dynamic antisymmetry and the syntax of noun incorporation

With bare nominal objects as in the progressive beim construction, only OV word order is found.. indirect object, goal, source, comitative or subject example 5, Mohawk, and ple 6, Ononda

DYNAMIC ANTISYMMETRY AND THE SYNTAX OF NOUN

INCORPORATION

VOLUME 84

Managing Editors

Marcel den Dikken, City University of New York

Liliane Haegeman, University of Ghent, Belgium

Joan Maling, Brandeis University

Editorial Board

Guglielmo Cinque, University of Venice

Carol Georgopoulos, University of Utah

Jane Grimshaw, Rutgers University

Michael Kenstowicz, Massachusetts Institute of Technology

Hilda Koopman, University of California, Los Angeles

Howard Lasnik, University of Maryland

Alec Marantz, Massachusetts Institute of Technology

John J McCarthy, University of Massachusetts, Amherst

Ian Roberts, University of Cambridge

For further volumes:

http://www.springer.com/series/6559

DYNAMIC ANTISYMMETRY AND THE SYNTAX OF NOUN

Springer Dordrecht Heidelberg London New York

Library of Congress Control Number: 2011929131

© Springer Science+Business Media B.V 2011

No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose

of being entered and executed on a computer system, for exclusive use by the purchaser of the work Printed on acid-free paper

Springer is part of Springer Science+Business Media ( www.springer.com )

This monograph is a revised version of my 2006 dissertation at the University ofToronto I have included more discussion on various issues in noun incorporation inNorthern Iroquoian languages based on changes in my understanding of the materialsince 2006 I have also added additional material gathered during postdocs at theUniversity of British Columbia and the University of Ottawa Much of the data

on Northern Iroquoian languages comes from my own fieldwork and from severalpublished sources The system of interlinear glosses varies significantly from onesource to the next For consistency, I have employed a uniform set of abbreviationsfor the interlinear glosses throughout

I wish to thank my consultants Daisy Elijah (Oneida), Nora Carrier and GloriaWilliams (Onondaga), Elizabeth Herrling (Halkomelem), and Beatrice Bullshields(Blackfoot) Without their assistance, I would not have been able to undertake thisresearch In addition, I wish to thank several colleagues of mine for their assis-tance with the relevant literature and/or data including Diane Massam (Niuean),Arsalan Khanemuyiour, Jila Ghomeshi, and Nick Pendar (Persian), Bettina Sprengand Martina Wiltschko (German), Will Seto, Lisa Seto and Max Mak (Cantonese),Yosuke Sato (Japanese), and Martina Wiltschko (Halkomelem)

Parts of this monograph have been presented at numerous conferences includingNELS35 and 36, the Linearization workshop at DGfS in 2010, the annual meetings

of the Canadian Linguistic Association and the Linguistic Society of America, andthe Workshop on Structure and Constituency in Languages of the Americas I wish

to thank the audiences there for insightful questions and suggestions that led me torevise and, hopefully, improve upon the details

Numerous people have listened to me talk about my ideas presented here orhave collaborated with me on related projects, and have given useful and sub-stantial feedback, or have simply asked vital, thought-provoking questions thatallowed me to improve my analysis This list includes (in no particular order)Elizabeth Cowper, Diane Massam, Alana Johns, Carrie Dyck, Martina Wiltschko,Gabriela Alboiu, Éric Mathieu, Joel Dunham, Bettina Spreng, Heather Bliss,Arsalan Khanemuyipour, Jila Ghomeshi, Kenji Oda, Yosuke Sato, plus others I havesurely forgotten

Finally, I wish to acknowledge the financial support of SSHRC doctoral and doctoral fellowships and a Killam post-doctoral research fellowship

post-v

DR bivalent direct (indicates that the object is lower on the thematic

hierarchy than the subject)

I.TH intransitive theme

JOIN joiner vowel, an epenthetic vowel in Northern Iroquoian languagesLCA Linear Correspondence Axiom

Abbreviations ixPST past tense

PUNC punctual aspect

1 Introduction 1

1.1 Outlook and Goals 2

1.2 Noun Incorporation 5

1.2.1 Properties of NI 5

1.2.2 On the Syntactic Nature of NI 8

1.3 Conclusion and Outline of Book 18

References 19

2 Theoretical Background 23

2.1 A Bit of History of Phrase Structure 23

2.2 Bare Phrase Structure 26

2.3 Antisymmetry 30

2.4 Head Movement 34

2.5 Alternative Accounts of Linearization 40

2.5.1 Demerge and Concatenate: SOV as Underlying Order 41

2.5.2 Parameterized Order 42

2.5.3 Departures from the LCA 47

2.6 Conclusion 48

References 48

3 Unifying Antisymmetry and Bare Phrase Structure 53

3.1 The Problem of Mutual C-Command 53

3.2 Previous Accounts 54

3.2.1 Guimarães (2000) and Self-Merge 57

3.2.2 Nunes and Uriagereka (2000) 60

3.2.3 Richards (2001) 61

3.2.4 Conclusion 61

3.3 The LCA and BPS 61

3.4 Proposal 71

3.5 Alternatives to Complement-to-Spec Roll-Up 77

3.5.1 Spec-To-Spec Movement and Romance Clitics 78

3.5.2 Avoidance of the Initial Merger Problem 81

xi

3.6 Linearization and Late Insertion 85

3.7 Summary 88

References 88

4 Noun Incorporation in Northern Iroquoian 93

4.1 Northern Iroquoian Languages 93

4.1.1 Clause Structure of Northern Iroquoian 94

4.1.2 Nominal Structure in Northern Iroquoian 97

4.2 Patterns of NI in Northern Iroquoian 100

4.2.1 Productivity of NI 101

4.2.2 Nominal Roots 106

4.3 Analysis 107

4.3.1 Previous Analyses 107

4.3.2 NI as Phrasal Movement 112

4.4 Properties of Iroquoian NI 118

4.4.1 NI in Ditransitives 118

4.4.2 NI and Overt DPs 121

4.5 Conclusion 123

References 123

5 Noun Incorporation and Its Kind in Other Languages 127

5.1 The Structure of Nominals and V/IN Order 127

5.2 NI in Sierra Popoluca 130

5.3 English Gerunds 132

5.3.1 Description of NI in English Gerunds 133

5.3.2 Analysis 135

5.4 German Progressives 138

5.5 Persian ‘Long Infinitive’ Constructions 143

5.6 Tamil Noun Incorporation and Coordination 147

5.7 Adverb Incorporation in Blackfoot 150

5.8 Conclusion 154

References 155

6 V+N Order 159

6.1 Polynesian Pseudo Noun Incorporation 160

6.1.1 The Structure of Niuean Nominals 161

6.1.2 Pseudo Noun Incorporation 163

6.1.3 Postlude on Tukang Besi Object Incorporation 165

6.2 Romance Compounds 167

6.2.1 Description of V+N Compounds 168

6.2.2 Analysis of Romance Compounds 170

6.3 Mapudungun 172

6.4 Upriver Halkomelem 176

6.5 Conclusion 181

References 183

Contents xiii

7 Conclusion 185

7.1 Summary 185

7.2 Limitations 186

7.3 Conclusions and Implications 190

References 194

Index 197

Chapter 1

Introduction

This monograph is a study of Noun Incorporation (NI) within the framework ofDynamic Antisymmetry It deals primarily with NI in Northern Iroquoian (repre-sented principally by Oneida and Onondaga, with some discussion on Mohawk);however, I also consider a number of other cases of NI and related phenomena in

a variety of languages I propose that NI is the direct result of the need to satisfythe Linear Correspondence Axiom (LCA) NI arises when the verb selects a barenominal root as a complement, thereby creating a point of symmetric c-command.This point of symmetry is resolved by the nominal root raising to the specifier of VP

My account is couched within two recent, well-received proposals on phrase ture, which I meld into a single cohesive framework Specifically, I examine BarePhrase Structure (BPS) (Chomsky1994) and Antisymmetry (Kayne1994) and pro-pose that the insights of both proposals can be maintained if we adopt a DynamicAntisymmetric view as proposed by Moro (2000;2004) Moro proposes that move-ment is driven by the need to satisfy the LCA, and furthermore that this movementtakes place overtly Thus, if two terms cannot be linearized because of their non-compliance with the LCA, something must move so that the LCA is satisfied Inparticular, I examine the case of two symmetrically c-commanding heads; a situa-tion that arises upon the initial merger of two heads (the Initial Merger Problem).This is precisely the situation in the case of NI as just described above Thus, NI

struc-is purely a result of the phrase structural properties of the VP Thstruc-is eliminatesthe need for a macroparametric mechanism, such as the Morphological VisibilityCondition (Baker1996) I will also show that not only does an LCA violation (inparticular, symmetric c-command) trigger movement, it can also trigger a cascade ofmovements often referred to as Complement-to-spec roll-up in the literature.1Thisapproach predicts that NI should typically appear with N+V order Indeed, this order

is significantly more common than V+N order (Caballero et al.2008); however, it is

by no means universal, as there are numerous counter-examples I discuss how NIwith V+N order arises in later chapters

1 These will come up later in the discussion where appropriate Representative examples include

snowballing (Aboh2004a) and intraposition movement (Rackowski and Travis2000 ).

1

M Barrie, Dynamic Antisymmetry and the Syntax of Noun Incorporation, Studies

in Natural Language and Linguistic Theory 84, DOI 10.1007/978-94-007-1570-7_1,

C

 Springer Science+Business Media B.V 2011

The remainder of this chapter is structured as follows Section1.1presents themajor goals of this monograph and gives a brief overview of how these goals will beaccomplished Section1.2introduces the phenomenon of NI and presents its prop-erties that are pertinent to the discussion Section1.3is an outline of the remainder

of the monograph

1.1 Outlook and Goals

I identify the following goals of this monograph:

• to present a thorough analysis of NI in Northern Iroquoian,

• to recast (Dynamic) Antisymmetry in a BPS framework,

• to argue for a Dynamic Antisymmetric analysis of NI, and

• to illustrate the Dynamic Antisymmetric analysis of NI with data from a variety

of other languages

In regards to the first goal I describe in detail the various properties of NI that areparticular to Northern Iroquoian and show how the proposed analysis for NI coupledwith an innovative structure for the doubling construction accounts for the specificproperties of NI in this language group A general description of NI is found in thefollowing section; however, I also discuss several additional properties of NI found

in Northern Iroquoian Specifically, I discuss the kinds of nominals that can undergo

NI and their morphological properties I also discuss how NI in Northern Iroquoian

interacts with wh-movement In short, an in depth description of NI in Northern

Iroquoian is presented along with the general properties of NI cross-linguistically.The facts presented here are based on a vast descriptive (Lounsbury1949,1953;Froman et al.2002; Woodbury1975b,1975a,2003; Michelson and Doxtator2002;Bonvillain1972) and theoretical (Michelson1991; Baker1988,1996; Postal1979)literature on this topic, as well as various field trips conducted from 2003 to 2005

on Oneida and 2005 to the present on Onondaga The discussion on the generalproperties of NI is based on various cross-linguistic surveys on the topic (Mithun1984; Caballero et al.2008; Gerdts1998; Massam2009)

Accommodating Antisymmetry and BPS in a unified framework is not a forward task I review and build on previous work on this topic (Guimarães2000;Oishi2003; Richards2001a; Uriagereka1999; Kayne2009a) and adapt a version ofthe LCA that is compatible with the machinery of BPS This comprises the bulk ofChapter 3 Briefly, BPS requires the initial merger to be between two heads, givingrise to a violation of the LCA as mentioned above To resolve this violation, theselected head raises to the specifier of the selector The next head that this com-plex merges with, however, will be in a symmetric c-command relation with thehead in specifier from the previous step, triggering another round of raising to sat-isfy the LCA This process sets off a cascade of roll-up type movement referred to

straight-1.1 Outlook and Goals 3above Of course, the derivation does not continue this way indefinitely I resolvethis situation by making the LCA sensitive only to heads with phonological content,

as suggested by Moro (2000) Thus, a phonologically empty head halts this type ofmovement I suggest that NI in particular highlights this fact about BPS and initialmerge since it typically involves merger of two lexical heads, hence two heads withphonological content Non-NI constructions involve initial merge between a lexi-cal head and a functional head or between two functional heads – both situations

of which involve the possibility of a phonologically null head, thus obscuring theInitial Merger Problem

I now discuss briefly how the current proposal captures the fact that a full DPcomplement to a verb and a bare noun2complement to a verb often exhibit differentlinearization properties The proposed analysis accounts for this asymmetry Whenthe verbal root, V, selects a full DP as a complement, V asymmetrically c-commandsthe material inside the DP in compliance with the LCA When the V selects a barenominal root as a complement, the two elements c-command each other, in viola-tion of the LCA Following Moro’s Dynamic Antisymmetry (2000), the nominalroot raises to SpecVP to satisfy the LCA By contrast, when V selects a full DP as

a complement, the verb asymmetrically c-commands the material inside the DP Inthe extended nominal projection alone a number of functional projections have beenidentified with syntactic and semantic effects (Ritter1992,1993; Szabolcsi1983;Abney1987; Cheng and Sybesma 1999; Ghomeshi 2003; Megerdoomian2008);however, whether the head is overtly realized with a particular morpheme is sub-ject to a great deal of cross-linguistic variation, and many heads are phonologicallyempty This gives rise to a great deal of variation in the surface realization of themorpheme order of DPs In any event, when the verb selects a full DP comple-ment, in the vast majority of cases, the verb will asymmetrically c-command all thematerial inside the DP

As already stated, the empirical foundation for this study consists of NI andrelated phenomena I offer an extensive discussion and analysis of NI in NorthernIroquoian inChapter 4, and I discuss similar facts in a variety of other languages

inChapters 5and6 In the following examples, the verbs are shown in boldface

and the complement nouns are italicized to show their relative positions The (a)

examples contain full DP complements and the (b) examples contain nominal plements that have undergone NI or a closely related phenomenon, which I laterargue to be essentially the same as NI Again, notice that the NI forms appearwith the order N+V (V+N is discussed in Chapter 6), while the order of V and

com-DP varies

2Throughout this monograph, I use the term bare noun to mean functionally bare (i.e., a bare root

or a bare nP), rather than morphologically bare in the sense of Giorgi and Longobardi (1991 ) This distinction is important because a morphologically bare noun in this other sense could contain a large number of functional projections with phonologically empty heads.

1 NI in Onondaga (Northern Iroquoian)

a wahahní:no˛TneTganakdaT

waT-ha-hnino˛-T neT ka-nakt-aT

FACT-3.SG.M.AG-buy-make-PUNC NE AGR-bed-NFS

‘I’m making a house for him.’

b wahanakdahní:no˛T

FACT- 3.SG.M.AG bed JOIN buy PUNC

‘He bought a bed.’

2 NI in English Gerunds

a Alice enjoys collecting stamps.

b Alice enjoys stamp-collecting.

3 German Progressives3,4

I eat.1.SG the apple.PL

‘I’m eating the apples.’

b Ich bin beim Äpfel- essen.

I am at.the apple.PL- eat.INF

‘I’m eating apples.’ / ‘I’m busy apple-eating.’

4 Persian Long Infinitives5

a sima æz xundæn-e in ketab xoš-eš mi-yad

Sima from reading-EZ this book good-3SG CONT-come.3SG

‘Sima likes reading this book.’

Sima from book reading good-3SG CONT-come.3SG

‘Sima likes reading books.’

In English, example 2, full DP complements appear to the right of the verb, andwhat is demonstrably a bare noun appears to the left of the gerundive form of theverb Gerunds are taken up in more detail inChapter 5 These data highlight theasymmetry between full DP objects and bare noun objects Specifically, whether

3 Of course OV word order is possible with full DP objects, as German is an SOV language The point here is that full DP objects can appear in either VO or OV word order, depending on whether the verb appears in 2nd position or sentence-finally With bare nominal objects as in the progressive

beim construction, only OV word order is found.

4 Unless otherwise stated, all German data are from Bettina Spreng and Martina Wiltschko (personal communication).

5 Unless otherwise stated, all Persian data are provided by Jila Ghomeshi, Arsalan Kahnemuyipour, and Nick Pendar (personal communication).

1.2 Noun Incorporation 5the object appears before or after the verb depends on whether the object is a full

DP or bare noun Of course, other factors may affect VO versus OV word order(see footnote 3), but the generalization still holds The theory of phrase structure Iadvocate here accounts for this asymmetry in a straightforward way

There are, of course, instances of NI with the order V+N I examine eral particular cases in detail inChapter 6: pseudo noun incorporation (PNI) inNiuean (Massam2001), V+N compounds in Romance (such as Spanish toca-discos

sev-‘record-player’ and French tranche-oeufs ‘egg slicer’), lexical suffixation in Upriver

Halkomelem, and NI in Mapudungun I argue that V+N order results from the ence of additional functional material in the IN that permits the LCA to linearize theverbal root with the elements inside the IN Thus, no movement needs to take placeand the order V+N surfaces

pres-The major scientific contribution of this proposal is twofold First, it accounts forthe robust cross-linguistic generalization that when a verb takes a bare noun as acomplement, we observe the order N+V more frequently than V+N (Caballero et al.2008), and when the verb takes a full DP complement, we get roughly a 50–50 splitbetween the orders V+DP and DP+V (Ruhlen 1975; Dryer 2008) InChapter 2,

I address the lack of a universal N+V order in NI constructions Second, NI and

a wide range of related phenomena fall out naturally from the theory of tion proposed here I do not need to posit any new syntactic mechanisms, so NIcomes for free, in effect As a consequence, we do not need a variety of differ-ent mechanisms to account for the phenomena in 2–4 Thus, rather than having toposit different mechanisms to account for NI and related phenomena, I propose thatexisting mechanisms accomplish this task

lineariza-1.2 Noun Incorporation

NI has been the subject of study for well over 100 years now (Kleinschmidt1852;Cuoq 1866; Kroeber1909; Sapir 1911) It in fact continues to be the subject ofmuch heated debate and controversy, especially concerning its status as a syntac-tic, morphological or lexical phenomenon (Baker1996; Baker et al.2005; Mithun

1984, 1986b; Sadock1980,1986; Rosen1989) The works of several authors cipally those in the references just mentioned) have converged on a common set ofprototypical properties of NI, which I describe below I also discuss the typologi-cal work of Caballero et al (2008), which describes in detail those properties of NIconcerning the order between the IN and the verbal root

(prin-1.2.1 Properties of NI

Northern Iroquoian languages illustrate many prototypical properties of NI (Rice1991; Baker1988,1996; Mithun1984; Woodbury1975b), which are described indepth inChapter 2and are analyzed inChapter 3 These include: (i) the ability toincorporate a direct object, instrument, path or subject of an unaccusative, but not an

indirect object, goal, source, comitative or subject (example 5, Mohawk, and ple 6, Onondaga),6(ii) the ability for the incorporated noun to appear with a full DPdouble or modifier (examples 7–8, Onondaga), (iii) the surface order N+V (observedthroughout, though see the forthcoming discussion on order in the following para-graphs), (iv) the optionality of NI (example 9, Oneida), (v) frozen scope (observed

exam-in 10), and (vi) idiomatic meanexam-ings I will show that properties (i) and (iii) boil down

to a situation of symmetric c-command triggering NI via movement Properties (ii)and (iv) are related to the structure for NI I propose inChapter 3 These are illus-trated in the following examples Note that the properties illustrated below applygenerally to all Northern Iroquoian languages (Baker1988,1996; Bonvillain1972;Woodbury1975b,1975a; Mithun1984; Abbott2000; Froman et al.2002)

5 NI in Mohawk Ditransitives (Baker1996: 207)

CIS-FACT- 1.SG.AG:3.SG.F.PAT-ball-NZLR- give-PUNC NE baby

‘I gave the ball to the baby.’

CIS-FACT- 1.SG.AG- baby- give-PUNC NE ball

‘I gave the baby to the ball.’ (NOT ‘I gave the ball to the baby.’)

3.PL.M.PAT- SRFL- path- run- PURP

‘They are walking on a path.’

b waThageTnhyayé˛hdaT (Woodbury,2003: 928)

FACT- 3.SG.M.AG:1.SG.PAT stick- JOIN- hit- PUNC

‘He hit me with a stick.’

c ohahaná:we˛h (Nora Carrier, Gloria Williams, speakers)

3.SG.NT.PAT- road- JOIN- wet- STAT

‘The road is wet.’

FACT- 1.SG.AG- animal- wash- PUNC NE pig

‘I washed the pig.’

6 See Öztürk ( 2009 ) and Cagri ( 2009 ), however, for discussions on so-called subject incorporation

in Turkish as well as Polinsky ( 1990 ) for subject incorporation in Chukchi.

1.2 Noun Incorporation 7

8 a waTkhní:no˛TneTne˛geTganakdaT (ibid.)

FACT- 1.SG.AG buy- PUNC NE DEM bed

‘I bought this bed.’

FACT- 1.SG.AG- bed- JOIN- buy- PUNC NE DEM

‘I bought this bed.’

a waPkhni:nú: ká:sleht

waP-k-hninu-´: ka-Psleht-Ø

FACT-1.SG.AG-buy-PUNC 3.SG.NT-car-NFS

‘I bought a car.’

b waPkePslehtahni:nú:

FACT- 1.SG.AG- JOIN- car- JOIN- buy- PUNC

‘I bought a car.’

10 Yah te-wake-nakt-a-hninu

No NEG-1.SG.PAT-bed-EPEN-buy.STAT

‘I didn’t buy the bed.’ Or ‘I didn’t buy a(ny) bed.’

=‘There is a bed that I didn’t buy.’ (Baker,2003: 116)

11 e˛dyo˛dwe˛nage˛Tsä:hgwaT (Onondaga, Gloria Williams ande˛-d-yo˛-ad-we˛n-a-ge˛TsR-hgw-aT Nora Carrier, speakers)

FUT-DUC-3.SG.F.AG-SRFL-word-EPEN-rest/pillow-INSTR-PUNC

‘They put it under a pillow’ / ‘They tabled the matter.’

Example 5 shows a ditransitive, in which the direct object, but not the indirect object,can undergo NI Example 6 shows NI of a path, an instrument and an unaccusative

subject Example 7 illustrates doubling of the IN (animal) with a hyponymous object (pig) Example 8 shows a stranded modifier (a demonstrative).7 Example

9 shows that NI is optional in Oneida Example 10 shows that the IN in Mohawkexhibits frozen scope with respect to a scope bearing element outside the verbalcomplex Finally, example 11 shows that NI often gives rise to idiomatic readings.Mithun (1984) also observes that NI constructions are often used in institutional-ized activities and are sometimes restricted to particular kinds of uses This is akin

7The use of the label modifier stranding to describe example 8 has theoretical implications that I

discuss in greater detail in Chapter 4

to the Canonical Use Constraint (CUC) discussed by Kiparsky (1997) All of theproperties just discussed hold in Northern Iroquoian generally.

Perhaps unsurprisingly so, it will become clear throughout the discussion, cially in the later chapters, that not all languages with NI exhibit the exact same set

espe-of properties for this phenomenon The landmark discussion on the types espe-of NI, itsdiscourse properties and effects on argument structure is Mithun (1984), a typologywhich maintains widespread currency in the literature today

I turn now to a discussion of the typological properties of NI as it relates to linearorder Based on Kayne’s (1994) principle of Antisymmetry, Baker (1996) proposedthat NI is universally N+V – a claim he backed down on (Baker et al 2005) inlight of the V+N order found in Mapudungun Caballero et al (2008) conductedthe first wide-scale cross-linguistic study on linear order in NI, where they showthat N+V order is significantly more prevalent in NI constructions than V+N order.Note that OV and VO order (with full DP objects) is pretty much evenly distributed,(Ruhlen 1975; Dryer 2008) Caballero et al report the following results to theirstudy (adapted from source document)

1.2.2 On the Syntactic Nature of NI

I conclude this section with some arguments in favour of maintaining a syntacticanalysis of NI, in spite of various counter-arguments I discuss the range of elementsthat can be incorporated, idiomatic readings associated with NI constructions, andfrozen scope

One argument against a syntactic treatment of NI lies in the fact that ments other than the direct object can be incorporated (Rosen1989; Spencer1995;

ele-8 Not all languages with productive NI were classified as OV or VO languages, hence the discrepancy in totals.

9 Note that I have nothing to say about the correlation between VO/OV order and V+N, N+V order despite the fact that this is significant.

1.2 Noun Incorporation 9

Di Sciullo and Williams 1987) In particular, paths, locations, instruments andadverbs can undergo NI (or adverb incorporation in the case of the latter), in severallanguages, as discussed above Excluded from NI constructions, however are goals,benefactives, sources and comitatives This seemingly disparate set of elements(direct objects, paths, instruments, locations, and adverbs) is actually a natural classunder a Larsonian (1988;2004) analysis of VP-internal elements Under Larson’sanalysis, these elements are introduced in the specifier or complement of a VPshell

posi-The inability for indirect objects to undergo NI has been noted in the literaturebefore (Mithun1984; Baker1996: 297ff.) The exact structure of ditransitives plays

an important role here Pylkkänen (2008), for instance, has argued that goal ments are introduced by a low applicative head (see Larson 2010for argumentsagainst this approach, though).10She also introduces other elements such as instru-ments by applicative heads; thus, this approach fails to capture the natural classdescribed above Larson’s (1988; 1990) original discussion of the double objectconstruction placed the indirect object and goal directly inside a VP shell, whichalso fails to capture the natural class described above Baker (1988; 1996) arguesthat the indirect object/goal/source is introduced in the VP shell by a phonologicallynull preposition Consider the following example (Baker1996: 298)

argu-10 Larson takes issue with the fact that Pylkkänen has divorced the goal argument from the event

semantics Thus, in a construction such as John baked Mary a cake, Mary is related to the cake

by a to.the.possession.of predicate, but is not related to the event Fatally, however, Larson argues

that this approach fails to capture the fact that the subject, John, must be responsible for Mary’s receipt of the cake Under Pylkkänen’s approach, the sentence above is compatible with a scenario

in which John baked a cake and someone else brought it to Mary Since I do not ultimately adopt Pylkkänen’s approach, I won’t pursue this line of reasoning further here.

Baker argues that the null preposition provides the semantic distinctions necessary

to encode direction in transfer of possession verbs (buy from vs give to) Note that

indirect objects in Northern Iroquoian languages can be either sources or goals, sothe distinction must be encoded somewhere (see also Michelson1991) Considerthe following Onondaga examples (Woodbury2003)

15 a waTho˛wahyahní:no˛T

waT-ho˛wa-ahy-a-hnino˛-T

FACT-3.AG:3.SG.M.PAT-fruit-JOIN-buy-PUNC

‘Someone bought fruit from him.’

b waTkhé:yo˛T

waT-khe-o˛-T

FACT-1.SG.AG:3.SG.F.PAT-give-PUNC

‘I gave it to her.’

It is not clear how Baker’s approach extends to paths, locations and instruments,however Under the approach above, a null P is necessary to give rise to the correctsemantics for the indirect object Assumedly, a null P would be necessary to give rise

to the correct reading (instrumental, locative, etc.) to the relevant nominal phrase Itwould be unclear, then, what blocks incorporation of the indirect object, but allows

it for paths, locations and instruments

Beck and Johnson (2004) also argue that indirect objects are introduced by a

functional projection based on its behaviour with the adverb again Crucially, ble object constructions allow both a repetitive and restitutive reading with again.

dou-Consider the following example, with the repetitive and restitutive readings spelledout, respectively (Beck and Johnson2004: 113)

16 Thilo gave Satoshi the map again

17 a Thilo gave Satoshi the map, and that had happened before

b Thilo gave Satoshi the map, and Satoshi had the map before

1.2 Noun Incorporation 11Beck and Johnson argue that this ambiguity arises as a result of an additional pred-icative head that relates the direct object to the indirect object (while still relatingboth to the main event, thereby avoiding the problems in Pylkkänen’s proposal, seenote 10) Thus, Beck and Johnson propose the following structures for double objectconstructions and DP + PP constructions The HaveP and PP serve as adjunction

sites for again, accounting for the restitutive readings.

18 a [vPThilo [v’ v [VPgive [BECOME[HavePSatoshi [Have the map]]]]]]

b [vPThilo [v’ v [VPthe map1give [BECOME[PPPRO1to Satoshi]]]]]Crucially, the ambiguity noted by Beck and Johnson is absent with locations, instru-ments and adverbs In the following examples, no restitutive reading is possible

(although the adverb again may take scope over different parts of the VP).11

19 a John is cooking in the kitchen again

b Joyce is trimming flowers with a knife again

c Suzanne is typing very quickly again

For the current purposes I will adopt the proposal of Beck and Johnson and assumethe structures in 18 for ditransitive constructions, though I leave out the BECOMEoperator as I do not discuss the formal semantic properties here.12

The proposal that adverbs, direct objects, instruments and locations form a ural class also receives interesting empirical support from an unexpected source.Non-canonical objects in Mandarin Chinese also comprise the same set of elements,namely paths, locations and instruments, but exclude goals, sources, benefactivesand comitatives (Li2010) The aforementioned elements can appear in preverbalPPs (the canonical position for adjunct PPs), while paths, locations and instrumentscan optionally appear in the canonical object position in Mandarin (immediately

nat-to the right of the verb) without a preposition Consider the following Mandarin

11 Note that a restitutive reading is available with 19b; however, this is independent of the instrumental phrase The same reading is available with the following sentence.

i Joyce is trimming flowers again (and they used to be trimmed before).

This reading is somewhat odd because cut flowers don’t typically re-grow their stems (thus requiring re-trimming) Nevertheless, it is available Under Beck and Johnson’s approach, we

would assume a stative BE predicate, which the adverb again could take scope over.

ii Joyce CAUSE flowers BE trim.

Crucially, the instrumental phrase is not introduced by an additional predicative head since or else we would predict yet another restitutive reading.

12 Many of the arguments for the various proposals for the structure of ditransitives rely on a ber of semantic distinctions Further details of these semantic properties in Northern Iroquoian languages will have to wait for future research.

num-Chinese data Examples 20 and 21 illustrate non-canonical objects with a locativeand instrumental, respectively Examples 22 – 24 show that the non-canonical objectconstruction is unavailable with benefactives, comitatives and sources.

‘He eats at the restaurant.’

he eat restaurant

‘He eats at the restaurant.’ / ‘He eats restaurant food.’

‘He likes to write with this brush pen.’

‘He likes to write with this brush pen.’

‘I make clothes for him.’

(‘I make things for him.’)

‘I make clothes with him.’

(‘I make things with him.’)

‘I borrow books from libraries.’

(‘I borrow from libraries.’)

It would be a strange coincidence indeed if NI in the languages under discussionand non-canonical objects in Mandarin Chinese picked out the same set of conceptswhich are eligible to participate in these constructions Thus, the non-canonical

1.2 Noun Incorporation 13objects in Mandarin Chinese coupled with the Larsonian shell hypothesis actu-ally argue in favour of a syntactic approach to NI of instruments, paths, locationsand adverbs.

NI constructions typically show idiomatic and unpredictable meanings This wasoften cited as an argument in favour of a lexical analysis for NI Consider thefollowing example (Woodbury2003: 304)

25 waTshagotshógwaks

FACT- 3.SG.M.AG:3.SG.F.PAT- SRFL- lip- eat- PUNC

‘He kissed her.’ (lit: ‘He ate her lips.’)

In 25, the idiomatic reading is available only in the NI construction.13 Marantz(1997) has argued against the notion of idiomatic meanings as an indicator of alexical component to the grammar based on phrasal idioms in a wide variety of

languages such as chew the fat, pull the wool over X’s eyes, make waves, and kick

the bucket These idioms have meanings which are not deducible from their syntax;

however, they participate in syntactic processes (such as passivization: The wool

was pulled over the customer’s eyes or focus fronting: Even the tiniest of waves I wouldn’t make with that administrator) and cannot be analyzed as syntactic atoms.

One may object that these idioms are of a different nature or character than theidiomatic readings of IN constructions Indeed the idioms in English (and manyother languages) include determiners on the nouns rather than bare nominals Giventhat I related the similarity of INs with bare nominals below, I consider yet additionaldata Consider the following Cantonese data (Barrie2008)

26 a caat3 haai4

polish shoe

‘flatter’ OR ‘polish shoes’

‘polish this pair of shoes’

#‘engage in this instance of flattery’

drink tea

‘have dim sum’ OR ‘drink tea’

‘drink this tea’

#‘have this (bit of) dim sum’

13 Note that this idiom is apparently found only in the Syracuse community in New York My consultants in Six Nations did not recognize this idiom.

With the bare nominal objects, both the idiomatic and compositional readings areavailable When the object contains higher functional material, only the composi-tional reading is available, exactly like in the instances of NI shown so far Note thatthe bare noun cannot be considered morphologically incorporated because it can beseparated from the verb by aspectual morphology and can even be focused in somecases, while retaining the idiomatic reading.

drink PRFV tea

‘have had dim sum’ OR ‘have drunk tea’

‘You’ve never even had dim sum before.’

Another similar illustration comes from light verb constructions in Japanese Again,consider the following data (Sato2010)

‘to cook’ NOT ‘to handle well’

Again, the idiomatic readings are available on the bare noun constructions only.Both the Cantonese and the Japanese data provide further illustrations that the avail-ability of idiomatic readings is correlated with the bareness of the noun rather thanwith any concept of word-hood

While idiomatic readings are found nearly always exclusively with NI tions, this is by no means universal Consider the following Chipewyan data (Cookand Wilhelm1998: 59)

construc-31 a na-j´ëth-the-Ø-Ø-da

ITER-hook-M/A-3.SG-VCL-sit

‘S/he is fishing again.’ (lit: sitting with a hook)

b j´ëth gh ˛a the-Ø-Ø-da

hook with M/A-3.SG-VCL-sit

‘S/he is fishing again.’ (lit: sitting with a hook)

1.2 Noun Incorporation 15Here, the idiomatic reading is found with both the incorporated and non-incorporated forms, again underscoring the lack of involvement of any notion ofword-hood in establishing idiomatic readings.

Related to the above is the fact that NI also subject to the CUC (Kiparsky1997), whereby pragmatically or culturally determined application of the incor-porated object is understood This information typically cannot be gleaned from

the syntax For instance in the phrase butter the toast, we understand a

particu-lar way of spreading butter on the toast (with a knife, only on one flat surface,not on the edges or on only half of the surface, for example) The follow-ing Oneida example (Michelson and Doxtator 2002) illustrates this phenomenonfor NI

32 wahathwistásheteP

FACT- 3.SG.M.AG- SRFL- money- count- PUNC

‘He counted money.’

The sentence in 26 cannot refer to counting kinds of money or numbers of bill orcoins It can only refer to counting the sum value of the money Harley (2008),however, showed that this is not a property of the lexicon, but a property of barenouns, based on the following English data

33 in bed in the bed

∗on bed on the bed

at church at the church

Unless we wish to posit that forms such as in bed and at church (and, in North American dialects of English, in the hospital) are syntactic atoms spit out from

a generative lexicon, the CUC cannot be used as an argument against a syntacticapproach to NI In fact, it shows that NI impinges on our understanding of the syntax

of bare or reduced nominal expressions

Finally, I address the issue of frozen scope NI constructions typically showfrozen scope as shown in example 10 above This has also been cited as a prop-erty of words, the idea being that words are syntactic atoms and thus morphemesinside words cannot take scope outside the word I argue that this is more a property

of bare nouns, however Consider the following Blackfoot example (Glougie2000)

34 a ihkan-ano-yi-a om-yi piita

PST.all-see.TR-3’-PL DEM-OBVeagle

‘They all saw the eagle.’ (∃ > ∀)

b ihkan-iyapi-ya piita

PST.all-see.INTR-PL eagle

‘They all saw an eagle.’ (∀ > ∃)

c piita ihkan-iyapi-ya

eagle PST.all-see.INTR-PL

‘They all saw an eagle.’ (∃ > ∀)

The difference to note between these two examples is that the object in 34a is a full

DP with a demonstrative, while the following two examples contain a bare nominal.Note also the intransitive marking on the b and c examples highlighting the fact thatthey contain bare nominal objects The bare noun in 34b, c has frozen scope despitebeing syntactically free in the sentence Again, it has been shown that a putativeproperty of NI, namely frozen scope, intended to argue for its status as a lexicalphenomenon, in fact is merely a property of bare nouns In fact, van Geenhoven(1998) has shown that frozen scope in West Greenlandic and bare plurals in Englishpattern identically in terms of scopal relations, thus highlighting the similarity inproperties between bare nouns and INs van Geenhoven gives the following examplefrom West Greenlandic (p 3, ex 3, Bittner1994: 118)

35 Junna Kaali-mit allagar-si-nngi-l-a-q

Junna.ABS Kaali-ABL letter-get-NEG-IND-TR-3.SG

‘It is not the case that Juuna got a letter/letters from Kaali.’

#‘There is/are a letter/letters from Kaali that Juuna didn’t get.’

As shown, the IN cannot take scope above negation Bittner also gives the ing English data showing that bare plurals in English also cannot take scope abovesentential negation (p 4, ex (5), citing Carlson1977: 19)

follow-36 John didn’t see spots on the floor

=It is not the case that John saw spots on the floor

=There were spots on the floor such that John didn’t see them

Related to the above discussion is Williams’ (2007) argument that word-internalmorphemes cannot take their resolution outside the word Consider the followingdata

37 a John described himself

b John told self-destruction stories

In 37a, the word himself finds its resolution higher up in the clause In other words,

himself must be coreferential with John In 37b, on the other hand, the self in destruction cannot look outside the word to find its resolution Thus this sentence

self-can refer only to generic self-destruction stories and not to self-destructions storiesabout John There are, however, some situations in which word-internal morphemessometimes do look outside the word for resolution First, consider the followingpiece of data that complements 37

38 I witnessed John’s self-destruction

Here, self must be coreferential with John, thereby taking its resolution outside the

‘word’ that contains it Consider also reflexives in Onondaga

1.2 Noun Incorporation 17

39 John waThadadaehsé˛thwaT

John FACT- 3.SG.M.AG- REFL - kick- PUNC

‘John kicked himself.’

Here, the reflexive morpheme atat must find its resolution outside the word that tains it Thus, atat must corefer to John One final example is found in a particular kind of Blackfoot wh-question.

con-40 tsa ááni anná Mary anní John omaanistsíkkamiyoowatahpi aniskayi

apastaminam?

tsa aani ann-wa Mary ann-yi John

how say DET-PROXMary DETJohn

om-aanist-ikkam-iy-oowat-a-hp-yi

3-DEG -fast-?-eat.VTA-EPEN-NZLR-INANIM.SG

DET-OBV-INVIS-INANIM.SG apple

‘How quickly did Mary say that John ate that apple?’ (embedded scope on

‘how’)

The embedded verbal complex contains a degree variable, aanist, which is bound

by the wh-element, tsa, in the left edge of the matrix clause These three lines of

evidence illustrate that the word does not act as a syntactic island for resolution ofanaphora or variable binding

This section has described the basic properties of NI, which I list below in bullets

It is important to note that not all languages with NI instantiate all the propertiesbelow to the same extent Thus, it is difficult to pin down a single set of definingproperties for NI

• NI acts on direct objects, locations, instruments and paths

• NI does not act on indirect objects, subjects, goals, sources or comitatives

• the IN can often appear with a full nominal double

• NI constructions typically appear with N+V order

• NI constructions typically alternate with non-NI constructions, in other words,

NI is often optional

• the IN often exhibits frozen scope

• NI constructions often give rise to special or idiomatic readings

This section has also given various arguments in favour of a syntactic approach to

NI These arguments are all based on blurring the distinction between autonomoussyntactic and morphological/lexical modules of the grammar Specifically, I showedthat many of the so-called morphological or lexical properties of NI are also

applicable to a variety of syntactic constructions These included non-canonicalobjects, idiomatic readings, and frozen scope In the case of non-canonical objects,

I showed that the range of incorporable elements (direct objects, locations, paths,instruments and adverbs) constitutes a natural syntactic class based on theoreticalconsiderations on Larson’s work on the VP shell and on empirical facts based on Li’sdiscussion of non-canonical objects in Mandarin Chinese In the case of idiomatic

readings, I reiterated Marantz’ observation that phrasal elements (kick the bucket)

can have idiomatic readings, too, and discussed more facts specifically related to NI.Finally, I showed, based on van Geenhoven’s work, that frozen scope is a property

of of INs and of bare nouns in many languages

1.3 Conclusion and Outline of Book

This chapter has given a brief outline of the goals of this monograph and describedhow these will be accomplished To recapitulate, I offer an in depth description

of NI in Northern Iroquoian, formulate a theory of phrase structure that maintainsthe core ideas of BPS and the LCA, relate this theory of phrase structure to NI inNorthern Iroquoian, and finally show how this theory can account for linear or in NIconstructions in a number of other languages Section1.2also described the gen-eral properties of NI found cross-linguistically based on a number of prior works Italso discussed the linear order of NI constructions What was found was the follow-ing: while VO and OV is quite evenly distributed cross-linguistically, N+V orderprevails over V+N order This order is most robust in non-productive or lexicalizedconstructions, but still holds in productive NI

The remainder of this monograph is structured as follows.Chapter 2 presentsthe theoretical background of this monograph It begins with a brief discussion onthe history of phrase structure within generative grammar and then presents thetheories of BPS and Antisymmetry necessary for an understanding of the subse-quent material The chapter ends with a discussion on Head Movement and itsabsence from UG Readers familiar with the basic tenets of Bare Phrase Structureand Antisymmetry may wish to skip some of this material

Chapter 3 starts off by outlining some of the theoretical problems inherent inbringing together BPS and Antisymmetry and discusses some earlier efforts in thisdirection.Section 3.2 discusses in more detail the problems raised at the begin-ning of the chapter, and explores the two logically possible ways of redefining theLCA in BPS terms, both of which have been previously discussed in the literature.This section also explores in detail the consequences that each of these approacheswould have for the rest of the grammar.Section 3.3discusses the core problem men-tioned above, which is the initial merger of two heads, which we recall is an LCAviolation This section discusses the ramifications of adopting a dynamic view ofAntisymmetry, which includes Complement-to-spec roll-up.Section 3.4discussessome other possibilities that are compatible with the approach adopted here andshows how they might be implemented in particular circumstances, such as clitic

References 19climbing Finally,Section 3.5discusses late vocabulary insertion and the DistributedMorphology framework, and how it bears on the current proposal.

Chapter 4is the main empirical illustration of the proposal put forth in the ous chapter, which is the analysis of NI in Northern Iroquoian It begins with a briefdefence of a syntactic approach to NI and illustrates the various patterns of NI found

previ-in Northern Iroquoian The followprevi-ing section presents an analysis of NI previ-in this guage group and discusses some previous analyses of NI in Iroquoian.Section 4.4discusses some other core properties of NI in Iroquoian, including NI in ditransitivesand doubling Finally,Section 4.5discusses some properties of Iroquoian DPs.Chapter 5discusses various other kinds of N+V incorporation including putative

lan-NI in gerunds in English, German and Persian Although the process looks similar

in all the constructions in these three languages, minor cross-linguistic differencesmanifest themselves in interesting ways While all three cases have N+V order,Persian optionally allows V+N order Also, both Persian and English allow only barenouns in the NI constructions while German allows number marking, but no otherkind of marking The chapter also discusses incorporation of conjoined nominals inTamil

Chapter 6discusses various kinds of V+N incorporation, which on the face of it

is problematic for the approach here In particular, I discuss Pseudo NI in the sense

of Massam (2001) Data from Niuean and Chamorro are discussed, along with thestructure of nominals in these languages The proposal made here will be shown not

to be fully compatible with Massam’s approach This chapter additionally discussesV+N compounds in Romance languages as well as NI in Mapudungun and lexicalsuffixation in Upriver Halkomelem

Chapter 7is a conclusion It includes a summary of the major empirical andtheoretical findings of this volume It also discusses some of the limitations of thisproposal with the hopes of stimulating future research Finally, it includes somebrief comments on some extensions of the proposals laid out here

References

Abbott, Clifford 2000 Oneida Munich: Lincom Europa.

Abney, Stephen 1987 “The English Noun Phrase and its Sentential Aspect.” PhD diss., MIT press, Cambridge, MA.

Aboh, Enoch 2004a The Morphosyntax of Complement-Head Sequences: Clause Structure and Word Order Patterns in Kwa Oxford: Oxford University Press.

Baker, Mark C 1988 Incorporation: A Theory of Grammatical Function Changing Chicago, IL:

University of Chicago Press.

Baker, Mark C 1996 The Polysynthesis Parameter Oxford: Oxford University Press.

Baker, Mark C 2003 Lexical Categories Cambridge: Cambridge University Press.

Baker, Mark C., Roberto Aranovich, and Lucía A Golluscio 2005 “Two Types of Syntactic Noun Incorporation: Noun Incorporation in Mapudungun and Its Typological Implications.”

Bittner, Maria 1994 Case, Scope, and Binding Dordrecht: Kluwer Academic Publishers Bonvillain, Nancy 1972 “Noun Incorporation in Mohawk.” In Papers in Linguistics from the

1972 Conference on Iroquoian Research, edited by Michael K Foster, 18–26 Mercury Series

Ethnology Division Ottawa, ON: National Museum of Man.

Caballero, Gabriela, Michael J Houser, Nicole Marcus, Teresa McFarland, Anne Phycha, Maziar Toosarvandani, Suzanne Wilhite, and Johanna Nichols 2008 “Nonsyntactic Ordering Effects

in Syntactic Noun Incorporation.” Linguistic Typology 12 (3):383–421.

Cagri, Ilhan M 2009 “Arguing Against Subject Incorporation in Turkish Relative Clauses.”

Chomsky, Noam 1994 Bare Phrase Structure MIT Occasional Papers in Linguistics Cambridge,

MA: MIT Department of Linguistics and Philosophy, MITWPL.

Cook, Eung-Do, and Andrea Wilhelm 1998 “Noun Incorporation: New Evidence from

Athapaskan.” Studies in Language 22 (1):49–81.

Cuoq, Jean-André 1866 Etudes philologiques sur quelques langues sauvages Montreal, QC:

English-Cayuga/Cayuga-Gerdts, Donna B 1998 “Incorporation.” In The Handbook of Morphology, edited by Andrew

Spencer, and Arnold M Zwicky, 84–100 Malden, MA: Blackwell.

Ghomeshi, Jila 2003 “Plural Marking, Indefiniteness, and the Noun Phrase.” Studia Linguistica

57 (2):47–74.

Giorgi, Alessandra, and Giuseppe Longobardi 1991 The Syntax of Noun Phrases Cambridge

MA: Cambridge University Press.

Glougie, Jennifer 2000 “Topics in the Syntax and Semantics of Blackfoot Quantifiers and Nominals.” MA Thesis, University of British Columbia, Vancouver, BC.

Guimarães, Maximiliano 2000 In Defense of Vacuous Projections in Bare Phrase Structure.

University of Maryland Working Papers in Linguistics 9:90–115.

Harley, Heidi 2008 “Bare Roots, Conflation and the Canonical Use Constraint.” University of Lund, Lund, Sweden, February 5–6, 2008.

Jiang-King, Ping 1998 “Sonority Constraints on Tonal Patterns.” In The Proceedings of the Seventeenth West Coast Conference on Formal Linguistics, edited by Kimary Shahin, Susan

Blake, and Eun-Sook Kim, 332–46 Stanford, CA: Center for the Study of Language and Information.

Kayne, Richard 1994 The Antisymmetry of Syntax Cambridge, MA: MIT Press.

Kayne, Richard 2009a Antisymmetry and the Lexicon Linguistic Variation Yearbook

8 (1):1–31.

Kiparsky, Paul 1997 “Remarks on Denominal Verbs.” In Complex Predicates, edited by A Alsina,

Joan Bresnan, and Peter Sells, 247–88 Stanford, CA: Center for the Study of Language and Information.

Kleinschmidt, Samuel 1852 Grammatik der grönländischen Sprache mit theilweisem einschluss des Labradordialects Berlin: Reimer.

Kroeber, Alfred L 1909 “Noun Incorporation in American Languages.” Paper presented at the XVI Internationalen Amerikanisten-Kongress, 2d, Vienna and Leipzig.

Larson, Richard K 1988 “On the Double Object Construction.” Linguistic Inquiry

19 (3):335–91.

References 21

Larson, Richard K 1990 “Double Objects Revisited: Reply to Jackendoff.” Linguistic Inquiry 21

(4):589–632.

Larson, Richard K 2004 “Sentence Final Adverbs and “Scope”.” In Proceedings of NELS 34,

edited by Keir Moulton, and Matthew Wolf, 23’43 Amherst, MA: GLSA Publications.

Larson, Richard K 2010 “On Pylkkänen’s Semantics for Low Applicatives.” Linguistic Inquiry

Lounsbury, Floyd Glenn 1953 Oneida Verb Morphology New Haven, CT: Yale University.

Marantz, Alec 1997 “No Escape from Syntax: Don’t Try Morphological Analysis in the

Privacy of Your Own Lexicon.” University of Pennsylvania Working Papers in Linguistics

Megerdoomian, Karine 2008 “Parallel Nominal and Verbal Projections.” In Foundational Issues

in Linguistic Theory: Essays in Honor of Jean-Roger Vergnaud, edited by Robert Freidin,

Carlos P Otero, and Maria Luisa Zubizaretta, 73–103 Cambridge, MA: MIT Press.

Michelson, Karin 1991 “Possessor Stranding in Oneida.” Linguistic Inquiry 22 (4):756–61 Michelson, Karin, and Mercy Doxtator 2002 Oneida-English/English-Oneida Dictionary.

Toronto, ON: University of Toronto Press.

Mithun, Marianne 1984 “The Evolution of Noun Incorporation.” Language 60 (4):847–94 Moro, Andrea 2000 Dynamic Antisymmetry Cambridge, MA: MIT Press.

Moro, Andrea 2004 “Linear Compression as a Trigger for Movement.” In Triggers, edited by

Anne Breitbarth, and Henk van Riemsdijk, 387–430 Berlin, New York, NY: Mouton de Gruyter.

Oishi, Masayuki 2003 “When Linearity Meets Bare Phrase Structure.” Current Issues in English Linguistics 2:18–41.

Öztürk, BalkIz 2009 “Incorporating Agents.” Lingua 119 (2):334–58.

Polinsky, Maria 1990 “Subject Incorporation: Evidence from Chukchee.” In Grammatical Relations: A Cross-Theoretical Perspective, edited by Katarzyna Dziwirek, Patrick Farrell, and

Errapel Mejias-Bikandi, 349–64 Stanford, CA: The Center for the Study of Language and Information.

Postal, Paul 1979 On Some Syntactic Rules in Mohawk New York, NY: Garland Publishing Pylkkänen, Liina 2008 Introducing Arguments Cambridge, MA: MIT Press.

Rackowski, Andrea, and Lisa Travis 2000 “V-Initial Languages: X or XP Movement and

Adverbial Placement.” In The Syntax of Verb Initial Languages, edited by Andrew Carnie,

and Eithne Guilfoyle, 117–42 Oxford: Oxford University Press.

Rice, Keren 1991 “Intransitives in Slave (Northern Athapaskan): Arguments for Unaccusatives.”

International Journal of American Linguistics 57 (1):51–69.

Richards, Norvin 2001a “A Distinctness Condition on Linearization.” In 20th West Coast Conference on Formal Linguistics, edited by Karine Megerdoomian Somerville, MA:

Cascadilla Press.

Ritter, Elizabeth 1992 “Cross-Linguistic Evidence for Number Phrase.” Canadian Journal of Linguistics 37 (2):197–218.

Ritter, Elizabeth 1993 “Where’s Gender?” Linguistic Inquiry 24 (4):795–803.

Rosen, Sara Thomas 1989 “Two Types of Noun Incorporation: A Lexical Analysis.” Language

65 (2):294–317.

Ruhlen, Merritt 1975 A Guide to the Languages of the World Stanford, CA: Language Universals

Project, Stanford University.

Sadock, Jerrold 1980 “Noun Incorporation in Greenlandic: A Case of Syntactic Word Formation.”

Language 56 (2):300–19.

Sadock, Jerrold 1986 “Some Notes on Noun Incorporation.” Language 62 (1):19–31.

Sapir, Edward 1911 “The Problem of Noun Incorporation in American Languages.” American Anthropologist 13:250–82.

Sato, Yosuke 2010 “Bare Verbal Nouns, Idiomatization and Icorporation in Japanese.” Paper presented at the Theoretical East Asian Linguistics, 6, Peking University: Beijing.

Spencer, Andrew 1995 “Incorporation in Chukchi.” Language: Journal of the Linguistic Society

of America 71 (3):439–89.

Szabolcsi, Anna 1983 “The Possessor that Ran Away from Home.” The Linguistic Review 3

(1):89–102.

Uriagereka, Juan 1999 “Multiple Spell Out.” In Working Minimalism, edited by Samuel Epstein,

and Norbert Hornstein, 251–82 Cambridge, MA: MIT Press.

van Geenhoven, Veerle 1998 “Semantic Incorporation and Indefinite Descriptions: Semantic and Syntactic Aspects of Noun Incorporation in West Greenlandic.” Dissertations in Linguistics (DiLi) Stanford, CA: Center for the Study of Language and Information.

Williams, Edwin 2007 “Dumping Lexicalism.” In The Oxford Handbook of Linguistic Interfaces,

edited by Gillian Ramchand, and Charles Reiss Oxford: Oxford University Press.

Woodbury, Hanni 1975a “Noun Incorporation in Onondaga.” PhD diss., Yale University, New Haven, CT.

Woodbury, Hanni 1975b “Onondaga Noun Incorporation: Some Notes on the Interdependence of

Syntax and Semantics.” International Journal of American Linguistics 41 (1):10–20 Woodbury, Hanni 2003 Onondaga-English/English-Onondaga Dictionary Toronto, ON:

University of Toronto Press.

Chapter 2

Theoretical Background

This chapter establishes the theoretical background in which the forthcominganalysis is couched In particular, I adopt the tenets of Bare Phrase Structure,Antisymmetry and Dynamic Antisymmetry Readers familiar with these conceptsmay wish to skip to the next chapter, where the theoretical proposal is presented, or

to skim the following sections The analyses of NI are presented inChapters 4,5,and6 However, before these analyses can be presented, I propose inChapter 3howBPS and the Linear Correspondence Axiom (LCA) can be made compatible witheach other In doing so, I will show that NI can be captured without appealing tohead movement, which has fallen out of favour in recent years

The remainder of this chapter is organized as follows Section2.1covers a briefhistory of phrase structure and how it relates to linearization Section2.2outlinesthe theory of BPS, and Section2.3outlines the theory of Antisymmetry Section2.4discusses the role of head movement and its elimination from UG Section 2.5covers various alternative theories of linearization Finally, Section2.6concludesthe chapter

2.1 A Bit of History of Phrase Structure

Chomsky (1957;1965) proposes that the grammar must distinguish between ciples that determine the structure of sentences in a given language and those thatderive the different word orders found for sentences in that language The formerare known as phrase structure (PS) rules and the latter are known as transforma-tions Phrase structure rules of the Standard Theory and its descendents gave way

prin-to the Headedness Parameter (Chomsky 1981; Stowell 1981; Travis 1989 interalia) in X-Bar Theory (Chomsky1970; Jackendoff1977) In its broadest form, theHeadedness Parameter allows for the four settings in 1; however, many formula-tions allow only the first two, with the specifier obligatorily on the left (Oishi2003;Richards2008)

23

M Barrie, Dynamic Antisymmetry and the Syntax of Noun Incorporation, Studies

in Natural Language and Linguistic Theory 84, DOI 10.1007/978-94-007-1570-7_2,

C

 Springer Science+Business Media B.V 2011

1 Four Structural Types of XPs

C-H-S order

The prevailing view, introduced nearly simultaneously by Stowell (1981) andChomsky (1981), was that the Headedness Parameter was a category-neutralspecification of the linear order of the specifier, head and complementizer in a givenlanguage The result of this approach was that clusters of word order properties(postpositions, post-nominal determiners, and OV order versus prepositions, pre-nominal determiners, and VO order; see Greenberg1963) could be easily accountedfor with one parametrically determined setting for headedness of all XPs In con-trast to this is the view that different syntactic categories can be assigned differentparameter settings, a view that leads to severe over-generation

There are several problems with the Headedness Parameter, however, many

of which have been previously discussed (Kayne 2003a, 2009; Kroch 2001) Idiscuss these only briefly here and leave the reader to consult the references indi-cated for a fuller discussion Kayne discusses at length several typological gapsthat the Headedness Parameter predicts should exist Further, Kroch (2001) pointsout that all SOV languages (even Japanese according to Kayne) depart from theGreenbergian correspondences in some way

Nakajima (1999) also points out that one of the goals of BPS is to pursue aderivational rather than a representational approach to phrase structure He suggests

2.1 A Bit of History of Phrase Structure 25that the representational nature of the Headedness Parameter is inconsistent withthe derivational approach of BPS Specifically, Nakajima argues that the headed-ness parameter is a condition on the representation of a syntactic structure and thatwhat is needed is a derivational approach to linearization.1,2 Additionally, Kayne(2009) points out that if the Headedness Parameter existed, it would be a verystrange parameter Ideally, parameters should be properties of functional elements

in the Lexicon, and the Headedness Parameter is not tied to any such functionallexical item Kayne draws the following analogy Most models of generative syntaxassume a bottom-up derivation; however, top-down models have also been proposed(Phillips2003) No one, though, has ever proposed that languages vary parametri-cally between these two choices such that some languages are bottom-up and someare top-down A Headedness Parameter, Kayne contends, would be no different thanthe bizarre parameter just described

Given the numerous problems with the Headedness Parameter, it comes as nosurprise that one would resurrect the Universal Base Hypothesis.3 The form ofthe Universal Base Hypothesis adopted here, of course, is Antisymmetry (Kayne1994), which is described in detail below Before continuing, however, I address acommon remark on Antisymmetry concerning the postulation of putatively unmo-tivated movements and functional projections to achieve the correct word order inthis framework (see Richards2008for recent discussion) I note that this approachlacks no more explanatory adequacy than an approach that assumes a HeadednessParameter with both left and right adjunction Headedness coupled with free left andright adjunction fails to prevent the same kinds of unattested structures that free gen-eration of functional projections and movements do One matter of ongoing research

is to understand the kinds of movements and functional projections available in UG,

in an attempt to constrain Antisymmetry

The next section outlines the theoretical approaches in which the proposal inChapter 3is framed: BPS and Antisymmetry The end of this section has a short dis-cussion of head movement and its proposed elimination from UG Readers familiarwith these concepts can skip to Section2.5

1 Nakajima offers such an approach, which we argue against later The current proposal does offer

a strongly derivational approach to linearization.

2 Wojdak ( 2008 ), in fact, does propose a derivational approach to word order assuming an updated version of the Headedness Parameter Again, however, such an approach suffers from many of the same problems as the traditional Headedness Parameter as described in this section, including over-generation.

3 The exact formulation of the Universal Base Hypothesis (Bach 1968 ) is still a matter of debate (Zwart 1997 ; Broekhuis 2006 ) On the basis of the prevalence of SVO and SOV patterns cross- linguistically, it comes as no surprise that the two contenders for this hypothesis are the S-H-C order (Kayne 1994, inter alia) and S-C-H order (Fukui and Takano 1998, inter alia) Still, others

have proposed that the directionality parameter specifies only the order between the head and the complement, with the specifier universally on the left (Oishi 2003 ; Ernst 2003 ; Richards 2008 ).

We address these issues at the end of this chapter.