Instant Ebook Access, One Click Away – Begin at ebookgate.

com

The Typology of Parts of Speech Systems The

Markedness of Adjectives 1st Edition David Beck

https://ebookgate.com/product/the-typology-of-parts-of-
speech-systems-the-markedness-of-adjectives-1st-edition-
david-beck/

OR CLICK BUTTON

DOWLOAD EBOOK

Get Instant Ebook Downloads – Browse at https://ebookgate.com

Click here to visit ebookgate.com and download ebook now
Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

The Theory of Fusion Systems 1st Edition David A. Craven

https://ebookgate.com/product/the-theory-of-fusion-systems-1st-
edition-david-a-craven/

ebookgate.com

Sentence Comprehension The Integration of Habits and Rules

Language Speech and Communication David J. Townsend

https://ebookgate.com/product/sentence-comprehension-the-integration-
of-habits-and-rules-language-speech-and-communication-david-j-
townsend/
ebookgate.com

The Drama of Speech Acts Joseph A. Porter

https://ebookgate.com/product/the-drama-of-speech-acts-joseph-a-
porter/

ebookgate.com

Seeking Recognition The Termination and Restoration of the

Coos Lower Umpqua and Siuslaw Indians 1855 1984 1st
Edition David R. M. Beck
https://ebookgate.com/product/seeking-recognition-the-termination-and-
restoration-of-the-coos-lower-umpqua-and-siuslaw-
indians-1855-1984-1st-edition-david-r-m-beck/
ebookgate.com
Common Nonsense Glenn Beck and the Triumph of Ignorance
1st Edition Alexander Zaitchik

https://ebookgate.com/product/common-nonsense-glenn-beck-and-the-
triumph-of-ignorance-1st-edition-alexander-zaitchik/

ebookgate.com

Markedness and Economy in a Derivational Model of

Phonology Andrea Calabrese

https://ebookgate.com/product/markedness-and-economy-in-a-
derivational-model-of-phonology-andrea-calabrese/

ebookgate.com

Karl and Carolina Uncover the Parts of a Book In the

Library Sandy Bridget Donovan

https://ebookgate.com/product/karl-and-carolina-uncover-the-parts-of-
a-book-in-the-library-sandy-bridget-donovan/

ebookgate.com

Multi Tiered Systems of Support Implementation Tools for

Speech Language Pathologists in Education 1st Edition
Sylvan
https://ebookgate.com/product/multi-tiered-systems-of-support-
implementation-tools-for-speech-language-pathologists-in-
education-1st-edition-sylvan/
ebookgate.com

A Brief History of Ancient Astrology 1st Edition Roger

Beck

https://ebookgate.com/product/a-brief-history-of-ancient-
astrology-1st-edition-roger-beck/

ebookgate.com
OUTSTANDING DISSERTATIONS
IN LINGUISTICS

Edited by
Laurence Horn
Yale University

A ROUTLEDGE SERIES
OUTSTANDING DISSERTATIONS IN LINGUISTICS
LAURENCE HORN, General Editor

NASALIZATION, NEUTRAL SEGMENTS MORPHOLOGICALLY GOVERNED

AND OPACITY EFFECTS ACCENT IN OPTIMALITY THEORY
Rachel Walker John Alderete

PROSODY AND Focus IN EUROPEAN MINIMAL INDIRECT REFERENCE

PORTUGUESE A Theory if the Syntax-Phonology
Phonolo,Rical Phrasing and Intonation Inteifacc
Sonia Frota Amanda Seidl

LAYERS IN THE DETERMINER PHASE DISTINCTIVENESS, COERCION AND

Rob Zamparelli SONORITY
A Unified Theory of Weight
PHONOLOGICAL RELATIONS BETWEEN Bruce Moren
WORDS
Laura Benua PHONETIC AND PHONOLOGICAL
ASPECTS OF GEMINATE TIMING
CONSONANT STRENGTH William H. Ham
PluJlJolo,l;ical Patterns and Phonetic
Manifi'stations VOWEL REDUCTION IN OPTIMALITY
Lisa M. Lavoie THEORY
Katherine Crosswhite
P,'\TTERNS OF REDUPLICATION IN
LUSHOOTSEED AN EFFORT BASED ApPROACH TO
Suzanne Urbanczyk CONSONANT LENITION
Robert Kirchner
THE SERIAL VERB CONSTRUCTION
PARAMETER THE SYNCHRONIC AND DIACHRONIC
Osamuyimen Stewart Thompson PHONOLOGY OF EJECTIVES
Paul D. Fallon
LONG-DISTANCE DEPENDENCIES
Mihoko Zushi GRAMMATICAL FEATURES AND THE
ACQUISITION OF REFERENCE
THE MORPHOSYNTAX OF THE A Comparative Study of Dutch and
ALGONQUIAN CONJUNCT VERB Spanish
A Minimalist Approach Sergio Baauw
Julie Brittain
AUDITORY REPRESENTATIONS IN
TURN-TAKING IN ENGLISH AND PHONOLOGY
JAPANESE Edward S. Flemming
Projcctability in Grammar, Intonation and
Scrf/antics
THE TYPOLOGY OF PARTS OF
SPEECH SYSTEMS
The Markedness of Adjectives

David Beck

i~ ~~O~~!~~~:UP
NEW YORK AND LONDON
Pubhshed in 2002 by
Routledge
270 Madison Ave,
New York NY 10016

Published in Great Britain by

Routledge
2 Park Square, Milton Park,
Abingdon, Oxon, OXI4 4RN

ROLltiedge is an imprint of the Taylor & Francis Group

Transferred to Digital Printing 20 I 0

Copyright © 2002 by Routledge

All rights reserved. No part of this book may be reprinted or reproduced or uti·
lized 111 any form or by any electronic, mechanIcal, or other means, now known
or hereafter invented, including photocopying and recording, or in any informa·
tlOll storage or retrieval system, without permission in writing from the publish
er.

Library of Congress Cataloging-in-Publication Data

Beck, David, 19113-
The typology of parts of speech systems: the markedness of adjetives /
DavId Beck.
p. cm. - (Outstanding dissertations in linguistics)
Includes bibliographical references and index.
ISBN 0-415-'.14155-5
1. Grammar, ComparatIve and general-Adjective. 2. Markedness
(Linglllstics) 3. Parts of speech. 4. Semantics. 5. Grammar, Comparative and
general-Syntax I. Title. II. Series.
P273.B4 2002
415-dc21 2002017'.100

Publisher's Note
The publisher has gone to great lengths to ensure the quality of this reprint
but points out that some imperfections in the original may be apparent.
 Contents
Acknowledgements ............................................................................................ vii
Abbreviations ...................................................................................................... ix
Note on phonological transcriptions ................................................................... xi

Chapter 1 Introduction ............................................................................................. 3

Chapter 2 Definitions of lexical classes ................................................................ 11

2.1 Semantic characterizations ........................................................................... 12
2.2 Morphological diagnostics ........................................................................... 14
2.3 Syntactic distribution .................................................................................... 18
2.4 Extended roles and syntactic markedness ................................................... 20
2.4.1 Criteria for markedness ....................................................................... 21
2.4.2 WFM and markedness .......................................................................... 24
2.4.3 Rigid versus flexible languages .......................................................... 28
2.4.4 Measures of contextual markedness: De- and recategorization ........ 31
2.4.5 Markedness and prototypical mappings ............................................. 36
2.5 The semantics of parts of speech ................................................................ .41
2.5.1 Prototypicality and peripherality in lexical classification ................ .42
2.5.2 Semantic NAMES .................................................................................. 45
2.5.3 Semantic predicates ............................................................................. 48
2.5.4 Property concepts ................................................................................ 52
2.5.5 HUMAN CHARACTERISTICS ................................................................. 54
2.5.6 Why semantic NAMEs are not linguistic predicates ........................... 63
2.5.7 Non-prototypical semantic predicates and implicit arguments ......... 65
2.6 Syntactic markedness and semantic prototypes .......................................... 71

Chapter 3 Semantics, syntax, and the lexicon ....................................................... 75

3.1 Some basic terminology ............................................................................... 76
3.2 Lexicalization and syntactic structure ......................................................... 80
3.3 Adjectives, markedness, and iconicity ........................................................ 83
3.4 Relations between semantic NAMEs: Attribution and possession ............... 85
3.5 Minor lexical classes .................................................................................... 91

Chapter 4 Types of lexical inventory .................................................................... 95

4.1 Verb-Adjective conflating inventories ...................................................... 101
4.1.1 Noun, verb, and adjective in Salishan .............................................. 103
4.1.1.1 Nominal predicates and nominal actants ................................ 104
4.1.1.2 Verbs as actants ....................................................................... 113
4.1.1.3 Verbs as unmarked modifiers .................................................. 122
4.1.1.4 Modification in Bella Coola .................................................... 125

v
vi Contents

4.1.2 Cora .................................................................................................. 131

4.1.2.1 Modification and relative clauses in Cora .............................. 132
4.1.2.2 Nouns and modification in Cora ............................................. 134
4.1.2.3 Flexibility and rigidity as syntactic parameters ...................... 137
4.2 Noun-Adjective conflating inventories ..................................................... 140
4.2.1 Quechua ............................................................................................. 142
4.2.2 Upper Necaxa Totonac ..................................................................... 149
4.2.2.1 Property concepts in Upper Necaxa ........................................ 150
4.2.2.2 Adjectives and nouns as syntactic predicates ......................... 153
4.2.2.3 Adjectives as actants ................................................................ 157
4.2.2.4 Nouns as modifiers .................................................................. 162
4.2.2.5 Secondary diagnostics: Quantification and pluralization ....... 166
4.2.3 Hausa ................................................................................................. 172
4.2.4 The N[A Vj inventory reconsidered .................................................. 185

Chapter 5 Conclusions ......................................................................................... 189

References ........................................................................................................ 205

Index ................................................................................................................. 213

Acknow ledgements
The requisite-but nonetheless genuine-thanks are due to the members of
my committee, Alana Johns, Jack Chambers, and Elan Dresher for their advice
and encouragement, and for keeping an open mind. I would also like to thank
Igor Mel' ('uk for his tireless enthusiasm, inspiration, and the copious donations
of his time that I have been the beneficiary of. Paulette Levy's hospitality and
friendship made my work in Mexico both pleasant and feasible. Without her
assistance and the support of the Seminario de Lenguas Indfgenas of the Univer-
sidad Nacional Autonoma de Mexico, none of it would have been possible.
Thanks also go to Keren Rice for her helpfulness, which went a long way in
making the pursuit of a doctoral degree a more rewarding task, and to Nila
Friedberg, whose companionship made it a more fulfilling one.
Financial support for this dissertation came from a number of sources, prin-
cipally a Doctoral Fellowship from the Social Sciences and Humanities Re-
search Council of Canada (#752-96-1718). Support for my fieldwork on Toto-
nac came from a PRA Fellowship from the Organization of American States, as
well as a travel grant from the School of Graduate Studies at the University of
Toronto and two separate General Research Grants from the Department of Lin-
guistics. Support for aspects of the fieldwork not directly related to this thesis
was also received from Alana Johns' SSHRC-funded project on grammatical
paradigms, which also funded what little work I was able to do on Inuktitut.
My final thanks go, of course, to the colleagues and consultants who are the
ultimate sources of much of the data used in this thesis. The Lushootseed in par-
ticular was provided by Dr. T.M. Hess, who has over the years given me gener-
ous access to his data and his insights; the facts from Cora were brought to my
attention by Veronica Vasquez, to whom lowe what understanding I have of
this material. Paul Newman was kind enough to provide some of the Hausa ex-
amples used in this dissertation, and I am grateful to Ida A wa for having enough
patience with me to provide some instruction in Inuktitut. The Mandinka data in
Section 2.5.7 are courtesy of Lamin Jabbi. My fieldwork on Totonac is owed
completely to my consultants, in particular Porfirio Sampayo Macfn, Longino
Barragan, and Luciano Romero Aguilar. Alvaro Barragan Alvarez provided both
language data and invaluable aid in setting up the practicalities of fieldwork; I
am grateful to Dofia Rosita for her words and her humour, and to Braulio
Cevedeo, Bartolome Garcfa, Artemio Perez, Juan Ramfrez Cortez, Zoyla Rivera
Alvarez, and Manuel Romero for contributing their time and their knowledge of
the language. Very special thanks are owed to Adolfo Gonzalez Amador and his
family-Linda, Diana, Yoffn, and Juan-for their friendship, and for making me
feel a part of their cIano

vii
Abbreviations
1,2, 3 =1st, 2nd, 3rd person INTR = intransitive
ABL = ablative IRR = irrealis
ABS = absolutive LC = lack of control
ACC = accusative LNK =linker
ACT = active LOC = locative
AGT = agentive MASC = masculine
ALN = alienable MD =middle
AP =antipassive NEG = negative
APPL = applicative NEU = neuter
ART = article NOM = nominative
AS =assertion NP =nominalizing prefix
ATRB = attributive NREL = non-human relative
CLS = classifier NS =non-subject
CMP = completitive OBJ = object
CMT = comitative PASS = passive
CNF = confirmative PERF = perfective
COL = collective PFT =perfect
CONT = continuative PL=plural
cOP =copula PNT = punctual
CS =causative PO = possessive
D =deictic PR =preposition
Df = feminine deictic PREP = prepositional
DAVC =DIMENSION, AGE, VALUE, COLOUR PRG = progressive
DEF = definite PRS = present
DET = determinative PST = past
DIM =diminutive PT =participle
DSD = desiderative QNT = quantifier
ECS = event-external causative QTV =quotative
EMPH = emphatic RC =relative clause
FEM = female RDP = reduplication
FUT =future REFL = reflexive
GEN = genitive REL =relativizer
HAB = habitual RES = resultative
HREL = human relative RHM = rhematizer
ICS = event-internal causative SBJ = subjunctive
IMPF = imperfective SBRD = subordinate
INALN = inalienable SG = singular
INSTR = instrumental STAT = stative
INTJ = interjection SUBJ = subject

ix
x Abbreviations

TOP= topic-marker
TRANS = transitive
TRM = transmutative
=
veN vicinity marker
VRB = verbalizer
=
WFM without further measures
Note on phonological transcriptions
For the sake of consistency, a single transcription system has been applied
to all the language data used here, independently of the system used in the
sources. The exceptions to this are languages with well-known, standardized
Latin orthographies (e.g. Russian, Spanish) and languages where the original
sources did not provide accurate enough phonological descriptions to allow reli-
able transliteration. Transcriptions therefore follow the standard Americanist
IPA where y =j, c = tJ, z =3,1:. = t1, and c = ts. In the data from Cora, cr is used
to represent a palatal stop. All other symbols have their normal IPA values, al-
though the Africanist system of accents has been used to mark tone in Hausa and
Mandarin, and the acute accent employed in the transcription of Totonac data is
used to indicate stress.

xi
THE TYPLOGY OF PARTS OF

SPEECH SYSTEMS
CHAPTER 1

Introduction

Ask a layperson what they know about grammar and you are likely to get an
answer that has something to do with parts of speech; ask a linguist what they
know about parts of speech and the answer is quite likely to be much less en-
lightening. Parts of speech systems or, as I will refer to them here, lexical
classes are among the most frequently overlooked aspects of linguistic analysis,
yet they are at the same time among the most fundamental elements of language.
Lexical classes playa key role in most-if not all-syntactic theories, they are
the cornerstones of lexicography and lexical semantics, and they are crucial
elements in morphological analysis, yet precise and rigorous definitions of these
classes have never been successfully formulated. More often than not, lexical
classes are treated merely as primitives, either in terms of input to rules, deter-
minants of underlying phrase structure, governors of inflectional patterns, or as
sources of valency and subcategorization frames. Thus, class-designations such
as "verb," "adjective," and "noun" are the linchpins of semantic, syntactic, and
morphological structure, but the terms themselves are rarely defined and their
properties, both formal and functional, are often taken for granted. While it is
certainly possible to carry out linguistic analysis without a clear definition of the
basic units involved (as long as the identity of these units can be agreed upon),
any theory which proceeds without a full understanding of its own primitives
rests on uncertain foundations.
Another serious drawback to building theoretical models based on tacit as-
sumptions about the properties and definitions of parts of speech is that, as mod-
ern linguistics expands its horizons and turns more and more to data from "ex-
otic" and previously undescribed languages, many of these assumptions are be-
ing challenged. Most current linguistic theories-whose main proponents are
speakers of and researchers in European languages-are built on the model of
what Sapir referred to as the "Standard Average European" language type with
its familiar three-way division of the lexicon into major open classes of verb,

3
4 The typology of parts of speech systems

noun, and adjective. Cross-linguistic investigation has shown, however, that not
all languages fit this pattern and at one time or another claims have been made
that certain languages lack distinctions between adjectives and verbs, adjectives
and nouns, or even between nouns and verbs (e.g. Kinkade 1983; Schachter
1985; Sasse 1993; Bhat 1994; Broschart 1997). When confronted with the neu-
tralization of parts-of-speech contrasts, any theory which relies on the three
major classes as primitives-or which, at least, has no clear idea of their ori-
gin-will have little success in dealing with this variation or of providing an
adequate explanation of its provenance.
In terms of cross-linguistic variability in parts of speech systems, perhaps
the most salient and widely-remarked upon point of divergence is the frequent
absence of the class of adjective. While nouns and verbs appear to be more or
less universal, languages that have no or only a few adjectives are a typological
commonplace. This seems to imply that there is something marked about the
adjectival category, and an investigation of its properties should shed some light
on the issues of typological variation in lexical classification and define lexical
classes in such a way as to motivate and constrain this variation. It is the aim of
the present work to deal with precisely these topics. The discussion will proceed
as follows: in Chapter 2, I will outline some previous, unary approaches to de-
fining lexical classes. Traditionally, these have taken three basic tacks-the se-
mantic characterization of parts of speech (Chapter 3), the development of mor-
phological diagnostics (mistakenly taken to be criterial definitions of the lexical
classes themselves) (2.2), and definitions based on simple syntactic distribution
(2.3). All of these approaches give more or less congruous results when applied
to the most typical members of each class and to languages that have Indo-
European style three-class parts of speech systems. When confronted with mar-
ginal cases and data from other types of language, however, none of them
proves adequate in terms of providing rigorous, criterial definitions.
More recent attempts to define parts of speech have tried to deal with varia-
tion in terms of syntactic markedness (2.4), a term that is discussed in some de-
tail in Section 2.4.1. One particularly good approach based on markedness is
that put forward by Hengeveld (1992a, 1992b) which formulates definitions of
parts of speech based on those syntactic roles that different lexical classes can
fiII without further morphological or syntactic measures being taken. As dis-
cussed in Section 2.4.2, when reformulated in terms of contrastive markedness,
Hengeveld's "without further measures" turns out to be a useful tool for identi-
fying the marked and unmarked roles of different parts of speech. Languages
differ from one another with respect to which lexical classes require further
morphosyntactic measures in order to occupy a given syntactic role, and lan-
guages that lack a specialized class of adjectives are divided by Hengeveld into
"rigid" and "flexible" types depending on the presence or absence of such meas-
Introduction 5

ures (Section 2.4.3). Further measures include, among other things, recategori-
zation (the acquisition by a word of the typical properties of another part of
speech) and decategorization (the loss of properties typical of a word's own
lexical class). These processes are described in Section 2.4.4. The major diffi-
culty with syntactic approaches based solely on distributional markedness, how-
ever, is that they in no way account for the common semantic core of meanings
that are consistently expressed by the same part of speech across so many lan-
guages. Section 2.4.5 looks at this issue in the light of work by Croft (1991),
which proposes that there are cross-linguistically unmarked mappings between
the semantic class and "pragmatic" roles typical of particular parts of speech.
Although Croft's proposal falls short of providing criterial definitions of
lexical classes and fails to properly constrain typological variation in parts of
speech systems, it does put forward the idea of leaving behind unary definitions
based on only one of semantics, syntax, or morphology. Instead, Croft's work
suggests a binary approach to the problem and looks at the mapping between the
typical properties of lexical classes at two levels of representation. The obvious
levels to choose for this are the syntactic and the semantic. The typical syntactic
properties of the three major parts of speech having already been discussed
(Sections 2.3 and 2.4), Section 2.5 goes on to examine their prototypical seman-
tic properties and shows how cross-linguistic variation in this domain can be
dealt with by a theory of semantic prototypes (2.5.1). This theory not only al-
lows the formulation of class prototypes for nouns (2.5.2), verbs (2.5.3), and
adjectives (2.5.4), it also predicts the most likely areas for cross-linguistic varia-
tion in class membership, meanings lying on the peripheries of the relevant se-
mantic categories being the most variable. Words referring to HUMAN CHAR-
ACTERISTICS represent a particularly variable class of meanings and are dis-
cussed in detail in Section 2.5.5. Following this some technical issues are dis-
cussed (Sections 2.5.6 and 2.5.7) before moving on to the next chapter.
The most important points contained in Section 2.5 are proposals for two
criterial semantic properties of nouns and verbs: nouns are said to be prototypi-
cally the expressions of semantic NAMEs (2.5.2) and verbs the expressions of
semantic predicates (2.5.3). Chapter 3 takes these two semantic criteria, com-
bines them with the unmarked syntactic roles of nouns and verbs identified by
Hengeveld (1992a, 1992b), and uses them to develop definitions of the two most
basic lexical classes. These definitions are spelled out in Section 3.1, which also
provides some terminology and outlines some elementary formalisms borrowed
from Meaning-Text Theory Cl.olkovskij & MeI'cuk 1967; MeI'cuk 1988). Sec-
tion 3.2 sketches the role played by the lexicon, or lexical inventory, in the
building of syntactic structures and iIIustrates where it is that lexical classes
originate (the lexicon) and how they interact with the rules mapping between the
semantic, syntactic, and morphological representations of sentences. Section 3.3
6 The typology of parts of speech systems

then demonstrates how the new definitions of verb and noun point to a clear and
criterial definition of the class of adjective. This definition accounts for the
cross-linguistic variability of the adjectival class (that is, for the fact that if a
language has only two major lexical classes, it has nouns and verbs and not ad-
jectives) in terms of the non-iconicity of the modificative relation, the unmarked
syntactic role of the adjective. The remainder of Chapter 3 deals with a number
of subsidiary issues, including the analysis of possessive and attributive con-
structions (Section 3.4)-both of which are shown to be distinct from modifica-
tion-and the potential application of the approach being developed here to the
definitions of two minor lexical classes (adverbs and adpositions-Section 3.5).
The next chapter, Chapter 4, examines the implications that these defini-
tions of lexical classes have for the typology of parts of speech systems, taking
as a starting point a common four-member typology current in the literature (e.g.
Schachter 1985; Bhat 1994). This typology proposes that, in terms of the three
major lexical classes, there are four possible types of lexical inventory: full
NA V inventories that distinguish three lexical classes, N[A V] languages where
words that are adjectives in three-class languages are verbs, [NA]V languages
where such words are nouns, and [NA V] languages which make no major-class
distinctions whatsoever. As it turns out, this typology is easily generated by a
feature system using the two criterial features, one syntactic and the other se-
mantic, that make up the definitions of parts of speech put forward in Chapter 3.
Thus, full inventory languages are said to be sensitive to both the syntactic and
the semantic parameters, the [NA V] inventories are sensitive to neither, the
N[A V] inventory is organized along purely semantic lines, and [NA]V inventory
would then be subdivided on syntactic grounds alone. The last type of inventory,
however, appears to present a problem in that, because the class of nouns inevi-
tably includes the expressions of all prototypical semantic NAMES, it is impossi-
ble to completely avoid semantic characteristics when considering the way in
which meanings are organized in the lexicon.
N[A V] inventories, however, are less problematic and two concrete exam-
ples of these are illustrated in Section 4.1, beginning with a discussion of lexical
classes in the Salishan family of languages (4.1.1). Salishan languages have ac-
tually been cited in the literature as making no major lexical class distinctions
(e.g. Kuipers 1968; Kinkade 1983), although based on the definitions for parts
of speech proposed here they can be shown to make the basic distinction be-
tween nouns and verbs. What most Salishan languages do not do, however, is
make a distinction between adjectives and verbs, all semantic predicates in these
languages being both unmarked syntactic predicates and unmarked modifiers.
Thus, the Salishan family (with the possible exception of Bella Coola, discussed
in Section 4.1.1.4) organizes its lexica on purely semantic grounds and does not
accord any special treatment to words expressing semantic predicates when they
Introduction 7

are used as modifiers. This is a very different situation from that found in Cora
(Section 4.1.2), which-like Salish--conflates adjectives and verbs but requires
that all modifiers of nouns appear inside relative clauses. Nonetheless, although
they differ syntactically, Cora and Salish are essentially identical in terms of
their parts of speech systems. As discussed in Section 4.1.2.3, the distinction
between the two arises from differences in the syntactic treatment of parts of
speech, not from any fundamental distinction in the number or type of lexical
classes distinguished in the lexicon.
Section 4.2 returns to the problem of [NA]V inventories. Such inventories
are quite common in the literature, the most frequent pattern being that repre-
sented by Quechua, where nouns and adjectives seem not to differ morphosyn-
tactically in either the role of modifier or the role of actant. Closer examination
of Quechua (4.2.1), however, reveals that both the use of nouns as "modifiers"
and the use of adjectives as actants are, in fact, marked uses when examined at
the semantics ¢::) syntax interface. In the former case, noun-noun "modifier"
structures can be shown to be attributive constructions as defined in Section 3.4.
Attributives are marked and non-iconic in that they involve the elision of an
underlying semantic predicate that is not realized in the syntax. Similarly, ad-
jectives used as actants appear to be elliptical constructions making anaphoric
reference to a nominal head whose identity is recoverable from discourse. This
implies that adjectival actants used out of context are ungrammatical, a hypothe-
sis that was put to the test in my fieldwork on Upper Necaxa Totonac, another
reputed [NA]V language of the Quechua type (Section 4.2.2). In addition, the
discussion of Upper Necaxa examines a number of other diagnostics for the
noun-adjective distinction. The analysis here demonstrates both the use of pri-
mary diagnostics for markedness in criterial syntactic roles and the application
of two secondary diagnostics, quantification and pluralization (Section 4.2.2.5).
These exemplify the ways in which such tests can be used to sort out lexical
class distinctions, as well as ways in which they can lead the investigator astray.
As a result of the discussions in Sections 4.2.1 and 4.2.2, languages of the
Quechua-Totonac type are shown not to be eligible for classification as [NA]V
languages because of the types of elision that occur both in noun-noun attribu-
tive constructions and when adjectives are used as actants. This leaves as the
only possible type of [NA]V language a language like Rausa (4.2.3) which does
not allow unmarked attributive constructions and avoids ellipsis by giving ab-
stract nominal readings to "adjectives" when these appear in actantial position.
Unfortunately, because of this last characteristic and one or two other features of
the syntax, it turns out that "adjectives" in Rausa are considered by speakers to
be the expressions of semantic NAMEs and are, therefore, abstract nouns. This
forces the reconsideration of the existence of the [NA]V inventory, which-as
shown in Section 4.2.4-turns out to be a logically impossible type of language.
8 The typology of parts of speech systems

This issue is discussed at some length in Chapter 5, which argues that the con-
straints on typological variation uncovered in the preceding chapter can be eas-
ily accounted for by replacing the free recombination of semantic and syntactic
features proposed at the beginning of Chapter 4 with an algorithm for the subdi-
vision of the lexical inventory that gives primacy to semantics over syntax. The
result is a sufficiently constrained theory of typological variation in parts of
speech systems based on rigorous and criterial definitions of each of the three
major lexical classes.
Before launching into this discussion, it is probably a good idea to say a few
words about the methodology applied here and some of the self-imposed limita-
tions of this study. While this investigation does aspire to being an essay in ty-
pology, it is of a substantially different nature than the broad-based typologies
inspired by Greenberg (1963) and others which attempt to take data from dozens
or even hundreds of languages and distill from them universal patterns and sta-
tistical tendencies (a particularly outstanding study of this type is that of Nichols
1992). While this is a feasible task when dealing with highly salient or superfi-
cial features of a language such as unmarked word-order or inflectional patterns
(although even features like these present a good number of problems), coming
to terms with the parts of speech system of an unfamiliar language is a far more
complicated task. As will be seen in some of the case-studies below, defining
the lexical classes of an individual language involves an understanding of a wide
variety of its semantic patterns, syntactic structures, and morphological features
and, in the most difficult cases, requires the investigator to go well beyond the
type of information included in most ordinary descriptive grammars. Given the
degree of familiarity required to sort out the parts of speech systems (at least in
problematic languages, which-after all-are the most interesting ones for this
type of study) and the type of information available, examining the lexical in-
ventories of hundreds of languages is the work of a lifetime. Instead, I have cho-
sen to let others do much of the inductive work and, based on their findings,
have identified four types of lexical inventory that seem to enjoy a great deal of
currency in the literature. From these, I examine a small number of languages
belonging to two of them.' The down side of this is, of course, that there may be
among the hundred of languages that I did not examine numerous exceptions or
even types that I have not anticipated. However, by focusing on a small number
of languages (five, if the Salishan family counts as one), it is possible to speak

, Of the other two types, the full inventory language is well-known and all-too-familiar,
and so really needs no special attention (naturally, it is discussed in passim at various
points in the dissertation). The fourth type, the [NAY] inventory is deliberately passed
over, given that the focus here is primarily the variation in the class of adjective. Some
thoughts on [NAY] inventories are found in Chapter 5.
Introduction 9

with a little more assurance and be confident that some measure of justice is
done both to the data and to its interpretation.
Also in aid of constraining the task at hand, I will limit the present discus-
sion to the three major classes of noun, verb, and adjective, and I will only be
concerned with the classification of "lexical meanings"-that is, with meanings
that refer to items or describe real-world qualities and events rather than ex-
pressing grammatical categories. Of course, this distinction is not always easy to
draw and theoretical approaches often vary widely in the criteria used to make it
(if they have any at all), but it will have to be adhered to here as a convenient
fiction. Another such fiction, one that is very common in the typological litera-
ture, is the notion of absolute language types. Hausa, for instance, is character-
ized in this introduction and throughout most of the discussion below as a po-
tential [NA]V language, implying that it has no adjectives, when in reality it has
a very small class of about a dozen of these. Similarly, Mandarin is character-
ized for heuristic purposes as having no adjectives in Section 2.4.2 (it actually
does have a few) and Tuscarora is said in Chapter 5 to completely lack un-
derived nouns, although it apparently does have a small number of nominal
roots. It is important to remember that when dealing with human language there
are no absolute types, and that its inherent variability and creativity will always
defeat those who want to speak in absolute terms on typological issues. This is
especially true of discussions of the lexicon which is, by definition, the reposi-
tory of the unsystematic, contradictory, and idiosyncratic. Thus, like so many
other linguistic classifications, these types must be taken as potentially gradient
categories and the languages discussed here treated as idealized versions of the
real thing. This is a necessary step in order to make generalizations, but should
not be allowed to obscure the fact that many languages may occupy intermediate
positions between the postulated types, or that they may conform to a type when
it comes to regularities but also have a lexicon chock-full of exceptions.
Another deliberate omission here has been discussion of the proper mor-
phosyntactic domain of lexical classification-that is, if lexical class distinctions
apply to lexical items or lexemes, what constitutes a lexeme? This problem is
especially perplexing in poly synthetic languages where it is often unclear what a
word is (on both the syntactic and phonological levels). In such languages, it
frequently appears either that parts of words belong to lexical classes, or that
entire phrases or even clauses can be legitimately treated as one or the other of
the major parts of speech. Resolving this problem (or even motivating a coher-
ent position on the issue) is, of course, the topic of a dissertation in itself and I
will make absolutely no attempt at it here. For the purposes of this investigation,
lexical class distinctions will be assumed to apply to words (whatever those are)
and, potentially, to set phrases and expressions in the lexicon. On this last point,
it should be kept in mind that the lexicon, at least until it is modeled by the lin-
10 The typology of parts of speech systems

guist, is not a dictionary and so is not constrained by lexicographic conventions.

Speakers have knowledge not only of words but of frequent combinations of
words and the conventionalized meanings these have, just as they have knowl-
edge of the conventionalized meanings of particular combinations of roots and
affixes. This last issue, that of conventionalization of meaning, will play only a
minor role in this discussion until it is addressed directly in the last chapter.
A final point that might need some clarification, particularly given the cur-
rent sociology of the field of linguistics, is the theoretical orientation of the pre-
sent work. This dissertation is intended as a typological study of the variation in
parts of speech systems and seeks to work out accurate and generally-applicable
definitions of parts of speech; it is hoped that these definitions and the approach
outlined here will be useful and accessible to as broad a cross-section of the field
as possible. Wherever feasible I have couched my definitions and arguments in
widely-accepted and generalizable terms that can be used (or at least under-
stood) by adherents of numerous theoretical approaches, and at several places in
the discussion I have spelled out how certain theoretical points might be ex-
pressed in the conventions of different models. Although I do draw on the for-
malisms of Meaning-Text Theory, I do so because it is my feeling that these
formalisms are straightforward and easily accessible, and that they elegantly
illustrate the larger points that I am making-however, it is not my belief that
anything I am arguing for depends crucially on the assumptions and conventions
of this particular theoretical framework. While the insights that I am trying to
capture here are often expressed in formal terms and, in the end, are considered
successful to the extent that they are accurately formalized, I draw very heavily
on the cognitive-functional literature, and the work as a whole is informed by
the belief, made explicit at various points in this work, that language is a system
for the expression of meaning, and that it is impossible to get to the heart of
many linguistic phenomena without taking into account how this meaning finds
expression in the morphosyntactic structures of natural language. Equally, I be-
lieve that it is useful to formalize such observations so as to allow the generali-
zations they imply to be applied and tested in a rigorous manner in a variety of
situations and languages. It is precisely this that I have set out to do here and it is
left to the reader to judge whether this has been successful and, hopefully, to
find something of use, whatever herlhis theoretical persuasion.
CHAPTER 2

Definitions of lexical classes

Lexical classes, or parts of speech, are the cornerstone of linguistic models

at a variety of levels of investigation. They are key elements in research on lexi-
cography, lexical semantics, syntactic theory, and morphological analysis, and
they traditionally play an essential role in the grammatical description of lan-
guages both familiar and exotic. At the same time, they are some of the least
clearly defined and least understood concepts in linguistics. While most people
have an intuitive sense of what constitutes a noun, a verb, or an adjective, based
largely on the characterizations of these classes in familiar Indo-European lan-
guages, to date no one has been able to satisfactorily define these classes in a
rigorous and criterial manner. To be truly useful and appropriate tools for lin-
guistic inquiry, definitions of lexical classes must necessarily perform two tasks.
They must accurately and unequivocally spell out what it means for a word to
belong to a particular lexical class and predict the properties (at whatever level
of description one chooses to formulate this definition) that all members of a
given lexical class will have. Such definitions must be universal in scope (that
is, they must apply to all words assigned to a given class in all languages), but
they must also be able to deal with the typological variation attested in the parts
of speech systems of the world's languages.
The definitions of lexical classes that have been in wide use to date have
tended to founder on both points. On the one hand, they have succeeded in de-
fining the core or focal areas of the classes, but have been notoriously unable to
deal with exceptional cases, leaving large numbers of words classified as be-
longing to a particular class but possessing none of the properties proposed as
criterial for that class. Typologically, on the other hand, such definitions are
either unable to deal with the differences languages show in the classification of
particular meanings-that is, why a word that is an adjective in language A is a
verb in language B-or they have given highly undesirable results which con-
flict with other generalizations drawn about a given class's behaviour or proper-

11
12 The typology of parts of speech systems

ties. This is clearly an untenable situation. Given that so much in linguistic the-
ory depends on lexical classes, it seems wise for us to take a look at some of
these earlier definitions to see why they went wrong and, at the same time, draw
on the insights contained in them in order to set out a newer, more rigorous set
of definitions. Traditionally, definitions of lexical classes can be divided into
three types-semantic, morphological, and syntactic, each of which will be dis-
cussed in turn in the following sections, after which I will turn to some more
recent developments in the syntactic and semantic characterizations of parts of
speech.

2.1 Semantic characterizations

The most familiar and intuitively appealing of the three traditional ap-
proaches to defining parts of speech is the semantic characterization, which
groups words into the three major classes-nouns, verbs, and adjectives-based
on their denotational or "contentive" meaning. Generally, in such an approach
nouns are said to be those lexical items denoting "people, places, and things,"
verbs are those which denote "actions and states," and adjectives are those
which denote "properties and qualities." This last group of meanings is also fre-
quently referred to as "property concepts" (Thompson 1988), defined as words
falling into one of the following seven categories identified by Dixon (1982):

(I) Classes of property concepts (with English examples)

DIMENSION - big. little. long. wide ...
PHYSICAL PROPERTIES - hard. heavy. smooth ...
COLOUR
HUMAN PROPENSITY - jealous. happy. clever. generous. proud...
AGE - new. young. old...
VALUE - good, bad, pure, delicious ...
SPEED - fast, slow. quick ...
(Thompson 1988: 168)

Of the three semantic characterizations of parts of speech, property concepts are

the most problematic: semantically, nouns and verbs are highly consistent across
languages (although, as seen in Section 2.5, there is some cross-linguistic varia-
tion even here). However, adjectives-or, more accurately, the expressions of
property concepts-show a great deal of intra-linguistic and cross-linguistic
variation, making a purely semantic definition very problematic.
On the cross-linguistic front, simple semantic definitions fail in that it is not
always possible to predict the lexical classification of a word in a given lan-
Definitions of lexical classes 13

guage from its meaning. For instance, in Hausa, the DIMENSION word 'wide' is
expressed as a noun f adi, though it is clearly a property concept, whereas in
Bemba the HUMAN PROPENSITIES 'strong', 'brave', and 'wise' are expressed by
the verbs ashipa, akosa, and aceenjela respectively (Schachter 1985). The
PHYSICAL PROPERTY 'hard' surfaces as a noun, taun:, in Hausa, as an adjective,
duro, in Spanish, and as a verb, xtadis, in Lushootseed, despite the fact that
'hard' is a property-concept and, hence, by a na'ive semantic definition, should
always be an adjective.
This problem is particularly obvious in languages that have a small, closed
class of adjectives and divide the remaining property concepts between nouns
and verbs. Such languages may have as few as half a dozen true adjectives and,
while all of these are typically expressions of property concepts, the remaining
members of this semantic class are not realized as adjectives. In the Bantu lan-
guage Venda, for instance, there are only twenty adjectives, listed in (2):

(2) Venda

hulu 'big' swa 'young, new' rema 'black'

!uku 'small' lala 'old' tshena 'white'
vhi 'bad' tete 'soft' tswu 'black'
lapfu 'long' khwivhilu 'red' hulwane 'important'
denya'thick' sekene 'thin' QU 'wet'
vhisi 'raw, green' pfufhi 'short' se!ha 'yellow'
vhuya 'good-natured' tswuku 'red'

(Dixon 1982: 4-5)

Dixon (1982) observes that in this type of language, the meanings of the reduced
class of adjectives seem to cluster consistently around notions relating to his
DIMENSION, AGE, VALUE, and COLOUR (DAVC) categories. Of the remaining
property concepts, PHYSICAL PROPERTIES in reduced-class languages tend to be
expressed by verbs and HUMAN PROPENSITIES tend to be expressed by nouns.
As useful as these observations are, they fall short of a criterial semantic
definition. In spite of the fact that adjectives in closed-class languages do tend to
express DA VC meanings, the number of adjectives in such languages can range
from dozens down to a mere handful. For example, Venda chooses-out of the
potentially much larger set of DAVC meanings-only twenty to be expressed as
adjectives. Igbo, on the other hand, realizes only a subset of seven of these
meanings (plus 'good', which is not an adjective in Venda) as adjectives:
14 The typology of parts of speech systems

(3) Igbo

ukwu 'large' Qh~'n} 'new' oj!'f 'black' Qma 'good'

rita 'small' ocye 'old' Qca. 'white' QjQ'o 'bad'

(Dixon 1982: 4; Schachter 1985: 15)

The remainder of the set of DAve meanings in (2) that are not expressed as ad-
jectives in (3) are thus presumably divided up in the Igbo lexicon between the
lexical classes of verb and noun. It seems impossible for any purely semantic
definition of adjective to be able to single out only the seven words shown in (3)
for Igbo and not include all twenty Venda words in (2), let alone for it to include
'good' in one of these languages and exclude it in the other. While it is true that
semantic characterizations of parts of speech help to identify likely candidates
for inclusion in particular lexical classes, work such as Dixon's shows that se-
mantics is not the whole story, particularly in the realm of adjectives. Semantic
characterizations of nouns and verbs seem to be reasonably, although not en-
tirely, accurate across languages, but even here they fail to achieve the rigour
required of a linguistic definition. As a result, may investigators have turned
away from semantic definitions of parts of speech altogether, while others have
tried to modify this approach by treating semantic domains as prototypes for
lexical classes, an idea to be examined in more detail in Section 2.5 below.

2.2 Morphological diagnostics

The second type of definition that has enjoyed wide currency is an essen-
tially morphological one that seeks to define parts of speech in terms of the
grammatical categories for which they are marked. The simplest approach along
these lines posits certain basic morphological categories that are purported to be
diagnostic of particular parts of speech, both within and across languages. Thus,
nouns are defined as those lexical items that have grammatical gender (Sp. perro
'dog MAsc ' : casa 'house FEM ') and are inflected for number (Eng. dog: dogs) and
case (Rus. masina 'carNOM ' : masinu 'car ACC '); verbs are inflected for tense (Eng.
he runs; he ran), aspect (Rus. kricat' 'shout'MPe' : kriknut' 'shoutPERe'), voice
(Bella Coola txis 'he cut it' : txim 'it was cut PAS; : txa 'he cut A;), and mood (Sp.
dices 'you speak' : diga 'speak!'); and adjectives are inflected for comparison
(Eng. big: bigger: biggest) and in many languages they show agreement for
number, gender, and/or case (Rus. novY}MASC:SG:NOM muze}MASC:SG:NOM 'new museum';
novyeMASC:PL:NOM muzeiMASC:PL:NOM 'new museums'; nova}aFEM:SG:NOM knigaFEM:SG:NoM
'new book'; v novo}FEM:SG:PREP knigeFEM:SG:PREP 'in the new book').
Definitions of lexical classes 15

Cross-linguistically, however, there is considerable variation with respect to

the inflectional categories encoded on lexical items belonging to all three
classes. For instance, nouns in Totonac and most Salish languages are not in-
flected for number, and recent research has suggested that, cross-linguistically,
plurality is a category potentially applicable to both nouns and verbs (Dolinina
& Beck 1998). Interior Salish languages and Totonac both lack nominal gender,
while the coastal Salish languages have gender but generally lack case, as do
Mandarin and (outside the pronominal paradigms) English and Spanish. Verbs
in Salish, Mandarin, and many other languages do not inflect for tense, and
verbs in Hebrew in all tenses are inflected for gender, as they are in the Russian
past tense. In Lushootseed, meanings corresponding to Indo-European tenses,
aspects, and moods can be applied to nouns, as in:

Lushootseed
(4) (a) tu+q'iya~.'~d ti tu+sc'istxw+s
PST+slug D PST+husband+3PO
'Slug had been her former (i.e. deceased) husband'
(Hess 1993: 84)

(b) xWi? kWi gW~+pispis

NEG D SB1+cat
'there are no cats'
(Hess 1993: 123)

(c) ~'u+l~t'+~d ti ~'u+t'is~d ?~ ti s~bad

HAB+flip+ICS D HAB+arrow PR D enemy
'he would flip the habitual arrows of the enemy away'
(Hess 1993: 83)

In (4a) the meaning 'past'-a tense in many languages-is applied to two dif-
ferent nouns, the predicate nominal q'iyaJ.:.,d 'slug' and the syntactic subject
sc'istx 'husband', whereas in (4b) the subjunctive, generally classified as a
W

mood, is applied to pispis 'cat' to indicate its non-existence (cf the use of the
subjunctive in negated subordinate clauses in Spanish: no conozco a nadie quien
SJ:.I2Q usarla 'I don't know anyone who knows sB ) how to use it'). (4c) illustrates
the application of what is usually glossed as the marker of habitual aspect to a
noun, t'is.,d 'arrow', the same marker that appears affixed to the verb I.,t' 'flip'.
Although there are no cross-linguistically universal morphological catego-
ries that can be used to define parts of speech, a more promising approach is to
define a set of grammatical categories that are cross-linguistically typical of one
or the other parts of speech and then to decide, on a language by language basis,
16 The typology of parts of speech systems

which of these is diagnostic of lexical classes in a given grammatical system.

While there is some difficulty with languages like Mandarin and Vietnamese
which have little or no morphology, definitions along these lines generally sin-
gle out the same core classes of meanings as do semantic and syntactic defini-
tions. This, however, is in itself an indication of the fundamental shortcoming of
a purely morphological definition of parts of speech: such definitions offer no
account of their own success. This success can, in fact, be attributed to a tacit
reliance on semantic and syntactic assumptions about the meanings and distri-
butions of parts of speech. Tense in Salish, for instance, might be dismissed as a
diagnostic for verbhood given its appearance on the nouns q 'iya~,:;)d 'slug' and
sC'istxW'husband' in (4a), but this presupposes the semantically-driven assump-
tion that these two words are, in fact, nouns. 2 In less problematic cases, say,
gender-marking of both verbs and adjectives in Russian, it may be possible to
devise more rigorous morphological criteria-i.e., there is a set of words which
always bear gender-marking and are marked for the case of their syntactic heads
(adjectives), while there is another set which bears marking for tense and aspect
(verbs) but can only be marked for gender in the past tense. However, even if
such morphological definitions can be crafted on a language specific-basis, on
the cross-linguistic front they do nothing to explain why it is that the bulk of
those words singled out as adjectives in Russian morphology express the same
meanings and have virtually the same syntactic distribution as those words sin-
gled out by the language-specific tests for adjectives in Hebrew, English, Toto-
nac, and Japanese.
More telling against a purely morphological definition of lexical classes is
the fact that, even intralinguistically, lexical-class boundaries drawn on purely
inflectional bases often give problematic results. Most languages, for instance,
have lexical items considered to be a member of a given class which do not have
all of the inflections that might be considered criterial for membership in that
class. Thus, the English word significance-which patterns syntactically and
semantically with nouns and does not share any inflectional categories with
verbs or adjectives-can be neither a plural (*significances) nor a possessor
(* significance's), whereas plurality and possessive inflections are commonly
cited as morphological indicators of nounhood in English (Lyons 1977: 426). In
Russian, a number of words such as piroznoe 'pastry' are declined as if they
were adjectives showing agreement with a neuter noun and thus pattern morpho-
logically with more run of the mill adjectives such as bol '.fo} 'big' and xorosij
'good', as shown in (5):

2 Alternatively, it could be argued that tu- is not a past tense marker because it appears
on both verbs and nouns-but again, this presupposes that words like q'iyaJ.:~ 'slug' and
sC'istxW'husband' are nouns and so preclude the application of a true tense marker.
Definitions of lexical classes 17

(5) Declension of Russian piroinoe 'pastry'

, ,
piroznoe 'pastryNEu bonoe 'big NEu xorosee 'goOd NEU '
NOM pirozn+oe bon+oe xoros+ee
ACC pirozn+oe bon+oe xoros+ee
GEN pirozn+ogo bon+ogo xoros+ego
DAT pirozn+omu bon+omu xoros+emu
INST pirozn+ym bon+im xoros+im
PREP pirozn+om boJ's+om xoros+em

Semantically, however, piroinoe is more closely related to morphological nouns

such as pirog 'pie' and tort 'cake'. Syntactically, piroinoe patterns as a noun as
well, undergoing quantification by numerals (pjat' piroinyx 'five pastries'),
serving as subject or object (Oni s"eli piroinye 'they ate the pastries'), the com-
plement of a preposition (Ja ne mogu iit' bez piroinyx 'I can't live without pas-
tries'), or the head of a relative construction (piroinoe, kotoroe ja kupi/ 'the
pastry that 1 bought'). What's more, unlike true adjectives but like nouns, pi-
roinoe can not serve as a modifier (*piroinoe testa 'pastry dough'), nor does it
have comparative (*piroinee) or superlative (*samoe piroinoe) forms. Given
that, first and foremost, lexical classes serve as input to syntactic rules-that is,
they characterize lexical items for the purpose of the rules used in the organiza-
tion of syntactic structures-classifying piroinoe as an adjective is at best in-
convenient, insofar as is it recognized by the syntax as a noun, as shown by its
functions and distribution.
Thus, while morphology often supplies important clues as to lexical class
membership, morphological definitions-like semantic characterizations-tend
to falter when confronted with cross-linguistic variation in lexical class member-
ship and intra-linguistic idiosyncrasies. As will become clearer in the course of
the discussion below, morphological properties, particularly inflectional catego-
ries, may reflect a word's underlying semantic and syntactic properties and so,
indirectly, may be an indicator of that word's lexical class. Given the existence
of a class of nouns in a language, for instance, it may turn out that all nouns in
that language must be inflected for singular or plural number, and so inflectional
marking for number can be used by the linguist as a indicator that a word may
belong to the class of nouns. This type of indicator, however, is not a definition
but a diagnostic, and is purely language-specific: as seen above, number is not a
universal inflectional category for nouns and in some languages it is an inflec-
tional category for other parts of speech as well. Indeed, the fact that number is
frequently marked on nouns follows from the semantics of the prototypical
noun-i.e. that nouns are prototypically discrete, countable objects-while the
fact that number may also be an inflectional category of verbs and adjectives
18 The typology of parts of speech systems

follows from their syntactics (noun-verb or noun-adjective agreement). Mor-

phology in this sense becomes a superficial phenomenon, depending not so
much on the universal characteristics of lexical classes as on how these classes
are treated by the morphosyntax of a given language.

2.3 Syntactic distribution

As noted in the previous section for words like Eng. significance and Rus.
piroznoe 'pastry', syntactic distribution is often more closely related to lexical
class membership than are inflectional patterns; this observation has led to a
number of attempts at defining parts of speech (often called "syntactic catego-
ries" in such definitions) in purely distributional terms. The most elementary of
these approaches defines each part of speech strictly on the basis of the syntactic
roles in which it is permitted to appear: nouns are defined as lexical items that
can be the subjects of a sentence, verbs can be syntactic predicates, and adjec-
tives are attributive modifiers (cf Chomsky 1965). Such definitions falter, how-
ever, when confronted with lexical items appearing in their non-prototypical or
extended uses. English nouns, for instance, can comfortably serve as attributives
of other nouns, as in jazz musician or gas stove, while certain adjectives can act
as syntactic subjects (e.g. The rich fear the poor; Louder is better).
By the same token, syntactic roles singled out as definitive of lexical classes
can be filled by complex, multi-word expressions such as non-finite VPs and
subordinate clauses. Thus, in the cat yowling in the backyard, the element which
fits the definition of adjective given above, "attributive modifier," is not a lexi-
cal adjective but a participial phrase, yowling in the backyard, and in the sen-
tence That she found him so quickly was a great surprise, the syntactic subject is
a finite clause, that she found him so quickly, rather than a noun. Even in simpler
sentences like The red squirrel sits in the park, as Lyons (1977: 429) observes,
"it is not nouns, but nominals, that function as subjects ... " and, by extension, it
is not necessarily adjectives, but members of the "expression class" adjectival,
which act as modifiers. On a micro-level this problem could be overcome by
simply adding the proviso "lexical item" to the syntactic definitions of parts of
speech given above (i.e. "an X is a lexical item whose syntactic distribution is
Y"), but solutions of this type gloss over the larger question of the relation be-
tween lexical classes and the corresponding expression class-namely, what is it
about nouns and nominals that accounts for their parallel distribution, and how
best to capture the semantic relationship between simplex (lexical) and complex
elements that fill the same syntactic role? From a cross-linguistic perspective,
these questions seem even more pressing when it becomes evident that the dis-
tributional parallels between simplex and complex elements found in languages
Definitions of lexical classes 19

like English (i.e. nouns have similar distributional patterns to finite complement
clauses, adjectives pattern with participles and relative clauses, etc.) are found in
a wide range of the world's languages.
An additional cross-linguistic difficulty with distributional definitions of
parts of speech comes from languages with reduced lexical inventories-that is,
languages which appear, on a distributional basis, to lack one or more lexical
class distinctions. Consider, for instance, the examples in (6) from the Salishan
language Lushootseed, which show the distributional overlap between verbs and
nouns «6a) and (b», and between verbs and adjectives «6c) and (d»:

Lushootseed
(6) (a) ?U+7~i~d ti7H piSpis 7~ ti7~7 s7uladx w
PNT+eat D cat PR D salmon
'that cat ate a salmon'

(b) pispis ti7H ?U+7~i~d 7~ ti7~7 s7uladx w

cat D PNT+eat PR D salmon
'that one eating the salmon [is] a cat'
(based on Hess 1993: 133)

(c) A,'u+l~k'w+~d ti7~7 ha7i s7~i~d

HAB+eat+ICS D good food
'[he/she] would eat the good food'

(d) han ti7~7 s7~i~d A,'u+l~k'w+~d

good D food HAB+eat+ICS
'the food [he/she] would eat [is/was] good'
(based on Bates et al. 1994: 105)

The example in (6a) illustrates a sentence whose predicate is the verb ?glgd
'eatINTR' and whose syntactic subject is tinl pispis 'that cat'; in (6b) the same
word, pispis 'cat', serves as the syntactic predicate. In (6d) the syntactic predi-
cate is the word ha?l 'good', which is shown acting as a modifier in (6c). A na-
rve distributional definition of a verb as "a lexical item that can act as a syntactic
predicate" would not only pick out the syntactic predicates of (6a)-?g1gd 'to
eatlNTR ' -and (6c)-lgk"gd 'eatrRANS' -as verbs, but would pick out the syntactic
predicates of (6b)-pispis 'cat'-and (6d)-ha?1 'good'-as well. This type of
argument can and has been used as evidence that Lushootseed, and Salish in
general, lacks an underlying lexical distinction between verbs, nouns, and ad-
jectives (e.g. Kuipers 1968; Kinkade 1983). As discussed in detail in Section
4.1.1 below, however, in the case of the noun-verb distinction this type of
20 The typology of parts of speech systems

methodology gives the wrong results. And, if Salish does indeed differentiate
between nouns and verbs, then clearly something other than straightforward
distribution has to be used in the definition of lexical classes.
Just as with semantics and morphology, then, syntactic distribution in and
of itself fails as an adequate means of defining parts of speech. Intralingui-
stically, words of a given lexical class are frequently capable of appearing in
syntactic roles which are typical of, or even diagnostic of, other parts of speech;
cross-linguistically, there is frequently variation with respect to the syntactic
roles open to different parts of speech, as illustrated by the Lushootseed exam-
ples in (6). Nevertheless, it is true that there are certain widespread commonal-
ties in the syntactic behaviour of lexical classes, just as there are prototypical
semantic domains and inflectional categories associated with them. This type of
observation has lead some researchers to treat syntactic variation in the distribu-
tion of parts of speech in terms of the markedness of a given syntactic role for
members of a particular lexical class: in essence, such approaches-like the na-
'ive syntactic definitions examined so far-identify certain syntactic roles as
being typical or unmarked for a given part of speech and then allow, in one way
or another, for the appearance of that part of speech in other, marked, roles in
the sentence. Thus, the appearance of the normally adjectival red as a syntactic
subject in Red is my favorite colour could be argued to be an example of a
marked or extended use of red in a basically nominal syntactic role. However, to
do this without recourse to purely stipulative definitions of parts of speech it is
necessary to show that the behaviour of an element in an extended position is in
some way marked (and that the criteria for markedness are not in some way
post-hoc or stipulative). These are complex issues to be taken up in more detail
in the section that follows.

2.4 Extended roles and syntactic markedness

As show above, neither semantic characterizations (Section 2.1) nor syntac-
tic distribution (Section 2.3) are sufficient in and of themselves to allow for a
criterial definition of lexical classes. Morphological properties of words tum out
to be useful as diagnostics for lexical-class membership on a language-specific
basis, but fail both as universal and intralinguistically comprehensive definitions
(Section 2.2). Of the three levels of linguistic description, it is the morphological
which shows the greatest cross-linguistic variation in terms of its correlation
with lexical class distinctions and so ultimately seems to be the least useful in
terms of finding a working definition. This leaves the semantic and the syntactic
levels, both of which have been the focus of more recent attempts to define parts
of speech. Different authors have dealt with the variation described in the sec-
Definitions of lexical classes 21

tions above in different ways, but two main types of definition have become
predominant in the literature. Semantic approaches have by and large moved in
the direction of treating variation in lexical classification in terms of prototypi-
cality and graded class membership. This will be discussed in detail in Section
2.5. As noted above, more recent syntactic approaches have dealt with distribu-
tional variability in terms of syntactic markedness-that is, they have sought to
define parts of speech in terms of their unmarked syntactic distribution. This
will be the focus of the remainder of Section 2.4, which begins with a discussion
of what markedness means and how it will be measured throughout the course
of this discussion (Section 2.4.1). Section 2.4.2 will introduce and slightly rede-
fine a term, WFM ("without further measures"), first proposed by Hengeveld
(1992a, 1992b). Following this I discuss another aspect of Hengeveld's work
that will playa major role in subsequent chapters, the distinction between rigid
and flexible languages (Section 2.4.3), after which I outline two important diag-
nostics of syntactic markedness, recategorization and decategorization (Section
2.4.4). Finally, in Section 2.4.5 I introduce the notion of markedness as the
measure of the proto typicality of certain types of mapping between semantic
class and "pragmatic" role as put forward by Croft (1991). While Croft's pro-
posal has some weaknesses, it does point in the direction of what seems to be the
correct approach to forming definitions of lexical classes that both account for
similarities in parts of speech systems and predict the attested variation in these
systems in the world's languages.

2.4.1 Criteria for markedness

Before undertaking a review of syntactic definitions of parts of speech

based on markedness, it is worth taking a little time to clarify what it is precisely
that is meant by markedness and what kinds of criteria will (and will not) be
allowed in the remainder of this discussion. Markedness is one of the most
widely, and wildly, used terms in linguistics, and its senses range from a very
narrow, structure-based notion of relative complexity to an extremely open
sense of "unusual" or "unnatural." A recent definition of markedness located
somewhere in the middle of the continuum is put forward by Giv6n (1995), who
writes that

three main criteria can be used to distinguish the marked from the unmarked
category in a binary grammatical contrast:

(a) Structural complexity: The marked structure tends to be more complex

(or larger) than the corresponding unmarked one.
22 The typology a/parts a/speech systems

(b) Frequency distribution: The marked category (figure) tends to be less

frequent, thus cognitively more salient, than the corresponding unmarked
category (ground).
(c) Cognitive complexity: The marked category tends to be cognitively more
complex-in terms of mental effort, attention demands or processing
time-than the unmarked one. (Givan 1995: 28)

Of these three criteria, (a) is the least controversial and the most universally ac-
cepted: given the contrast between two (comparable) elements A and B, the
more complex of the two is the marked one. The second and third items on
Giv6n's list, however, are much less straightforward. Frequency is a very com-
monly cited criterion for markedness, due largely to the intuitive feeling that the
unmarked is the most usual or standard form. While this may often be the case,
it is not always so, and Trubetskoy (1969: 262ff) argues explicitly against fre-
quency as a reliable indicator of markedness, offering a number of examples of
phonological segments which are marked (in terms of their complexity, etc.) but
are statistically more frequent than their unmarked counterparts. The unreliabil-
ity of frequency as a measure of markedness also becomes obvious if examined
in concrete terms. In phonology, for instance, the appearance of a marked pho-
neme in a high-frequency word (say, a function word, a common morpheme, or
a usual expression) could potentially make the instances of that phoneme more
frequent than those of its unmarked counterpart. In the domain of lexical classes,
it turns out that in English the predicative use of adjectives is textually more
frequent than the attributive use (Thompson 1984)-yet clearly, judged in terms
of structural complexity (adjectival predicates require a copula), the former is
the more marked of the two constructions. Thus, while frequency in a textual
sense may tend to correlate with markedness, it is neither a necessary nor a suf-
ficient criterion for markedness and so will not be used in the course of this dis-
cussion.
There is, however, a type of markedness that will be used here that is, at
least intuitively, related to the notion of frequency. Consider the following
situation: in a particular language, words belonging to the lexical class X appear
in six structural environments {EI' E2, ... Ed. In three of the six environments,
X displays a set of properties {PI' P 2, ... P7 } (e.g. inflectional categories, refer-
ential meaning, etc.), but in E3 and E5 X displays a reduced set of these proper-
ties {PI' P 3, P7 } and in E6 it has only one of these {P2 }. Environments~, E 5, and
E6, then, can be considered marked structural configurations for X with respect
to the remainder of environments {E I, E 2 , E 4 } in which X displays the greatest
range and most consistent set of properties. The markedness of a given envi-
ronment, then, can be determined by a reduction in number of typical properties
of X, which are those which X displays in the largest number of environments.
Definitions of lexical classes 23

This is referred to as decategorization. Markedness can also result from the ac-
quisition of a new property, Pg , in one or more of the environments open to X,
provided that either a) the number of environments in which X has the set of
properties {PI' P2, ... P7 } is greater than the number of environments where X
has the set of properties {PI' P2 , ... Pg} or b) X has, in addition to Ps, only a re-
stricted subset of the other properties {PI' P2, .. , P7 } in the marked environment.
This is frequently referred to in the literature as recategorization. 3 This measure
of markedness seems like a kind of frequency in that it is determined based on
the "frequency" with which a certain set of properties is associated with the
members of the set of environments {EI> E2 , ... E6 } open to X. Because of the
dangers inherent in the term "frequency," however, it is safer to refer to this type
of markedness as contextual markedness, a term which has the added advantage
of reminding us (as noted by Givan above) that the markedness of a given item
must not only be determined relative to some other item of a comparable nature,
but also must be determined for a specific context. Both types of contextual
markedness will be discussed in more concrete terms under the headings of de-
and recategorization in Section 2.4.4.
Givan's third criterion, cognitive complexity, is also somewhat problematic,
although if used judiciously it turns out to be a useful one. Terms such as "men-
tal effort, attention demands or processing time" are frequently used in a hand-
waving fashion without regard to the fact that-as real-world, neurological
events-they are subject to empirical verification. Failing psycholinguistic
measurement of complexity based on the criteria proposed by Givan, then, it is
important to be very clear what is meant by "cognitive complexity" and to pro-
vide plausible reasons to think that this complexity would indeed correspond to
increased effort, attention, or processing time. To this end, I wish to propose
one, specific type of cognitive complexity that will playa role in the discussion
below, something that I will refer to as non-iconicity. According to this criterion,
a linguistic sign a <'a', A> (that is, the sign a having the signified 'a' and the
signifier A) is more marked than a sign b <'b', B> if A is a less direct reflection
of 'a' than B is of 'b'. On its own, of course (like all uses of the term "iconic-
ity") this can be very open-ended. In the context of this discussion, it will be put
to a single, highly constrained and specific use in Section 3.3 below (to which
the reader is referred for a specific example). The rationale for this criterion is
simply that a non-iconic sign will be harder to process than an iconic sign, and

3 On its own, recategorization is trickier to establish than decategorization, particularly

for lexical items that have a very limited number of syntactic roles. Generally, it is only
invoked in cases where P 8 is felt to be marked in its application to X for other reasons
(e.g., that P8 is typical of another lexical class, Y, or the construction is marked in terms
of complexity with respect to some other environment in which X appears).
24 The typology of parts of speech systems

hence is cognitively more complex. Straying from the field of linguistics, an

illustration of this might be the mental effort required to recognize a picture of a
familiar object (a direct representation matched to visual information) versus
recognizing it from a description (which requires lexical access and linguistic
processing). In terms of writing systems, an ideographic system is more difficult
to learn in that the representations of words contain no information about their
phonological shape, whereas an alphabetic system allows learners to match
written representations to spoken words. This last example probably gives as
good a formulation of the notion of "direct representation" as we are going to
get: the more direct a representation is (that is, the more iconic it is) the more
information it contains about the underlying meaning it represents. Thus, if the
signifier B contains more information about 'b' than A does about 'a', b can be
said to be less marked (and more iconic) than a.
The criteria for markedness that will be used here, then, differ somewhat
from those put forward by Givan (although, unlike Givan, I have in no way tried
to be comprehensive in my formulations). This dissertation makes use (implic-
itly and explicitly) of three criteria for syntactic markedness:

(7) (a) Structural complexity: An element X is marked with respect to

another element Y if X is more complex, morphologically or syn-
tactically, than Y
(b) Contextual markedness: An environment E is a marked one for an
element X if E is not a member of the largest subset of environments
of X where X displays the greatest number of common properties
(c) Cognitive complexity: An element X is marked with respect to
another element Y if the representation of X is a less direct expres-
sion of X' s meaning than the representation of Y is of Y' s meaning

An important point to be made about all of these criteria is that they are formu-
lated in terms of contrast-that is, it is not enough to say that X is marked, it is
necessary to specify what it is that X is marked in contrast to. Thus, it is essen-
tial to keep in mind that markedness is always contrastive, an issue which is
central to the discussion in the section that follows.

2.4.2 WFM and markedness

Definitions of parts of speech in terms of unmarked syntactic roles typically

start from a position similar to the naive syntactic definitions of lexical classes
examined in Section 2.3. The unmarked syntactic role of nouns is thus claimed
to be that of syntactic actant (subject or object) of a verb, verbs are said to be
Definitions of lexical classes 25

unmarked syntactic predicates, and adjectives are unmarked modifiers of

nouns. 4 All of these lexical classes, however, are said to have additional-
marked or "extended" (Dik 1978)-uses which overlap with the unmarked dis-
tribution of the other classes. The task of the linguist then becomes sorting out-
in an unambiguous, non-tautological manner-which uses of a given lexical
item constitute extended uses and which are unmarked, and, hence indicative of
that item's lexical class membership. In one of the best examples of this meth-
odology to date, Hengeveld (l992a, 1992b) makes reference to the "additional"
grammatical machinery required to allow a lexical item to appear in an extended
syntactic role. Hengeveld refers to such mechanisms as "further measures" and
uses this notion to arrive at the definitions of the major parts of speech in (8):5

(8) verb-a lexical item which, without further measures being taken (WFM),
has predicative use only
noun-a lexical item which WFM can be used as the actant of a syntactic
predicate
adjective-a lexical item which WFM can be used as the modifier of a
noun

For Hengeveld, "further measures" are defined as those morphosyntactic means

which "derive" Functional Grammar predicates from constituents that are not
already predicates (1 992a: 58). Hengeveld (l992a: 58) illustrates this with the
folIo wing set of English attributive constructions:

Enelish
(9) (a) the intelligent detective
(b) the singing detective
(c) the detective who is singing

AIl of the italicized words in (9) are, in syntactic terms, modifiers of detec-
tive-however, only the first one, intelligent, is used "without further measures
being taken" and so fits into the category of adjective. The modifier in (9b) is a
verb, sing, suffixed with the participial/gerundive suffix -ing (a morphological

4 In Meaning-Text Theory CZiJlkovskij & Mel'cuk 1967; Mel'Cuk 1988) and various
other dependency-based grammars, an actant is the equivalent of a syntactic argument in
generative phrase-structure grammars. I will continue to use this term throughout in order
to avoid confusion with the term "argument," which I will restrict to the semantic sphere.
5 Note that I have re-formulated Hengeveld's (1992a: 58) definitions-which in the
original are couched in the terms of Functional Grammar (Dik 1978)-to make them
more accessible to those unfamiliar with the framework.
26 The typology of parts of speech systems

measure), whereas in (9c) the modifier is the same verb contained within a rela-
tive clause (a syntactic measure). These examples show that verbs in English
can be modifiers, but not without further measures being taken. Also included
under the rubric of further measures would be derivation-the hairy detec-
tive-and the use of syntactic elements such as complementizers, particles, etc.,
which allow lexical items to appear in contextually marked syntactic roles.
In the examples in (l0), the same procedure helps to establish the fact that
in the Salishan language Lushootseed there is a conflated class of adjec-
tive/verbs, which are both WFM modifiers and syntactic predicates:

Lushootseed
(10) (a) k'wi+ax· ti?~? q·u? dx·ca?k·
trickle+now D water seaward
'this water trickled down to the sea'
(Bierwert 1996: 77, line 86)

(b) ti?~? ha?i ?u+k'·ik'·~i q·u?

D good PNT+(RDP)trickle water
'this nice trickling water'
(Hess 1993: 117)

(c) b~q· stubs

fat man
'fat man'

(d) h~la?b+~x· c~d b~qW

really+now Iso fat
'I [am] really fat'
(Bates et al. 1994: 38)

Sentence (lOa) shows the word k"'i 'trickle' acting WFM as a syntactic predicate
(the morpheme _ax· is a clitic and tends to appear associated with the first word
in a sentence), while in (lOb) the same word appears as a modifier of the noun
q wu ? 'water'. The use of k"'i 'trickle' as a modifier requires no further measures,
and the fact that it appears in (lOb) reduplicated for locative distributivity and
marked for punctual aspect indicates that it has not lost the inflectional possi-
bilities open to verbs in predicate position (a sign of decategorization and,
hence, of markedness in a particular syntactic role-see Section 2.4.4). (lOc)
shows the property-concept word bgqW 'fat' acting as a modifier, while in (lOd)
it is shown acting as a syntactic predicate. In neither role does bgqW 'fat' require
further morphosyntactic measures and, as discussed in more detail below (Sec-
Definitions of lexical classes 27

tion 4.1.1), in neither role are there further morphological, syntactic, or semantic
measures invoked that would allow for one use of b~qW to be declared to be
marked with respect to the other. Thus, there seems not to be a distinction in
Lushootseed between adjectives and verbs, and words that fall into either of
these classes in English may be either actant or modifier without further meas-
ures being taken.
The utility of Hengeveld's approach to lexical classes, then, is that it allows
for the extended use of lexical items without creating exceptions to putatively
universalist definitions based on syntactic distribution. Unfortunately, although
Hengeveld's (l992a, 1992b) formula "without further measures" appears to be a
useful and accessible one for expressing a measure of morphosyntactic
(un)markedness in terms of structural complexity, WFM appears to miss a crucial
aspect of the notion of markedness. 6 For Hengeveld, WFM seems to be simply a
measure of the "amount" of morphosyntactic machinery implemented to allow a
lexical item to appear in a given syntactic role, the "unmarked" case being
where no additional measures are needed (this being the criterion for deciding
that a given syntactic role is diagnostic of lexical class). For example, English
allows big to modify a noun WFM, as in big boy, but does not allow run to mod-
ify boy without the implementation of a morphological measure, as in running
boy: thus, big is an adjective (a WFM syntactic modifier) and run is not. The dif-
ficulty is that if "further measures" to include all morphosyntactic markers that
appear with lexical items in a given syntactic role, Hengeveld's definitions break
down in what should be fairly straightforward and obvious instances. Consider
the case of a language like Russian, where actants of verbs bear morphological
case-markers and are ungrammatical without them:

Russian
(11) (a) Citaj+u knig+u
read:IMPF+ ISG:PRS book+FEM:SG:ACC
'I read the book'

(b) *citaj+u knig

read:IMPF+1SG:PRS book

6 In all fairness to Hengeveld and his insightful treatment of lexical classes, it is not clear
to me that the formula "without further measures being taken" is intended as a measure of
markedness per se, if it is simply meant as a principle in its own right, or if the objections
that I am about to raise are precluded by some aspect of the Functional Grammar formal-
ism that I am not conversant with. Whatever Hengeveld's intentions, given the role that
markedness will play in the discussion below, it is nonetheless important to sort out in
what respects WFM does, and does not, correspond to it.
28 The typology of parts of speech systems

(c) knig+a xoros+aja

book:FEM+SG:NOM goOd+FEM:SG:NOM
'the book is good'

(d) *knig xoros+aja

book good+FEM:SG:NOM

If a naYve definition of further measures is used (something along the lines of

"additional morphosyntactic machinery"), the case-marker in (11) would cer-
tainly seem to qualify. Since the case-marker is essential for the noun to appear
in any actantial role in a sentence, this seems to show that in Russian there is no
class of lexical item which can serve WFM as the actant of a syntactic predicate.
In other words, by the definitions in (8) above, Russian has no nouns.
Clearly, this is an undesirable result. The problem lies in the unconstrained
way in which "further measures" have been defined up to this point and the way
in which the formula WFM departs from traditional accounts of markedness such
as that in Trubetskoy (1969), which are by and large based on the notion of con-
trast. As seen in (7), markedness is always a relative term. In the case of Russian
nouns, keeping this in mind eliminates case-endings as signs of markedness in
that nouns in Russian never appear without these endings and a case-marked
noun can therefore never be contrasted with a non-case-marked form. Thus, it is
important to remember that the formula "without further measures" covers only
those measures that are contrastively marked, in this case marked in a given
syntactic role relative to some other environment in which these measures are
not taken. The suffix -ing in running alluded to earlier is a good example of this:
the form running can be compared to run, with respect to which it is marked (by
structural complexity), making -ing a contrastively marked further measure.
Linking further measures to markedness disqualifies case-markers and other
inflectional morphemes (such as zero-person morphology on predicative uses of
run) and allows the investigator to focus on those morphosyntactic devices that
are truly contrastive for different parts of speech in the same syntactic role.

2.4.3 Rigid versus flexible languages

Another very valuable observation made by Hengeveld (1992a, 1992b) is

that languages vary not only with respect to whether or not they have parts of
speech that conform to all three of the definitions in (8), they also vary in terms
of the syntactic behaviour of these parts of speech, creating a distinction be-
tween what he calls rigid and flexible languages. The Lushootseed examples in
(10) above illustrate a flexible language in that the conflated class of verbs and
Definitions of lexical classes 29

adjectives in Lushootseed can be used WFM as either syntactic predicates or as

modifiers. An often-cited case of a rigid language is Mandarin (Schachter 1985;
Hengeveld 1992a, 1992b), which conflates the classes of adjective and verb and
marks both in the role of modifier with the relati ve particle de, shown in (12): 7

Mandarin
(12) (a) neige ntihaizi piflOliang
D girl beautiful
'that girl is beautiful'

(b) piaoliang+de ntihaizi

beautiful+ATRB girl
'a beautiful girl'

(c) neige nUhaizi liaojie

D girl understand
'that girl understands'

(d) liaojie+de ntihaizi

understand+ATRB girl
'a girl who understands'
(lit. '[an] understanding girl')
(Schachter 1985: 18)

Not only verbs and words denoting property concepts take de when acting as
modifiers-nouns also take de when used in attributive constructions with other
nouns, giving a possessive reading:

Mandarin
(13) xiansheng+de shu hen duo
teacheHATRB book very many
'the teacher has many books' (lit. 'the teacher's book [are] very many')
(Modern Chinese 1971: 92)

7 According to Modern Chinese (1971: 92f f), however, Mandarin does have adjectives
-i.e. predicates which appear WFM as modifiers of nouns. This seems to be a small class
limited to frequently-used monosyllabic modifiers. Given that the focus here is the open,
productive class of adjectives these can be disregarded for the moment for heuristic pur-
poses. Note also that while the examples in (12) are from Schachter (1985), I have added
marking for tone and changed the orthography slightly to conform with the pinyin system
used in Modern Chinese. Thanks are due to Zhou Hong for help with this.
30 The typology of parts of speech systems

The same particle de is used in conjunction with personal pronouns to form pos-
sessives (wo 'I' > wode 'my') and with complex verbal expressions to form
relative-clause equivalents and modifying phrases:

Mandarin
(14) (a) ta shi yfge gongzua jijf+de tongzhi
he is CLS work enthusiastic+ATRB comrade
'he is a comrade who works enthusiastically'

(b) zhongwen bu hao+de ren bli hui yang

Chinese not goOd+ATRB person not able use

zheiben zidian
CLS dictionary
'those who are not good at Chinese can not use this dictionary'
(lit. 'not-good-at-Chinese people not able to use this dictionary')
(Modern Chinese 1971: 214)

In short, de is a syntactic marker of attribution which is used in every case where

some element of a sentence acts as the modifier or attribute of a noun, irrespec-
tive of the lexical or expression class of that element.
What is shows it that no part of speech in Mandarin is marked part of
speech when used as modifiers, although the use of all words is marked in terms
of structural complexity relative to their forms in other environments. Thus,
liaojie+de 'understand+REL' can be compared with its predicative form, liaojie
'understand', with respect to which it is structurally marked. The same is true of
the word piilOliang 'beautiful'-expressed in English as an adjective (and an
unmarked modifier)-which was also seen in (12) to be an unmarked syntactic
predicate. This allows, in accordance with the definitions in (8), both words to
be classified as verbs: they both are unmarked syntactic predicates, and both are
structurally marked (relative to their predicative forms) when used as modifiers
of nouns, disqualifying them as adjectives. The same is true for (almost) all
property concept words in Mandarin, making it a verb-adjective conflating lan-
guage. Note, however, that exactly the opposite line of reasoning leads to the
same conclusion for Lushootseed. As seen in (10) above, in this language nei-
ther verbs nor adjectives are marked as syntactic predicates, both classes thus
meeting the criteria for verbhood; because the same words are unmarked modi-
fiers, the conclusion is that Lushootseed is also a verb-adjective conflating lan-
guage on the grounds that Lushootseed makes no syntactic distinction between
the two classes (and that it does not have another class of words which are un-
marked modifiers and marked syntactic predicates). Lushootseed and Mandarin
Definitions of lexical classes 31

have the same type of parts of speech system, but differ with regards to the
treatment of the conflated verb-adjective class in modifier position in the syn-
tax. The rigid/flexible distinction, then, is really a difference in the morphosyn-
tax of the two languages and not in the organization of their lexica. This issue
will be come a little clearer as the discussion progresses and lays out in more
concrete terms what it is exactly that constitutes a parts of speech system, and
how these are generated in the lexicon. The problem will be taken up again in
the context of a concrete example in Section 4.1.2.3 below.

2.4.4 Measures of contextual markedness: De- and

recategorization

As discussed in the previous section, the formula "without further meas-

ures" contains some pitfalls, although it can be salvaged by redefinition in terms
of the contrastive markedness. For example, the use of a verb as a syntactic
subject in an English sentence like Running is good for your health is marked in
that the verb run requires the gerundive -ing in subject position. The use of a
noun, as in Food is good for your health, however, is unmarked in that food ap-
pears on its own, and subject position can be shown to be a contextually un-
marked environment for nouns. The behaviour of these two items can then be
contrasted in other syntactic roles and with other lexical items to arrive at the
following conclusions: 1) run is not WFM a syntactic actant (llike *runlrunning)
but it is WFM a syntactic predicate (l run)-therefore, by the definitions in (8),
run is a verb; 2) food is not WFM a syntactic predicate (l *0/am food), but it is
WFM an actant (/ like food)-thus, by the definitions in (8),food is a noun.
Similarly, the existence of unmarked modifiers like big (the big boy) shows that
run is not WFM a modifier of nouns (the *runlrunning boy) and therefore is not
an adjective.
Note, however, that the same is not obviously true for the noun food-that
is, given the right circumstances, the expression food boy might be considered
grammatical and so, apparently, there is no basis to claim that food is marked as
an attributive expression with respect to big. The issue of the attributive use of
nouns is an intricate one and will be dealt with at some length below-first in
semantic terms, later on in this section, and then in more formal terms in Section
3.4-and so for the moment it will have to be set aside. There are, however,
many more straightforward instances where words appear in extended syntactic
roles without overt morphosyntactic further measures being taken. A frequently
cited example (and one that will also playa major role in this discussion below)
is the appearance of adjectives as the prima facie heads of noun phrases in lan-
guages like Spanish, which commonly make use of expressions such as el rojo
32 The typology of parts of speech systems

'the red (one)'. Adjectives used this way show no overt morpho syntactic mark-
ing and seem to be WFM actants of verbs in expressions such as me gusta el raja
'I like the red one', thus apparently qualifying as nouns. However, evidence for
the markedness of adjectives used in this way can be adduced from the semantic
shift that adjectives undergo when used as actants in discourse. An expression
such as los rajas is an anaphoric reference to some object (made overt in the
English gloss 'the red ones') whose identity is known to the speaker and the
hearer, whereas the meaning of raja 'red' when used as a modifier includes only
the notion of colour, the object being specified by the modified noun. Thus, even
though there are, in morphosyntactic terms, no overt further measures used with
Spanish adjectives in actantial positions, the resulting shift in meaning shows
them to be cognitively complex and, hence, marked-such semantic shifts being
a frequent consequence of what is called the "extended use" of a lexical item.
As noted by Hopper & Thompson (1984), parts of speech have typical dis-
course functions which are commensurate with their unmarked syntactic roles:
verbs typically describe actions and narrate events, nouns introduce and name
participants in events, and adjectives attribute properties to these participants.
These discourse functions are typical of each of the lexical classes, but they are
by no means the only functions open to different parts of speech: English nouns
can, for instance, attribute properties (a pant leg) and verbs may be used refer-
entially (the best thraw of the day). When used in these atypical roles, however,
Hopper & Thompson point out that such lexical items show signs of extended
use in that they frequently lack many of the properties that they have in their
more prototypical roles-in other words, they show signs of lower categoriality
or "decategorialization" (henceforth, "decategorization"). When used attribu-
tively nouns can no longer take deixis (*a the pant leg) or inflection for number
(*a pants leg), whereas verbs used as actants do not take inflection for tense,
aspect, or mood, nor do they show agreement for person or number. A given
part of speech will show the greatest range of inflectional possibilities when it is
used in its unmarked role, whereas the same part of speech will have fewer in-
flectional possibilities when it appears in a marked or extended role.
This loss of inflectional possibilities by a word in an extended syntactic
role, as well as the concomitant loss of the semantic notions these inflections
express, are clear signs of decategorization. The change in inflectional possibili-
ties is one potential measure of what was termed in Section 2.4.1 "contextual
markedness." Words in extended roles which have been decategorized can be
considered contextually marked in comparison with their attestations in non-
extended uses. More importantly, decategorized words of one lexical class can
also be considered as marked with respect to words of another class for which
the same role is not an extended one and Which, therefore, have not undergone
 Another Random Document on
Scribd Without Any Related Topics
damaged disk or other medium, a computer virus, or computer
codes that damage or cannot be read by your equipment.

1.F.2. LIMITED WARRANTY, DISCLAIMER OF DAMAGES - Except for

the “Right of Replacement or Refund” described in paragraph 1.F.3,
the Project Gutenberg Literary Archive Foundation, the owner of the
Project Gutenberg™ trademark, and any other party distributing a
Project Gutenberg™ electronic work under this agreement, disclaim
all liability to you for damages, costs and expenses, including legal
fees. YOU AGREE THAT YOU HAVE NO REMEDIES FOR
NEGLIGENCE, STRICT LIABILITY, BREACH OF WARRANTY OR
BREACH OF CONTRACT EXCEPT THOSE PROVIDED IN PARAGRAPH
1.F.3. YOU AGREE THAT THE FOUNDATION, THE TRADEMARK
OWNER, AND ANY DISTRIBUTOR UNDER THIS AGREEMENT WILL
NOT BE LIABLE TO YOU FOR ACTUAL, DIRECT, INDIRECT,
CONSEQUENTIAL, PUNITIVE OR INCIDENTAL DAMAGES EVEN IF
YOU GIVE NOTICE OF THE POSSIBILITY OF SUCH DAMAGE.

1.F.3. LIMITED RIGHT OF REPLACEMENT OR REFUND - If you

discover a defect in this electronic work within 90 days of receiving
it, you can receive a refund of the money (if any) you paid for it by
sending a written explanation to the person you received the work
from. If you received the work on a physical medium, you must
return the medium with your written explanation. The person or
entity that provided you with the defective work may elect to provide
a replacement copy in lieu of a refund. If you received the work
electronically, the person or entity providing it to you may choose to
give you a second opportunity to receive the work electronically in
lieu of a refund. If the second copy is also defective, you may
demand a refund in writing without further opportunities to fix the
problem.

1.F.4. Except for the limited right of replacement or refund set forth
in paragraph 1.F.3, this work is provided to you ‘AS-IS’, WITH NO
OTHER WARRANTIES OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR ANY PURPOSE.

1.F.5. Some states do not allow disclaimers of certain implied

warranties or the exclusion or limitation of certain types of damages.
If any disclaimer or limitation set forth in this agreement violates the
law of the state applicable to this agreement, the agreement shall be
interpreted to make the maximum disclaimer or limitation permitted
by the applicable state law. The invalidity or unenforceability of any
provision of this agreement shall not void the remaining provisions.

1.F.6. INDEMNITY - You agree to indemnify and hold the Foundation,

the trademark owner, any agent or employee of the Foundation,
anyone providing copies of Project Gutenberg™ electronic works in
accordance with this agreement, and any volunteers associated with
the production, promotion and distribution of Project Gutenberg™
electronic works, harmless from all liability, costs and expenses,
including legal fees, that arise directly or indirectly from any of the
following which you do or cause to occur: (a) distribution of this or
any Project Gutenberg™ work, (b) alteration, modification, or
additions or deletions to any Project Gutenberg™ work, and (c) any
Defect you cause.

Section 2. Information about the Mission

of Project Gutenberg™
Project Gutenberg™ is synonymous with the free distribution of
electronic works in formats readable by the widest variety of
computers including obsolete, old, middle-aged and new computers.
It exists because of the efforts of hundreds of volunteers and
donations from people in all walks of life.

Volunteers and financial support to provide volunteers with the

assistance they need are critical to reaching Project Gutenberg™’s
goals and ensuring that the Project Gutenberg™ collection will
remain freely available for generations to come. In 2001, the Project
Gutenberg Literary Archive Foundation was created to provide a
secure and permanent future for Project Gutenberg™ and future
generations. To learn more about the Project Gutenberg Literary
Archive Foundation and how your efforts and donations can help,
see Sections 3 and 4 and the Foundation information page at
www.gutenberg.org.

Section 3. Information about the Project

Gutenberg Literary Archive Foundation
The Project Gutenberg Literary Archive Foundation is a non-profit
501(c)(3) educational corporation organized under the laws of the
state of Mississippi and granted tax exempt status by the Internal
Revenue Service. The Foundation’s EIN or federal tax identification
number is 64-6221541. Contributions to the Project Gutenberg
Literary Archive Foundation are tax deductible to the full extent
permitted by U.S. federal laws and your state’s laws.

The Foundation’s business office is located at 809 North 1500 West,

Salt Lake City, UT 84116, (801) 596-1887. Email contact links and up
to date contact information can be found at the Foundation’s website
and official page at www.gutenberg.org/contact

Section 4. Information about Donations to

the Project Gutenberg Literary Archive
Foundation
Project Gutenberg™ depends upon and cannot survive without
widespread public support and donations to carry out its mission of
increasing the number of public domain and licensed works that can
be freely distributed in machine-readable form accessible by the
widest array of equipment including outdated equipment. Many
small donations ($1 to $5,000) are particularly important to
maintaining tax exempt status with the IRS.

The Foundation is committed to complying with the laws regulating

charities and charitable donations in all 50 states of the United
States. Compliance requirements are not uniform and it takes a
considerable effort, much paperwork and many fees to meet and
keep up with these requirements. We do not solicit donations in
locations where we have not received written confirmation of
compliance. To SEND DONATIONS or determine the status of
compliance for any particular state visit www.gutenberg.org/donate.

While we cannot and do not solicit contributions from states where

we have not met the solicitation requirements, we know of no
prohibition against accepting unsolicited donations from donors in
such states who approach us with offers to donate.

International donations are gratefully accepted, but we cannot make

any statements concerning tax treatment of donations received from
outside the United States. U.S. laws alone swamp our small staff.

Please check the Project Gutenberg web pages for current donation
methods and addresses. Donations are accepted in a number of
other ways including checks, online payments and credit card
donations. To donate, please visit: www.gutenberg.org/donate.

Section 5. General Information About

Project Gutenberg™ electronic works
Professor Michael S. Hart was the originator of the Project
Gutenberg™ concept of a library of electronic works that could be
freely shared with anyone. For forty years, he produced and
distributed Project Gutenberg™ eBooks with only a loose network of
volunteer support.
Project Gutenberg™ eBooks are often created from several printed
editions, all of which are confirmed as not protected by copyright in
the U.S. unless a copyright notice is included. Thus, we do not
necessarily keep eBooks in compliance with any particular paper
edition.

Most people start at our website which has the main PG search
facility: www.gutenberg.org.

This website includes information about Project Gutenberg™,

including how to make donations to the Project Gutenberg Literary
Archive Foundation, how to help produce our new eBooks, and how
to subscribe to our email newsletter to hear about new eBooks.
Welcome to Our Bookstore - The Ultimate Destination for Book Lovers
Are you passionate about books and eager to explore new worlds of
knowledge? At our website, we offer a vast collection of books that
cater to every interest and age group. From classic literature to
specialized publications, self-help books, and children’s stories, we
have it all! Each book is a gateway to new adventures, helping you
expand your knowledge and nourish your soul
Experience Convenient and Enjoyable Book Shopping Our website is more
than just an online bookstore—it’s a bridge connecting readers to the
timeless values of culture and wisdom. With a sleek and user-friendly
interface and a smart search system, you can find your favorite books
quickly and easily. Enjoy special promotions, fast home delivery, and
a seamless shopping experience that saves you time and enhances your
love for reading.
Let us accompany you on the journey of exploring knowledge and
personal growth!

ebookgate.com