A ~.

.~ ~.>.~4.

~

~ ~

~ ~

J .,~ ~

2~ <. . .<2.222«2<2<.. < ««<~< ~«2< 2<2<2<22<22<2 «2~22<~ <<A«

2 2 <<2

2 2<2< ~ <222<2<

222<2.2 2<2< ~<2~22«2<2<>2<2<22A<42<2<P<2 <22 2< ««222<2>22<<>««22<2«2i2<2<2<2<22<2> K2~2<42424 2< <2< <<<< <2< 2<~j~2<22<2«2<2«A«P<~4 2< ~ < *<<<2< 24< «« <. 2«<2<2« 2.

2< 2 *..<«<4«~ <2 « «2 222< <<«<2< «<4>2 <24< «<2< <>«<<2222< <2<2<2

«<<<2< <2<

A 2222~< « «

2 2<2<2««< <2«22<<2« < <2<4«AAX>< I 2< <2<2« 2 <>2

2(2< <44<2< 2222<2 « 2<2<22< 22<2<2<2<4<2<2

2>22< 4 <22<2<22 2 < <<~~A«<~>.~ <22<.2<<< « <2 2 2<<<<<< 22<222....2<~< ><~<2«2<~4

<222<2 2<2<2< 2<>« <2< <<~ «<«22 ~r~<«<.2

< 2 2

4~442~L~i2«7<2< 2< «<««'<«2 <7< <2<<2«< 22<2 «<« 2«22<2«<, <22 <~ <22<<

22<2<2< «<2< <~< 22<2<2<22<2<22<2< 2<2<2<*<4<. 242<2< .<2«2<2<2<2<22«<2< <<<«< >2

2 22<2<22222<22 <2<. <2< <«<<<2< 247<2< 22< <4>24<2«2««2«« >~~<<~2<«< «<22<

22< <<2 < <« <2<< .2..

<2 <<«2<2<2<2<22«2<2<<2< <2 <2

2<2< «<<2<2<2>.«~ «< 2

<<2< <2< < < «<2<2<2<22<22< «2<2<2<2< 2<.<22«< <22<2<2«2«««2>«<22< «2<2««.2<2«44 2<2< 2 <22.<.«.2««2<<

2 < <2 2< 2 «2<2<<<«<«2< <22<%2<<2<2< <<....<>22~2<< .<<2 <<2 «2 «'<<2 <2< ~< «2««<«««««<2«>22« <

<<<2 «2«2< 2 2<24222<2<2<2< A2<2<24<22<22<2<2~<2<2<22<2< <2 <.2<2<2<2«.2<2.2<4<24 ««««<

<<««<2<22<2<2<~2< «<2< .<.«2.<2«.2<22..<2< 2.222<2<2224<22 <2< < <22«<2<2«,« <242<22<2 2I >< 7<2<2<22<22<2<2<2<22< <

2< .«.. 2<22<2 <2«<*222<2«22<2<2<2«<4«4< <, <22<2<2<2<2<42,2.24<22 <>< 42<2«<2<22<2<2..2.22<2.2«2«2.2.<22<22<22. 2<2< ««<2« 22.2<21<2<.. <2 <2 «2..«< <42<2<2<2<2 <22<222<2 «<~ < .2<2< .. 2«<.«.«<2<2<44 <> 2<44 4 2<2<4422« 222<2< 2«<2«'~«2<.« < < ~$2« «

<222< 2 2 22 «2<2«2.I2<4 <2. 2~. ~ 22 2<~2 < <2<2« < 2«22<222 «< < 2< <«<22«2«)$4.2242<2<2< ~ 2< <4~4444«><

2<2<2 2<2<2<4<.<22<2«2<~2 <<2<2<222 .. < </«« 2«<«< 22<<2 «« 4~2<4<$4<4 2<2< <

2 2 ~ 2~2>2«[. <224< 242<2<2<2<2<2<2<4 2 2 2< 222>22 2

2<2< < 42<2< 2<2<422<

«<2< 2<<« <2<22** <2<2<2«««««««2«2««2«<.2.<22<2«<2«22 < <2< <2<22<22 «2<2«<2«< 2« <2

<22222<2 «2<2<2«~ 2««<~«<2<22<.~.2«22<2 «2<7< >22<2<22<7<2<2<2<2 22 «2««««<

>2< <222<22<2< «<2< ««<2>

2<2<2~ «<A«~

2«p22p«.«< 2< 2 2 222 2<2

42.2 2<2 <2222;2<22<.22<22< 242< 2< 222<<2 2 <2 . , <««.22<22

.2<2.2<2. 2<2.<2«2<2>2 «<2 >~42<22242< <<«<«2<2 ««<2<222< «2 ««2<4..2<2<.2«<; < 2< <222<2< 2< « < ««~p<422<44....<.«<.««« 2

<2<22«22<2««2«2 «2<4<22«<2~ «2< 2<<22 <22<42<2< <2< 22 2<2< «<

«<«<2< 24<2<2«<2<«< / <<22 <2>2< 2< «<«2<22<2< 2 « 2< <<<<< <2< 2 2<2«22««42<~2<2< 222 2<2<«<2<22<222 «<2 ~ <2 2<<2<222 22 <2<2«2«2««« 2<

2< <<2<222 2<2<22<22 2< «2<2<22<2< «2 2<2< <2 2

«<2 ~

<2< <<<2 < 2< < <2

2< .22<2<44« 2 <2<2 2

<2 2<2<222<22 2 2<2<2 2< 2 2222 2< 2<2<22< <2<2<

2222 <<2<2< 2<22<22222<

2<222

222<

2<2< 2< 2<<72 <2<2< 2<2< <22< 2<22 2 2< <<«<2<4 22 2<2 <<<«22222 «2<2 2<2..2<.<«.<<.«<2< <4< <<2< 222<2< <2 2 2<2<2 <22222

2<2<2< <22< <>224<. «<22<2.2<22<2<222<22<2

2< 2<2< <2 2 2 2 « «<2<.«.<22<2222.2<4< 2<2<2< <<«<A 2 «<2< 2 2<222<2<2<2< 2<2<42<2 «««<<<<.«<2<2<2«22<2<2<2j«2< <2< <>2<2< <2 ..2<2<2<2««««< <2<4422<42<2<. <.2«..2<2< 2 < < < <2<2< <2<2<42 2< 2<4<4<242<2<2 «««~«~44<~«««««~< «22<2< «.~:2<2<'7< 22 <2< 22<2

<4< «2. . <22 2< «<2<2<2<2<2< «<2<.22>224<24<22<22<22<442<2.2<2... 2<22 2<2<22<2<2<22 <<2< 2<2<2««< 22<2<2<2< 2 2<2<2< 2 2 <'2< «<2<2<.« 22 <<.««< 22<2 2«<«<2< 2< 2 2< ~~2 <2<< 222<222222222<222.222222<2222222222<222.22<22222<22< I.222<22«.«< 2<2<2 < <«222<22<2 <22 « < <22 4<22 2 22< 222< «<2<~~«2 <2<2 2<222<2<2<2<2222<2<2242 2422 <2<2 242<2<4i~2<2<2<2<2<2««2< 22 2<22.2 «2<2 <2 22<2<22<2 <<<<2< 2«2<2«2«22««2<2<2<2<22<22<222242<2<22«<<<2< 22<2<22< <2<«««<2<>~'« <4 .,«2««2 «2<4<4<222 22<22< <2<2 «2 2 >242«<:2«< 22< 2< «<~<i«<< <<«2 <2 2<4

...........................2222<222<22<22<2 2< «.2<.«<<2<2«<2«A<2222«<2«< «2<2<2<2<2<2< <2222 «««<2<- <2 ~ <~<

424 RLE Progress Report Number 140

Chapter 1. Linguistics

Chapter 1. Linguistics

Academic and Research Staff

Professor Morris Halle, Professor Noam A. Chomsky

1.1 Introduction

The work of the Linguistics group is directed toward abetter understanding of the mental capacities ofhuman beings through the study of the nature, acqui-sition, and use of language. Language is a uniquelyhuman faculty: only humans appear to be capable oflearning and using a language, and every normalhuman acquires knowledge of one or more lan-guages.

We are trying to understand how this linguistic knowl-edge is represented in the speaker's mind. The cen-tral issues of linguistics research are:

What is the nature of linguistic knowledge? What dospeakers of a particular language such as Latvian,Spanish or Walpiri know, and how does knowledge ofone language resemble or differ from that of anotherlanguage? How do speakers acquire this knowl-edge? How do speakers put this knowledge to use inproducing and understanding utterances? What arethe physiological mechanisms that provide the mate-rial basis for storage, acquisition and utilization of lin-guistic knowledge?

Our ability to answer these questions differs consid-erably, and our research reflects these differences.At present, we have progressed further with regard toanswering the questions posed by item one and havemade less progress with item four. Currently, ourresearch is heavily concentrated on issues con-cerned with the nature of the knowledge that charac-terizes fluent speakers of various languages.However, we are making a significant effort to solvethe other questions also.

We are studying these topics along a number of par-allel lines. Linguists have investigated the principlesby which words are concatenated to form meaningfulsentences. These principles have been the primarydomain of inquiry into the disciplines of syntax andsemantics. Phonology studies the sound structure ofwords while morphology examines the manner inwhich different languages combine different mean-ing-bearing units (specifically, stems, prefixes, suf-fixes and infixes) to form words. The latter topic has

attracted increasing interest in recent years and willprobably become more prominent in our researchefforts in the future.

1.2 Abstracts of Doctoral Dissertations

The following are abstracts of dissertations submittedin 1997 to the Department of Linguistics and Philoso-phy in partial fulfillment of the requirements for thedegree of Doctor of Philosophy in Linguistics.

1.2.1 Verb Raising and A/A-Bar Distinction:Evidence from Exeptional CaseMarking

Judy Yoo-Kyung Baek

Abstract

This thesis investigates the nature of A-movement onthe basis of the facts regarding ECM constructionsacross languages. More specifically, this thesis dealswith two types of illicit A-'chains found in ECM con-structions in Korean: A-movement out of CP/finiteclauses.

To account for the impossibility of A-movement out ofCP in the theory of grammer, a locality condition onchain (=LCC) is proposed. The intuition behind theLCC is that A-chain is truly local in its nature in thesense that it needs to go through every interveningspecifier whether actual feature checking takes placeor not. In contrast, A-chain only goes through theposition that is required for feature checking.

A reconsideration and a new formulation of thenotion of A/A'-distinction has also been made in thisthesis,which depends on the property of the categorythat occupies the head of the specifier. A specifier ofa head that contains a lexial category (=N, V, A, P) ora trace of lexial category counts as an A-position,while specifiers of a functional head counts as an A-position. A significant consequence of this reformula-tion of A/A'-distinction is that verb raising cruciallyhinges on that A-status of a specifier of the functionalcategory to which the verb raises and adjoins.

425

Chapter 1. Linguistics

A crosslinguistic study of A-movement in this thesisalso shows that a generalization can be establishedthat a language that allows A-movement out of afinite clause lacks overt infinitival constructions. I pro-pose that in laguages that do not have overt infiniti-vals a finite T serves a dual function of both finite andinfinitival T in the sense that finite T has strong nomi-native case feature that can be checked against DPwith undeleted case feature regardless of its caseproperty. The strong nominative case feature of T willattract the closest DP into its Specifier position in theovert syntax. A crucial consequence of this claim fol-lows that a case feature can also enter into a multiplechecking relation due to feature mismatch. ECM isexeptional in these languages in the sense that itinvolves multiple feature checking of the accusativecase.

1.2.2 Metrical Theory and English Verse

David McKay

Abstract

I propose a generative linguistic theory of rhythmicstructure in English verse based on principles of met-rical phonology, in which metrical grids are built up bynatural metrical rules on the basis of a phonologicalrepresentation and are then subject to various con-straints, including important constraints on phonolog-ical phrasing. I use this theory to analyze poems byYeats, Hopkins, Longfellow, Swinburne, and Shakes-peare. I show that theory can account for a greatvariety of verse rhythm in a natural way, includingsome which have not been previously analyzed, andI show that it allows an analysis of Hopkins' verse inSprung Rhythm which is more accurate than, andtheoretically preferable to, earlier analyses in the tra-dition of generative metrics. Finally, I discuss somehypotheses about the parameters of variation inEnglish verse rhythm.

1.2.3 Jingulu Grammar, Dictionary, andTexts

Robert J. Pensalfini

Abstract

This dissertation is primarily intended as a throughdescription of the Jingulu language of North-CentralAustralia. The first part describes Jingulu's phonolog-ical, morphological and syntactic structure, illustrat-ing this with numerous examples collected by theauthor in the field in 1995 and 1996. There is a sec-

ondary focus on what Jingulu can contribute to anunderstanding of language from a theoretical per-spective. Chapter 1 focuses on the socio-historicalcontext in which the language is spoken, with alengthy excursus on the state of endangerment of themajority of the world's languages. Chapter 2 isdevoted to Jingulu phonolgy, with in-depth theoreticaldiscussions of Australian stress systems and Jinguluregressive vowel harmony. Chapter 3 outlines thearchitecture of the language faculty and the theory ofmorphology that underlie the description and analy-ses of the following chapters. Chapter 4 discussesJingulu syntax, focusing on the question of noncon-figurationality, and includes a syntatic typology of thevarious types of nonconfigurationality found amongthe world's languages. Chapters 5 and 6 are exposi-tions of the morphology of Jingulu nominal and ver-bal words, respectively. The theory outlined inchapter three is applied in detail to the complex andapparently bizarre morphological systems of Jingulu,and this complexity is seen to follow from a smallnumber of principles governing how formal featurescan be spelled out an the surface. Chapter 7 con-tains 34 glossed and translated texts collected by theauthor. Throughout the dissertation, I have preferredto provide more text and sentence examples ratherthan fewer, so that future researchers can test mygeneralizations, examine the data to find their own,and refute or affirm my analyses.

Part II of the dissertation is a Jingulu to English dic-tionary with an English to Jingulu word finder. EachJingulu entry in the dictionary is accompanied bygrammatical, morphological, and cultural informationin addition to an English translation. Most Jinguluentries also include examples of the word used in asentence. The dictionary is the latest stage in a col-laboration that has involved many people over sev-eral decades.

1.2.4 Aspects of A

Orin J. Percus

Abstract

This thesis is about interpretation of indefinites. Itargues that English singular indefinites are individ-ual-denoting expressions that encode a dependencyon situations and carry a uniqueness presupposition.In particular, it argues that once we accept that theindefinite determiner imposes a uniqueness presup-position we gain an understanding of the phenom-enom of quantificational variability At the heart of the

426 RLE Progress Report Number 140

Chapter 1. Linguistics

explanation is the idea that uniqueness presupposi-tion provides a clue on the basis of which we identifythe domain of an adverbial quantifier.

1.2.5 What Moves Where When in WhichLanguage?

Norvin W. Richards Ill

Abstract

Much work in syntax has used the properties of wh-movement as a probe into the nature of the deriva-tion. One perennial issue is the nature of wh-in-situ.Is wh-in-situ related to its scopal position by an oper-ation like movement or by an entirely different pro-cess? If wh-in-situ does undergo invisible movement,why is this movement invisible? If we assume a deri-vational model, what is the relation between overtand covert movement in the derivation?

In this thesis I will investigate the properties of multi-ple-wh questions in a number of languages (particu-larly Bulgarian, Serbo-Croatian, Chinese, andJapanese), in an attempt to find evidence for a partic-ular answer to these questions. I will argue that theclassic model assumed by the Extended StandardTheory is essentially correct: there is covert move-ment, and all covert movement follows all overtmovement in the derivation (and is therefore invisiblebecause it takes place after the point in the derivationat which the representation is interpreted by the pho-nological component).

One crucial aspect of the argument will involve inves-tigation of the nature of additional-wh effects. I willclaim that additional-wh effects only appear whencertain structural and derivational conditions on therelation between the wh-movements involved aremet, and additional-wh effects can therefore be usedto determine which wh-movement operations pre-cede which others.

Chapter 1 is an overview of some competing claimsabout the architecture of the grammar and a discus-sion of the nature of the evidence that might help usto choose among these claims. In chapter 2, I dis-cuss the distribution of the wh-island effects in anumber of languages, arguing that the overt/covertdistinction is in fact irrevelant to the distributuion tothe wh-islands. Chapter 3 is a discussion of thenature of Superiority effects in several languages. Inchapter 4, I investigate the nature of feature strengthand develop a version of Procrastinate which is

empirically distinct in several desirable ways fromthat developed by Chomsky (1993). Finally, Chapter5 discusses additional-wh effects in some detail.

1.2.6 INFL in Child and Adult Languages:

Agreement, Case, and Licensing

Carson Theodore Robert Scutze

Abstract

I propose an analysis of the inflectional system ofclauses that captures both cross-linguistic variationand differences between adult speakers and youngchildren learning a given language. The phenomenaof interest fall into two classes: (1) case marking andsubject-predicate agreement; and (2) tense markingand the licensing of overt and null subjects. Themajor goals are:

* to motivate the complete separation of licensing;

. to argue that agreement is exclusively responiblefor case, and tense exclusively for (subject)licensing;

* to propose a theory of case and agreement, moti-vated by child as well as adult laguage data;

* to argue that children's "root infinitive" utterancesviolate no principles of syntax in either domain-rather, children differ from adults in their choicesamong convergent structures.

I argue that structural case marking is a reflex of thesame syntactic feature-checking relation as agree-ment; I label this conglomeration Accord. The pres-ence in a clause of features involved in an Accord isnot an absolute convergence requirement. Rather, itis due to a preference among convergent deriva-tions, expressed as the accord maximization princi-ple (AMP), which compares structures that differ onlyon uninterpretable features (in the sense of Chomsky1995). Among those that meet all convergencerequirements, only those with the most Accord rela-tions are admissible. Children do not always suc-cessfully enforce this preference, sometimesreverting to representation where structural case fea-tures have not been introduced. When this happens,arguments appear in the default case of the languge,supplied in the spelled out component. Evidencefrom child corpus studies (both normal and Specifi-cally Language Impaired) shows that children knowboth that case and agreement must be checkedtogether and that default case must be suppliedwhen case is absent.

427

Chapter 1. Linguistics

With regard to subject licensing, I show that theabsence of Tense features is often compatible withboth PRO and overt subjects. Children's overt sub-jects of nonfinite clauses are thus consistent withadult grammar. The relationship between the distri-bution of syntactic Tense features and the meaningof clauses is governed by interface conditions onwhich adults and children apparently differ.

1.2.7 Specificity and Agreement in StandardWestern Armenian

Michele Sigler

Abstract

This thesis is a study of specificity and agreement inStandard Western Armenian (SWA) within the frame-work of the Minimalist Program (Chompsky 1993).Because it is a language that has rich nominal andverbal morphology, SWA provides us with overtsigns, in both the nominal and verbal domains, of theunderlying sturctural relations that constitute agree-ment as it is understood in this theoretical model.

The thesis has two parts. In the first part, I examinethe distribution and interpretation of nominal suffixes,paying particular attention to bear singular countnoun phrases, mass indefinites, bear plurals, andspecific noun phrases, which bear the definite articlesuffix. I show that the definite article is, in fact, amarker of specificity and attribute this to its beingassociated with the -feature person. I argue thatbare (singular count and mass) NPs lack 0-featurealtogether. Assuming the split DP structure proposedby Ritter (1992), I argue that 0-feature are checkedwithin DP and propose a feature-based characteriza-tion of the types of noun phrases distinguished bynominal suffixes.

In the second part I discuss the nonagreement con-struction, a construction in which nonspecific pluralnoun phrases do not trigger plural agreement onnontransitive verbs. I show that this can beaccounted for using the feature checking mechanismof the Minimalist Program, by assuming that the sub-ject is specified for number only and not the personand that number features are checked in the specifierposition of TP. In this derivation, AgrP is absent,because neither the subject nor the verb has personfeatures to check there. Positioning an Agr-less deri-vation allows us to account for the fact that transi-tives are unacceptable in non-agreementconstructions: In a derivation whose sole functionalprojection is TP, there is no place for a DP object to

check its Case features, hence transitive non-agree-ing derivations do not converge. Nonagreeing uner-gatives are ruled out on the assumption that theirsubjects are licensed in a position external to thepredicate (following Hale and Keyser 1993). Byassuming that the predicate is represented by the TPrather than VP, I conclude that the subject of an uner-gative or transitive is licensed in an external positiononly, where this means specifier of AgrP. We see thatthe proposed Agr-less analysis of nonagreeing sen-tences permits an account of their interpretationbased on Diesing's (1992) Mapping Hypothesis, ageneral account of the mapping of syntactic repre-sentation onto semantic representation.

428 RLE Progress Report Number 140