in fact adjoin within the vP - this is tied to the property of semantic saturation and the fact that the subject can remain within vP in German (much like in HindiUrdu). There is still the question of temporal adverbials, which Stepanov and Tsai assume are licensed via covert LF-movement/Agree with CP. In our account this would translate to the operation Agree with a [Q] -bearing C, and would require that the v not possess interrogative features at all. Of course, if the wh -adverb is introduced outside of the vP-phase entirely, this is perhaps not surprising. Stepanov and Tsai (2008) provide an elaborate set of empirical tests for manner and purpose adverbials and distinct typological predictions about the correlations between interpretation of these adverbs and wh-in-situ constructions. While a more nuanced account of a range of types ofwh-adverbials in Hindi-Urdu (and Kashmiri), their adjunction points, and their relationship to this wider typology is beyond the scope of our investigation here, this is certainly an important area for future research. 15.

107

108

Peripheries in Kashmiri and Hindi-Urdu

uninterpretable [wh] features in the matrix clause. The full wh-phrase in (2) is too deeply embedded for this to be possible, because two phase boundaries intervene (that defined by the matrix v head and that defined by the subordinate C head). The account of wh-in-situ as Agree immediately encounters a challenge here. If the wh-expletive kyaa found in Hindi-Urdu is assumed to have the same characteristics as the wh-expletive found in Kashmiri, it also cannot be the whexpletive in the matrix clause in (2) which values the features on the matrix C. The wh-expletive has only a single, uninterpretable [Q] feature, and no [wh] feature at all. It can only function to satisfy the EPP on a head. So the question remains: how do the features on the full wh-phrase and the features on the matrix C head get valued? There is a deeper question here as well, which concerns the very nature of the wh-expletive. In the case ofKashmiri, the wh-expletive serves the role of satisfying the EPP on the phase defining, [Q]-bearing head. However, it can hardly be the case that the wh-expletive in (2) satisfies the EPP on the matrix C head- it seems instead to be occupying the same preverbal position that full wh-phrases in HindiUrdu occupy. If it can neither value the features on the C head, nor satisfy the EPP on that head, the wh -expletive seems to have no purpose whatsoever. Finally, the Kashmiri wh-expletive construction has also been termed "partial movement': because the full wh-phrase in the lower clause has dearly moved from its base position to the left periphery of that clause. However, it is unclear in what sense (2) is a case of partial movement in Hindi- Urdu. Certainly the wh-phrase in the embedded clause has not moved to the clause edge. These questions suggest that the analysis reviewed in Section 4.4.1 will require adjustment to be extended to Hindi-Urdu. In what follows, I will combine the observations from this section with the evidence examined in Section 3, which indicated that Hindi-Urdu wh-material can be understood as located in the specifier position of vP. 4·5 4·5-1

Wh-dependencies in Hindi-Urdu: The vP phase

Wh-movement in Tagalog: A case for [Q] -bearing v

Among those who have argued that the vP phase is a possible stopping-off point for long-distance wh-movement are Rackowski and Richards (2005). They offer supporting evidence for the phase theory of movement by claiming that the vP is actually the only phase relevant to successive-cyclic wh-movement cross-linguistically. Rackowski and Richards show that specific arguments in Tagalog must move to the specifier of vP to receive the appropriate semantic interpretation. Evidence

Chapter 4. Wh-expletives in Hindi-Urdu

for this claim includes overt morphology on the verb indicating agreement in case with the shifted argument. Arguments that must shift include wh-phrases. Rackowski and Richards compare this shift to object shift in Germanic languages. On this interpretation, in the Tagalog sentence in (33) the verb shows agreement in case with the shifted wh-word (Tagalog data from Rackowski and Richards). (33)

Sino [ang binigy-an ng lalaki ng bulaklak _ ]? who ANG give-DAT cs man cs flower 'Who did the man give the flower to?'

The verb can also agree with a CP complement: (34)

Sa-sabih-in ng kalabaw na masarap ang bulaklak ASP-.say-Acc cs water-buffalo that delicious ANG flower 'A/the water buffalo will say that the flower is delicious'

In the case of long-distance wh-dependencies, in order for extraction to be possible from the embedded clause, the verb must agree with that clause: (35)

Kailan [sasabih-in ng sundalo [na 0-u-uwi ang pangulo e]] when ASP. say-Ace cs soldier that nom-ASP-go.home ANG president 'When will the soldier say that the President will go home?'

(36) *Kailan [m-agsa-sabi ang sundalo [na 0-u-uwi ang pangulo e]] when NOM-ASP-say ANG soldier that NOM-ASP-go.home ANG president 'When will the soldier say that the President will go home?' This fact leads Rackowski and Richards to propose that both v and interrogative C have features that must be valued in the process of wh-movement, and may also possess the EPP property, causing Goals to appear in their specifiers. We have seen this account at work in the analysis of Kashmiri wh-movement and wh-expletive constructions in 4.4.1 above. However, in a significant break with previous approaches, Rackowski and Richards claim that it is unnecessary for non-interrogative C to Agree with any wh-material at all. For example, for a long-distance wh-extraction in English such as in (37), Rackowski and Richards propose the derivation in (38). 16 (37)

Who did you say Obama appointed_?

(38)

[C[QJ [ who v [C [whe v whe]]]]

L_j

Rackowski and Richards (2005) use different names for the features involved in whmovement. I will instead use the feature names used in previous chapters of this book. The outcomes are equivalent

16.

109

no Peripheries in Kashmiri and Hindi- Urdu

In (38) who moves first to the specifier of the lower v. After the matrix v has agreed with the embedded C, the embedded C phase is transparent to the matrix v probe (we will consider this in more detail below). This probe can then find and interact with who, which moves into the matrix clause. In the final step of the derivation, who moves into the specifier of the interrogative matrix C. It is crucial for Rackowski and Richards to assume that once a probe P has Agreed with a clausal goal G, P can ignore G for the rest of the derivation. I will not examine this "transparent goal" claim in more detail at this juncture, though Rackowski and Richards point us to other resources (Richards 1998, Hiraiwa 2001). Rackowski and Richards acknowledge that languages that show evidence for wh-related morphology on intermediate C, or wh-expletives in the specifier of intermediate C, will prove challenging for this view. If non-interrogative CPs are not phases which force movement through their specifiers, we require alternative explanations of these facts. For instance, the morphology of Irish complementizers has been analyzed as indicative of successive cyclic wh-movement through Spec, CP (McCloskey 1990, 2001). Complementizers exhibit a distinguished form if an A-bar binding relation (actually a movement relation) holds between a position within the CP they head and a position external to that CP. Rackowski and Richards claim this morphology simply indicates that the C in question has entered into an Agree relation with a [Q] -bearing v, which has in turn Agreed with a wh-phrase. An approach which depends solely on the specifier of vP as a stopping-off point for long-distance wh-movement will also encounter difficulty in accounting for wh-expletives and instances of partial movement such as those found in languages like German and Kashmiri. Rackowski and Richards acknowledge that in order to understand why wh-material would seem to appear overtly in the specifier of a non-interrogative CP in these languages, they would need to assume a version of the indirect dependency approach. That is, they would need to claim that any CP in which a full wh-phrase or wh-expletive appears is in fact an interrogative CP, and that the scope properties of that wh-phrase are a result of a complex process of coinde:xation and/or covert clausal pied-piping. Of course, in the case in which the indirect dependency approach is not found to be tenable due to reasons discussed above (Beck and Berman 2000, Fanselow and Mahajan 2000, Bayer 1996), Rackowski and Richards' approach becomes problematic for languages like Kashmiri. In what follows, I will propose an alternative position that could admit Rackowski and Richard's approach without sacrificing our current understanding of Kashmiri: for some languages the specifiers of non-interrogative CPs are wh-positions and are stopping-off points for wh-phrases and wh-expletives. For other languages, the specifier of vP is the only stopping- off point for long

Chapter 4. Wh-expletives in Hindi-Urdu

wh-movement. If this is the case, we might expect to see some languages in which wh-phrases and wh-expletives can remain in Spec, vPP I will claim here that Hindi- Urdu is just such a language. Insofar as this line of reasoning is correct, we should be able to understand the contrasts between Hindi-Urdu and Kashmiri. 45.2

An account of Hindi-Urdu wh-dependencies

At this point we will bring together two strands of research. The first is the account of partial movement described above, built on the intuition that the A-bar movement system is shaped by the same mechanisms that shape the A-movement system. The second is Rackowski and Richards' account of Tagalog, built on the intuition that the specifier position of the vP may play the role traditionally ascribed to the specifier position of CP. Let us now pursue a detailed analysis of the Hindi-Urdu wh-expletive construction in (2) combining these two approaches. The first step is to assume that a full wh-phrase and a wh-expletive in HindiUrdu will possess exactly the same features as they do in Kashmiri. Specifically, as before, we assume that a wh-phrase will have an uninterpretable [Q] feature and an interpretable [wh] feature (which provides its interpretation as a choice function variable). A wh-expletive will have an uninterpretable [Q] feature, but no interpretable features at all. The groups of features motivating a long-distance wh-dependency in HindiUrdu will be more or less identical to those in Kashmiri. The primary difference is that the features are not located C heads but on v heads. This causes wh-material to appear not in the specifier positions of CP, but in the specifier positions of vP. In the matrix clause the interrogative C head will bear an interpretable [Q] feature and an uninterpretable [wh] feature, just as it does in single-clause Hindi-Urdu sentences, and in Kashmiri. In this way, the [iQ] will be interpreted by the semantics as an unselective binder of choice functions. This matrix C head, however, will not have the EPP property. This means that the features on the C head will then be valued by virtue of the C head's relation with a wh -phrase in some accessible position within its domain (its own phase or at the edge of the immediately

17. Breuning (2007) argues for an account of Passamaquoddy relative root wh -dependencies in which it is not the scope marker (his term) that remains in Spec, vP but instead a relative root which attaches to the verb phrase. In this approach, the entire wh -element + relative root complex passes through Spec, vP in the matrix clause on its way to the matrix CP. The whexpletive is spelled out in Spec, CP, however the relative root is spelled out in the lower Spec, vP. We could view this as the intermediate result, in some sense, in which some evidence of the presence of the wh -element remains in Spec, vP, though not the wh -element itself (Breuning's

account is one of multiple spellout, an idea not addressed here).

111

111

Peripheries in Kashmiri and Hindi-Urdu

lower phase). In this case, this phase will be the vP phase in the matrix clause. The schema in (39) illustrates the featural content of each of the relevant heads in the wh-expletive construction in (2), as well as the features of the wh-material. An explanation of the derivation follows. (39)

[ep

C ... lvP wh-expl [v.. __ ] [ep C ... [vP wh-XP [ v ... _ [u.wh] [uQ] [uwh] [uQ] [uwh] [iQ] [EPP] [iwh] [uQ] [EPP]

]]]]]

There are two different types of heads that bear features relevant to wh-movement in (39). The first is v, which bears wh-features in both the subordinate and matrix clauses. 18 The second is C, which only bears features relevant to wh-movement in the main clause. Non-interrogative C heads do not bear a [wh] feature in HindiUrdu under this analysis, and neither interrogative nor non-interrogative C heads bear the EPP. As a consequence, this approach correctly predicts that no whmaterial is ever found in the specifier of interrogative or non-interrogative CP in Hindi-Urdu. Such elements are found, rather, in lower positions. Beginning in the lower clause in (39), the wh-phrase kis-ko originates in object position. The v head in the lower clause is a probe possessing the EPP property, and interacts with and raises kis-ko into its second specifier position. In this interaction the [uwh] feature on vis valued, as is the EPP. The [uQ] of the wh-XP is also valued, and it will move no further. Note, however, that this v head has no relevant interpretable features - the wh-phrase in its specifier will not be interpreted in this position. If this account of Hindi-Urdu were precisely like the account of Kashmiri, we would expect the subordinate C head to have an uninterpretable [wh] feature and the EPP. That is, it would be one in the sequence of C and v probes interacting with the full wh-phrase. However, under the approach of Rackowski and Richards (2005), non-interrogative C heads have no wh-related features at all and they do not participate in wh-movement. In the hybrid view adopted here, though Kashmiri is a language in which non-interrogative C heads do have wh-features (as

Note the symmetry between the derivation of partial wh-movement in Kashmiri in (26) and partial wh-movement in Hindi-Urdu in (39), in that there is an uninterpretable [Q] feature on the head that hosts the full wh-phrase in its specifiier (the embedded C in Kashmiri and the embedded v in Hindi-Urdu). The presence of this uninterpretable Q feature arrests the movement of the full wh-XP. If the wh-XP's uninterpretable features are not fully valued at these points respetively, the derivation will crash, because it is the wh -expletive that is encountered by and interacts with any higher probes. t8.

Chapter 4. Wh-expletives in Hindi-Urdu 113

is evidenced by the presence of wh-material in the specifier of non-interrogative CPs), Hindi-Urdu is a language in which they do not. Moving up in the structure, the next head with relevant features is the v head in the matrix clause. As discussed above for Kashmiri, the wh-expletive kyaa originates within in the domain ofv. The matrix v first interacts with the wh-expletive. This interaction values the EPP property on v. However, the [uwh] on the v head remains unvalued and so the head continues to act as a probe. Following the approach of Rackowski and Richards (2005), the matrix v head must have some features that require it to Agree with the embedded C, just as it might interact and agree with a direct object. In Tagalog, the presence of this feature has overt morpho-phonological consequences; in Hindi-Urdu, it does not. A consequence of this relation is that the phase boundary of the embedded C becomes transparent to v, and v can continue probing down to the next phase edge. The matrix v must then probe to the edge of the lower v phase, finding the wh-phrase kis-ko in its specifier. In interacting with the wh-phrase, the v head values its [uwh] feature. 19• 20 At this point all of the features on the matrix v have been valued. The only remaining head with unvalued features is the matrix C, which probes its domain up to the edge of the lower phase, or the matrix v. Note that this is as far as the matrix C can probe; the v head possesses interrogative features and is not 'transparent' in the sense ofRackowski and Richards. The matrix C head values the [uQ] of the wh-expletive, and subsequently its own [uwh] feature with the [wh] feature on the matrix v. Now the derivation is complete and licit, with all features valued and no unvalued features remaining on any wh -material or in any head. In the interpretive component, as before, the C bearing interpretable [ Q] triggers the

19. Ifthis type of cross-clausal probing is generally available, why we do not see long -distance agreement across finite clause boundaries in Hindi-Urdu? Bhatt (2005:777) suggests that, assuming that embedded subjects are lower than Spec, TP (as we have here), a Minimalitybased explanation that does not reference phase boundaries can prohibit LDA. He suggests that a matrix finite T 0 cannot 'look past' another finite T 0 (the embedded one) while looking for an argument with which to value features. Therefore adopting Rackowski and Richard's approach here does not pose a problem for our understanding of limits on LDA in HindiUrdu. 20. There is evidence that intermediate heads in a wh-movement sequence should have interrogative features of some kind. In particular Henry (1995) observes that in Belfast English subject -auxiliary inversion takes place not only in highest C head in a wh-movement sequence, but in intermediate heads as well. She takes this as an indication that at least in this context these intermediate heads share the interrogative status of the matrix C-head. In the present account, this notion is refiected in the [wh] which appears on intermediate C and v heads in Kashmiri and on v heads in Hindi-Urdu.

114 Peripheries in Kashmiri and Hindi-Urdu

introduction of the unselective binding operator that binds the choice-function variable of the wh-phrase. Let us take a moment here to recall the basic intuition that this account is attempting to capture. We have seen a body of empirical evidence that suggests that the surface position for wh-material in Kashmiri is the specifier position of CP, but that in Hindi-Urdu such material occupies the specifier position of vP. We have also seen that in Hindi- Urdu, unlike in Kashmiri, wh-material can never appear in the specifier position of any intermediate CP, and we have no specific morphological evidence that it has ever appeared in this position. Fundamentally, what we wish to claim is that the same clausal topology surfaces in both Kashmiri and Hindi- Urdu, but at the CP layer in Kashmiri, and the vP layer in Hindi- Urdu. As I see it, there are a number of ways of capturing this intuition in this framework. One possible view is that in Hindi-Urdu, CP is not a phase at all. This view would claim that phases are parametrized across languages, and defined in some language-specific manner. Of course, this approach would require a major rethinking of our understanding of phases. To this point phases (CP, vP, DP), have been identified as universal processing units of the derivation. It is unclear what it would mean to say that in a certain language, a certain projection does not constitute a phase, or possibly only constitutes a phase for one type of movement but not another. A second possible way to capture the basic intuition described above is to assume that wh-material moves through the specifier of every phase-defining head (C and v) on its way to its ultimate position, regardless of language. In this view, the fundamental distinction between Hindi- Urdu and Kashmiri is that the surface position of wh-material in Hindi-Urdu is in the specifier position of vP, and in Kashmiri is in the specifier position of CP. Regardless of whether there is any overt evidence that wh-rnaterial moves through the specifier of a head, we would assume that it must do so as part of the fundamental design of phases. The disadvantage of this approach is, for those languages like Hindi-Urdu that have no evidence of wh-material ever appearing in the specifier position of CP, we must posit that it does pass through these positions. On the other hand, the advantage of this view is a certain uniformity in the understanding of A-bar movement across different language types. Every C and v head in every language possesses features relevant to wh-movement. What primarily varies by language is which of these heads possess the EPP property. A third possible way to capture the v/C distinction between Hindi-Urdu and Kashmiri is to adopt the view advocated in Rackowski and Richards (2005). Specifically, when a v Probe agrees with a clausal phase head, it can probe beyond that phase boundary to the edge of the next lower phase. This approach requires us to assume that completed phases are not immediately exported to

Chapter 4. Wh-expletives in Hindi-Urdu 115

the interfaces (Chomsky 2004), but instead remain in the workspace throughout the derivation. At this point, I see no clear empirical test for distinguishing among these proposals. For the sake of familiarity I will adopt the Rackowski and Richards view. In addition, the basic claim Rackowski and Richards make (that the vP-phase can play a crucial role in wh-processes) correlates well with our empirical observations about Hindi-Urdu. 21 The analysis proposed here could be expressed under any of the sets of assumptions above (and possibly others). What is crucial is that the basic intuition about the relative clausal organization of A-bar movement in Hindi- Urdu and Kashmiri is captured. Returning to the diagram in ( 39), if the w h-phrase in the lower clause were not frozen in place, but instead was forced to continue raising, and if no wh-expletive was available in the numeration to satisfy the EPP property on the higher v, the result would be displacement, an additional strategy for forming a long-distance wh-dependency in Hindi-Urdu. Let us reconsider ( 1), repeated here. (1)

Sita-ne kis-ko soc-aa ki Ravii-ne _ __ Sita-ERG who-Ace think-PRF.M that Ravi-ERG dekh-aa? see-PRF.M

[Hindi- Urdu]

'Who did Sita think that Ravi saw?' Recall that (1) is nearly identical to the wh-expletive construction in (2), except that the wh-word that originates in the lower clause appears in the matrix clause. There is significant debate in the literature about the nature of (and even the grammaticality of) the displacement in (1). Some researchers contend that this is just another example ofwh-movement (Simpson 2000, Simpson & Bhattacharya 2003), familiar from a wide range of languages. Others (Dayal 1996) contend

21. Given the assumptions Rackowski and Richards make about 'transparency: it may seem that the v in Kashmiri need not have any features associated with wh-movement (these features could be on C head alone). However, Rackowski and Richards' account is concerned with the way in which verbs agree with clauses, not the way in which probing heads agree with one another. In their view, it is the agreement of verbs with their complement clauses (which has a morphological reflex in Tagalog) that permits transparency. It would be an unmotivated extension of this account to assume that clauses also agreed with their verbs, and that vis then transparent to a probing C. For this reason, we would need to maintain that v heads have whfeatures in Kashmiri On the other hand, an account which privileges a 'mirror-image' view of Hindi-Urdu and Kashmiri wh-features would need to take the form of the alternative suggested in Section 4. 5.2, in which every C and v head in both languages has features associated with wh-movement.

116

Peripheries in Kashmiri and Hindi-Urdu

that this is a form of scrambling and entirely unrelated to either the wh-expletive construction in (2), or to the usual form of wh-movement visible in other languages. Given the current theoretical framework, there is a way of viewing (1) in which it is less relevant what name is given to the movement. It~ as is claimed by Chomsky (2004), all movement is driven by the interaction of features and/or by the EPP property, then movement in both a full wh-movement language like Kashmiri and in a so-called wh-scrambling language like Hindi-Urdu must be driven by the same basic mechanisms. The upshot of this notion is that the same features that induce wh-expletive constructions might also induce displacement of the wh-word in (1). Whether this displacement is in fact termed scrambling or movement becomes less important. The diagram in (40) depicts the feature bundles that would appear on each head in the long-distance wh-displacement in (1). Of course, the wh-phrase kisko has the same features it has in a sentence like (2)/(39). An explanation of the derivation follows.

(40)

[ep

C ... lvP wh-XP [v ... ] [ep C ... lvP _ [u.wh]

[iQ]

[uQ] [uwh] [iwh] [EPP]

[v ... _ [uwh] [EPP]

]]]]]

Just as in (39), in (40) the full wh-phrase kis-ko has an uninterpretable [Q] feature and an interpretable [wh] feature. The probe v in the lower clause interacts with kis-ko and raises it to an outer specifier position. This interaction values the [u.wh] on v as well as the EPP. At this point kis-ko is at a phase edge. The embedded C head has no features relevant to wh-movement. and therefore does not interact with it. According to Rackowski and Richards' approach, the v in the matrix clause must probe and Agree with the C head. When this occurs, the embedded CP phase becomes transparent to the v head, and it can probe for material beneath it. If there is no wh-expletive in the numeration, the matrix v head must probe to the edge of the embedded vP-phase, and it will find and interact with the wh-phrase kis-ko. The wh -phrase will raise into an outer specifier of the matrix vP, valuing the [uwh] on the matrix v and satisfying the EPP property on that head. Note that the whphrase does not yet have all of its features valued (namely [uQ]). As in (39), the C head in (40) has an uninterpretable [wh] feature and an interpretable [Q] feature. The wh-phrase is on the edge of the immediately lower phase (the matrix vP), and so the C Probe can interact with it, and all features are mutually valued. That is, the [uwh] on the C head and the [uQ] on the whphrase. At this point all features in the derivation have been valued, and the derivation is licit. Whether the wh-displacement in (40) is termed wh-movement or

Chapter 4. Wh-expletives in Hindi-Urdu 117

scrambling, the wh-phrase raises into a higher clause, attracted by features of a higher, phase-defining head.22 Let us now address at what point the wh-expletive is first introduced in the clause. Above I argued that in Kashmiri the wh-expletive is base generated in a position in which it can be assigned case by transitive v, in the specifier of AspP. Here I will suggest that the same holds true for Hindi-Urdu. As in Kashmiri, this correctly predicts that sentences in which another wh-DP or the expletive object yeh co -occurs in a single clause with the wh -expletive are ungrammatical in HindiUrdu (data from Simpson 2000) (41) *Sita-ne yeh kyaa soc-aa ki Ravii-ne kis-ko dekh-aa? Sita-ERG EXPL (wH)EXPL think-PRF.M that Ravi-ERG who-Ace see-PRF.M 'Who did Sita think that Ravi saw?' [Hindi- Urdu] (42) *Sita-ne kis-ko kyaa dekh-aa? Sita-ERG who-Ace EXPL saw-PRF.M Intended: 'Who did Sita see?'

[Hindi-Urdu]

In this view, Hindi-Urdu sentences like (41) and (42) 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. Wh-dependencies mediated bywh-expletives that extend across three or more clauses have been widely discussed in the literature (for instance, McDaniel1989 for German and Romani; Dayal1994 for Hindi-Urdu; Horvath 1997 for Hungarian). In Hindi-Urdu, a wh-expletive must appear in the preverbal position in every clause intervening between the clause containing the full wh-phrase and the

Note that if the matrix v in either (39) or (40) happened to have a [uQ] feature it would have no discernable effect on the derivation. In either case the feature would be valued by the [uQ] on the wh-XP. It is whether or not a [uQ] feature appears on the embedded vthat distinguishes a wh-expletive construction form a wh-displacement construction. Once the wh-XP has been frozen in place (all features have been valued) in the lower clause, if no wh-expletive happens to be in the numeration to be merged into the matrix clause, the EPP in the matrix clause cannot be satisfied and the derivation will crash. Note that this represents the solution to a longstanding puzzle: although wh -in-situ phrases in a single clause appear to take scope over the clause as a whole, wh-ln-situ phrases in subordinate clauses in Hindi-Urdu do not. The ID approach has been to force adjunction of the subordinate clause to the matrix CP/IP to create islandhood. However, in the wh-expletive account presented here, there is a purely syntactic requirement that sentences with a wh-phrase from the subordinate clause taking matrix scope are those with overt wh-movement/scrambling or with a wh-expletive in the matrix clause. :21.

118

Peripheries in Kashmiri and Hindi-Urdu clause at which the wh-phrase is interpreted (Dayal1994, Mahajan 2000). This is illustrated in (43): (43)

a.

Raam-ne kyaa soc-aa ki ravii-ne kyaa kah-aa ki Ram-ERG EXPL think-PRF.M that Ravi-ERG EXPL say-PRF.M that

konsa aadmii aay-aa which man

come-PRF.M

'Which man did Ram think that Ravi said came?'

[Hindi- Urdu]

b. *Ram-ne kyaa soc-aa ki ravii-ne kah-aa ki kon sa aadmii aay-aa Accounting for wh-expletive constructions spanning three or more clauses in Hindi-Urdu is a simple extension of the account for two clause structures ((44) is abridged for clarity). (44)

[ep C [vP wh-expl [v] b C [vP wh-expl [v] [cp C [vP wh-XP [ v ... _ [uwh] [uQ] [uwh] [uQ] [uwh] [uQ] [uwh] [uQ] [iwh] [uQ] [iQ] [EPP] [EPP]

]]]]]

[EPP]

This extension is relatively trivial because it is not the embedded full wh-phrase that values the features of the interrogative C head in the wh-expletive analysis of sentences like (43a) (see (39)), but instead a local interaction with the closest v head. In fact, for each intermediate v head with wh-features (as well as the interrogative C head), uninterpretable wh-features are always valued by features at the edge of the next lower vP. Importantly, this means that interrogative C head can be an unlimited distance from the full wh-phrase provided that the EPP on each intervening v head is valued by a wh-expletive. If the necessary wh -expletives are not present in the numeration, the derivation will crash, accounting for the ungrammaticality of (43b) above.23 This account of wh-expletive constructions in Hindi-Urdu offers answers to a number of questions that up until now remained mysterious. First and foremost, the combined strands of research have suggested that the syntactic position of whphrases and wh-expletives that preferentially appear is consistent with the specifier of vP. In Section 4.3 I presented a range of empirical support for this claim, and

13. In (44), the v head in the intermediate clause has the same feature specification as the lowest v head, as opposed to the features of the matrix v. An expletive is required in the intermediate clause to satisfy the EPP on v. The [uQ] feature on the intermediate v head values the [uQ] on the wh -expletive merged in its specifier. Of the two interrogative v heads available in the lexicon of Hindi-Urdu (both of which bear EPP), the head with [uQ] must appear in the numeration of the intermediate clause, or a the resulting derviation will not converge. This is analgous to the situation in two clause wh-expletive structures.

Chapter 4. Wh-expletives in Hindi-Urdu 119 we have seen here that the claim is well-integrated into an account oflong-distance

wh-dependencies in the language. 24 Second, this analysis correctly predicts that no wh-phrases or wh-expletives are ever found in Spec, CP in Hindi-Urdu. In contrast with Kashmiri, interrogative CPs interact with wh-material via Agree at a distance, and do not require that wh-material move to Spec, CP. Further, under the view we have adopted from Rack.owski and Richards, this is due to the fact that non-interrogative CPs in Hindi-Urdu do not participate in wh-movement. Non-interrogative C heads possess no wh-related features, and therefore do not interact with wh-material at all (though see above for other ways to implement this view). Third, and perhaps most importantly, this account provides a way of understanding the role of the wh-expletive in Hindi-Urdu. The wh-expletive serves to value the EPP on the attracting head for long wh-movement, and in Hindi-Urdu this happens to be the v head. The features on that head may then be valued via Agree with some other goal in its domain. Thus, wh-expletives in HindiUrdu and Kashmiri can be viewed as being governed by exactly the same set of mechanisms. Finally, Hindi-Urdu can now be considered a language that exhibits a true case of "partial" wh-movement, in that the full wh-phrase in the lower clause moves from its base position into the specifier of the embedded vP. This resolves a puzzling mismatch, in that other languages with wh-expletive constructions also

24· This approach must also account for the matrix scope ofwh-phrases inside temporal adjuncts as in (i). (i)

Ram [kyaa khaa-te hue] ghar ga-yaa? Ram what eat-HAB while home go-PFV.MSG 'What did Ram go home while eating?'

[Hindi-Urdu] (Dayal1996:33)

Dayal (1996) suggests that these infinitival phrases are gerunds base-generated in complement position. Under the indirect dependency approach, the wh-phrase inside the gerund moves at LF to Spec, CP to produce a matrix wh -question. In the account proposed here the gerund, like any other complement of the verb with interrogative features, moves to Spec, vP to statisfy the EPP on the v head. The interrogative matrix C then probes its domain and interacts with the gerund in Spec, vP, and the derivation is licit. The only additional assumption needed is that the entire gerund is pied-pied (much like the content of any wh-DP in complement position). Keep in mind that the full wh-phrase inside the non-finite clause would move to the edge of the internal verbal domain, whatever form that functional v head may take (see, for instance, Pyllk!tnnen 2002, Moulton 2004). It is therefore accessible to higher probes (namely the matrix C head), just as the arguments of non-finite clauses are accessible to processes like Long Distance Agreement (Bhatt 2006). As we might expect under our account, overt movement of argument wh -phrases from within infinitival adjuncts into the matrix clause is also possible (Bhatt 2003).

110

Peripheries in Kashmiri and Hindi-Urdu

tend to exhibit "partial" wh-movement constructions. Now Hindi-Urdu can also be understood to fall under both of these categories. 4·5·3

Comparison with other accounts

There are previous approaches to questions in Hindi-Urdu (Dayal1994, 1996; Lahiri 2002) and related languages such as Bangla (Simpson & Bhattarcharya 2003) that are worth addressing in the context of the current effort. Simpson and Bhattacharya's (2003) analysis of wh-expletive constructions in Bangia assumes that Bangia, though typically analyzed as SOY and wh-in-situ just like Hindi-Urdu, is in fact SVO and has obligatory overt wh-movement. According to Simpson and Bhattacharya, this word order, as well as overt wh-movement, is typically disguised by a number of factors, but is manifest in certain contexts. Although the account I have proposed in this chapter stops short of assuming underlying SVO word order for Hindi-Urdu, I do claim that finite CPs are complements in the language, and that the displacement of wh-phrases that we see in Hindi- Urdu is in fact wh-movement. In this way, the data presented for Bangia in Simpson and Bhattacharya, as well as the account given for that data, informs the analysis I ultimately adopt. The most recent alternative account for some of the phenomena addressed in this chapter is found in the indirect dependency approach ofLahiri (2002), so in the following section I will briefly summarize this analysis and show that one of its core assumptions is unworkable. 25 I claim that the account presented here provides us with greater empirical coverage at less theoretical cost. Lahiri (2002) proposes a variation of the indirect dependency approach (ID) for wh-expletive structures in Hindi-Urdu. His analysis consists of a semantic elaboration on and revision of Dayal's (1994, 1996) approach. It shares with this (and all ID accounts) two core assumptions: (a) the wh-word kyaa which we have called a wh-expletive here is not an expletive at all but instead the full HindiUrdu wh-word kyaa meaning 'what: and is base generated in an argument position in the matrix clause; and (b) the apparently subordinate clause is not an argument of the verb but instead has some other syntactic status as an adjoined element. Lahiri provides several empirical arguments against various versions of the direct dependency approach (DD), or any account that requires that the embedded full wh-phrase move to the position of the wh-expletive at LF. He successfully shows that presuppositions of wh-expletive constructions, wh-expletive

15. For a dicussion of direct and indirect dependency approaches in more detail, see Chapter 3 (3.3.3).

Chapter 4. Wh-expletives in Hindi-Urdu

constructions with embedded yes-no questions, and the "scope freezing" of wh-expletive structures with amount questions cannot be easily understood in aDD account in which the full wh-phrase must move into the matrix clause at LF, to be interpreted separately from the remainder of the embedded CP. It is unsurprising that these arguments do not extend to the wh-expletive account presented here, since it does not depend on LF movement of the embedded wh-phrase. For instance, Lahiri (2002) shows that Hindi-Urdu wh-expletive structures with embedded how many phrases exhibit scope freezing, in that they permit only narrow scope of the embedded how many phrase (not the ambiguity available in English). Accounting for this under the ID approach is quite natural since the question "how many books did Ram read" remains intact at LF and does not permit wide scope of the quantifier. However, a DD account would require a mechanism by which only the wh-word would move to the matrix clause at LF, somehow leaving the quantifier to be interpreted in situ. Under the wh-expletive account presented here, on the other hand, the wh-indefinite in the embedded clause is unselectively bound at a distance and the quantifier many remains in the embedded clause. Lahiri himself mentions (Footnote 2) that the unselective binding of choice functions would be one way of interpreting structures in which the wh-phrase remains embedded, but makes a decision not to evaluate this type of analysis. I now turn to the ID claim that apparently subordinate CPs in Hindi-Urdu are not arguments of the matrix verb, but are in some manner adjoined to the matrix clause. I will show here that given this assumption, it is impossible to actually generate the LF required in the ID approach. An often-repeated fact arguing against this basic assumption is that quantifiers in the matrix clause of a wh-expletive structure (or any noninterrogative embedded clause structure) in Hindi- Urdu appear to be able to bind pronouns in the embedded clause in the normal way. In (45a-b) the quantifier har aadmii 'each man' in the matrix clause can bind the pronoun us-ne 'he (erg)' or wo 'he' in the second CP (Mahajan 2000). (45) a.

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

b.

Aap-nei har aadmii-ko kyaa kah-aa ki woi kis-ko dekh-eegaa? 2PL-ERG each man-ACC what say-PRP.M that 3SG WhO-ACC see-PUT.M 'Who did you tell every man he would see?' [Hindi- Urdu]

The bound interpretation of these pronouns would be unexpected under any version ofiD because the pronoun would not be in the scope of the quantifier at

121

112

Peripheries in Kashmiri and Hindi- Urdu

LF. Take for instance the LF for a wh-expletive construction proposed by Lahiri (2002) in (46). (46)

---------CP

DP·

~

DPi

CP

I~

kyaa what

kis-sej ramaa who-INS ramaa

C'

~

Raam 4. soctaa hai (Ram 4. thinks) t;

baat talk

kar-egii do-FUT

(Lahiri's (42))

In order to solve this problem, Lahiri develops a semantic workaround that mimics the effect of variable binding through keeping the pronoun in the restriction of the wh-quantifier and allowing it to be closed off by a universal closure operation. The bound reading of ( 45a) under this view would then be: (47)

A.p3f ['ii'x [Qx(f(x))] & p =" 'ii'x (man'(x) --+ think'(x, f(x))] (Lahiri 2002: (114))

Applied to (45a), the restriction in (47) says that for everyx, f(x) (the proposition x thinks) is of the type x saw person y for some y. Lahiri uses the availability of this mechanism to argue for his version of the ID approach, in which the second wh-clause is in the restriction of kyaa at LF. This seems to address the apparent problem of variable binding in wh-expletive structures. Of course, the fact that quantifiers in a matrix clause of the wh -expletive structure can bind pronouns in the lower clause is expected under the wh-expletive account presented here, as they c-command the variable in the familiar way. For this reason, under my account no special workaround is required. However, there are other facts which pattern with the binding data, suggesting that material in the subordinate clause truly is in the command domain of material in the matrix clause at the level of interpretation. Negative Polarity Items (NPis) in Hindi-Urdu have been carefully investigated (Mahajan 1990, Lahiri 1998, Kumar 2006). It is well known that the class of weak NPis, when embedded, can be licensed by negation in a higher clause. ki koi bhii aay-aa. [Hindi-Urdu] Sariitaa-ne nahiiN kah-aa Sarita-ERG NEG said-PRF.M that someone even come-PRF.M 'Sarita did not say that anybody came.' (Kumar 2006: 144) (49) *Sariitaa-ne kah-aa ki koi bhii aay-aa. (48)

It is not possible to demonstrate the same effects in a wh-expletive structure, because wh-expletives are incompatible with sentential negation (Mahajan 2000,

Chapter 4. Wh-expletives in Hindi-Urdu 123

Lahiri 2002, Malhotra & Chandra 2007). However, because both Dayal (1994, 1996) and Lahiri (2002) propose that no subordinate clause in Hindi-Urdu (whether in a wh-expletive structure or not) is ever a complement to the verb, (48)-(49) are crucial to our understanding of the relationship between matrix and subordinate clauses in Hindi-Urdu. Dayal (2000) suggests that apparent embedded clauses are in fact basegenerated as adjuncts to IP (discontinuous from any scope marker or expletive), since nominal complements appear to the left while alleged sentential complements appear to the right. Under the view, as Lahiri ( 1998: 80) expresses it, that "the conditions on NPI licensing be statable as LF conditions, involving c-command of the NPI by the negative element at LF~ it would be impossible to explain the facts in (48)-(49) given the structure Dayal proposes. The NPI in (48) would never be in the c-command domain of the negative element at LF, and so the contrast between (48) and (49) would remain mysterious. Lahiri (2002) proposes another possible structure (though he does not choose between this and Dayafs alternative), following Herburger (1994), in which kyaa and the interrogative CP form a DP at D-structure before the CP is extraposed and right-adjoined to IP (to arrive at the appropriate surface word order). To create Lahiri's version of the LF under the ID approach (in (47) above), it seems that the CP must then be reconstructed back into the restriction of kyaa and then the entire DP must move to Spec, CP at LF for scope reasons. So for non-interrogative sentences such as (48), the second CP would in principle be reconstructed back into the DP containing the overt or covert demonstrative pronoun, and in this way the NPI could be reconstructed back into the c-command domain of the negation. 26 However, there is a challenge for this approach that ultimately prevents us from being able to construct the LF in (47). Bhatt and Dayal (2007) argue for independent reasons that the only type of constituent that can scramble to the right in Hindi-Urdu is a VP remnant. For instance, their account derives a sentence with a rightward scrambled IO as in [S DO V Aux IO] through having the verb undergo head movement into a higher aspectual projection, and the DO leftward scramble out of the VP. Then the VP projection containing only the IO (as well as a trace of the DO) can be moved rightward. Bhatt and Dayal speculate (see their Section 3.3), that apparently right-adjoined CPs are actually part of a rightward-scrambled VP remnant However, in order to arrive at a VP projection that exhaustively contains the second CP from a syntactic structure like that which

26. Although NPis aren't always licensed under reconstruction (*Any professors weren't available), it is attested (see Aoun & Benmamoun 1998, for example).

114 Peripheries in Kashmiri and Hindi-Urdu

Lahiri proposes, the direct object DP kyaa would need to be scrambled leftward out of the VP first. Since reconstruction would recreate the post-leftward scrambling but pre-rightward scrambling order (Bhatt and Dayal), we could never arrive at the LF Lahiri has proposed in (47). The CP will never be able to be recombined with the DP kyaa, since kyaa has scrambled leftward outside the VP. Moreover, the VP remnant is a scope island, so covert movement of the CP after reconstruction of the VP containing it is not possible (Bhatt and Dayal). This means that we cannot actually construct an LF like (46) in which the apparently embedded CP is in the restriction of kyaa at the appropriate scope position. We have arrived at something of a catch-22 with respect to deriving the ID LF in (47). Recall that the syntactic approach under the first version ofiD (Dayal 2000) did not explain the licensing ofNPis in Hindi-Urdu as in (48)-(49). However the syntax proposed under the second version of ID (Lahiri 2002) requires a rightward-scrambling operation restricted to an otherwise voided VP-remnant. This seems to force a series of operations that precludes the creation of the very LF this analysis proposes. On the other hand, under the wh-expletive account presented in this chapter, none of these issues arise. Because the second CP is in fact the complement of the main clause verb, NPI licensing proceeds as expected (via c-command). Since no scrambling in any direction is required, we need not be concerned with reconstruction possibilities. Lahiri's (2002) version of the ID approach has clear empirical advantages over DD approaches, as he demonstrates in detail. However these criticisms of DD accounts do not to extend to the wh-expletive account presented here, which in fact predicts the very facts he discusses. Further, given a set of independently justified assumptions about Hindi-Urdu syntax. it is unclear that the LF proposed in the ID approach can even be constructed. On the other hand, the wh-expletive account effortlessly accommodates phenomena such as variable binding and NPI licensing that must be treated exceptionally under any ID approach. In this section, I have argued that any account requiring the complement clause to be generated adjoined to the matrix clause, or to be extraposed to this position, is ultimately untenable. Of course, we could attempt to rescue the ID account by claiming that the CP sister of kyaa is not extraposed, but postsyntactically aligned to the right edge of the matrix clause, as I propose in Section 4.2. This version would avoid the problems of LF formation associated with extraposition. However, this presents us with an empirical puzzle. If the apparent finite clause complement to the verb were actually the sister of the DP kyaa in Hindi-Urdu, and were obligatorily linearized to the right edge of the matrix clause at PF. we would expect that all finite clauses that are DP complements should behave similarly.

Chapter 4. Wh-expletives in Hindi-Urdu

However, finite clause complements of DPs in Hindi-Urdu can optionally remain in-situ (Bhatt 2003).

(SO)

a.

mujhe [yah khabar [ki ye log nahiiN aa lsG.DAT this news that these people NEG come mil-ii paa-enge)] kal able-FUT.PL yesterday find-PRF.F

(51)

'I got the news that those people won't be able to come yesterday'

[Hindi- Urdu]

b.

mujhe [yah khabar] kal milii [ki ye

paa-eNge]

a.

Mona jaan-tii hai [ki Rohit chanT hai) [Hindi-Urdu] Mona know-HAB.F AUX.PRS that Rohit cunning be.PRS 'Mona knows that Rohit is cunning.'

b.

Mona yeh jaan-ti hai. Mona EXPL know-HAB.F AUX 'Mona knows this.'

c. *Mona [ki Rohit chanT

log

nahiiN aa

[Hindi-Urdu]

hai] jaantii hai.

(52) *Sita-ne kyaa [ki Ravii-ne kis-ko dekh-aa] soc-aa?(compare with (2)) Sita-ERG EXPL thatRavi-ERG who-Ace see-PRF.M think-PRF.M 'Who did Sita think that Ravi saw?' [Hindi- Urdu) b. *Mona yeh [ki Rohit chanT hai] jaanti hai. Mona EXPL that Rohit cunning be.PRS know-HAB.F AUX.PRS The sentences in (50)-(52) illustrate that CP complements of DPs can optionally appear in situ, but apparent CP complements of verbs cannot (irrespective of whether they appear with the wh-expletive kyaa or clausal expletive yeh). It would then seem that the second CP in wh-expletive constructions in HindiUrdu (in (52)) does not pattern with CP complements of DP, posing a problem for this patch of the ID approach. The solution is to assume that the second CP in these structures is in fact what it appears to be - a clausal complement to the verb. This brings us to the core issue: expletives are not necessarily theoretically desirable objects. However, in this case we are forced to accept that kyaa is best understood as an A-bar system expletive, akin to the well-attested A-system expletives. Mahajan (2000) claims that the wh-expletive kyaa seems to have prosodic properties that differ from the wh-XP kyaa - unsurprising if it is in fact a whexpletive. Overall the wh-expletive account presented here best captures both the nature of kyaa in long-distance wh-dependencies, and the complement status of embedded finite clauses in Hindi- Urdu.

115

12.6

Peripheries in Kashmiri and Hindi-Urdu

4·5·4

Conclusion

Despite important differences between Hindi-Urdu and Kashmiri, the preferred position in both languages for interrogative and non-interrogative focus is immediately preverbaL The proposal in this chapter suggests that if wh-material (whphrases and wh-expletives) is found at the edge of the CP phase in some languages (like Kashmiri), we should also expect to find it at the edge of the vP phase in others. I claim that Hindi-Urdu is such a language, and that we can understand the systematic set of contrasts between Kashmiri and Hindi- Urdu as evidence for the same clausal topology in both languages, lower in Hindi-Urdu (on the v head) and higher in Kashmiri (on the C head). I have made use of this notion to construct a new, unified account of the various strategies of forming long-distance wh-dependencies in the two languages.

CHAPTER5

Sluicing in Hindi-Urdu

Sluicing is a well-studied form of ellipsis in which a constituent question is reduced to a stranded wh-phrase. One of the crucial processes claimed to underlie sluicing, the movement of the wh-remnant, is precisely the process I investigated in Chapter 4. The sluicing construction in Hindi-Urdu provides a further test of the account proposed above and has much to tell us about both wh-movement and the structure of the clause in Hindi- Urdu. In this chapter I explore sluicing in Hindi-Urdu, and provide a new account of this syntactic phenomenon based on the approach to the periphery constructed in this book. I will show here that the precise form sluicing takes in Hindi-Urdu can be attributed to the manner in which wh-dependencies are formed in the language and the properties of the phase-defining functional head C.

5.1

Sluicing in Hindi-Urdu and Kashrniri

Whether or not we would expect to find sluicing at all in Hindi-Urdu depends on the account of sluicing we subscribe to. If sluicing only appears in languages with systematic wh-movement to the specifier of CP, then we would not expect to see sluicing in Hindi-Urdu, since under every analysis the language has no wh-movement to this position. In fact, Hindi-Urdu is one of a number oflanguages traditionally understood as wh-in-situ that do exhibit what looks like sluicing, including Japanese (Takahashi 1993, 1994), Korean (Kim 1997), and Turkish (Ince 2006). 1 [Hindi-Urdu]

(1)

Aisha-ne ek ciiz khariid-ii lekin mujhe nalliiN pa-taa Aisha-ERG a thing buy-PERF.F but lSG.DAT NEG know-HAB.M

(ki) kyaa. (that) what 'Aisha bought something, but I don't know what'

(Mahajan 2005)

Merchant (2001) divides these languages into a set that is truly wh-in-sit:u (Japanese, Chinese, and Korean) and those like Hindi-Urdu and Thrltish that exhibit what seems to be a kind of focus movement.

1.

12.8 Peripheries in Kashmiri and Hindi- Urdu

(2)

Khadija jaldi calii jaa-egii lekin mujhe nahiiN Khadija early leave go-FUT but 1sG.DAT NEG pa-taa (ki) kyoN know-HAB.M (that) why 'Khadija will leave early, but I don't know why.'

(Mahajan 2005)

On the other hand, is not surprising that Kashmiri exhibits sluicing, as it is alanguage featuring systematic wh-movement to Spec, CP. To the best of my knowledge, sluicing data from Kashmiri was not reported in the literature prior to Manetta (2006). The data below was produced and judged grammatical by nativespeaker informants (JC (2/16/06) and PK (3/31/06)):2 [Kashmiri] (3)

k;lNsi kh-yav bati magar me chu-ni someone.ERG eat-MSG.PST food but 1SG.DAT AUX.MSG-NEG pataa bm. know who.ERG 'Someone ate the food, but I don't know who.'

(4)

(JC 2/16/06)

Tim aas' yi kath zaanaan ki Cheney-an mor 3PL.NOM AUX.PST.3PL this story know.IMPERF that Cheney-ERG kill.PST kaNh, magar me chu-ni pataa kus. someone.NOM but 1SG.DAT AUX.MSG-NEG know who.NOM 'They were believing the story that Cheney killed someone, but I don't know who.'

Sluicing in Hindi-Urdu has been relatively under-researched and is still poorly understood. This section first seeks to aggregate and summarize what is known about the construction in Hindi-Urdu. I will then propose an account of sluicing in Hindi-Urdu (and Kashmiri) following in the spirit of Toosarvandani (2007), but taking advantage of the approach to the periphery presented in this book. I will argue against the account presented in Manetta (2006), which suggests that Hindi- Urdu sluicing patterns more with English VP-ellipsis and is in fact the deletion of the sister of Spec, vP. Instead, I claim here that Hindi-Urdu sluicing is an exceptional instance ofwh-movement to Spec, CP in Hindi-Urdu, driven by a C head possessing the ellipsis feature E (Merchant 2001).

1.

Thanks also to Rakesh Bhatt (p.c.) for confirming these judgments.

Chapter 5. Sluicing in Hindi-Urdu 129

5.2

Accounts without movement to Spec, CP

There are a number of ways we could try to understand Hindi- Urdu sluicing that do not require the wh -phrase preceding the sluice to appear in the specifier of CP. 3 This is appealing, since in the analysis developed here the wh-phrase occupies the specifier of vP, not CP. Kizu (1997, 2005) claims that what appears to be sluicing in Japanese and other wh-in situ languages is in fact a kind of reduced cleft (also see Kuwahara 1996). This structure has also been called pseudosluicing (Merchant 2008). This account claims that apparent sluices like that in (5) are actually derived from the structure in (6). (5)

Taroo-ga nanik.a-o yon-da rashii ga, watasi-wa Taro-NOM something-Ace read-PAST 1-heard but. I-TOP nani-(da)-ka wakaranai what-coP-Q don'tknow 'I heard that Taro read something, but I don't know what'

(6)

... watasi-wa [ep2 [ep 1 opi I-TOP

[IP

Taroo-ga t. yon-da]-noJwa Taro-NOM read-PAST-NM-TOP

[nanicda ]-k.a] wakaranai what-coP-Q don't know 'I don't know what it is that Taro read.' According to Kizu's proposal, the topicalized CP 1 in (5) becomes empty when it has a linguistic antecedent, and the result is (6), in which only the wh-phrase in focus position and (optionally) the copula appear. Because Japanese is a null-subject language that also allows the copula to be omitted, a sentence like (5) resembles English-style sluicing when it in fact has the structure in (6). Hindi- Urdu does exhibit cleft constructions (7a), and as in English the pivot of the cleft can be a wh-phrase (7b ):4 (7)

a.

Phone hai jo mez kii daayii taraf hai Phone be.PRS REL table GEN right side be.PRS 'It is the phone that is to the right of the table.'

3· This section in particular has benefited from the attention of two anonymous reviewers. All remaining errors are my own. 4· Moreover, as Shaher et al. (2008) note in an eyetracking study, Hindi clefts serve to facilitate retrieval of the defied element.

130

Peripheries in Kashmiri and Hindi-Urdu

b.

kyaa he jo mez kii daayii taraf hai what be.PRS REL table GEN right side be.PRS 'What is it that is to the right of the table?'

At first glance, this approach to sluicing in Hindi-Urdu might look promising. Kizu offers additional support that sluicing can be reduced to pseudosluicing in Japanese, including the observation that Japanese sluices optionally permit the copula to appear (as in (5)), just as in cleft constructions. It also follows that clefts in Japanese, like Japanese sluices, are sensitive to islands. (8) *Taroo-ga Hanako-ga nanika-o kat-ta toyuu uwasa-o Taro-NOM Hanako-NOM something-Ace bought COMP rumor-Ace shinjiteiru ga, watashi-wa Taroo-ga Hanak.o-ga ei kat-ta toyuu believe but I-TOP Taro-NOM Hanak.o-ACC bought COMP uwasa-o shinjiteiru-norwa nanika-ka nani-ka wakranai rumor-Ace believe-NM-TOP something-Q what-Q don't know 'Taro believes the rumor that Hanako bought something, but I don't know what it is that Taro believes the rumor that Hanako bought.' Although this may be a coherent account of the facts in Japanese, there are a numher of reasons why this approach is less than adequate for Hindi- Urdu. In Japanese, the wh-phrase remnant of the sluice typically resists case -mar king: (9)

Dareka-ga Taroo-o nagut-ta ga watasi-wa Someone-NOM Taro-Ace hit-PAST but I-TOP dare(* -ga )-ka siranai who-NOM-Q don't know. 'Someone hit Taro, but I don't know who.'

(10)

Taroo-ga dareka-o nagut-ta ga watasi-wa dare(*-o)-ka siranai Someone-NOM Taro-Ace hit-PAST but I-TOP who-NOM-Q don't know. 'Taro hit someone, but I don't know who.'

This falls into place if this wh-phrase is actually the pivot of a cleft. (11)

Taroo-o nagut-ta-no-wa dare(*-ga)-ka siranai Taro-Ace hit-NM-TOP who-NOM-Q dmit know '(I) don't know who it is that hit Taro'.

However, in Hindi-Urdu Gust as Ince (2006) demonstrates for Turkish), the whphrase remnant must be inflected for the same case as it would be were it in a non -elided structure. (12)

Aisha-ne kisi-ko dekhaa, lekin mujhe nahiiN Aisha-ERG someone-Ace see-PRF.M but lsG.DAT not

Chapter 5. Sluicing in Hindi- Urdu

pa-taa kis-ko/""kaun know-HAB.M who-ACC/""NOM 'Aisha saw someone, but I don't know who.' (13)

Kisi-ne Aisha-k.o dekh-aa, lekin mujhe nahiiN Someone-ERG Aisha-Acc see-PRRM but lsG.DAT not pa-taa kis-ne/*kaun knoW-HAB.M who-ERG/*NOM 'Someone saw Aisha, but I don't know who.'

In addition, Merchant ( 1999, 2008) argues against the reduction of English sluicing to pseudosluicing by exploring differences between English sluices and English clefts with wh-pivots. Here I will apply several of these tests to Hindi-Urdu sluices and clefts. First, sluicing with adjunct wh-phrases is grammatical in Hindi-Urdu, but clefting with adjunct wh-pivots is not (unlike with arguments, as in (16)). (14)

Us-ne gaRi-ko fix kiy-aa, magar mujhe nahiin pa-taa He-ERG car-Ace fix do-PRRM but 1sG.DAT NEG know-HAB.M kese (*thaa). how (was) 'He fixed the car, but I don't know how (*it was).' (e.g. with what tool)

(15)

Subhan Ali aay-aa, magar mujhe nahiiN pa-taa Subhan Ali come-PRF.M, but lsG.DAT NEG know-HAB.M kyooN (*thaa). why (was) 'Subhan Ali came, but I don't know why (*it was)'

(16)

magar mujhe nahiiN pa-taa Us-ne koi gaRi fix kii, He-ERG some car fix do.PRF.F but lsG.DAT NEG know-HAB.M kaunsii (thii). which one (was) 'He fixed some car, but I don't know which one (it was).'

The same contrast holds for implicit arguments. (17)

Aisha-ne kah-aa ki us-ne khay-aa, magar mujhe nahiiN Aisha-ERG say-PRF.M that 3PL-ERG eat-M.PRF but lsG.DAT NEG pa-taa kyaa (*thaa) know-HAB.M what (was) 'Aisha said that they ate, but I don't know what (*it was).'

131

132 Peripheries in Kashmiri and Hindi-Urdu

For clefts with wh-pivots, only an exhaustive reading is available (18B). On the other hand, sluices are compatible with a 'mention-some' or non-exhaustive interpretation (lSB'). (18)

A: Aap-ko kisi officer se baat k.arnii caahiiye 2PL-ERG some officer with talk do-INF want 'You should speak with an officer.' B: *Kaun hai, masail ke tor par? Who is example manner as 'Who is it, for example?' B': Masail ke tor par, kis-se? Example manner as who-INSTR? 'For example, who?'

In arguing against the pseudo sluicing account for Malagasy. Potsdam (2007) suggests out that pseudo sluicing may well be an instance of what Hankamer & Sag (1976) call deep anaphora. Material that seems to be missing should be recoverable pragmatically, not necessarily under linguistic identity. 5 In Hindi-Urdu, as in English, sluicing appears to be surface anaphora, requiring a linguistic antecedent. (19)

[Shown a picture of an unknown woman] a. ..1 don't know who. b.

I dont know who she is.

kaun. c. *Mujhe nahiiN pa-taa know-HAB.M who lSG.DAT NEG 'I dontknowwho.' d.

Mujhe nahiiN pa-taa kaun hai. know-HAB.M who be.PRS lSG.DAT NEG 'I dont know who it is.'

These tests indicate that Hindi-Urdu clefts with wh-pivots have properties distinct from sluices. It therefore seems undesirable to claim that Hindi-Urdu sluices are

5· Potsdam (2007) claims that Malagasy sluicing is in fact the elision of material (an IP) following a sentence-initial wh-phrase in a pseudodeft. The characteristics ofwh-movement are different enough in Malagasy to make the pseudocleft analysis of sluicing inapplicable to Hind-Urdu. Potsdam claims that all wh-questions with initial wh-phrases in Malagasy (which has unmlllked predicate-initial order) are derived by pseudoclefting more generally (not by wh-movement to Spec, CP). This strategy is supposedly employed in a number of other Austronesian languages. Also, only subject wh -phrases can be extracted in this way (and therefore only subjects can be sluicing remnants). Since none of these features are shared with Hindi-Urdu, I will not pursue the pseduoclefting account for Hindi-Urdu sluicing here.

Chapter 5. Sluicing in Hindi-Urdu 133

pseudosluices, or reduced clefts. While this analysis may account successfully for the facts of Japanese, we will need to look elsewhere to better understand sluicing in Hindi-Urdu. In Manetta (2006) I proposed that Hindi-Urdu sluicing could be analyzed as the deletion of the sister of the specificer ofvP. In this view, sluicing in Hindi-Urdu involves movement of the wh -phrase to the specifier of vP in the manner proposed in Chapter 4, followed by elision of its sister, as in (20): (20)

Aisha-ne kisi-ko dekhaa, lekin mujhe nahiiN pa-taa ... Aisha-ERG someone-ACC see-PRF.M but 1SG.DAT NEG know-HAB.M 'Aisha saw someone, but I don't know ... .' CP

~ ~T TP

C

vP

kis-~----

who Aisha-ne ______ .... _____ _ t;

dekhaa

In (20) all material below the circled node is sluiced, producing (1). 6 This account of sluicing in Hindi-Urdu is able to derive the properties of the construction discussed above. If sluicing in Hindi-Urdu represents a case of elision following normal wh-movement (to Spec, vP), then it is not surprising that the wh-remnant must receive the same case marking it would in a non -elided structure as in (12)-(13). Similarly, it is not surprising that an adjunct phrase or implicit argument can be a wh-remnant as in (14)-(15) and (17). Since sluicing structures contain only an unexceptional movement of the wh-phrase, we should expect the wh-material to behave precisely as though it were in a non-elided sentence. In this approach there is no need to hypothesize any exceptional movement to Spec, CP. If wh-movement is to Spec, vP in Hindi-Urdu more generally, then sluicing does not present an exceptional case. Instead, sluicing is an elision process in an unremarkable structure, just as it is in Kashmiri. The difference between the two sluices is that in Kashmiri the wh-remnant is in Spec, CP, while in Hindi-Urdu it is in Spec, vP.

6. Under the account presented in Manetta (2006), I claimed that the subject Aisha-ne is still located in the first specifier of vP, and for this reason is elided.

134 Peripheries in Kashmiri and Hindi- Urdu

Another argument for this analysis of sluicing in Manetta (2006) centers on the question of whether island repair in Hindi-Urdu sluicing is more restricted than in English. Mahajan (2005) has controversially suggested that Hindi-Urdu sluicing is much like Japanese sluicing (Takahashi 1993, 1994) in that it does not appear to repair island violations. Complex NP island: (21)

a.

*MaiN jaan-taa huuN ki Ravii-ne [Salmaa-ki ek lsG.NOM know-HAB.M AUX.PRs.lsG that Ravi-ERG Salma-GEN a hui kitaab phar dii lekin mujhe larke-ko dii boy-DAT give.PRF.F AUX book tear give.PRRF but lSG.DAT nahiiN pa-taa (ki) kis-ko NEG know-HAB.M (that) who-DAT 'I know Ravi tore up the book Salma gave to a boy, but I don't know to whom.' (Mahajan 2005)

b...MaiN is kahani-ko maan-taa hooN ki Ram-ne lsG.NOM this story-Ace believe-HAB.M AUX.PRS.lSG that Ram-ERG kisi-ko mar-aa, lekin un-ko nahiiN pa-taa know-HAB.M someone-ACC kill-PRF.M but 3PL-ERG NEG (ki) kis-ko. (that) who-Ace. 'I believe the story that Ram killed someone, but they don't know who.' Others have claimed that Hindi-Urdu sluicing does exhibit island insensitivity - at least in some cases. For instance, Chandra & Ince (2007) attribute the ungrammaticality of (21a) to the fact that a genitive subject DP c-commands the wh-phrase. Their claim is that genitive subjects are found in A-bar positions, and therefore intervene between the wh-phrase in the sluice and an attracting C head. Chandra & Ince cite two examples from Mahajan (2005) (a complex-NP island and an adjunct island) with genitive subjects c-commanding the wh-phrase, and offer alternatives without a genitive subject that they judge grammatical: (22)

a.

MaiN [yeh baat ki John apnii kitaab roz kisi-ko lsG [this fact that John self's book daily someone-DAT de-taa hai] jaantii huuN, par pa-taa nahiiN kis-ko give-HAB.M AUX) know be, but knoW-HAB.M NEG who-DAT. 'I know the fact that John gives his book to someone daily, but I don't know to whom.'

Chapter 5. Sluicing in Hindi-Urdu 135 b.

MaiN dukhii huuN [kyuunki John kisi-k.o roz 1sG sad be.PRS.1SG because John someone-Ace daily maartaa hai], par pa-taa nahiiN kis-ko. hit-HAB AUX] but knoW-HAB.M NEG who-ACC. 'I am sad because John hits someone daily, but I don't know who.'

Though this explanation does not appear to extend to account for the ungrammaticality of (21b), nor for the ungrammaticality of other island violating sluices that Mahajan reports which do not contain genitives, it is indicative of the broader disagreement in the literature surrounding island violation repair in Hindi-Urdu sluicing. For the sake of argument, Manetta (2006) asks the following question: if it is the case that Hindi-Urdu sluicing is restricted in ways that sluicing in English, Kashmiri, and other full wh-movement languages is not, does Hindi-Urdu sluicing pattern more with English VP-ellipsis (VPE), which also does not regularly amnesty island violations? Chung, Ladusaw, and McCloskey (1995) offer the following ungrammatical examples ofVPE: (23) *What did you leave before they started playing t? (24) We left before they started playing party games. *What did you leave before they did? Merchant (2001) also compares the ungrammatical VPE below with sluicing: (25)

They want to hire someone who speaks a Balkan language, but I don't knowwhich (*they do).

Though there are grammatical examples in which VPE repairs island violations (see Kennedy and Merchant 2000, Fox and Lasnik 2003), in most VPE cases islands are not repaired, while the analogous sluicing constructions are quite natural. Merchant (2002, 2008) and Fox and Lasnik have offered various forms of the claim that it is the size of the ellipsis site that is crucial in the ability of the ellipsis to consistently amnesty island violations. That is, ellipsis sites ofVP size or smaller tend not to repair island violations, while larger, TP-sized ellipsis sites do. 7

7· The facts of Hindi-Urdu do not seem compatible with the specifics of Merchant's (2008) proposal Merchant suggests that the reason the smaller ellipsis sites (VPE-sized) do not permit island violation repair is that an illicit intermediate trace of wh-movement remains following deletion. In the case of an English sluice, Merchant claims that wh-movement to Spec, CP moves through VP and TP/IP, and leaves a trace in Spec, VP and Spec, IP. If either of the traces of movement remains, Merchant claims the result will be ungrammaticality. IP (clause-sized) ellipsis (sluicing) would remove both of these offending traces, but smallerVPE would leave the trace in Spec, IP. It is unclear how this approach could be applied to the

136 Peripheries in Kashmiri and Hindi-Urdu Manetta ( 2006) does not attempt to apply the specific details of either of these proposals to Hindi- Urdu, but instead to point to a possible correlation between ellipsis site size and island repair. The proposal in Chapter 4 is that Hindi-Urdu wh-material is located in the specifier of vP. If this is the case, deletion of the sister of the wh-phrase in Hindi- Urdu would leave an ellipsis site of a size significantly smaller than a full clause (TP/IP). Instead, the ellipsis site would be similar in size to that left by English VPE. It would seem, then, that what appears to be sluicing in Hindi-Urdu might actually be quite different from sluicing in English, which would both support the syntactic position posited for wh-material in this account, as well as explain any restrictions on island violation repair. However, there are a number of reasons why the vP-deletion approach to sluicing in Hindi-Urdu cannot be right. We must first note that it is based on unstable (and shifting) empirical ground. The status of island violation repair in HindiUrdu sluicing is subject to ongoing debate even among native-speaker linguists, and therefore cannot serve as the sole empirical motivation for the theoretical argument above. In addition, the vP-deletion account makes empirical predictions that are not borne out. This account predicts that any material that is not contained within the vP should remain following the sluicing operation. However as we see in the sluice in (26a), verbal material which is outside of vP is clearly elided, including the auxiliaryverb hai, located in T (the elision is in (26b): (26)

a.

Salim koi kitaab khariid-naa caah-taa Ham-eN hai. Salim some book buy-INF want-HAB.M AUX.PRS. we-DAT nahiiN pa-taa kaunsii. NEG know-HAB.M which.R 'Salim wants to buy a book. We don't know which one.'

b.

Elided Material: [Salim __ khariidnaa caahtaa hail

Regardless of our understanding of the position of the subject in Hindi- Urdu (see above discussion in 4.5.2 and Bhatt 2005 for claims that the Hindi- Urdu subject need not be found in Spec, TP), the elision of the auxiliary verb in (26) suggests this elision site is at least TP-sized. Finally, since scrambling outside the vP (as discussed in Chapter 4) should in principle be possible in a sentence in which sluicing will occur, we would expect

Hindi-Urdu case. If the wh-phrase in Hindi-Urdu appears in Spec, vP and a VP-sized site is elided, under Merchant's view, all offending traces of movement to this position should be eliminated It is, howeve1; worth speculating whether Merchant's approach might be better applied to the account of Hindi-Urdu sluicing below in Section 5.3.

Chapter 5. Sluicing in Hindi- Urdu 137

non-wh material to be able to escape a VP-sized sluice via scrambling. However, this is not the case. (27)

dekh-aa, Mujhe pa-taa hai ki Sita-ne kisi-ko lsG.DAT know-HAB.M AUX.PRS that Sita-ERG someone-Ace see-PRF.M lekin mujhe nahiiN pataa (*Sita-ne) kis-ko but lSG.DAT NEG know-HAB.M (Sita-ERG) who-ACC 'I know that Sita saw someone, but I don't know who'.

1his section has demonstrated that neither the clefting approach nor the vP deletion approach seem to provide an adequate account of sluicing in HindiUrdu. In what follows I will pursue an alternative account that assumes that a sluice in Hindi-Urdu is a clause-sized ellipsis, just like sluicing in English and in Kashmiri.

5·3

A new account: Movement to spec, CP

If we are to understand sluicing in Hindi- Urdu as the ellipsis of a clause-sized constituent, we must determine why and how the wh-remnant appears in Spec, CP. In the account ofwh-movement in Hindi-Urdu proposed in this book, I have located the wh-phrase in a dedicated position for interrogative focus at the edge of the verbal domain: Spec, vP. Any subsequent movement to Spec, CP must be driven exceptionally, outside of the normal processes of wh-displacement There are a number of accounts of sluicing in languages understood to be wh-in-situ that propose a type of discourse-driven movement to the clause edge. For instance, Takahashi (1993, 1994) proposes that sluicing structures in Japanese are derived following an exceptional instance of optional (scrambling-like) whmovement to the specifier of CP. 8 Toosarvandani (2007) claims that Farsi sluicing is an instance of fronting to a focus projection (FP) dominating T, driven by the presence of the feature that defines the ellipsis site (E). In what follows I will pursue a similar account of Hindi-Urdu sluicing, with crucial differences motivated by the analysis of Hindi-Urdu presented in above chapters. The approach to the periphery explored in this book suggests that the position for both interrogative and non-interrogative focus in Hindi-Urdu is the preverbal position, and that there is not a dedicated focus position at the clause edge. In contrast to Toosorvandani's approach to Farsi, I have not claimed here that there is

8. See also Hiraiwa and Ishihara (2002), who propose that sluicing structures feature exceptional movement to a focus position above CP.

138 Peripheries in Kashmiri and Hindi- Urdu

any Focus Phrase below (or above) CP in Hindi-Urdu. However, in an interrogative clause in Hindi-Urdu the C head does possess wh-features and does serve as a probe interacting with the wh-phrase goal in the specifier of vP. According to the account presented in Chapter 4, the reason that the wh-phrase does not raise to the clause edge (and an important point of variation between Kashmiri and Hindi-Urdu) is that C does not possess the EPP in Hindi-Urdu. So two questions remain: first, is the syntax of sluices in Hindi-Urdu consistent with the view that the wh-remnant is Spec, CP? If so, how is movement to this position motivated? Let us first confirm that it is in fact movement of the wh-phrase that precedes the deletion of the remainder of the constituent question. Diagnostics which Merchant (2001) terms 'form-identity generalizations' can help determine whether sluicing appears to be movement-derived in Hindi-Urdu. First, the facts presented above with respect to case-marking of the wh-remnant in (12)-(13) suggest that a move-and -delete analysis is appropriate. That is, thew h-remnant is inflected for the same case that it bears in the non-elided counterpart. A second diagnostic involves prepositional phrases. Hindi-Urdu is a language that requires that post-positions be pied-piped. If a sluice is a structure derived via wh-movement, post-positions should be obligatorily pied-piped in sluices as well. (28)

sooc-taa he ki Sita khaana Ali kis-k.e liyee/*kis/*kaun Ali who-for/who.OBL/who.NOM think-HAB.M AUX that Sita food pakaa rahii hai cook PROG.F AUX.PRS 'For whom does Ali think Sita is cooking'?

(29)

hai, magar Ali-ko nahiiN Sita khaana pakaa rahii Sita food cook PROG.F AUX.PRS but Ali-dat NEG pa-taa kis-ke liyee/*kis/*kaun know-HAB.M who-for/who.OBL/who.NOM 'Sita is cooking but Ali doesn't know for whom'.

It seems that the wh-phrase in Hindi-Urdu is in fact undergoing movement to the position it occupies before sluicing occurs. We now address the first question above: whether the wh-remnant in the Hindi- Urdu sluice could be located in Spec, CP. We can return to the account presented in Chapter 2 of the Kashmiri element ki, which optionally precedes embedded clauses with both declarative and interrogative force. Toosarvandani suggests that the wh-remnant in a Farsi sluice appears in Spec, FP below CP in part because the Farsi complementizer appears to the left of the wh-remnant, and this complementizer is presumably located in C. While the facts in Hindi-Urdu appear similar to those in Farsi, our account of the homophonous ki in Kashmiri in Chapter 2

Chapter 5. Sluicing in Hindi-Urdu 139

extends seamlessly to Hindi-Urdu, in that ki in Hindi-Urdu does not seem to be selected by the verb, appears to have no specific semantic content. and is transparent to selection. I claimed in Chapter 2 that ki is a morphological marker of the phase edge - one that is inserted following spell-out. Given this account, the optional appearance of ki before the wh-remnant in a Hindi-Urdu sluice need not be taken as evidence that the wh-remnant is in a position below C. (30)

Aisha-ne ek ciiz khariid-ii lekin mujhe nahiiN pa-taa Aisha-ERG a thing buy-PERF.F but lSG.DAT NEG knoW-HAB.M (ki) kyaa. (that) what 'Aisha bought something, but I don't know what'

(Mahajan 2005)

The facts at hand are so far consistent with an account of Hindi-Urdu in which the wh-remnant is located in Spec, CP. If, as I claimed in Chapter 4, the C head also possesses features relevant to wh-movement, we can now ask how movement to this position can be motivated. Toosarvandani proposes for Farsi that the ellipsis feature (E) is crucial in driving the exceptional instance of fronting. According to Merchant (2001, 2004, 2008), theE feature on a head triggers the non-pronunciation of its complement provided that the complement has an (appropriately) identical antecedent. Toosarvandani suggests that in Farsi, the E feature is only available in the lexicon bundled with a [uwh] and the [EPP]. This is a formal encoding of the observation that in sluicing, the remnant must be an interrogative phrase. In English (and presumably in Kashmiri), the E feature is only found on C. Although Toosarvandani locates theE feature in Farsi on the Focus head (heading the Focus Projection below CP) I propose here that theE feature in Hindi- Urdu is also located on C. In 2.5.1 I introduced an account of the periphery in Kashmiri that claimed that the selection and organization of features into lexical items is a principal locus of grammatical variation, and identified the C heads available in the lexicon of Kashmiri. Similarly. in the analysis of Hindi- Urdu in Chapter 4, we determined the C and v heads in the lexicon of Hindi-Urdu, and demonstrated that it is the properties of specifically these phase-defining functional heads that produces the basic contrasts between the A-bar systems of Kashmiri and HindiUrdu. To the lexicon of Hindi-Urdu, the account of sluicing adds an additional C head. It is precisely like the existing interrogative C of Hindi-Urdu (31a), except that it also possesses theE feature bundled with the EPP (31b ): (31)

a.

C b. C [iQ, uwh] (iQ, uwh] [E, EPP]

140

Peripheries in Kashmiri and Hindi-Urdu

This proposal restricts sluicing in Hindi-Urdu to constituent questions, and forces fronting of wh-material to Spec, CP only in sluicing constructions. We can now offer an account of Hindi-Urdu sluicing in a sentence like (32) that features movement of the wh-remnant to Spec, CP. The wh-phrase first moves to Spec, vP in the way discussed in Chapter 4. Since the C in this case happens to bear the [E, EPP] feature bundle, the wh-phrase will continue to move to Spec, CP. This will permit the valuing of the uninterpretable [wh] feature on Cas well as the the EPP bundled with theE feature. This interaction also values the uninterpretable [Q] feature on the wh-phrase. TheE feature then triggers non-pronunciation of the complement of the C head (represented by strikethrough). (32)

Aisha-ne ek ciiz khariid-ii lekin mujhe nahiiN pa-taa know-HAB.M Aisha-ERG a thing buy-PERF.F but lSG.DAT NEG (ki) kyaa Aisha-ne khariid-ii (that) what Aisha-ERG buy-PERF.F

'Aisha bought something, but I don't know what Aisha bought' (33)

[CP kyaa [u.Q] [iwh] •

«[C:T-..........--h[v"""P---f[v-v.......,....-.---t [uwh] [u.wh] [iQ] [EPP] [E, EPP]

Tc____

___,'j'------'

The analysis schematized in (33) captures the basic properties of sluicing in HindiUrdu, and makes cr udal use of the understanding of the structure of phase-defining functional heads in the lexicon as addressed in Chapter 2. Note that this analysis of sluicing also supports the core proposal concerningwh-material in Hindi-Urdu in Chapter 4 - the specifier of vP is the dedicated position for interrogative focus. In a non-sluicing interrogative sentence, wh-material moves into Spec, vP, on the edge of the vP phase. Only in this position is it accessible to interact via Agree with the C head. If it did not move to Spec, vP, it would be inaccessible to the C head as it would not be on the edge of the phase defined by v. The same is true of the structure in (33). Were the wh-phrase not already in Spec, vP, it could not Move to Spec, CP, and therefore could not be the wh-remnant in a sluicing construction. The uninterpretable features on C would not be valued, and the derivation would crash. And because the E feature (bundled with the EPP) is can only be found on the C head in a Hindi-Urdu sluicing construction (as only TP-sized constituents may be sluiced), the wh-phrase must be in the domain of the C probe in order to Move to Spec, CP and form a licit sluicing derivation. Notice that the account in (33) extends to Kashmiri sluices, such as those in (3)-(4). The only difference is that the C head in a Kashmiri interrogative

Chapter 5. Sluicing in Hindi-Urdu 141

already possesses the EPP feature, requiring a wh-phrase to appear in its specifier. It is therefore not necessary for the EPP to be bundled with E in Kashmiri (or in English, for that matter), since the movement to the clause edge in these languages is not exceptional. What I have not tried to capture with the analysis in (33) is the controversial island repair status of sluicing in Hindi-Urdu. Because this is such an emblematic feature of English sluicing it has been a focus of sluicing investigations in a wide range of languages. However, the facts in Hindi-Urdu are simply not wellestablished. Without a clear set of empirical observations, I can only leave an account of the behavior of island constraints in Hindi-Urdu sluicing to future research, with the hope that this data can be used to further assess the account presented here. A final objection to the account in (33) comes from Merchant's (2008) and Kizu's (1997) criticism of the approach to Japanese sluicing which also hinges on movement to Spec, CP. Given that this would be the only instance of required wh-movement to the clause edge in Japanese, they assert that we should first look to other possible analyses. However, in this chapter I have reviewed the available alternatives in the case of Hindi-Urdu (including the proposal in Manetta 2006), and have determined that they do not offer a coherent account of HindiUrdu sluicing. Moreover, as Toosarvandani (2007) points out, the proposal that a lexical item exists across a number of languages which groups together wh-features, the E feature, and the EPP is evidence of considerable regularity across lexicons. Given the account proposed here, we can further observe that these features seem coincide on the phase-defining functional head C crosslinguistically (in languages as diverse as English, Hindi- Urdu, and Kashmiri). Proposals of this kind directly contradict the core premise of Merchant's (2001) influential approach to sluicing- that the syntax of sluicing is simply the syntax of an ordinary wh-question. In the case of Hindi-Urdu, it seems that sluicing cannot be derived simply from a typical wh-interrogative. Accounts requiring exceptional movement to derive sluicing mark the continued unraveling of sluicing as a unified syntactic phenomenon (pointed out by Hoyt & Teodorescu 2004, van Craenenbroek & Liptak 2006, Potsdam 2007). On the other hand, the lexical consistencies identified here, particularly with respect to the phase-defining functional heads, speak to a set of universal extra-syntactic principles guiding the composition of the lexicon.

CHAPTER

6

Conclusions The analysis of the peripheries of Kashmiri and Hindi- Urdu presented in this book not only represents an advancement in empirical coverage, but also provides new insight into the way in which key components of the grammar are structured. In this concluding chapter we will look specifically at ramifications for phase theory, our understanding of displacement, and the source of crosslinguistic variation.

6.1

A theory of the periphery

The object of inquiry in this book has been the A-bar systems of Kashmiri and Hindi- Urdu. More specifically, I examined the rich left edge of the Kashmiri clause and long-distance wh-dependencies in Kashmiri and in Hindi-Urdu, with special focus on wh-expletive constructions. The micro-comparative approach adopted here allowed us to construct an account of the periphery, its internal structure, and its role in the derivational process. Specifically, the investigation pursued in this book has yielded three key results. First, we have arrived at an improved understanding of the organization of the left periphery of the clause: a single phase-defining functional head, C, containing an ordered set of features. This permits even a relatively rich left periphery to be accommodated in a way that is consistent with the notion of the phase. Second, this book proposes an approach to wh-dependencies that posits that the same basic mechanisms drive both the A-movement system (the system of movement for the purposes of case and agreement) and the A-bar movement system (movement not for the purposes of case and agreement, but instead for reasons of discourse-status). This approach afforded a clear account oflong-distance wh-dependencies in Kashmiri. Third, I have advanced an account of parametric variation in terms of the properties of the phase-defining heads C and v. This provides an understanding of a systematic set of contrasts between the A-bar systems ofKashmiri and HindiUrdu, and permits a unified approach to the strategies of forming long-distance wh-dependencies in the two languages.

144 Peripheries in Kashmiri and Hindi-Urdu

These results contribute to the development of a theory of the periphery, or the dedicated positions in the syntax to which discourse-driven movement takes place. Our analysis here suggests that there is not only a single periphery, at the left edge of the clause, but also potentially clause-internal peripheries at which we find A-bar elements. This work has specifically identified a periphery at the clause edge (C), and a periphery clause-internally (v), and has furthered an understanding of how the properties of these peripheral heads are organized and structured. Moreover we have established the connection between these peripheries and phasehood and can thereby assess how these heads mediate seemingly unbounded w h-dependencies.

6.2

Wh-expletives and the role of expletives in the grammar

The developments discussed above are in a large part made possible by an investigation ofwh-expletives. Analysis of the DP-expletive has been crucial to our development of a theory of the A-system. The possibility that a similar syntactic object might exist in the A-bar system sparks a range of questions, both about the nature of A-bar dependencies and the purpose of expletives themselves. In this section I will briefly address what is completely new in the account ofwh-expletives in this book, and why this provides new insights into the processes of the grammar. Many previous accounts of what I have called wh-expletive constructions do not view an invariant, minimal wh-word (represented in (lb) as what), as an expletive at all: (1)

a.

Wh-displacement [wh-phrase i ... [ep ... __ i]

b.

Wh-dependency with embedded wh-phrase [what ... [ep wh-phrasei ... --il

Because this invariant, minimal wh-word seems to appear at the position at which the scope of full wh-phrase was to be interpreted it has also been called a scopemarker.1 As is discussed in more detail in Chapter 3, there have been two core families of approaches to the construction as in (lb). The first was to assume that the minimal wh-word is truly contentful and an argument of the verb in the matrix clause, and that the crux of the analysis was to determine how the embedded wh-phrase or the embedded clause itself was linked to that scope-marker

1. Except in the case of wh-dependendes spanning more than two clauses, which in many languages require these minimal wh -words to appear in each clause intervening between the full wh-phrase and the position at which it is interpreted

Chapter 6. Conclusions 145

(Indirect Dependency). The second was to maintain the non-contentful nature of the matrix wh-word and propose that its position be occupied (or replaced) by the fully contentful wh-phrase at the level of interpretation (Direct Dependency). The analysis proposed in this book represents a significant departure from these previous approaches. Importantly, I assume that the minimal wh-word in the matrix clause in a structure like (lb) is in fact an expletive. This means that it lacks any interpretive content (or in our analytical terms any interpretable features). However, the full wh-phrase in the embedded clause need not move out of that clause at any level of the grammar to be licensed or correctly interpreted. Through the operation Agree and unselective binding, both processes can proceed at a distance. In this analysis of (lb), wh-expletives are in no way linked to the whphrases whose position of interpretation they indicate, and have no independent interpretation of their own. Instead, I propose that wh -expletives play a syntactic role analogous to that played by DP-expletives; that is, they serve to satisfy the requirement of a head that it have an additional specifier. That head in turn interacts with the full wh-phrase via the operation Agree. It is the formal encoding of this analogy that yields the major results of this book. The existence of a wh-expletive provides special insight into the basic design of the grammar. Particularly mobilized in this book is the observation that expletives represent a striking parallel between the A and A-bar movement systems. In this view, both the system of displacement for case and agreement and the system of discourse-driven displacement feature non-contentful minimal elements that appear in the positions to which contentful phrases normally move. This suggests that it is not necessarily movement itself that satisfies criteria of the syntax, but instead that there are fixed positions which must be filled. In the A-bar movement system, these positions are at the peripheries. Therefore the peripheral phase-defining head in a wh-depedency possesses a property requiring its specifier to be filled with overt wh-material. This is the property we have encoded here as the EPP. Wh-expletives permit a wider range of wh-dependencies to be constructed alongside full wh-movement, including varying degrees of partial movement (displacement to a position that is not the base position but still more embedded than the position of interpretation). However, not all languages have a wh-expletive in the lexicon, meaning that only some exhibit multiple ways of forming a longdistance wh-dependency. If the understanding of wh -expletives presented here is correct, the existence of a wh-expletive in a given language points to the existence of a peripheral head with a requirement that its specifier be filled. This is why the very presence of the wh-expletive construction in Hindi-Urdu is such a puzzle under the wh-in situ view. If there are no dedicated peripheral heads which must have their specifiers filled by overt wh-material in a long -distance wh -dependency,

146 Peripheries in Kashmiri and Hindi-Urdu

why would there be constructions in Hindi-Urdu with wh-expletives? The very presence of the wh-expletive suggests that such a dedicated head does exist, and in fact leads us to investigate whether that head is v. Although we cannot answer here the question of why expletives exist, nor why there should be heads that require certain kinds of overt material to appear in their specifiers, we can broaden the scope of the question itself. To understand expletives we cannot focus only on the system of case and agreement; we must also examine the system of discourse-motivated displacement.

6.3

Phases and their edges

This book has constructed a revised understanding of the periphery. The periphery is a dedicated space for elements with certain kinds of information-structure roles found at the edge of each derivational unit, or phase. The properties of the phasedefining head, or in more analytical terms, the features that it bears, determine the shape the periphery will take. The phase-defining head can potentially Agree with an element present in its domain, cause the element to Move to and surface in its specifier, or serve as an intermediate host for an element that will ultimately move into subsequent phases. In this view, the universal properties of the periphery define A-bar processes generally, and the extent to which these properties may vary determines the extent of crosslinguistic variation in A-bar systems. If we probe the notion of the phase itself, we find that these conclusions concerning A-bar systems say something important about derivational processes. The phase itself is understood to be the basic computational unit constructed by the derivation before transfer to the interfaces. When we commit to naming a phrase as a phase, we are committing to the shape and size of what is fundamentally a computational object. The analysis explored here commits us at least to the phases smaller than the clause. The proposal CP and vP- importantly the phase is here suggests that there is a crucial link between elements that will ultimately take on certain information-structure roles (that is, A-bar elements) and the size and nature of the computational object. Many types of A-bar elements are characterized by the ability to be displaced from their point of introduction. The structure of the grammar is such that the only way for such an element to be displaced outside of a phase is to move to the edge of that phase. Therefore, the head that we postulate to define a minimal computational unit will be the head that defines the very nature of the A-bar system. Or perhaps conversely, we should investigate whether those very heads that seem to play a defining role in the A-bar system have met other, independent criteria that would lead us to believe that they represent the size and type of object that should be considered a computational unit.

tar

Chapter 6. Conclusions 147

6.4

Displacement and formal features

Displacement of linguistic objects- that is, the surfacing of an object at a position distinct from the one at which it is introduced into the syntax- seems to be a core property of human language. Partial wh-movement constructions such as those found in Kashmiri, and I claim here, in Hindi-Urdu, are examples of displacement par excellence. The partially moved wh-phrase is located not in its base-generated position, nor is it found at its point of semantic interpretation. It is instead located at a fixed syntactic position somewhere in between. (2)

[CP what-expl ... [cP wh-phrasei ... --d

The account presented here has analyzed this unique form of displacement in terms of uninterpretable features, or purely syntactic requirements. In other words, it is the syntactic requirement of the matrix C head in (2) that its specifier be filled in order for the derivation to converge; this is not needed for the purposes of semantic interpretation. Zeijlstra (2009) claims that all dislocation results from a mismatch between the semantic and phonological components, and that it is purely formal, uninterpretable features which license the required dislocation. In this book I have shown that the grammar exhibits a certain flexibility in pursuit of a solution to this mismatch (and the satisfaction of purely formal criteria). Alongside instances of pure displacement (full wh-movement) we see the use of noncontentful syntactic objects (wh-expletives) to satisfy the same formal features and achieve the effects of long-distance dislocation without full displacement of the linguistic object

6.5

New research opportunities

The work done here opens several important paths of inquiry. While this book has dealt extensively with wh-movement and wh-expletive constructions in Kashmiri and Hindi-Urdu, we have not addressed relative and correlative clauses in these languages. There is a wealth of previous research on these structures in Hindi-Urdu (Kachru 1973, Dayal1996, Mahajan 2000, Bhatt 2003, Davison 2009, Bhatt & Munshi 2009, among others). While less work has been done in Kashmiri (most recently see Bhatt & Munshi 2009), it is well-known that these clauses have a special status in the language, as they are an instance in Kashmiri of verb final finite clauses. (3)

Swa kitaab [yos tsl raath paraan oosuk] The book [REL 2SG yesterday read-IMPERF AUX.PST-2SG The book that you were reading yesterday....

148 Peripheries in Kashmiri and Hindi-Urdu

It is clear that the periphery of the relative clause is a unique one, and the micro-comparative investigation of relative clauses in Indic languages remains an active area of current research. A second avenue of inquiry is suggested by the discussion of sluicing in Kashmiri and Hindi-Urdu in Chapter 5. Further comparative investigation of sluicing, gapping, and other ellipsis operations in the two languages should continue to yield interesting results, and a great deal of careful empirical work remains to be done. Ideally these processes have much to tell us about the shape of the clause and the nature of the movements of constituents that escape ellipsis. The third opportunity that is opened up by the theoretical developments concerns the wh-expletive construction. A wide range of languages outside of Indic exhibit this construction alongside full wh-displacement (for instance German, Hungarian, Iraqi Arabic, and Japanese). It will be important to return to these well-researched versions of the wh-expletive construction to determine whether the approach presented here can offer new insight. In particular, the interaction between wh-expletive constructions and verb-second (as in Germanic) and whexpletive constructions and pre-verbal focus (Hungarian), should prove especially relevant in light of the progress made here.

6.6

Summary

The account featured in this book represents increased economy on several fronts. I have first proposed an account of a relatively rich left periphery that employs a single functional head, while preserving the empirical coverage of the cartographic effort. I have also pursued an understanding of A-bar movement that assumes the underlying mechanisms of A-movement. This is akin to saying that the grammar has only a very limited set of feature-driven operations that drive both flavors of displacement and agreement. Finally. I have shown that two languages such as Kashmiri and Hindi-Urdu, whose peripheries seem radically different on the surface (verb-second vs. verb-final wh-movement vs. wh-in-situ), can be understood to exhibit nearly identical clausal topology, just at different phase boundaries. However, economy of derivation is not advanced in this analysis for its own sake, but instead in pursuit of a theory of the A-bar movement system and A-bar dependencies that provides a better overall understanding of the inter-language variation at hand. I began by suggesting that this book would be guided by a series of longstanding research questions concerning the periphery: how is the periphery itself structured (and to what degree is this fixed by universal principles)? What are the mechanisms that drive displacement to the periphery? And finally, how does the

Chapter 6. Conclusions 149

periphery mediate instances apparently unbounded displacement? The results of this investigation claim that these dependencies are mediated by the peripheries, clause-edge (C), and clause-internal (v). The structure and organization of heads permits us to assimilate our understanding of these seemingly unbounded dependencies with our understanding of clause-bound interactions.

References Adger, David. 2003. Core Syntax: A Minimalist Approach. Oxford: Oxford University Press. Anderson, Stephen. 1977. On the formal description of inflection. Chicago Linguistic Soctety Papers 13: 15-44. Anderson, Stephen. 1982. Where's morphology? Linguistic Inquiry 13: 571-612. Aoun, Joseph & Elabbas Benmamoun. 1998. Minimality, reconstruction, and PF movement. Linguisticlnquiry 29 (4): 569-597. Bayer, Jose£ 1996. Directionality and Logical Form: On the Scope of Focusing Particles and Wh-in-situ. Dordrecht: Kluwer. Beck, Sigrid & Steven Berman. 2000. Wh-Scope Mar~ Direct vs Indirect Dependency. In Cyclicity in Wh-Scope Marking, Uli Lutz. Gereon MUller, Arnim von Stechow (eds ), 17-44. Amsterdam: John Benjamins. Belletti, Adriana (ed). 2002. Structures and Beyond. The Cartography of Syntactic Structures. Oxford: Oxford University Press. Beninca, Paola 200 l. The position of Topic and Focus in the Left Periphery. In Current studies in Italian Syntax, G. Cinque & G. Salvi (eds), 39-64. Amsterdam: Elsevier. Bhatt, ~jesh. 2003. Lecture Handouts: Topics in the Syntax of Modern Indo- Aryan Languages. MIT, Spring 2003. Bhatt, Rajesh. 2005. Long Distance Agreement in Hindi- Urdu. Natural Language and Linguistic Theory 23: 757-807. Bhatt, ~lesh & Veneeta DayaL 2007. Rightward Scrambling as Rightward Remnant Movement. Linguistic Inquiry 38:287-301. Bhatt, Rajesh, & Sadaf Munshi. 2009. Two locations for negation: evidence from Kashmiri. Linguistic Variation Yearbook 9: 205-240. Bhatt, Rakesh. 1993a. Nominative Objects in Kashmir!. Paper presented at the Linguistic Soctety of American meeting. Los Angeles, Jan. 7-10. Bhatt, Rakesh. 1993b. Psyched out-Analysing quirky constructions. Papers from the 29th Regional Meeting of the Chicago Linguistic Society, pp. 77-88. Bhatt, Rakesh & James Yoon. 1992. On the Composition ofComp and Parameters ofV2. In Proceedings ofWCCFL X., D. Bates (ed). Stanford: CLSI, 41- 52. Bhatt, Rakesh. 1999. Verb Movement and the Syntax ofKashmiri. Dordrecht: Kluwer. Biberauer, Theresa. 2005. Splitting not Spreading: A new perspective on the Cff connection. GLOW in Asia, New Delhi. Biberauer, Theresa & Ian Roberts. 2005. Subjects, Tense, and verb-movement in Germanic and Romance. GLOW in Asia, New Delhi. Bobaljik. Jonathan & Hoskuldur Thrainsson 1998. Two Heads Arent Always Better than One. Syntax 1(1): 37-71. Boeckx, Cedric. 2003. Intricacies oflcelandic Agreement. ms. University of Maryland/Harvard University. Bo8k.ovic, 2eljko 2000. Sometimes in Spec CP, sometimes in situ. Step by step: Essays on minimalism in honor of Howard Lasnik. Roger, Martin, David Michaels, and Juan Urtagereka (eds), 53-87. Cambridge, Mass.: MIT Press.

152

Peripheries in Kashmiri and Hindi-Urdu Bo8k.ovi.C, :leljko. 2002. A-movement and the EPP. Syntax 5 (3): 167-218. Brody, Michael. 1995. Remarks on the order of elements in the Hungarian focus field In Approaches to Hungarian 5. ed Istvan KeneseL pp. 29-44. Szeged: JATE Press. Bruening, Benjamin. 2006. Differences between the Wh-scope-marking and Wh-copy constructions in Passamaquoddy. Linguistic Inqutry 37: 25-39. Bruening, Benjamin. 2007. Wh-in-Situ Does Not Correlate with Wh-Indefinttes or Question Particles. Linguistic Inquiry 38: 139-166. Burzio, Luigi 1986. Italian Syntax. Dordrecht : D. Riedel. Butt, Miriam & Tracy King. 1996. Structural Elements and Focus Wrthout Movement. Proceedings of the First LFG Conference. Grenoble: XRCE. Chandra, Pritha. 2005. Hindi-Urdu Long Distance Agreement: Agree, AGREE or Spec-Head? ms. University of Maryland, College Park. Chandra, Pritha & Ince, Atakan. 2007. Genitive Islands and Sluicing in Hindi-UrdiL Paper presented at Western Conference on Linguistics, San Diego, November 2007. Chatterji, S.K. 1963. Languages and Literature of MDdern India. Calcutta: Bengal Publishers Private Ltd Chomsky, Noam. 1981. Lectures on Government and Binding. Dordrecht: Foris. Chomsky, Noam. 1986. Barriers. Cambridge: MIT Press. Chomsky, Noam. 1995. The Minimalist Program. Cambridge, Mass: MIT Press. Chomsky, Noam. 2000. Minimalist Inquiries: the Framework. In Step ey Step, eds. Roger Martin et a1, 89-155. Cambridge, Mass: MIT Press. Chomsky, Noam. 2001. Derivation by Phase. In Ken Hale: a life in language, M. Kenstowicz (ed), 1-52. Cambridge, Mass: MIT Press. Chomsky, Noam. 2004. Beyond explanatory adequacy. In Structure.s and Beyond, A. Belletti (ed),104-131. Oxford: Oxford University Press. Chomsky, Noam. 2008. On Phases. In Foundational Issues in Linguistic Theory: Essays in Honor ofJean-Roger Vergnaud, Robert Freidin. Carlos Peregrfn Otero & Maria Luisa Zubi.zarreta (eds), 133-166. Cambridge, MA: MIT Press. Chung, Sandra, Bill Ladusaw, & James McCloskey. 1995. Sluicing and Logical Form. Natural Language Semantics 3:239-282. Cinque, Guglielmo. 1990. Tjpes ofA'-dependencies. Cambridge, MA: MIT Press. Cinque, G. 1993. A null theory of phrase and compound stress. Linguistic Inquiry 24, 239-298. Cinque, Guglielmo. 1999. Adverbs and Functional Heads. Oxford: Oxford University Press. Cowper, Elizabeth. 2005. The Geometry oflnterpretable Features: Infi in English and Spanish. Language 81: 10-46. Davison, Alice. 1988. Operator binding, gaps, and pronouns. Linguistics 26 (2): 181-214. Davison, Alice. 2003. Dependent Structural Case and the Role of Functional Projections. Presented at the 23rd SALA Roundtable, Austin. Davison, Alice. 1988. Operator binding, gaps, and pronouns. Linguistics 26: 181-214. Dayal, Veneeta. 1994. Binding facts in Hindi and the Scrambling phenomenon In Theoretical Perspectives on Word Order Issues in South Asian Languages B. Butt, T. King, & G. Ramchand, 237-261 (eds). Stanford, California: CSLI. Dayal, Veneeta. 1996. Locality in Wh-Quantijictaion. Dordrecht: Kluwer. Dayal, Veneeta. 2000. Scope Marking: Cross-linguistic Variation in Indirect Dependency. In Wh-Scope Marking, lniLutz, Gereon Miiller, Arnim von Stechow (eds ), 157-194. Amsterdam: John Benjamins.

References 153 Dayal, Vaneeta 1994. Scope-marking as Indirect Wh-Dependency. Natural Language Semantics 2: 137-170.

De Hann, Germen & Fred Weerman 1985. Finiteness and Verb Fronting in Frisian. In Verb Second Phenomena In Germanic Languages, eds Hubert Haider & Martin Prin.7horn. Dordrecht: Forts. De Villiers, Jill, Peter de Villiers, & Thomas Roeper. 2007. Wh-questions: Moving beyond the first Phase. ms. Smith College. Diesing, Molly. 1990. Verb Movement and the Subject Position in Yiddish. In Natural Language and Linguistic Theory. 8: 41-79. Dwivedi, Veena. 1994. Syntactic Dependencies in Relative Phrases in Hindi. Ph.D. dissertation, University of Massachusetts, Amherst. Edgeworth, M.P. 1841. Grammar and Vocabulary of the Kashmiri Language. Journal of the Royal Asiatic Society ofBengallO: 2, 1038-1064. Fanselow, Gilbert & Anoop Mahajan. 1996. Partial movement and Successive Cyclicity. In Papers on Wh-Scope Marking, eds. Uli Lutz et al., 131-161. Arbeitspaperie de Sonderforschungsbereicb.s 340, University of Stuttgart and University ofTubingen. Fanselow, Gilbert & Anoop. Mahajan. 2000. Towards a Minim.alist Theory of Wh-expletives, Wh-copying, and Successive Cyclicity. In Wh-Scope Marking, Uli Lutz, Gereon Mi.iller, Arnim Von Stechow (eds.), 195-230. Amersterdam: John Benjamins. Folli, R. & Heidi Harley. 2004. On obligatory obligation: The composition of Italian causatives. A Castro et al. (eels). Romance, Op. 47: collected papers on Romance syntax. 87-113. MITWPL 47. Cambridge, MA: MIT Press. Fox, Danny. 2000. Economy and Semantic Interpretation. Cambridge, Mass: MIT Press. Fox, Danny & David Pesetsky. 2005. Cyclic Linearization of Syntactic Structure. Theoretical Linguistics 31: 1-45. Fox, Danny & Howard Lasnik. 2003. Successive-Cyclic Movement and Island repair: The Difference between Sluicing and VP-Ellipsis. Linguistic Inquiry 34 (1): 143-154. Fox, Danny & David Pesetsky. 2005. 'Cyclic linearization of syntactic structure.' Theoretical Linguistics 31 (1/2): 1-46. Fussman, Gerard 1972. Atlas Linguistique des parlers Dardes et Kafirs. Ecole Francaise d'Extreme Orient, voL 86. Gambhir, Vijay. 1981. Syntactic restrictions and discourse functions of word order in standard Hindi. Doctoral Dissertation, University of Pennsylvania: Philadelphia, PA. Georgi, Doreen & Gereon Miiller. 2010. Noun-phrase structure by reprojection. Syntax 13(1): 1-36.

Greybenyova, Lydia. 2006. Sluicing Puzzles in Russian. In Proceedings of the Annual Workshop on Formal Approaches to Slavic Linguistics 14. Ann Arbor.: Michigan Slavic Publications. Grierson, G.M. 1919. The Linguistic Survey of India. 7:2, Calcutta. Harley, Heidi. 2008. On the Causative Construction In Miyagawa, S. & S. Mamoru (eds). The Oxford Handbook ofJapanese Linguistics. OUP: Oxford Hayes, Bruce & Lahiri, Aditi (1991). Bengali intonational phonology. Natural Language & Linguistic Theory 9, 47-96. Henry, Alison. 1995. Belfast English and Standard English: Dialect Variation and Parameter setting. Oxford: Oxford University Press. Herburger, E. 1994. A Semantic Difference between Full and Partial Wh-Movement in German. LSA Annual Meeting, Boston, MA

154 Peripheries in Kashmiri and Hindi-Urdu

Hiraiwa, Ken. 2001. Multiple Agree and the Defective Intervention Constraint in Japanese. In MITWPL 4fJ: the Proceedings of HUMIT 2000, 67-80. Cambridge, Mass: MIT Working Papers in Linguistics. Hiraiwa, Ken & Shinichiro Ishihara. 2002. Missing links: deft, sluicing, and 'no da' construction in Japanese. MIT Working Papers In Linguistics 43: 35-54. Hook, P.E. 1976. Is Kashmiri an SVO Language? Indian Linguistics 37:137-142. Hook, Peter & Omkar Koul. 2004. Case as Agreement In Non-nominative Subjects, eds. Bhaskarararoa, Peri & K. V. Subbarao, 101-114. Amsterdam: John Benjamins. Hook, Peter & A. Manaster-Ramer. 1985. The Position of the Finite Verb in Germanic and Kashmiri. Towards a Typology ofV2 Languages. In Germanic Linguistics, ed J.T. Faarlund, 46- 58. Indiana University Linguistics Club, Bloomington. Hornstein, Norbert. 1995. Logical Form. Oxford: Blackwell Horvath, Julia. 1997. The Status of'Wh-expletives' and the Partial Wh-movement Construction of Hungarian. Natural Language and Linguistic Theory 15: 509-572. Hoyt, Frederick & Alexandra Teodoresc\L 2004. How many kinds of sluicing and why? Single and multiple sluicing in Romanian, English, and Japanese. ms., University of Texas, Austin. Ince, Atakan. 2006. Pseudo-Sluicing in Thrkish. In University of Maryland Working Papers in Linguistics 14. eds. N. Kazanina, U. Minai, P. Monahan & H. Thylor (eds), 111-126. College Park, MD: UMWPiL. Ishihara, S. 2001. Scrambling and its interaction with stress and focus. MIT Working Papers

in Linguistics 38: Proceedings of the Twelfth Student Conference in Linguistics (SCIL12). Cambridge, MA: MITWPL, 95-110. Kachru, Braj. 1969. A Reference Grammar ofKashmiri. Urbana: University of Illinois Press. Kachru, Y., B.B. Kachru & T.G. Bhatia. 1976. The Notion of "Subject": A note on Hindi-Urdu, Kashmiri and PunjabL In The Notion of Subject in South Asian Languages, ed M. Berma, 79-108. University of Wisconsin, Madison. Karttunen, L. 1977. Syntax and semantics of questions. Linguistics and Philosophy 1: 3-44. Kaula, Ishvara. 1898. The Kashmir Shabdamrita A Grammar in SanskritofKashmiri, G.A. Grierson (ed), Calcutta: The Asiatic Society. Kennedy, Christopher & Jason Merchant. 2000. Attributive Comparative Deletion. Natural Language and Linguistic Theory 18: 89-146. Kidwai, Ayesha 1999. Word Order and Focus Positions in Universal Grammar. In The Grammar of Focus, ed Goerges Rebuschi & Laurice Tuller (eds), 213-245. Amsterdam: John Benjamins. Kidwai, Ayesha 2000. XP-Adjunction in Universal Grammar: Scrambling and Binding in I-nndiUrdu. Oxford: Oxford University Press. Kim, Jeong.-Seok. 1997. Syntactic Focus Movement and Ellipsis: A minimalist approach, Ph.D. dissertation. University of Connecticut. Kiss, Katalin, E. 1987. Configurationality in Hungarian. Dordrecht: Kluwer. Kizu, Mika 1997. A note on sluicing in wh-in-situlanguages. Ms., in ProceedingsofSCIL, MIT Working Papers in Linguistics. Kizu, Mika 2005. Cleft Constructions in Japanese Syntax. New York: Palgrave MacMillan. Ko, Heejeong. 2007. Assymetries in Scrambling and Cyclic Linearization. Linguistic Inquiry. 38:49-83. Koul Omkar N., & Ruth Lalla Schmidt. 1983. Kashmir!: A Sociolinguistic Survey. IlLS, Patiala Kumar, Rajesh. 2006. Negation and Licensing of Negative Polarity Items in Hindi Syntax. New York: Routledge.

References 155 Kuroda, S.- Y. 1988. Whether we agree or not: a comparative syntax of English and Japanese. In Papers from the Second International Workshop on Japanese Syntax, W. Poser (ed), 103-143. Stanford: CSLI. Kuwahara, K. 1996. Multiple wh-phrases in elliptical clauses and some aspects of clefts with multiple foci. Formal approaches to Japanese linguistics 2.MITWPL 29, 97-116. Iatridou, Sabine & Anthony Kroch. 1992. The Licensing of CP-recurston and its Relevance to the Germanic Verb-Second Phenomenon. Working Papers in Scandinavian Syntax 50:1-24. Lahiri. Utpal.. 1998. Focus and Negative Polarity in Hindi Natural Language Semantics 6: 57-123. Lahiri, Utpal.. 2002. On the proper treatment of"expletive wh" in Hindi Lingua 112: 501-540. Lahne, Antje. 2009. A multiple specifier approach to left-peripheral architecture In Romance Languages and Linguistic Theory 2006, ed Torck, Daniele & W. Leo Wetzels. Amsterdam: John Benjamins. Leech, R. 1844. A Grammar of the Cashmeeree Language. Journal of the Asiatic Society, 13 (1): 397-421. Mahajan, Anoop. 1990. The A/A-bar distinction and movement theory. Ph.D. dissertation, MIT, Cambridge, Mass. Mahajan, Anoop. 1994. Toward a Unified Theory of Scrambling. In Studies on Scrambling: MiJvement and Non-movement approaches to Free Word-Order Phenomena, ed N. Corver & H. C. van Riemsdijk. Berling: Mouton de Gruyter, 301- 330. Mahajan, Anoop. 1997. Rightward scrambling. In Rightward movement, eds. Dorothee Beerman, David LeBlanc, & HenkvanRiemsdijk.186-214. Amsterdam: John Benjamins. Mahajan, Anoop. 2000. Towards a Unified Treatment ofWh-Expletives in Hindi and German. Wh Scope Marking. Eds. Uli Lutz, Gereon Mi.iller, Arnim von Stechow, 317-332. Amsterdam: John Benjamins. Mahajan, Anoop. 1994. The Ergativity Parameter: have-be Alteration, Word Order and Split Ergativity. In Proceedings of the North Eastern Linguistic Society, ed Merce Gonzalez. 317-331. Mahajan, Anoop. 2005. Gapping and Sluicing in Hindi. GLOW in Asia, New Delhi. Malhotra, Shitt & Pritha Chandra. 2007. A Short Note on Wh-scope-marking. University of Maryland Working Papers in Linguistics 16, ed A. Omaki, I. Ortga-Santos, J. Sprouse & M. Wagers, College Park, MD: UMWPiL. Manetta, Emily. 2005a Feature Structure on Functional Heads: the Kashmirl Clause Edge. Presented at GLOW in Asia, New Delhi. Manetta, Emily. 2005b. The Minimalist A-bar System. LSA Annual Meeting, Oakland, CA Manetta, Emily. 2006. Wh-expletives in Hindi-Urdu: the vP-phase. LSA Annual Meeting, Albuquerque, New Mexico. Manetta, Emily. 2006. Peripheries in Kashmiri and Hindi-Urdu. Ph.D. dissertation. University of Caifornia, Santa Cruz. Manetta, Emily. 2010. Wh-expletives in Hindi-Urdu: The vP phase. Linguistic Inquiry 41: 1-34. Manetta, Emily. To appear. Reconsidering rightward scrambling: Postverbal constituents in Hindi-Urdu. Linguistic Inquiry. Mascia, Colin. 1991. The Indo-Aryan Languages. Cambridge: Cambridge University Press. May, Robert. 1977. The grammar of quantification. Ph.D. dissertation. Massachusetts Institute of Technology.

156

Peripheries in Kashmiri and Hindi-Urdu McCloskey, James. 1990. Resumptive PronoWJS, A-bar Binding, and Levels ofRepresentation in Irish. In the Syntax ofMDdern Celtic Languages, ed. Randall Hendrick. 199-248. San Diego: Academic Press. McCloskey, James. 2001. The Morphosyntax of Wh -extraction in Irish. Journal of Linguistics 37: 67-100.

McCloskey, James. 2005. Questions and Questioning in a Local English. In Cross-Linguistic Research in Syntax and Semantics: Negation, Tense, and Clausal Architecture. Zanuttinl. R., H Campos, E. Herburger & P. Portner (eds). Washington D. C.: Georgetown University Press. McDaniel, Dana 1989. Partial and Multiple WH-Movement. Natural Language and Linguistic 1heory 7: 565-604. Merchant, Jason. 1999. The syntax of silence: Sluicing, islands, and identity in ellipsis. Ph.D. dissertation, University of California, Santa Cruz. Merchant, Jason. 2001. The syntax ofsilence: Sluicing, islands, and the theory of ellipsis. Oxford: Oxford University Press. Merchant, Jason. 2002. PF Output Constraints and Elliptical Repair in SAl Comparatives. Proceedings of the 21st West Coast Conference on Formal Linguistics, Mikkelson, Line and Chris Potts (eds). Somerville, MA: Cascadilla Merchant, Jason. 2008. Variable Island Repair Under Ellipsis. In Topics in Ellipsis. Johnson, Kyle (ed), 132-153. Cambridge: Cambridge University Press. Mikkelsen, Line. 2009. 0 n what comes first in a verb-second language. Syntax Circle, University of California, Berkeley. April2009. Miyagawa, Shigeru. 1997. Against optional scrambling. Linguistic Inquiry 28: 1-25. Miyagawa, Shigeru. 2001. The EPP, scrambling, and wh-in-situ. In Ken Hale: A life in language, ed.. by Michael Kenstowicz, 293- 338. Cambridge, MA: MIT. Moulton, Keir. 2004. External arguments and gerunds. Toronto Working Papers in Linguistics 22: 121136.

Miiller, Gereon 20 10. 0 n deriving CED effects from the PIC. Linguistic Inquiry 41 (1 ): 35-82. Newmeyer, Frederick. J. 2004. On Split-CPs, Uninterpretable Features, and the "Perfectness" of Language. ZAS Papers in Linguistics 35. Nissenbaum, Jon. 2000. Investigations of convert phrase movement Ph.D. Dissertation. MIT, Cambridge, MA. Pesetsky, David, & Esther Torrego. 2001. T-to-C movement: Causes and consequences. In Ken Hale: A life in language, Michael Kenstowicz (ed), 355-426. Cambridge, MA: MIT Press. Pesetsky, David & Esther Torrego. 2007. The Syntax of Valuation and the Interpretability of Features. In Phrasal and Clausal Architecture, S. Karimi, V. Samiian & W. Wilkins (ed), 262-294. Amsterdam: John Benjamins. Pollock, Jean- Yves. 1989. Verb movement, universal grammar, and the structure ofiP. Linguistic Inquiry 20(3): 365-424. Potsdam, Eric. 2007. Malagasy sluicing and its consequences for the identity requirement on ellipsis. NLLT23, 577-613. Puskas, Genoveva 2000. Word Order in Hungarian. Amsterdam: John Benjamins. Pylkkiinen, Liina 2002. Introducing Arguments. Ph.D. Thesis. MIT. Cambridge, MA. Pylkkiinen, Liina 2008. Introducing Arguments. Cambridge, Mass: MIT Press. Rack.owski, Andrea & Norvin Richards. 2005. Phase edge and extraction: a Tagalog case study Linguistic Inquiry 36: 565-599. Raina, Achla 1991. An S-Selection Approach to Grammar: Some Issues in Kashmiri Syntax. Ph.D. Dissertation, Indian Institute of Technology, Kanpur.

References 157 Reinhart, Tanya 1998. Wh-in-situ in the framework of the Minimalist Program. Natural Language Semantics 6: 29-56. Richards, Norvin 1998. The Principle of Minimal Compliance. Linguistic Inquiry 29: 599-629. Richards, Norvin 2004. Against bans on lowering. IJnguistlc Inquiry 35: 453-463. Richards, Norvin. 2001. Movement In language: Interactions and architectures. Oxford: Oxford University Press. Richards, Norvin. 2005. Agreement and Successive Cyclic Movement in Tagalog and English. GLOW in Asia, New Delhi Rizzi. Luigi 1992. Argument/adjunct (a)symmetries. In North Eastern Linguistics Society 22, 365-81. University of Massachusetts, Amherst Rizzi. Luigi 1996. Residual verb second and the Wh-Criterion. In Parameters and functional heads 2, Bellettl. A. & Rizzi. L (eds), 63-90. NY: Oxford University Press. Rizzi. Luigi 1997. The fine structure ofthe left periphery. In Elements of Grammar: Handbook in Generative Syntax, L Haegeman (ed), 281-337. Dordrecht: Kluwer. IDzzi, Luigi 2001. On the position "'nt(errogative)" in the Left Periphery of the Clause. In Current Studies in Italian Syntax, Cinque, G. & G. Salvi (eds), 287-329. Amsterdam: Elsevier. Rizzi. Luigi (ed.). 2004. The Structure ofCP and IP (vol2). Oxford University Press, Oxford. Roberts, Ian. 2000. Clitics and Agreement. Ph.D. dissertation, MIT, Cambridge. Ross, John Robert 1969. Guess who? In Papers from the Fifth Regional meeting of the Chicago IJnguistics Society, Binnick, Robert et al. (eds), 252-286. Chicago: University ofChkago. Rudin. Catherine. 1988. 0 n multiple questions and multiple wh- fronting. Natural Language and IJnguistic 1heory6: 445-501. Sabbagh, Joseph. 2007. Ordering and Linearizing Righwartd Movement Natural Language and IJnguistic Theory. 35: 349-401. Schmidt, Ruth. 1999. Urdu: An Essential Grammar. New York: Routledge. Shaher, Rukshin, Pavel Logacev, Felix Engelmann, Narayanan Srinivasan. & Shravan Vasishth. 2008. Clefting, topical!zation and the given-new preference: Eyetracking evidence from Hindi In Proceedings of AMLaP Conference, Cambridge, UK: Cambridge. Simpson, Andrew. 2000. Wh-Movement and the Theory of Feature Checking. Philadelphia: John Benjamins. Simpson, Andrew. 2005. Sluking. Class Handout LSA Summer Institute: Syntactic Analyticity. Simpson, Andrew & Tanmoy Bhattacharya. 2003. Obligatory Overt Wh-movement in a Wh-in-situ Language. Linguistic Inquiry: 127-142. Stepanovic, Arthur. 200 l. Late Adjunction and Minimalist Phrase Structure. Syntax 4(2): 94-125. Stepanovk, Arthur & Wei-Tien Dylan Tsai 2008. Cartography and licensing of wh-adjuncts: a cross linguistk perspective. Natural Language and Linguistk Theory 26: 589-638. Subbarao, K.V. 1984. Word Order in Kashmir!. some further evidence. In Hook, Peter & Koul, Omkar (eds ), Apects of Kashmiri Linguistcs. Bari Publications, New Delhi. Subbarao, K.V. 2001. Agreement in South Asian Languages and Minimalist Inquiries: The Framework. in The Yearbook of South Asian Languages. P. Bhaskararao & K. V. Subbarao (eds). London/New Delhi: Sage Publications/Thousand Oaks. Subbarao, K. V. & S. Munshl. 2000. Agreement in Kashmiri and Minimalist Inquiries: the Framework. unpublished ms, The University of Delhi. Svenonius, Peter. 2003. On the edge. In Peripheries. eds. D. Adger, C. de Cat & G. Tsoulas Kluwer, Dordrecht pp. 259-287. Takahashi, Dalko. 1993. Movement ofwh-phrases in Japanese. Natural Language and Linguistic Theory 11: 655-678.

158 Peripheries in Kashmiri and Hindi-Urdu

Takahashi, Daiko. 1994. Sluicing In Japanese. Journal ofEast Asian Linguistics 3: 265-300. Travis, Lisa 1984. Parameters and Effects of Word Order Variation, Ph.D. dissertation, MIT, Cambridge, MA. Travis, Lisa. 2010. Inner Aspect: the articulation ofVP. Dordrecht: Springer. Tsai, Wei-Tien Dylan. 1994. 0 n economizing the theory of A' -dependencies. Ph.D. dissertation, MIT. [Published in 1999, New York: Garland]. Ura, Hiroyuki. 2000. Checking Theory and Grammatical Functions in Universal Grammar. New York: Oxford University Press. van Craenenbroek, Jeroen & Anik6 Liptak. 2006. The crosslinguistic syntax of sluicing: evidence from Hungarian relatives. Syntax 9: 248-274. Wali, Kashi. 2002. Wh-Questions in Marathi and Kashmiri. in Topics in Kashmiri Linguistics. eds. 0 mkar Koul & Kashi Wali Creative books, New Delhi, pp. 1-16. Wali, Kashi. & Omkar Koul. 1997. Kashmiri. New York: Routledge. Webel.huth, G. 1989. Syntactic Saturation Phenomena and the MiJdern Germanic Languages. Ph.D. dissertation, University of Massachusetts-Amherst, Amherst, MA. Zakharyin, Boris. 1984. Kashmiri and the Typology of South Asian Languages. in Aspects of Kashmiri Linguistics. N. Koul & Peter E. Hook (eds). 27-46. New Delhi: Bahri. Zeijlstra, Hedde. 2009 Dislocation Effects, Uninterpretable Features, Functional Heads, and Parametric Variation: Consequences of Conflicting Interface Conditions. In InterPhases: Phase-Theoretic Investigations of Linguistic Interfaces. ed. Kleanthes Grohm.ann. Oxford: Oxford Univeristy Press. Zubizarreta, M.L 1998. Prosody, Focus, and Word Order. Cambridge, Mass: MIT Press. Zwart, C. Jan Wouter. 1997. Morphosynta:x of Verb Movement: A Minimalist Approach to the Syntax of Dutch. Dordrecht: Kluwer. Zwicky, Arnold. 1990. Inflectional morphology as a (sub )component of grammar. In Contemporary Mi:lrphology. Dressler, Wolfgang Hans C. Luschiitzky, Oskar E. Pfeiffer, & John R Rennison (eds), 216-236, Berlin: Walter de Grutyer.

Index

complementizer (particle

A

A-bar movement system 8-11., 48. 53-56, 82-8:;, 98-101, 143-146 adverbs 23, 51-5l, 96-99 Agree 5-7, 34-35, 54-60, 73-74.. 76,99-104,107, 109-110,119 agreement,long-distance 113 anti-locality 60-63, 79, 103 aspect 16, 18 head 62,102 auxiliary verb 21-2l, 52, 136

ki) 40-43. 45· 49. 138-139

complementizer 40-43. 45,

Bangla 92, no Bengali 19 see also Bangla Bhatt, Rajesh 91, 103. 123-125,

see also auxiliary verb covert movement 87, 103, 107,

121-124

unselective 56-57, 99, 103, 111, 114 121, 145

Bruening, Benjamin 65, 90 Butt, Miriam and Tracy Holloway King 84> 96

c 'cartographic' approach l, 12, 24-34 case 8-9,16,60-64.79-80, 102, 117, 138 choice function 55-56, 69-70, 77. 99. 111, 121 Chomsky, Noam 2., 5-10, 29, 34> 38, 44> 46, 51, 89-90, 115-116 Chung, Sandra 64, 135 deft 129-131 137

ditic see pronominal ditic complement clause, finite see subordinate clause

G

German 10, 60-63. 66-67,

110, 123-124

73.107

D

H

Davison, Alice 4l, 92, 102, 147 Dayal. Veneeta 64-66, 91-92,

Horvath, Julia 10, 6o, 62, So, 117 Hungarian 10, 62, 79-80

117-125

68-69, 77. 119-120, 145

see also indirect dependency discourse 1, 8, 29, 96, 137, 143-146

147

Bhatt, Rakesh 18-23, 3l, 51, 52 binding, of variables 66-?l,

38-40, 45-46. 52, 90, 105, 140-141

57-58 copula 129-130

deep anaphora 133 direct dependency 48, 62,

B

Fox, Danny 7, 93, 99, 135 functional heads 24-27, 33-35,

E E (ellipsis) feature 4 137-141 echo question 95 economy 148 ellipsis 4 12, 127, 135-137 verb phrase 135-136 EPP property 9-10,36-38, 54-56, 62, 71· 78-79. 100, 102,141 ergative 15-19 expletive (DP) 10-11, 6:;, 79, 82-83, 144-145 clausal yilyeh 61, 117, 125

I Impenetrability condition 6, 29 indirect dependency 48, 64-69, 7l, no, 119-120, 145

see also direct dependency interpretable feature 5-7, 10, 54-62, 69-70, 75-78, 99-102, 111-118,147 island 117, 124 repair 134-136

strong 66 weak 73 wh- 9

J Japanese 129-130, 141 K Kidwai, Ayesha 9, 91-95,

see also wh-expletive

104-105

F

L

feature bundling 141-142 focus 22-26, 28-31, 36-37,

Lahiri, Utpal 38, 120-124 lexicon 5-6,33-40,44-46,84 Logical Form (LF) 5, 48,

44-51, 95-96, 98-99. 105, 137-139 particle -ti (Kashmiri) 22,95 force 1, 6, 24-26, 28-33. 41-4.2. 138

65,103

M Mahajan, Anoop 9-10, 122-125,127-128,134-135

160 Index

McCloskey, Jim 45, no, 135 Merchant, Jason 8, 127-129, 135-141 Merge 5-6, 10, 34-35, 38 of wh-expletives 58-64 micro-comparative 1, 7, 11 Minimalist program 5-6, 51, 68,78 multiple-wh questions 50,56 Munshi, Saddaf 37, 147

N negation 21, 122-123 negative polarity items 67, 122-123

0 operator 70, 74> 103. 114 p parametric variation 33. 39, 44-46. 107, 114> 143 partial wh-movement 53-54. 58-59. 84> 108, 112, 119-120, 147 periphery, clause-internal 4> 48,143,148-149 Pesetsky, David 56, 81, 93. 101 phase 6-8, 29-32, 42-46, 55. 75-76, 99. 113-114, 146 vP as a 89-90, 108-110, 113-114 phase-defining heads 4> 6-8, 12, 29-31. 46. 89-90, 139-141,146

T

Phonological Form (PF) 8, 43. 124> 147 pied-piping 138 pivot 129-132 post-position 138 postverbal 43. 49, 95, 96 pronominal ditic 17-18 pseudoslulcing 129-133

Tagalog 1o8-113 Toosarvandani, Maziar 4> 137-139 topic 22-23. 25-26, 28, 30, 32-33. 81, 129 tri-dausal wh-questions 73-83. ll7-120

R

u

Rackows.ki, Andrea and

uninterpretable feature 5-7, 10, 54-62, 69-70, 75-78,89,99-102, 111-n8, 147

Norvin Richards 108-110 Reinhart, Tanya 55-56, 69-70, 99,10 Rizzi, Luigi 13. 24-25, 33 Romani 10, 68, 117

s scope-marker 10, 64-69, 144 see also wh-expletive scrambling 9, 91-92, 104-106, 116-117, 124> 136-137 rightward 193, 123-124 Simpson, Andrew 54> 59-64 specifiers, multiple 27-29, 32, 37. 45.55 subject-auxiliary inversion 77, 113 subordinate clause 23-24. 32. 40-41. 49-50, 57-59, 65-66, 71-72. 100-101, 122-123 surface anaphora 133 Subbarao, K. V. 19-20

v verb-final18, 37, 93. 148 verb-second 1-4, 18-24. 27, 36-37. 49-52, 92-93 verb-third 1, 27, 89

w wh-adverbs 107 wh-expletive 2, 3. 10-11, 58-82, 101-103, 108-110, 144-146 multiple 72-83 wh-in-situ 12, 49. 65, 69-70, 90-91, 94-95, 97, 125, 127,137