DEMONSTRATIVES FORM, FUNCTION, AND GRAMMATICALIZATION
DEMONSTRATIVES
TYPOLOGICAL STUDIES IN LANGUAGE (TSL)
A companion series to the journal "STUDIES IN LANGUAGE"
Honorary Editor: Joseph H. Greenberg
General Editor: Michael Noonan
Assistant Editors: Spike Gildea, Suzanne Kemmer
Editorial Board:
Wallace Chafe (Santa Barbara) Ronald Langacker (San Diego)
Bernard Comrie (Leipzig) Charles Li (Santa Barbara)
R.M.W. Dixon (Canberra) Andrew Pawley (Canberra)
Matthew Dryer (Buffalo) Doris Payne (Oregon)
John Haiman (St Paul) Frans Plank (Konstanz)
Kenneth Hale (Cambridge, Mass.) Jerrold Sadock (Chicago)
Bernd Heine (Köln) Dan Slobin (Berkeley)
Paul Hopper (Pittsburgh) Sandra Thompson (Santa Barbara)
Andrej Kibrik (Moscow)
Volumes in this series will be functionally and typologically oriented, covering specific topics in
language by collecting together data from a wide variety of languages and language typologies. The
orientation of the volumes will be substantive rather than formal, with the aim of investigating
universals of human language via as broadly defined a data base as possible, leaning toward crosslinguistic,
diachronic, developmental and live-discourse data.
Volume 42
Holger Diessel
Demonstratives
Form, function, and grammaticalization
JOHN BENJAMINS PUBLISHING COMPANY
AMSTERDAM/PHILADELPHIA
DEMONSTRATIVES
FORM, FUNCTION, AND
GRAMMATICALIZATION
HOLGER DIESSEL
Max Planck Institute for Evolutionary Anthropology
The paper used in this publication meets the minimum requirements of
American National Standard for Information Sciences - Permanence of
Paper for Printed Library Materials, ANSI Z39.48-1984.
8
TM
Library of Congress Cataloging-in-Publication Data
Diessel, Holger.
Demonstratives : form, function, and grammaticalization / Holger Diessel.
p. cm. -- (Typological studies in language ISSN 0167-7373 ; v. 42)
Includes bibliographical references and index.
1. Grammar, Comparative and general--Demonstratives. I. Title II. Series.
P299.D46 D54 1999
415 99-046743
ISBN 90 272 2942 2 (Eur.) / 1 55619 656 3 (US) (Hb; alk. paper) CIP
ISBN 90 272 2943 0 (Eur.) / 1 55619 657 1 (US) (Pb; alk. paper)
© 1999 - John Benjamins B.V.
No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other
means, without written permission from the publisher.
John Benjamins Publishing Co. • P.O.Box 75577 • 1070 AN Amsterdam • The Netherlands
John Benjamins North America • P.O.Box 27519 • Philadelphia PA 19118-0519 • USA
Table of Contents
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
CHAPTER 1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Preliminary remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Towards a definition of demonstratives . . . . . . . . . . . . . . . . . . . . . 2
1.3 Outline and literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.4 Language sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
CHAPTER 2
Morphology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.1 Demonstratives in Guugu Yimidhirr, Ambulas, Ewondo, and Korean . 13
2.1.1 Guugu Yimidhirr . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.1.2 Ambulas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.1.3 Ewondo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.1.4 Korean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.2 The morphology of demonstratives . . . . . . . . . . . . . . . . . . . . . . . . 22
2.2.1 Demonstrative clitics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.2.2 The inflection of demonstratives . . . . . . . . . . . . . . . . . . . . . 25
2.2.3 Demonstrative stems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
CHAPTER 3
Semantics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.1 The semantic features of demonstratives . . . . . . . . . . . . . . . . . . . . 35
3.1.1 Deictic features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.1.2 Qualitative features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.1.3 Summary: the semantic features of demonstratives . . . . . . . . 50
3.2 The features of demonstratives: a systematic overview . . . . . . . . . . 50
vi TABLE OF CONTENTS
CHAPTER 4
Syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.1 Demonstrative pronouns and demonstrative determiners . . . . . . . . . . 59
4.1.1 Adnominal demonstrative pronouns . . . . . . . . . . . . . . . . . . . 60
4.1.2 Adnominal demonstratives in English . . . . . . . . . . . . . . . . . 62
4.1.3 Pronominal demonstrative determiners . . . . . . . . . . . . . . . . . 71
4.1.4 Pronominal and adnominal demonstratives: an overview . . . . 73
4.2 Demonstrative adverbs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.3 Demonstrative identifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
4.3.1 Phonological evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.3.2 Morphological evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
CHAPTER 5
Pragmatic use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
5.1 The exophoric use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
5.2 The anaphoric use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
5.3 The discourse deictic use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
5.4 The recognitional use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
5.5 The special status of exophoric demonstratives . . . . . . . . . . . . . . . . 109
5.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
CHAPTER 6
Grammaticalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
6.1 Some general principles of grammaticalization . . . . . . . . . . . . . . . . 116
6.2 Criteria for the grammaticalization of demonstratives . . . . . . . . . . . 118
6.3 The grammaticalization of pronominal demonstratives . . . . . . . . . . . 119
6.3.1 Third person pronouns . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
6.3.2 Relative pronouns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
6.3.3 Complementizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
6.3.4 Sentence connectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
6.3.5 Possessives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
6.4 The grammaticalization of adnominal demonstratives . . . . . . . . . . . 128
6.4.1 Definite articles and noun class markers . . . . . . . . . . . . . . . 128
6.4.2 Linkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
6.4.3 Boundary markers of postnominal relative clauses/attributes . . 132
6.4.4 Determinatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
6.4.5 Number markers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
6.4.6 Specific indefinite articles . . . . . . . . . . . . . . . . . . . . . . . . . 138
TABLE OF CONTENTS vii
6.5 The grammaticalization of adverbial demonstratives . . . . . . . . . . . . 139
6.5.1 Temporal adverbs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
6.5.2 Directional/locational preverbs . . . . . . . . . . . . . . . . . . . . . . 140
6.6 The grammaticalization of identificational demonstratives . . . . . . . . 143
6.6.1 Nonverbal copulas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
6.6.2 Focus markers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
6.6.3 Expletives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
6.7 The diachronic origin of demonstratives . . . . . . . . . . . . . . . . . . . . . 150
6.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
CHAPTER 7
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
7.1 Major findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
7.2 Future research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
APPENDIX A
Data sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
APPENDIX B
The inflectional features of pronominal demonstratives . . . . . . . . . . . . 171
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Language Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
Name Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
Acknowledgments
This book is a revision of my doctoral dissertation, Demonstratives in Crosslinguistic
and Diachronic Perspective, written at the State University of New York
at Buffalo. I would like to thank my thesis advisors, Matthew Dryer, Karin
Michelson, Jean-Pierre Koenig, and David Zubin for their guidance and intellectual
stimulation. I also would like to acknowledge the support and inspiration that
I received from my friends and fellow graduate students David Kemmerer,
Martha Islas, Wendy Baldwin, Alissa Melinger, Matthew Davidson, David
Houghton, and Cori Grimm. Special thanks goes to Eve Ng, who proofread the
entire manuscript. Finally, I am very grateful to Nikolaus Himmelmann and Edith
Moravcsik, who provided insightful comments on earlier versions of this book.
Parts of Chapter 4 and Chapter 6 first appeared, in different form, in the
following articles and conference papers: "The grammaticalization of demonstratives
in crosslinguistic perspective", Chicago Linguistic Society 33 (1997);
"Predicative Demonstratives", Berkeley Linguistics Society 23 (1997) (to appear);
"The morphosyntax of demonstratives in synchrony and diachrony", Linguistic
Typology 3 (1999). I thank the publishers for permission to include revised
material from these publications in the present monograph.
Abbreviations
ABL ablative
ABES abessive
ABS absolutive
ACC accusative
ADE adessive
ADJ adjective
ADV adverbial/adverb
AFF affirmative marker
AGR agreement marker
ANA anaphor
ANIM animate
AOR aorist
ART article
ASS assertive marker
ATT attribute marker
BEN benefactive
CIRC circumstantive
CLASS classifier
COLL collective
COM comitative
COMP complementizer
CONTRF contrafactive
COP copula
DAT dative
DEF definite
DEIC deictic
DEM demonstrative
DET determiner
DIST distal
DP determiner phrase
DTM determinative
DU dual
EMPH emphatic
ERG ergative
EXCL exclamative marker
EXIST existential marker
FACT factive
FEM/F feminine
FOC focus
FUT future
GEN genitive
G1S⁄G2S noun classes
H hearer
HAB habitual aspect
IDENT identificational/
identifier
ILL illative
IMP imperative
INANIM inanimate
INDEF indefinite
INE inessive
INSTR instrumental
INTJEC interjection
INVIS invisible
LAT lative
LK linker
LOC locative
MASC/M masculine
MD modal particle
MED medial
xii ABBREVIATIONS
N noun
NC noun class
NEG negative marker
NEUT/N neuter
NLZ nominalizer
NOM nominative
NONFACT nonfactive
NONPAST nonpast
NP noun phrase
NPIP noun phrase initial
NUC nucleus
NUM number
OBJ object
OBL oblique
P preposition
PART particle
PASS passive
PERF perfective aspect
PF present perfect
PL plural
POSS possessive
PRED predicative marker
PRES present
PRO pronominal/pronoun
PROX proximal
PAST past
PURP purposive
Q question
QNT quantifier
REF reference
REL relative marker
S speaker
SG singular
SPEC specific indefinite article
SUB subordinate marker
SUBJ subject
STA stative aspect
TEMP temporal
TNS tense marker
VIS visible
1SG first person singular
2SG second person singular
3SG third person singular
3PL third person plural
CHAPTER 1
Introduction
1.1 Preliminary remarks
All languages have demonstratives, but their form, meaning and use vary
tremendously across the languages of the world. Some languages have only a few
demonstrative particles, which they employ in a variety of syntactic contexts for
a wide range of semantic and pragmatic functions. Other languages have demonstratives
that are morphologically complex (i.e. not merely particles), syntactically
restricted, and semantically and pragmatically very specific in function.
This work provides the first large-scale analysis of demonstratives from a
crosslinguistic and diachronic perspective. It is based on a sample of 85 languages
from a wide range of genetic groups and geographical areas. The first part of
this book analyzes demonstratives form a synchronic point of view. It examines
their morphological structures, semantic features, syntactic functions, and
pragmatic uses. The second part is concerned with diachronic aspects of demonstratives,
in particular with their grammaticalization. Across languages, demonstratives
provide a common historical source for definite articles, relative and
third person pronouns, copulas, sentence connectives, directional preverbs, and
many other grammatical items. I describe the different mechanisms by which
demonstratives grammaticalize and argue that the evolution of grammatical markers
from demonstratives is crucially distinct from other cases of grammaticalization.
The main purpose of this study is to provide a source of reference for both
field workers and theoretical linguists who are interested in demonstratives and
their grammaticalization. The book provides a systematic overview of all empirical
aspects of demonstratives and addresses a number of theoretical issues that are
of more general interest in typology, syntax and grammaticalization theory.
This introductory chapter presents a preview of the major results of my
investigation and discusses previous work dealing with demonstratives and their
grammaticalization. I begin with a brief definition of the notion of demonstrative.
2 DEMONSTRATIVES
1.2 Towards a definition of demonstratives
There are three criteria that are relevant for the notion of demonstrative that I
have used in this study. First, demonstratives are deictic expressions serving
specific syntactic functions. Many studies confine the notion of demonstrative to
deictic expressions such as English this and that, which are used either as
independent pronouns or as modifiers of a coocurring noun, but the notion that
I will use is broader. It subsumes not only demonstratives being used as pronouns
or noun modifiers but also locational adverbs such as English here and there.
Second, demonstratives generally serve specific pragmatic functions. They
are primarily used to focus the hearer's attention on objects or locations in the
speech situation (often in combination with a pointing gesture), but they may also
function to organize the information flow in the ongoing discourse. More
specifically, demonstratives are often used to keep track of prior discourse
participants and to activate specific shared knowledge. The most basic function
of demonstratives is, however, to orient the hearer outside of discourse in the
surrounding situation.
Finally, demonstratives are characterized by specific semantic features. All
languages have at least two demonstratives that are deictically contrastive: a
proximal demonstrative referring to an entity near the deictic center and a distal
demonstrative denoting a referent that is located at some distance to the deictic
center. There are, however, a few languages in my sample in which some
demonstratives are distance-neutral. For instance, though German has three
adverbial demonstratives-hier 'here', da 'there', and dort 'there'-it employs
a single demonstrative pronoun: dies 'this/that' (cf. 3.1.1). Demonstratives such
as German dies are typologically uncommon and one might argue that they are
indistinct from third person pronouns and/or definite articles (cf. Anderson and
Keenan 1985: 280), but I will treat them as demonstratives for two reasons: first
like distance-marked demonstratives, distance-neutral demonstratives are commonly
used to orient the hearer in the surrounding situation, and second they can
always be reinforced by demonstratives that are marked for distance if it is
necessary to differentiate between two or more referents (see 3.1.1).
1.3 Outline and literature
Demonstratives have been the subject of numerous investigations in linguistics
and philosophy.1 The vast majority of studies concentrates on their pragmatic use
and meaning. The scope of the current investigation is broader. It does not only
INTRODUCTION 3
examine their use and meaning, but also their morphological structures, syntactic
functions and grammaticalization. One of the major challenges of this work is to
integrate the analysis of different aspects of demonstratives into one coherent
presentation. This section provides an overview of my investigation, which is
organized into five chapters.
Chapter 2 examines the morphology of demonstratives. It is divided into two
sections. The first section describes the demonstrative systems of four languages.
It illustrates the extent of formal variation among demonstratives in different
languages. The second section examines the morphological structures of demonstratives
more systematically. It discusses the properties of demonstrative clitics,
the inflectional behavior of demonstratives in different syntactic contexts, and the
formation of demonstrative stems. There is to my knowledge no previous work
on the morphology of demonstratives in typological perspective. The current
study is the first investigation in this domain.
Chapter 3 investigates the semantic features of demonstratives. The meaning
of demonstratives and other deictics has been studied extensively. The collections
of articles edited by Weissenborn and Klein (1982), Jarvella and Klein (1982),
and Rauh (1983) present an overview of the research in this domain. Some of the
articles in these collections describe the semantic features of demonstratives from
a comparative or crosslinguistic perspective (e.g. Fillmore 1982; Ehlich 1983), but
most of them concentrate on the demonstratives in one particular language
(usually English). There are two studies that are especially important to my
investigation: Fillmore (1982) and Anderson and Keenan (1985). Both studies
examine the semantic features of demonstratives from a typological perspective.
My analysis is inspired by their approach, but it is based on a much larger
language sample and distinguishes systematically between semantic features of
two different domains: (i) deictic features, which indicate the location of the
referent in the speech situation, and (ii) qualitative features, which classify the
referent (cf. Lyons 1977; Rauh 1983). The deictic features comprise features that
indicate whether the referent is near, away or far away from the deictic center,
whether it is visible or out of sight, at a higher or lower elevation, uphill or
downhill, or moving toward or away from the deictic center. The qualitative
features indicate whether the referent is an object, person or place, whether it is
animate or inanimate, human or non-human, female or male, a single entity or a
set, or conceptualized as a restricted or extended entity. All of these features are
directly encoded by several demonstratives in my sample.
Chapter 4 examines the syntactic properties of demonstratives. It is argued
that one has to distinguish between the distribution and the categorial status of
demonstratives. The categorial status of a demonstrative is defined by the
4 DEMONSTRATIVES
combination of two features: (i) a certain distribution and (ii) a specific form.
Two demonstratives belong to different categories if they are distributionally and
formally distinguished. I use the attributes pronominal, adnominal, adverbial, and
identificational in order to indicate the syntactic context in which demonstratives
occur (i.e. their distribution); and I use the nominals (demonstrative) pronoun,
determiner, adverb, and identifier in order to indicate their categorial status.
Table 1 presents an overview of these terms.
The distinction between the distribution and categorial status of demonstratives
Table 1. Demonstratives: distribution and category
Distribution Category
pronominal demonstrative
adnominal demonstrative
adverbial demonstrative
identificational demonstrative
demonstrative pronoun
demonstrative determiner
demonstrative adverb
demonstrative identifier
is crucial because some languages use demonstratives of the same grammatical
category in more than one syntactic context, while other languages employ
categorially (i.e. formally) distinct demonstratives in each position.
Pronominal demonstratives are independent pronouns in argument position
of verbs and adpositions. They are often formally distinguished from adnominal
demonstratives, which accompany a cooccurring noun. For instance, French uses
the demonstratives celui, celle, ceux, and celles as independent pronouns and ce,
cette, and ces as modifiers of a subsequent noun. Since pronominal and adnominal
demonstratives are formally distinguished in French, I assume that they
belong to different categories: celui, celle, ceux, and celles are demonstrative
pronouns, while ce, cette, and ces are demonstrative determiners.
Unlike French, many languages do not distinguish between demonstrative
pronouns and demonstrative determiners. Tuscarora, for instance, has two
demonstrative particles, hèník6˜ 'this/these' and kyèník6˜ 'that/those', which are
either used in isolation as independent pronouns or together with a cooccurring
noun. Following Mithun (1987), I assume that pronominal and adnominal demonstratives
belong to the same category in Tuscarora. Unlike adnominal demonstratives
in French, adnominal demonstratives in Tuscarora do not function as
determiners; rather, they are used as independent pronouns that are only loosely
adjoined to a coreferential noun in apposition. In other words, Tuscarora does not
have demonstrative determiners; it has only demonstrative pronouns that are used
INTRODUCTION 5
in two different syntactic contexts. There are also languages in my sample that
do not have a class of demonstrative pronouns. These languages use demonstrative
determiners in combination with a third person pronoun or a classifier in
contexts where languages such as English employ demonstrative pronouns.
Adverbial demonstratives are usually distinguished from pronominal and
adnominal demonstratives. There are only a few languages in my sample in
which adverbial demonstratives have the same form as demonstratives that are
used as independent pronouns or noun modifiers. For instance, in Ngiyambaa it
is possible to refer to a location by a demonstrative pronoun in locative case, as
illustrated in (1).
(1) Ngiyambaa (Donaldson 1980: 317)
yaba=lugu ]a-ni-la guri-nja
track=3SG.GEN that-LOC-GIVEN lie-PRES
'His tracks are there.'
The demonstrative in (1) is semantically equivalent to English there, but
from a morphological and syntactic perspective it can be viewed as a demonstrative
pronoun with a locative case marker.
In addition to demonstrative pronouns, determiners and adverbs, there is a
fourth demonstrative category, which is almost entirely unknown in the typological
literature (but see Himmelmann 1997). Many languages use special
demonstrative forms in copular and nonverbal clauses, as in example (2) and (3)
from French and Ponapean.
(2) French
C' est Pascal.
this/it is Pascal
'It/this is Pascal.'
(3) Ponapean (Rehg 1981: 143)
Iet noumw pinselen.
here.is your pencil
'Here is your pencil.'
The demonstratives in these examples function to identify a referent in the
speech situation. They are usually considered demonstrative pronouns, but many
languages distinguish ordinary demonstrative pronouns from demonstratives in
copular and nonverbal clauses. Both French and Ponapean employ demonstratives
with a different form as independent pronouns in other syntactic contexts. I
distinguish therefore between demonstratives in copular and nonverbal clauses
and demonstratives that occur in other sentence types. As shown in Table 1, I use
6 DEMONSTRATIVES
the term identificational demonstrative for demonstratives in copular and nonverbal
clauses regardless of their categorial status; and I use the term demonstrative
identifier when I refer to demonstratives in copular and nonverbal clauses
that are categorially (i.e. formally) distinguished from demonstratives in other
contexts.
Chapter 5 is concerned with the pragmatic uses of demonstratives. As
pointed out above, demonstratives are primarily used to draw the hearer's
attention to entities in the speech situation, but they may also serve a variety of
other pragmatic functions. Following Halliday and Hasan (1976: 57-76), I use the
notion exophoric for demonstratives that are used with reference to entities in the
surrounding situation, and I use the term endophoric for all other uses. The
endophoric use is subdivided into the (i) anaphoric, (ii) discourse deictic, and (iii)
recognitional uses, as shown in Figure 1.
pragmatic uses
exophoric endophoric
anaphoric discourse deictic recognitional
Figure 1. The pragmatic uses of demonstratives
The exophoric, anaphoric, and discourse deictic uses are discussed in studies
by Lyons (1977), Levinson (1983), Webber (1991), Fillmore (1997), Himmelmann
(1996, 1997), and many others. Exophoric demonstratives refer to nonlinguistic
entities in the speech situation; they focus the hearer's attention on
persons, objects or locations in the outside world. Anaphoric demonstratives are
coreferential with a noun phrase in the preceding discourse; they keep track of
prior discourse participants. Discourse deictic demonstratives refer to a chunk of
the surrounding discourse; they express an overt link between two propositions.
In addition to the exophoric, anaphoric and discourse deictic uses, Himmelmann
(1996, 1997) describes another usage, which he calls the recognitional use.
In the recognitional use, demonstratives function to indicate that speaker and
hearer are familiar with the referent due to shared experience. The demonstrative
in the following sentence exemplifies this use.
INTRODUCTION 7
(4) English
Do you still have that radio that your aunt gave you for your birthday?
Although the radio is not present in the speech situation and is mentioned
for the first time, it occurs with the distal demonstrative that. The speaker uses
the demonstrative in order to indicate that the hearer is able to identify the
referent based on specific shared knowledge. Aspects of the recognitional use
have been described in several recent investigations (cf. Chen 1990; Gundel et al.
1993), but Himmelmann (1996, 1997) provides the only systematic account of
this use. His work is especially significant for my treatment of the pragmatic uses
in Chapter 5. Though I adopt Himmelmann's distinction of the four pragmatic
uses that I have mentioned, I challenge his hypothesis that all four uses have
equal status. In accordance with most previous work (cf. Bühler 1934; Lyons
1977), I contend that the exophoric use represents the basic use from which
all other uses derive. I support my hypothesis with evidence from language
acquisition, markedness theory, and grammaticalization.
The final chapter of this study investigates the diachronic reanalysis of
demonstratives as grammatical markers. I argue that the process by which
demonstratives grammaticalize is crucially determined by the syntactic context
in which the demonstrative occurs. More specifically, I show that pronominal,
adnominal, adverbial, and identificational demonstratives provide the source for
four different sets of grammatical markers. Pronominal demonstratives are
frequently reanalyzed as third person pronouns, relative pronouns, complementizers,
sentence connectives, pronominal determinatives, possessive pronouns,
and verbal number markers. Adnominal demonstratives may develop into definite
articles, linkers, nominal number markers, adnominal determinatives, specific
indefinite articles, and boundary markers of postnominal attributes. Adverbial
demonstratives provide a common source for temporal adverbs, directional
preverbs, and sentence connectives. Finally, identificational demonstratives may
evolve into copulas, focus markers, and expletives.
Grammaticalization is a gradual process (Lichtenberk 1991). At the initial
stage of this process, grammatical items often have the same form as their
historical source. The distinction between demonstratives and grammatical
markers such as definite articles is therefore not always immediately obvious. In
Chapter 6, I argue that the grammaticalization of demonstratives can be viewed
as a cline ranging from exophoric demonstratives used to orient the hearer in the
outside world to grammatical markers serving specific syntactic functions.
Endophoric demonstratives are somewhere in between the two ends of this cline;
8 DEMONSTRATIVES
they often mark the initial stage of a grammaticalization process, giving rise to
definite articles and many other grammatical items.
The diachronic reanalysis of demonstratives has been the subject of numerous
investigations, in both traditional historical linguistics and recent work in
grammaticalization. Some of the important older work includes Brugmann (1904),
Brugmann and Delbrück (1911), Paul (1920), and Behaghel (1923-1932). These
studies are exclusively concerned with the reanalysis of demonstratives in Indo-
European languages. More recent studies have shown that the grammaticalization
of demonstratives is a common historical process in virtually every language.
Heine and Reh (1984), Heine, Claudi and Hünnemeyer (1991a), Hopper and
Traugott (1993), Lehmann (1995a), and Harris and Campbell (1995) describe the
reanalysis of demonstratives as definite articles, relative and third person pronouns,
copulas, and complementizers in a wide variety of languages. Other
studies concentrate on the grammaticalization of one particular item. For instance,
the reanalysis of demonstratives as definite articles has been described in studies
by Christophersen (1939), Heinrichs (1954), Krámský (1972), Ultan (1978a),
Harris (1978, 1980), Greenberg (1978, 1991), Lüdtke (1991), Vogel (1993), Cyr
(1993a, 1993b, 1996), Leiss (1994), Epstein (1994, 1995), Laury (1995, 1997),
and Himmelmann (1997, 1998). There are also several studies that examine the
development of copulas from demonstratives: Berman and Grosu (1976), Li and
Thompson (1977), Schuh (1983a), Hengeveld (1990), Gildea (1993), and Devitt
(1994). Other studies that are relevant to my investigation include Lehmann
(1984), who describes the development of relative pronouns in Ancient Greek,
Old High German, and several other languages; Sankoff and Brown (1976), who
examine the emergence of boundary markers of postnominal attributes in Tok
Pisin; Lockwood (1968), who discusses a number of grammatical items that
developed from demonstratives in German; and Frajzyngier (1997), who shows
that nominal and verbal number markers in many Chadic languages originate
from demonstratives. Chapter 6 summarizes much of this work and discusses
several other grammaticalization channels that have only been described in
reference grammars.
Finally, I consider the question: where do demonstratives come from -
what is their historical source? Demonstratives are usually considered grammatical
items. Grammaticalization theory claims that all grammatical items are
ultimately derived from lexical expressions, but in the case of demonstratives
there is no evidence from any language that they developed from a lexical source
or any other source, for that matter, that is non-deictic (Himmelmann 1997: 20).
Based on this finding, I advance the hypothesis that demonstratives might form
a class of deictic expressions that belong to the basic vocabulary of every
INTRODUCTION 9
language (cf. Plank 1979a; Traugott 1982). This would not only explain why demonstratives
cannot be traced back to lexical items, it would also account for the
fact that demonstratives serve a particular pragmatic function that sets them apart
from all other linguistic expressions (cf. Bühler 1934; Peirce 1955; Ehlich 1979,
1982, 1983, 1987). Furthermore, it might explain why demonstratives are among
the very few items that exhibit a non-arbitrary relationship between phonetic form
and meaning (cf. Woodworth 1991). If demonstratives are not derived from
lexical items they would present a second source domain from which grammatical
markers may emerge, and that would undermine one of the central assumptions
of grammaticalization theory.
1.4 Language sample
My study is based on a sample of 85 languages. The bulk of the data comes from
reference grammars and other published sources, supplemented by information
obtained from native speakers and language specialists. For each language, I
gathered information on the morphological structures, semantic features, syntactic
functions, and grammaticalization of demonstratives. I used two criteria for
selecting the languages of my sample: (i) genetic diversity and (ii) geographical
distance.2 For the genetic classification, I consulted two sources: Ruhlen (1991)
and the Ethnologue (Grimes 1997). With two exceptions, my sample includes at
least one language of every major language family that Ruhlen suggests. The two
exceptions are Miao-Yao and Chukchi-Kamchatkan. Both families include fewer
than half a dozen languages in Ruhlen's classification. My sample also includes
several language isolates (Burushaski, Basque, Korean, Ainu) and an Englishbased
creole (Tok Pisin). Large and diverse phyla are better represented than
small and homogeneous phyla. For instance, Niger-Congo, which comprises
several hundred languages, is represented by five languages in my sample, while
Khoisan, which subsumes 30 languages, is only represented by one.
In order to ensure geographical diversity I used the geographical divisions
suggested by Dryer (e.g. 1992a). He distinguishes six major geographical areas:
North America, South America, South East Asia and Oceanic, Africa, Eurasia,
and Australia and New Guinea (Dryer 1989c distinguishes only five areas). My
sample includes at least eight languages of each of the six geographical areas that
Dryer suggests. Below, I have given a list of all 85 languages that are included
in my sample. The sources that I have used for each language are listed in
Appendix A.
10 DEMONSTRATIVES
North America
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
Halkomelem
Kiowa
Lealao Chinantec
Mam
Mojave
Oneida
Passamaquoddy-Maliseet
Picurís
Quileute
Slave
Tümpisa Shoshone
Tuscarora
Tzutujil
Ute
West Greenlandic
Salishan
Kiowa-Tanoan
Oto-Manguean
Mayan
Hokan
Iroquoian
Algonquian
Kiowa-Tanoan
Chimakuan
Athapaskan-Eyak
Uto-Aztecan
Iroquoian
Mayan
Uto-Aztecan
Eskimo-Aleut
South America
16.
17.
18.
19.
20.
21.
22.
23.
Apalai
Barasano
Canela-Krahô
Epena Pedee
Hixkaryana
Urubu-Kaapor
Wari'
Yagua
Carib
Tucanoan
Ge-Kaingang
Choco
Carib
Tupi-Guarani
Chapakuran
Peba-Yaguan
INTRODUCTION 11
South East Asia and Oceanic
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
Acehnese
Ao
Byansi
Khasi
Kusaiean
Lahu
Manam
Mandarin Chinese
Mulao
Nùng
Pangasinan
Ponapean
Santali
Vietnamese
West Futuna-Aniwa
Western Malayo-Polynesian
Tibeto-Burman
Tibeto-Burman
Mon-Khmer
Central-Eastern Malayo-Polynesian
Tibeto-Burman
Central-Eastern Malayo-Polynesian
Sinitic
Daic
Daic
Western Malayo-Polynesian
Central-Eastern Malayo-Polynesian
Munda
Mon-Khmer
Central-Eastern Malayo-Polynesian
Africa
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
Duwai
Ewondo
Gulf Arabic
Izi
Karanga
Koyra Chiini
Kunuz-Nubian
Lango
Logbara
Margi
Modern Hebrew
Nama
Nandi
Ngiti
Supyire
Swazi
Turkana
Western Bade
Chadic
Niger-Congo
Semitic
Niger-Congo
Niger-Congo
Songhay (wider affiliation unknown)
Eastern Sudanic
Nilotic
Central Sudanic
Chadic
Semitic
Khoisan
Nilotic
Central Sudanic
Niger-Congo
Niger-Congo
Nilotic
Chadic
12 DEMONSTRATIVES
Eurasia
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
Ainu
Basque
Burushaski
Czech
Finnish
French
Georgian
German
Japanese
Kannada
Korean
Lezgian
Punjabi
Swedish
Turkish
Isolate
Isolate
Isolate
Slavic
Finno-Ugric
Romance
Kartvelian
Germanic
Japanese
Dravidian
Isolate
North Caucasian
Indo-Aryan
Germanic
Turkic
Australia and New Guinea
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
Alamblak
Ambulas
Dyirbal
Guugu Yimidhirr
Hua
Ngiyambaa
Nunggubuyu
Tauya
Tok Pisin
Urim
Usan
Wardaman
Yankunytjatjara
Yimas
Sepik
Sepik
Pama-Nyungan
Pama-Nyungan
East New Guinea Highlands
Pama-Nyungan
Gunwingguan
Madang-Adelbert Range
Creole
Torricelli
Madang-Adelbert Range
Gunwingguan
Pama-Nyungan
Nor-Pondo
CHAPTER 2
Morphology
This chapter is concerned with the morphological structures of demonstratives.
Some languages have only a few demonstrative particles; they are uninflected and
do not combine with any other morpheme. Other languages employ demonstratives
that are marked for gender, number and/or case and may combine with
derivational affixes or other free forms. Languages of this sort can have several
hundred demonstrative forms. For instance, Denny (1982: 372) reports that the demonstrative
system in Inuktitut comprises 686 different forms, formed from
twelve demonstrative roots and a wide variety of inflectional and derivational
morphemes. Another language with an extremely complex demonstrative system
is Santali, a Munda language spoken in northeastern India. According to
Bodding's description (1929: 118-147), there are well over 200 distinct forms in
the demonstrative system in Santali.
In this chapter, I first describe the demonstratives of four individual languages,
and then I examine the morphological structures of demonstratives more
systematically from a typological perspective.
2.1 Demonstratives in Guugu Yimidhirr, Ambulas, Ewondo, and Korean
The following four sections describe the demonstratives in Guugu Yimidhirr,
Ambulas, Ewondo, and Korean, respectively. The purpose of these descriptions
is to illustrate the extent of formal variation among demonstratives in different
languages.We will see that the demonstratives of some languages are morphologically
invariable, while the demonstratives of other languages are formed from a
demonstrative root and a wide variety of other morphemes.
2.1.1 Guugu Yimidhirr
Guugu Yimidhirr is a Pama-Nyungan language spoken in northeastern Australia.
14 DEMONSTRATIVES
The language has two demonstrative roots, yii 'proximal' and nhaa 'distal', which
are inflected for case:
The demonstratives in Table 2 refer to "things", "places" and "times" (Haviland
Table 2. Demonstratives in Guugu Yimidhirr (Haviland 1979: 73)
PROXIMAL DISTAL
ABS
ERG/INSTR
LOC/ALL
ABL/CAUSAL
PURP
COM
yii, yiyi
yiimuun
yiway, yuway, yiimu
yiimunganh
yimudhirr, yimidhirr
nhaa, nhaayun
nhaamuun
nhaway, nhaamu
nhaamunganh, nhaawanun (ABL only)
nhaamuu
nhamudhirr, nhamidhirr
1979: 72). They may function as independent pronouns (1a), noun modifiers (1b),
locational deictics (1c), or identificational demonstratives in nonverbal clauses
(1d). That is, Guugu Yimidhirr does not distinguish between demonstrative
pronouns, determiners, adverbs, and identifiers; rather, it uses the same demonstrative
forms in various syntactic contexts:
(1) Guugu Yimidhirr (Haviland 1979: 73, 165, 92, 54)
a. ngayu nhinaan yiimuun gunda-l
1SG.NOM 2SG.ACC PROX.INSTR hit-NONPAST
'I'll hit you with this (thing I have here).'
b. nhayun nambal bada gada-y iii
DIST.ABS rock.ABS down come-PAST ...
'That rock dropped...'
c. ngayu yiway nhin.ga-l
1SG PROX.LOC sit-NONPAST
'I'll stay here.'
d. yii yugu yalmba-aga
PROX.ABS tree.ABS sandhill-ABES
'This is a tree of the sandhill.'
The demonstratives in absolutive case have plural forms: yinharrin 'these/this
kind' and nhanharrin 'those/that kind'. They seem to be the only demonstratives
in plural. An example is given in (2).
MORPHOLOGY 15
(2) Guugu Yimidhirr (Haviland 1979: 74)
yinharrin bama binaal-mul
these.ABS people.ABS know-PRIVATE
'These (sort of) peope don't know (about it).'
The absolutive form nhaayun 'proximal' is often glossed as 'it' and might be
considered a third person pronoun:
(3) Guugu Yimidhirr (Haviland 1979: 73)
buligi gada-y, nyulu nhaayun gunda-y
bullock.ABS come-PAST 3SG.NOM DIST.ABS kill-PAST
'The bullock came and he killed it.'
The purposive form nhaamuu refers to propositions and is usually translated by
'therefore':
(4) Guugu Yimidhirr (Haviland 1979: 74)
nyulu wawu-murrgarra bama-agal yirrga-nda
3SG.NOM breath-unable man-ADES speak-CONTRF
guugu wangaarr-ga-m-i nhaamuu nyulu
speech.white man-GEN-mu-DAT DIST.PURP 3SG.NOM
guugu yi-mi-dhirr maa-ni
speech.ABS PROX-mu-COM take-PAST
'He was unable to talk to Aboriginals in the white man's language,
and therefore he learned Guugu Yimidhirr.'
There are two other deictic expressions in Guugu Yimidhirr that one might
classify as demonstratives: yarra 'yonder' and yarrba 'there, that way, that's the
way'. Unlike the demonstratives in Table 2, these two forms are uninflected.
Haviland (1979: 72) points out that they normally accompany gestures, but their
status and function remain somewhat unclear.
2.1.2 Ambulas
Ambulas, a Sepik language spoken in New Guinea, has three demonstrative roots:
kén 'proximal', an 'medial' and wan 'distal'. In contrast to Guugu Yimidhirr,
where the same demonstrative forms occur in a variety of syntactic contexts,
Ambulas has several categories of demonstratives serving specific syntactic
functions.
To begin with, the demonstrative roots, kén and wan, are used as demonstrative
identifiers in nonverbal clauses.
16 DEMONSTRATIVES
(5) Ambulas (Wilson 1980: 454)
kén bakna walkamu taalé
this just little place
'This is just a little place.'
Demonstrative pronouns are formed by combining kén and wan with the third
person pronouns dé 'he', lé 'she', bét 'they (dual)', and de 'they (plural)',
yielding the forms in Table 3.
Ambulas has a second series of demonstrative pronouns, which is used only with
Table 3. Demonstrative pronouns in Ambulas (1) (Wilson 1980: 56)
PROXIMAL DISTAL
SG.M
SG.F
DU
PL
dé-kén
lé-kén
bét-kén
de-kén
dé-wan
lé-wan
bét-wan
de-wan
inanimate referents. The forms are given in Table 4; they are unmarked for
gender and number. Note that the inanimate demonstratives express a three way
contrast while the (animate) demonstratives in Table 3 have only two deictic
roots: proximal and distal. The medial demonstrative an does not combine with
third person pronouns.
Ambulas also has two series of demonstrative determiners, shown in Table 5.
Table 4. Demonstrative pronouns in Ambulas (2) (Wilson 1980: 56)
PROXIMAL
MEDIAL
DISTAL
kénikénan, kénikinan
anikénan, anikinan
wanikénan, wanikinan, waninan
Both series distinguish three degrees of distance. The demonstratives of the first
series are also used as "an introducer and closure marker of a discourse" and as
temporal adverbs meaning 'now' and 'then' (Wilson 1980: 56). Demonstrative
determiners always precede the noun and other noun modifiers, as shown in (6a-b).
MORPHOLOGY 17
(6) Ambulas (Wilson 1980: 433, 86)
Table 5. Demonstrative determiners in Ambulas (Wilson 1980: 57)
SERIES 1 SERIES 2
PROXIMAL
MEDIAL
DISTAL
kéni
ani
wani
kénina
aniké, aniki
waniké, waniki, wanina
a. kéni kudi
this talk
'This talk'
b. kénina bét-ku jébaa
this 3DU-POSS work
'This work of theirs'
Table 6 shows the demonstrative adverbs. Wilson (1980: 58) argues that they
might have developed from two morphemes: the demonstratives kén and wan and
a locational adverb or adposition. However, from a synchronic perspective these
forms are monomorphemic.
Finally, there is a series of manner demonstratives, shown in Table 7.
Manner demonstratives are commonly used to refer to a chunk of the surrounding
Table 6. Demonstrative adverbs in Ambulas (Wilson 1980: 57)
PROXIMAL
DISTAL
TO.PROXIMAL
TO.DISTAL
kéba
waba
kénét
wanét
Table 7. Manner demonstratives in Ambulas (Wilson 1980: 57)
LIKE THIS (forward referring)
LIKE THIS (forward referring)
LIKE THAT (backward referring)
kéga
aga
waga
discourse (cf. 4.2). They are discourse deictics that indicate an overt link between
two propositions. The distal form waga refers back to an element of the preceding
discourse, while the proximal and medial forms anticipate upcoming information.
Example (7) shows the distal demonstrative waga referring back to the
preceding clause.
18 DEMONSTRATIVES
(7) Ambulas (Wilson 1980: 464-5)
gayéba kutdu méné gayéba kutkéyo ména
in.village he.catch.and you in.village must.catch your
naawi wale waaru waatbete waga de yo
peer with argue be.angry.and thus they do
'If he catches a pig in the village, you must catch one in the village.
When you quarrel with your peer, that is what they do.'
2.1.3 Ewondo
Ewondo is a Northwest Bantu language spoken in southern Cameroon. It
distinguishes three distance categories - proximal, medial, and distal - and six
noun classes:
The demonstratives in Table 8 can be used as independent pronouns or as noun
Table 8. Demonstrative pronouns/determiners in Ewondo (Redden 1980: 67-70)
NEAR S NEAR H AWAY FROM S+H
SG PL SG PL SG PL
NC1
NC2
NC3
NC4
NC5
NC6
\f'
\u¯
dı¯
dzı¯
\ı¯
\u¯
bá
mı¯
ma¯
bı¯
ma¯
dı¯
\f'lo¯
\u¯lú
dı¯lí
dzı¯lí
\ı¯lí
\u¯lú
bála¯
mı¯lí
ma¯lá
bı¯lí
ma¯lá
dı¯lí
\f'lí
\u¯lí
dílí
dzílí
\ílí
\u¯lí
bálí
mı¯líí
ma¯lí
bı¯líí
ma¯lí
dı¯líí
modifiers. That is, Ewondo does not distinguish between demonstrative pronouns
and determiners.3
Both pronominal and adnominal demonstratives always cooccur with the
definite article é. If a demonstrative is used adnominally, the article and the demonstrative
frame the noun as in the following example:
(8) Ewondo (Redden 1980: 67)
é mod \f'
ART man DEM
'this man'
Occasionally the definite article is repeated before the demonstrative as in (9):
MORPHOLOGY 19
(9) Ewondo (Redden 1980: 67)
é mod é \f'
ART man ART DEM
'this man'
The second article in this example can be omitted, but the initial é is obligatory
if the demonstrative after the noun functions as a noun modifier. If the noun is
not preceded by é the demonstrative does not form an NP with the cooccurring
noun; instead, it functions as an identificational demonstrative in a nonverbal
clause. Compare the following two examples:
(10) Ewondo (Redden 1980: 67, 67)
a. é kádá \f'
ART crab DEM
'this crab'
b. kádá \f'
crab DEM
'This is a crab.'
Due to the definite article that precedes the noun in (10a), \f' is interpreted as a
demonstrative noun modifier; that is, \f' is an adnominal demonstrative in this
case. Since the noun in (10b) does not occur with the definite article é, the
following demonstrative is interpreted as an identificational demonstrative rather
than a noun modifier.
Like adnominal demonstratives, pronominal demonstratives are generally
accompanied by the definite article é:
(11) Ewondo (Redden 1980: 178-9)
\g6' wa-yi é ]f'l6 te w-ee-ya]ga kóám
if you-want ART DEM you-FUT-wait a.long.time
'If you want that one you'll wait quite a while.'
Identificational demonstratives are formally indistinguishable from pronominal
(and adnominal) demonstratives, but adverbial demonstratives have a different
phonological form. There are two sets of adverbial demonstratives in Ewondo,
shown in Table 9.
While adnominal and pronominal demonstratives comprise three deictic
forms, adverbial demonstratives express a four way deictic contrast. They indicate
a location (i) near the speaker, (ii) near the hearer, (iii) away from speaker and
hearer, and (iv) far away from speaker and hearer. The last category is probably
also used for locations out of sight. The demonstratives of set 1 are used to
indicate an "exact location", while the demonstratives of the second set indicate
20 DEMONSTRATIVES
that the location is somewhere "around" or "in the vicinity of" the speaker, hearer
Table 9. Demonstrative adverbs in Ewondo (Redden 1980: 145)
SET 1 (precise) SET 2 (vague)
NEAR S
NEAR H
AWAY FROM S+H
FAR AWAY FROM S+H
vá
vála¯
válí
álí
mú
múlu¯
wóé
múlí
or some other reference point (cf. Redden 1980: 145-6). The adverbial demonstratives
of both sets may coalesce with the preposition à 'in/on/at/to', yielding the
following forms:
Table 10. Complex demonstrative adverbs in Ewondo (Redden 1980: 147)
SET 1 (precise) SET 2 (vague)
NEAR S
NEAR H
AWAY FROM S+H
FAR AWAY FROM S+H
ává
ávála¯
áválí
álí
ámú
ámúlú
áwóé
ámúlí
2.1.4 Korean
Korean has three demonstrative particles: i, referring to an object or person near
the speaker, ku, indicating a referent near the hearer, and ce, referring to an
object or location that is away from both speaker and hearer. The Korean demonstratives
are determiners; they cannot be used as independent pronouns. The
semantic equivalent of a pronominal demonstrative in English is a noun phrase
consisting of i, ku or ce and a "defective noun", which indicates the type of
referent (Sohn 1994: 294). The examples in (12a-c) show i, ku and ce followed
by the defective nouns kes 'thing/fact', i 'person' and il 'thing/fact'.4
(12) Korean (Sohn 1994: 294-6)
a. i kes
this thing/fact
'this (one/thing/fact)'
MORPHOLOGY 21
b. ku i
that person
'that (one/person)/he/she/it'
c. ce il
that.away thing/fact
'that (one/thing/fact)'
The demonstratives in (12a-c) are determiners; they have the same syntactic
function as demonstratives that precede a regular noun, as in the following examples:
(13) Korean (Sohn 1994: 114, 114, 114)
a. i cip
this house
'this house'
b. ku cha
that car
'that car (near you)'
c. ce san
that mountain
'that mountain over there'
I, ku and ce are morphologically invariable. Number and case is indicated through
affixes on the (defective) noun:
(14) Korean (Sohn 1994: 297)
i kes-tul-i
this THING-PL-NOM
'these (things/facts)'
Korean has three demonstrative adverbs that correspond to i, ku and ce:
Synchronically yeki, keki and ceki are monomorphemic, but historically they are
Table 11. Demonstrative adverbs in Korean (Sohn 1994: 296)
NEAR S
NEAR H
AWAY FROM S+H
yeki
keki
ceki
composed of the base form eki meaning 'place' and the demonstrative
determiners i, ku and ce (cf. Sohn 1994: 296). In place of the demonstrative
adverbs, the demonstratives i, ku and ce may be used with the defective noun kos
'place' in order to indicate a location:
22 DEMONSTRATIVES
(15) Korean (Sohn 1994: 295)
ku kos-ey kathi kaca
that place-to together go
'Let's go there together.'
Finally, there are three diminutive demonstratives in Korean: yo 'near speaker',
ko 'near hearer', and co 'away from speaker and hearer' (Sohn 1994: 114). Like
i, ku and ce, the diminutive demonstratives are determiners; they cannot occur
without a subsequent noun.
2.2 The morphology of demonstratives
The four previous sections exemplified the extent of formal variation among demonstratives
of different languages. We saw that the demonstratives of some
languages are morphologically invariable (e.g. Korean), while the demonstratives
of other languages inflect for gender, number and/or case (e.g. Guugu Yimidhirr,
Ewondo). We also saw that some languages distinguish demonstrative pronouns,
determiners, adverbs, and identifiers (e.g. Ambulas), while other languages have
only a single series of demonstratives that they employ in various syntactic
contexts (e.g. Guugu Yimidhirr). And finally we saw that the semantic features
of demonstratives vary across languages: some languages have only two deictic
terms (e.g. Guugu Yimidhirr), others have three (e.g. Korean), and yet others
have four (e.g. demonstrative adverbs in Ewondo). In what follows I examine the
form, meaning and syntax of demonstratives more systematically. In the remainder
of the current section, I investigate their morphological structures. I describe
in turn the properties of demonstrative clitics (2.2.1), the inflectional features of
demonstratives in different syntactic contexts (2.2.2), and the formation of demonstrative
stems (2.2.3). The semantic and syntactic features will be examined
in Chapter 3 and 4, respectively.
2.2.1 Demonstrative clitics
Most demonstratives are phonologically unbound. There is, however, a substantial
number of languages in my sample in which some demonstratives may cliticize
to an element in their environment: Acehnese, Ponapean, Lango, Margi, Nandi,
and several others. All of the demonstrative clitics that are included in my sample
are enclitics (i.e. clitics that follow their host); demonstrative proclitics do not
occur, but they seem to occur in other languages. An example from Yagaria is
shown in (16b):
MORPHOLOGY 23
(16) Yagaria (Renck 1975: 64, 66)
a. ma'i nina
this water
'this water'
b. m=ígopa
this=ground
'this ground'
Example (16a) shows an unbound demonstrative, which is commonly replaced by
a proclitic, as in (16b), if the referent is not emphasized (clitic boundaries are
indicated by the equal sign =).
All demonstrative clitics that are included in my sample are used adnominally;
pronominal, adverbial and identificational demonstratives are always
unbound. There are, however, languages in which the latter are clitics. Consider,
for instance, the following examples from Kilba (Chadic), in which both adnominal
and identificational demonstratives are attached to a preceding noun:
(17) Kilba (Schuh 1983b: 315, 318)
a. kí=nà
house=this
'this house'
b. k'6t'6]=ná
sheep=this
'This/it is a sheep.'
The demonstrative clitics in (17a-b) have the same segmental shape, but they are
tonally distinguished. The demonstrative in (17a) is a determiner marked by
falling tone, while the demonstrative identifier in (17b) carries a rising tone.
Like adnominal and identificational demonstratives, pronominal demonstratives
may cliticize to an element in their environment. The following example
from Kawaiisu (Numic) shows a pronominal demonstrative that is attached to the
first word of the clause. The pronominal enclitics are usually translated by third
person pronouns, but since they are marked for distance (=ina 'proximal' vs.
=ana 'distal') one might consider them demonstratives (cf. Zigmond et al.
1991: 47-48).
(18) Kawaiisu (Zigmond 1991: 172)
mee-gG-pGgadG=ina una wiigara
say-BEN-PERF=DEM.PROX.SG.ANIM that red.racer
'Red Racer said to her.'
24 DEMONSTRATIVES
Unlike adnominal, pronominal and identificational demonstratives, adverbial demonstratives
are always unbound. I am at least not aware of any language in
which adverbial demonstratives are clitics.5
While demonstratives may cliticize to an element in their environment, they
are probably never bound to a specific word. Some of the sources that I consulted
refer to bound demonstratives as suffixes (e.g. Creider and Tapsubei Creider
1989: 40; Noonan 1992: 86), but they seem to use the term suffix in a broad
sense subsuming all bound forms including enclitics. Among other things,
suffixes and enclitics differ in their syntactic behavior: while suffixes are
associated with a specific word, enclitics are attached to a phrase (cf. Zwicky
1977; Klavans 1985; Anderson 1992: 198-223).6 Although the distinction between
suffixes and enclitics is theoretically well defined, it is empirically often difficult
to decide whether an (adnominal) demonstrative is bound to a word or a phrase.
The distinction is especially problematic if adjectives and other noun modifiers
precede the noun. In such a case, it is usually impossible to determine whether
a bound demonstrative is a suffix of the noun or an enclitic of the noun phrase.
If, on the other hand, the modifiers follow the noun, it is immediately obvious
whether a bound demonstrative is used as an enclitic or suffix. Consider, for
instance, the following examples from Lango.
(19) Lango (Noonan 1992: 155, 155, 156)
a. gwók=k'I
dog=this
'this dog'
b. gwôkk à dwó]=]'I
dog ATT big.SG=this
'this big dog'
c. gwóggî à d'f]'f àryf'=n'I
dogs ATT big.PL two-this
'these two big dogs'
The demonstratives in (19a-c) attach to the last free form of the noun phrase: in
(19a) the demonstrative follows a noun, in (19b) it follows an adjective, and in
(19c) a numeral. Noonan (1992: 86) refers to the adnominal demonstratives in
Lango as suffixes, but these examples show that they are enclitics of the noun
phrase rather than suffixes of the noun (as Noonan points out in a footnote). I
suspect that all demonstratives being referred to as suffixes in my sources are in
fact enclitics. Suffixes tend to be obligatory in the contexts in which they occur,
while enclitics are often interchangeable with unbound (stressed) forms. Since demonstratives
are commonly used to emphasize a referent, I hypothesize that
MORPHOLOGY 25
bound demonstratives can always be replaced by an unbound stressed form. If
this is correct, it would suggest that all bound demonstratives are clitics and that
demonstrative affixes do not exist.
2.2.2 The inflection of demonstratives
Most languages included in my sample have at least some demonstratives that are
inflected for gender, number and/or case. The inflectional features of demonstratives
vary with their syntactic function. Pronominal demonstratives are more
likely to inflect than adnominal and identificational demonstratives, which, in
turn, are more often inflected than adverbial demonstratives. Table 12 shows the
number and percentage of pronominal demonstratives that are inflected for
number, gender and case in my sample; a detailed overview of the inflectional
features of pronominal demonstratives is given in Appendix B.
There are 68 languages in my sample in which pronominal demonstratives are
Table 12. The inflectional features of pronominal demonstratives (cf. Appendix B)
Inflected Uninflected Number Gender Case Total
Number
Percentage
68%
80%
17%
20%
64%
75%
38%
45%
25%
30%
850%
100%
marked for gender, number, and/or case, and 17 languages in which pronominal
demonstratives are uninflected. In some of the latter, (pronominal) demonstratives
are always accompanied by a nominal such as a defective noun (e.g. Korean; cf.
2.1.4), a third person pronoun (e.g. Kusaiean; cf. 4.1.3), or a classifier (e.g. Nùng;
cf. 4.1.3). Such demonstratives are strictly speaking not pronominal. They are demonstrative
determiners that are embedded in a noun phrase. Demonstratives of
this sort will be discussed in Section 4.1.3.
Apart from such 'pronominal NPs', there are only ten languages in my
sample in which pronominal demonstratives are uninflected, namely Acehnese,
Izi, Koyra Chiini, Mulao, Oneida, Tuscarora, Tok Pisin, Urim, Urubu-Kaapor, and
Usan. In all of these languages, nouns are in general uninflected (or, at least, they
are not regularly inflected as in English). The demonstratives of these languages
behave therefore just like other nominals and can be considered pro-nouns even
though they lack the trappings of a typical pronoun.
The most common inflectional feature of pronominal demonstratives is
number followed by gender and case. There are sixty-four languages in my
sample in which pronominal demonstratives are inflected for number, thirty-eight
26 DEMONSTRATIVES
languages in which they are inflected for gender, and twenty-five languages in
which they are case-marked. The case endings of pronominal demonstratives are
usually the same as, or very similar to, the case endings of nouns, while the
gender and number features are either expressed by special endings or, more
frequently, by stem alternations (see below).
Like pronominal demonstratives, the majority of adnominal demonstratives
is uninflected. However, there is a significant number of languages in my sample
in which adnominal demonstratives are morphologically invariable while pronominal
demonstratives are inflected for gender, number, and/or case. In fact, if a
language employs adnominal demonstratives that are marked for gender, number
and/or case, one can predict that the pronominal demonstratives are marked for
the same feature(s). Table 13 lists some of the languages in my sample in which
adnominal demonstratives are uninflected while pronominal demonstratives are
marked for gender, number and/or case (note that some of the case endings might
be enclitics of the noun phrase rather than suffixes of the noun).
In all of the languages shown in Table 13, adnominal demonstratives cooccur
Table 13. Inflected DEM PROs vs. uninflected DEM DETs
Pronominal Adnominal
Ambulas
Byansi
Duwai
Epena Pedee
Hua
Kannada
Lezgian
Nama
Tauya
Turkish
number
number
number
number
number
number
number
number
gender
gender
gender
gender
case
case
case
case
case
case
case
Ø
Ø
Ø
Ø
Ø
Ø
Ø
Ø
Ø
Ø
Total: 8 4 7 10
with an inflected noun. The grammatical features of the noun phrase are thus
sufficiently marked by the inflectional endings of the noun. This is illustrated by
the following example from Lezgian.
MORPHOLOGY 27
(20) Lezgian (Haspelmath 1993: 259)
a insan-ar
that human-PL
'these people'
Like adnominal demonstratives, identificational demonstratives are often morphologically
invariable in languages in which pronominal demonstratives are marked
for gender, number and/or case. Table 14 shows that there are (at least) nine
languages in my sample in which pronominal demonstratives have certain
inflectional features while identificational demonstratives are morphologically
unmarked. Languages in which pronominal demonstratives are uninflected while
identificational demonstratives are marked for gender, number and/or case do not
occur in my sample.
Demonstrative adverbs are almost always uninflected. In Section 2.1.1, we saw
Table 14. Inflected DEM PROs vs. uninflected DEM IDENTs
Pronominal Identificational
Ambulas
Czech
Duwai
French
German
Margi
Pangasinan
Swedish
T. Shoshone
number
number
number
number
number
number
number
number
number
gender
gender
gender
gender
gender
case
case
case
Ø (kén PROX)
Ø (to)
Ø (n'6mù)
Ø (ce)
Ø (das)
Ø (]ú)
Ø (nía PROX)
Ø (det)
Ø (isün PROX)
Total: 9 5 3 9
that the semantic equivalents of 'here' and 'there' in Guugu Yimidhirr are demonstratives
with certain (locational) case endings, but since they belong to the
same paradigm as demonstratives that are used as independent pronouns and noun
modifiers, they cannot be classified as adverbs. Guugu Yimidhirr does not have
a particular class of demonstrative adverbs.
There are, however, a few languages in my sample in which demonstrative
adverbs are case-marked and categorially distinguished from demonstrative pronouns.
One of them is Hua, in which both demonstrative pronouns and adverbs
are inflected for case. Demonstrative adverbs consist of two morphemes in Hua,
a demonstrative root and one of four locative case markers, while demonstrative
28 DEMONSTRATIVES
pronouns are formed from three elements: a demonstrative root, a (non-locative)
case marker, and the suffix -bo', which Haiman (1980: 259) analyzes as a
nominalizer (see next section). Compare the following two forms:
(21) Hua (Haiman 1980: 259, 259)
a. ma-bo'-mamu'
PROX-NLZ-ERG
'this (one)'
b. ma-roga
PROX-LOC
'here/hither'
The demonstrative pronoun in (21a) consists of three morphemes: the demonstrative
root ma, the nominalizer -bo', and the case marker -mamu'. (21b) shows a
demonstrative adverb formed from two morphemes: a demonstrative root and a
locative case marker. Unlike the demonstrative pronoun in (21a), the demonstrative
adverb does not include the nominalizer -bo', and therefore it cannot be
classified as a demonstrative pronoun in locative case. Both pronominal and
adverbial demonstratives are case-marked in Hua, but they are categorially
distinguished due to the absence and presence of -bo'.
2.2.3 Demonstrative stems
Having described the inflectional endings of demonstratives, I now examine the
formation of demonstrative stems. A demonstrative stem is a demonstrative
without its inflectional endings. It consists of a demonstrative root (i.e. a deictic
element) and possibly some other morpheme: a derivational affix or another free
form. In the following I concentrate on the stem formation of demonstrative
pronouns and adverbs; demonstrative determiners and identifiers will not be
considered because they have either the same stems as pronominal demonstratives
or they consist only of a demonstrative root.
In the literature, it has often been argued that demonstrative pronouns and
determiners are derived from demonstrative adverbs (cf. Anderson and Keenan
1985: 279; Greenberg 1985: 277; Himmelmann 1996: 246), but Brown (1985) and
Woodworth (1991) present evidence that challenges this view. Their studies show
that demonstrative adverbs are often morphologically more complex than demonstrative
pronouns and determiners, which seems to suggest that they are derived
from demonstrative pronouns or noun modifiers that combined with some other
morpheme. My data include examples that would support either one of these
hypotheses. I assume therefore that there is no unidirectional pathway leading
MORPHOLOGY 29
from demonstrative adverbs to demonstrative pronouns/determiners or vice versa.
Both developments seem to occur. Furthermore, it is conceivable that in some
languages demonstrative pronouns, determiners, and adverbs developed independently
of one another from a deictic particle with no specific syntactic function
(see below).
The stems of demonstrative pronouns are often formed from a demonstrative
root and a nominalizing affix, a third person pronoun, or a classifier. In the
previous section we saw that the stems of demonstrative pronouns in Hua are
formed from a demonstrative root and the suffix -bo'. Haiman (1980: 259)
characterizes -bo' as a "nominalizing suffix", used to derive pro-nominals from
a demonstrative root. Similar stem formations occur in several other languages
in my sample. Lezgian, for instance, has two demonstrative roots, i 'proximal'
and a 'distal', which may function as noun modifiers (e.g. a insan-ar that human-
PL 'those people'; Haspelmath 1993: 259). The demonstrative pronouns are
formed by combining i and a with the suffix -da 'ergative/oblique' (-di 'absolutive'),
which is otherwise used to derive nouns from adjectives. Haspelmath
(1993: 110) characterizes the demonstrative pronouns in Lezgian as "substantivised
forms":
(22) Lezgian (Haspelmath 1993: 110, 111)
a. -da deriving a noun from an adjective
˜qacu > ˜qacu-da
'green' '(the) green one'
b. -da deriving a DEM pronoun from a DEM root
i > i-da
'this' 'this one'
In other languages, demonstrative pronouns are composed of a demonstrative root
and a third person pronoun. Consider, for instance, the demonstrative in (23) from
Margi, which consists of the third person pronoun nàjà, the plural marker -'yàr,
and the demonstrative root -tà.
(23) Margi (Hoffmann 1963: 86)
ná1à nàjà-'yàr-tà
give.me 3-PL-that
'Give me those!'
Apart from Margi, there are six other languages in my sample in which demonstrative
pronouns are formed by combining a demonstrative root with a third
person pronoun: Acehnese, Ambulas, Ao, Barasano, Kokborok, and Khasi. In
some of these languages, the combination of a demonstrative root and a third
30 DEMONSTRATIVES
person pronoun is optional (e.g. Acehnese), but in others it is obligatory (e.g.
Ao). Table 15 shows the demonstrative pronouns in Ao. They are formed from
two demonstrative roots, ya 'proximal' and ci 'distal', and one of five third
person pronouns: pá 'he', lá 'she', ipá 'it', pa»lhnok 'they (human)', and item
'they (non-human)':
The demonstratives in Table 15 can be viewed as '(grammatical) compounds'
Table 15. Demonstrative pronouns in Ao (Gowda 1975: 34)
PROXIMAL DISTAL
SG.M.HUMAN
SG.F.HUMAN
SG.NONHUMAN
PL.HUMAN
PL.NONHUMAN
pá-ya
lá-ya
ipá-ya
pa»lhnok-ya
item-ya
pá-ci
lá-ci
ipá-ci
pa»lhnok-ci
item-ci
formed from two independent words. Since they are only slightly different from
demonstratives with inflectional endings, I suspect that third person pronouns are
a common historical source for gender and number markers on demonstratives.
The grammaticalization path that I suggest is schematized in (24a-b).
(24) a. 3.PRO DEM > 3.PRO-DEM > AFFIX-DEM
b. DEM 3.PRO > DEM-3.PRO > DEM-AFFIX
At the initial stage of the grammaticalization process, demonstratives and third
person pronouns are two independent forms that commonly cooccur. Such
pronominal NPs occur, for instance, in Kusaiean (cf. 4.1.3). At the next stage, demonstratives
and third person pronouns coalesce and form a complex word
consisting of two elements that are still used as independent words in other
contexts. This stage is exemplified by the demonstratives in Ao in Table 15. At
the final stage, demonstratives and third person pronouns must cooccur and can
no longer be used in isolation. At this point, the third person pronouns have
basically assumed the function of gender/number markers.7
Finally, pronominal demonstratives are quite frequently formed from a demonstrative
root and a classifier, which may evolve into a noun class marker.
Examples from Mandarin Chinese, Yagua and Barasano are shown in (25) to
(27), respectively.
MORPHOLOGY 31
(25) Mandarin Chinese (Li and Thompson 1981: 533)
nèi-ge
that-CLASS
'that (one)'
(26) Yagua (Payne and Payne 1990: 374)
jiy-nù
this-CLASS:ANIM:SG
'this (one)'
(27) Barasano (Jones and Jones 1991: 57)
ti-a-re
that-CLASS-OBJ
'that (one)'
The stems of demonstrative adverbs are usually distinguished from the stems of
demonstrative pronouns, determiners and identifiers. They are often formed from
a demonstrative root and a locative or directional affix, as in the following
example from Kiowa:
(28) Kiowa (Watkins 1984: 189)
é˛-dé Ø-df'-+Რg'f ó-y
here-toward 3SG-move-come and over.there-widely.bounded
Ø-ph '˛ f
3SG-stop
'He was coming here and stopped over there.'
Example (28) includes two demonstrative adverbs: one is marked by the locative
suffix -y 'location.widely.bounded' and the other by the directional marker -dé
'toward speaker'. The corresponding demonstrative pronouns/determiners are built
on the same demonstrative roots ( ˛ é-'proximal' and ó- 'distal'), but they combine
with other suffixes.
Similar demonstrative adverbs occur in many other languages in my sample.
Demonstrative adverbs in Yimas, for instance, are formed from three deictic
roots, k 'proximal', m 'medial', and n 'distal', and a locative affix. As shown in
Table 16, the proximal and distal roots combine with the prefix ta-, while the
medial form occurs with the suffix -nti.
Other languages form demonstrative adverbs from a demonstrative root and
a noun meaning 'place'. For instance, in Section 2.1.4 we saw that the demonstrative
adverbs in Korean developed from a demonstrative particle and the defective
noun eki 'place' (cf. 29).
32 DEMONSTRATIVES
(29) Korean (Sohn 1994: 296)
Table 16. Demonstrative adverbs in Yimas (Foley 1991: 114)
PROXIMAL
MEDIAL
DISTAL
ta-k
m-nti
ta-n
yeki < i eki
'here (NEAR S)' 'this place'
keki < ku eki
'there (NEAR H)' 'that place'
ceki < ce eki
'there (AWAY FROM S+H)' 'that place'
Similar demonstrative adverbs seem to occur in Kokborok (cf. Gowda 1975: 33).
2.3 Summary
In this chapter, I first described the demonstrative systems of four individual
languages and then I examined the morphological features of demonstratives more
systematically. The major findings of my investigation can be summarized as follows:
1. Though the demonstratives of most languages are independent words, there
are some languages in which certain demonstratives may cliticize to an
element in their environment.
2. The occurrence of demonstrative clitics is largely restricted to adnominal demonstratives;
pronominal, adverbial and identificational demonstratives are
almost always free forms.
3. Though (adnominal) demonstratives may cliticize to an adjacent element,
they are probably never bound to a specific word; that is, there are probably
no demonstrative affixes.
4. The inflectional features of demonstratives vary with their syntactic function:
pronominal demonstratives are more likely to inflect than adnominal and
identificational demonstratives, which, in turn, are more often inflected than
adverbial demonstratives. The latter are usually uninflected unless they occur
with a set of locational case markers.
MORPHOLOGY 33
5. The most frequent inflectional feature is number, followed by gender and
case.
6. In languages in which nouns are inflected for gender, number and/or case,
pronominal demonstratives are always marked for the same features whereas
adnominal and identificational demonstratives are often uninflected. In fact,
if adnominal and/or identificational demonstratives are inflected, one can
predict that the pronominal demonstratives of the same language occur (at
least) with the same inflectional features.
7. Though the stems of most demonstratives are monomorphemic, there are demonstratives
whose stems are composed of multiple morphemes. The stems
of demonstrative pronouns may consist of a deictic root and a nominalizer,
a third person pronoun or a noun classifier, and the stems of demonstrative
adverbs may be formed from a deictic root and a locative/directional affix
or a locational noun. Demonstrative determiners and identifiers have either
the same stems as demonstrative pronouns or their stems consist solely of
a deictic root.
CHAPTER 3
Semantics
The previous chapter examined the morphological properties of demonstratives.
This chapter investigates their meaning. The meaning of demonstratives comprises
two kinds of features: (i) deictic features, which indicate the location of the
referent relative to the deictic center, and (ii) qualitative features, which characterize
the referent (cf. Lyons 1977: 648; cf. also Fillmore 1982; Rauh 1983; Hanks
1989, 1990). The deictic features indicate whether the referent is near or removed
from the deictic center, whether it is at a higher or lower elevation, uphill or
downhill, or whether it is moving toward or away from the deictic center. These
features are primarily encoded by demonstrative roots. The qualitative features
provide classificatory information about the referent. They indicate, for instance,
whether the referent is animate or inanimate, female or male, or human or nonhuman.
These features are usually expressed by morphemes that attach to a demonstrative
root, but in some languages the root itself is classifying. The
following two sections investigate the deictic and qualitative features of demonstratives,
respectively. The final section provides an overview of all features -
semantic, pragmatic and syntactic - that are commonly encoded by demonstratives.
3.1 The semantic features of demonstratives
3.1.1 Deictic features
Demonstratives are deictics. Deictic expressions are linguistic elements whose
interpretation makes crucial reference to some aspect of the speech situation. As
Levinson (1983: 54) puts it, "deixis concerns the ways in which languages encode
or grammaticalize features of the context of utterance or speech event, and thus
also concerns ways in which the interpretation of utterances depends on the
analysis of that context". Deictic expressions are traditionally divided into three
semantic categories: person, place and time (cf. Bühler 1934: 102). Person deixis
36 DEMONSTRATIVES
comprises the personal pronouns I and you, which denote the speech participants;
place deictic expressions refer to objects, locations or persons (apart from the
speech participants); and time deictic expressions indicate a temporal reference
point relative to the time of the speech event. Demonstratives are place (or
spatial) deictics. They indicate the relative distance of an object, location or
person vis-à-vis the deictic center (also called the origo), which is usually
associated with the location of the speaker.
In addition to person, place and time deixis, Levinson (1983: 61-96) discusses
two other deictic categories: "social deixis", which concerns the social status
of the speech participants (for which some languages employ honorifics), and
"discourse deixis", which applies to deictic elements that refer to aspects of the
surrounding discourse (cf. Fillmore 1997: 103-125). In Chapter 5, I show that demonstratives
are often used as discourse deictics, which can be seen as an
extension of their primary use as spatial deictics.
All languages have at least two demonstratives locating the referent at two
different points on a distance scale: a proximal demonstrative referring to an
entity near the deictic center, and a distal demonstrative indicating a referent that
is located at some distance to the deictic center. English, for instance, has such
a two-term deictic system, consisting of the proximal demonstratives here and
this and their distal counterparts there and that. There are many other languages
in my sample that have a two-term deictic system. Consider, for instance, the demonstratives
in Table 17 from Vietnamese. Like English, Vietnamese has two demonstrative
pronouns/determiners and two demonstrative adverbs. Note that the
proximal and distal forms of the demonstrative pronouns/determiners are only
distinguished by tone; they have the same segmental features.
Table 17. Demonstratives in Vietnamese (Thompson 1965: 142)
DEM PROs/DETs DEM ADVs
PROXIMAL
DISTAL
này
no»
dây
day
Both English and Vietnamese indicate the contrast between proximal and distal
referents through different demonstrative roots. Alamblak expresses the same
contrast through bound morphemes that attach to a demonstrative root unmarked
for distance.
As shown in Table 18, the deictic suffixes -ar 'proximal' and -ur 'distal' are
not obligatory to form pronominal/adnominal demonstratives in Alamblak. Bruce
SEMANTICS 37
(1984: 81-82) does not explain when and why a demonstrative root is used
Table 18. Demonstrative pronouns/determiners in Alamblak (Bruce 1984: 81)
NEUTRAL PROXIMAL DISTAL
SG.M
SG.F
DU
PL
Gnd-r
Gnd-t
Gnd-f
Gnd-m
Gnd-ar-r
Gnd-ar-t
Gnd-ar-f
Gnd-ar-m
Gnd-ur-r
Gnd-ur-t
Gnd-ur-f
Gnd-ur-m
without a distance marker, but there are many examples in his grammar in which
Gnd does not occur with -ar or -ur. In such a case, Gnd is often translated by a
definite article, but it is always glossed as 'DEM'. It seems that Gnd serves a
pragmatic function that is somewhere in between a definite article and a demonstrative.
Like Alamblak, French uses two bound morphemes, ci 'proximal' and là
'distal', to indicate the relative distance of the referent to the deictic center. The
demonstrative roots themselves are distance-neutral. Table 19 shows that ci and
là are attached either to a demonstrative pronoun or to a noun that is preceded
by a demonstrative determiner.
The distance markers ci and là are usually obligatory to form a demonstrative
Table 19. Demonstrative pronouns/determines in French
DEM PROs DEM DETs
PROXIMAL DISTAL PROXIMAL DISTAL
SG.M
SG.F
PL.M
PL.F
celui-ci
celle-ci
ceux-ci
celles-ci
celui-là
celle-là
ceux-là
celles-là
ce livre-ci
cette maison-ci
ces livres-ci
ces maisons-ci
ce livre-là
cette maison-là
ces livres-là
ces maisons-là
pronoun; they can only be omitted if celui, celle, ceux, or celles are modified by
a relative clause or a prepositional phrase (cf. Calvez 1994: 62). The demonstrative
determiners ce, cette and ces, on the other hand, are frequently used without
ci or là. Harris (1978, 1980) argues that ce, cette and ces can be viewed as
definite articles rather than demonstratives when they occur without a distance
marker. This raises the interesting question whether distance is a necessary
feature of the category demonstrative. Are demonstratives generally marked for
38 DEMONSTRATIVES
distance or are there reasons to consider an item a demonstrative even if it does
not indicate the relative distance of its referent to the deictic center? Anderson
and Keenan (1985: 280) argue that a deictic expression unmarked for distance
"would be little different from a definite article" or third person pronoun (cf. Frei
1944: 119). In their view, demonstratives are generally distance-marked.
Himmelmann (1997: 53-62) takes a different view. He argues that demonstratives
do not always encode a deictic contrast. His hypothesis is primarily
based on data from colloquial German. There are two expressions in colloquial
German that one might consider demonstratives: dies and stressed das. Dies is
almost always used adnominally, but das can be both an independent pronoun and
a modifier of a cooccurring noun. Dies and das do not contrast deictically: both
forms may occur with proximal and distal meaning. In order to indicate that dies
or das are used contrastively, they are commonly accompanied by a demonstrative
adverb (e.g. das da 'this/that there', das Haus da 'this/that house there').
Himmelmann argues that at least one of these forms, dies, functions as a demonstrative.
8 He shows that dies serves the same pragmatic functions as demonstratives
that are deictically contrastive. Like this and that in English, dies focuses
the hearer's attention on entities in the speech situation, often in combination with
a pointing gesture. Since definite articles and third person pronouns do not
function to orient the hearer in the surrounding situation, dies must be a demonstrative
despite the fact that it does not encode a deictic contrast. Similar demonstratives
seem to occur in other languages. Supyire, for instance, has only one
series of demonstratives which, according to Carlson (1994: 160), is "used with
both proximal and distal meaning". I assume therefore, with Himmelmann, that
demonstratives are not generally distance-marked. Some languages have demonstratives
that do not indicate a deictic contrast. The occurrence of distance-neutral
demonstratives is, however, crosslinguistically infrequent. Apart from German and
Supyire there are only five other languages in my sample in which some demonstratives
are distance-neutral: Alamblak, French, Czech, Koyra Chiini, and Tok
Pisin. Moreover, even though pronominal and adnominal demonstratives are not
always deictically contrastive, adverbial demonstratives are generally distancemarked
(cf. Himmelmann 1997: 49). All eighty-five languages included in my
sample have at least two adverbial demonstratives that indicate a deictic contrast.
The occurrence of distance-neutral demonstratives is thus restricted to certain
syntactic contexts. All languages employ at least some demonstratives that are
distance-marked, and, as the examples from Alamblak, French, and German have
shown, distance-marked demonstratives are often used to reinforce demonstratives
that are distance-neutral (also in Koyra Chiini; cf. Heath 1999: 61). Distance is
SEMANTICS 39
thus after all a feature that occurs in the demonstrative system of all languages
even though individual elements of the system may lack a distance feature.
Turning to languages with three deictic terms, one has to distinguish between
systems in which the middle term refers to a location in medial distance relative
to the deictic center, and systems in which the middle term denotes a referent
close to the hearer. Anderson and Keenan (1985: 282-286) refer to these two
systems as distance-oriented and person-oriented systems, respectively (cf.
Fillmore 1982: 49-50). Spanish, for instance, has a distance-oriented system,
consisting of the demonstratives este 'proximal', ese 'medial' and aquel 'distal'
(Anderson and Keenan 1985: 283-5), while Japanese has a person-oriented
system, in which the middle terms (based on the deictic root so-) refer to a
location near the hearer: sore 'that (near hearer), soko 'there (near hearer)' etc.
(cf. Kuno 1973; Imai 1996). Two other examples from Yimas and Pangasinan are
shown in Table 20 and 21, respectively.
The demonstratives in Yimas are built on three demonstrative roots: k 'proximal'
Table 20. Demonstrative pronouns/determiners in Yimas (Foley 1991: 112)
SG DU PL
PROXIMAL
MEDIAL
DISTAL
p-k
m-n
p-n
pla-k
mpl
pla-n
pia-k
m-ra
pia-n
Table 21. Demonstrative pronouns in Pangasinan (Benton 1971: 88)
SG PL
NEAR S
NEAR H
AWAY FROM S+H
(i)yá
(i)tán
(i)mán
(i)rá-ya
(i)rá-tan
(i)rá-man
and n 'distal', which take number prefixes, and m 'medial', which takes number
suffixes. They form a distance-oriented system, while the demonstratives in
Pangasinan are person-oriented: the middle term tan refers to a location near the
hearer. Anderson and Keenan (1985) point out that in both distance-oriented and
person-oriented systems the middle term is often the preferred form for anaphoric
reference.
Not every deictic system that includes three deictic terms is either a distance
40 DEMONSTRATIVES
or a person-oriented system. Nama, for instance, has three deictic terms which,
according to Anderson and Keenan (1985: 285-286), are basically used as a
variant of a two-term system. Nama uses the demonstrative nee to indicate a
referent near the deictic center, and it uses nãá ( is a click) in order to refer
to objects or persons that are not included in the domain that is conceptualized
as the deictic center. However, the latter term is only used in "neutral deictic
settings" and it does not occur in contrast to nee. That is, nee and nãá are never
used within the same construction to indicate that one of two referents is closer
to the deictic center than the other. In order to express a deictic contrast between
two referents, Nama employs a third demonstrative, náú 'distal', which is used
only in contrast either to nee or nãá; it never occurs in sentences without one
of the other two forms. Examples are given in (1a-b).
(1) Nama (Anderson and Keenan 1985: 286, 286)
a. nee kxòep tsı˜í náú kxòep
this man and that man
'this man and that one'
b. nãá kxòep tsı˜í náú kxòep
that man and that (other) man
'that man and that other one'
Returning to the contrast between distance and person-oriented systems, it is
interesting to note that distance-oriented systems tend to have fewer deictic terms
than person-oriented systems. A distance-oriented system is usually confined to
three deictic terms. Anderson and Keenan (1985: 286-295) report languages
having four, five or even more demonstratives distinguished by pure distance, but
such systems do not occur in my data. Based on the languages in my sample, I
would support Fillmore (1982: 48-9), who maintains that "that there are never
really more than three [distance categories]" and that all larger systems either
involve the hearer as a point of reference or other deictic dimensions such as
visibility or elevation.
Unlike distance-oriented systems, person-oriented systems may involve four
deictic terms. In Section 2.1.3 we saw, for instance, that Ewondo has four
adverbial demonstratives: vá 'near speaker', vála¯ 'near hearer', válí 'away from
speaker and hearer', and álí 'far away from speaker and hearer'. The corresponding
pronominal and adnominal demonstratives have only three deictic terms in
Ewondo, but as shown in Table 22, Quileute uses four distance categories
throughout the entire deictic system.
Note that the location of the hearer is only relevant to the first and second
terms within this system. The third and fourth distance categories relate the
SEMANTICS 41
referent to a domain that includes both speaker and hearer. In other words, the
Table 22. Demonstratives in Quileute (Andrade 1933: 246, 252)
DEM PROs/DETs
NON-FEM FEM
DEM ADVs
NEAR S
NEAR H
NEAR S+H
AWAY FROM S+H
yü'x» o
yi'tca
sa''a
ha
yü'ko
yi'tca
ksa'
ha
xo''a
so''o
sa''a
átca'a
deictic center is conceptualized in two different ways in this system: it is
conceptualized as the sole domain of the speaker (excluding the hearer) for the
first and second deictic terms (i.e. NEAR S and NEAR H); and it is conceptualized
as the common domain of speaker and hearer for the two other terms (i.e. NEAR
S+H and AWAY FROM S+H). None of the demonstratives in this system involves
more than three reference points: (i) the referent, (ii) the deictic center, and,
depending on the term, (iii) another reference object. A person-oriented system
including four deictic terms is thus basically a variant of a three-term deictic
system with an additional category for referents near the hearer.9
In addition to distance, demonstratives may indicate whether the referent is
visible or out of sight, at a higher or lower elevation, uphill or downhill, upriver
or downriver, or moving toward or away from the deictic center. Fillmore
(1982: 51) considers these features non-deictic, but based on the information that
I have gathered from my sources I would contend that they have a deictic
character. Like the features 'proximal' and 'distal', these features are deictic
because they indicate the location of the referent relative to the deictic center: the
referent is 'out of sight' or 'downriver' from the perspective of the speaker (or
some other reference point to which the deictic center has been shifted). It is
conceivable that some of the morphemes that encode these feature can also be
used non-deictically (e.g. uphill from the village), but in all of the sources that I
consulted their meaning is described in such a way that it involves the speaker
as the unmarked point of reference. This suggests that these features are deictic
and that the non-deictic use of these morphemes (if it exists) is due to pragmatic
extensions. In the remainder of this section, I discuss examples of demonstratives
that encode features of such semantic categories as visibility or elevation that
receive a deictic interpretation.
Visibility is a common deictic category in Native American languages. In my
sample, there are seven American Indian languages that have particular demon42
DEMONSTRATIVES
strative forms for invisible referents: West Greenlandic, Halkomelem, Quileute,
Passamaquoddy-Maliseet, Tümpisa Shoshone, Ute, and Epena Pedee. Table 23
shows the demonstrative pronouns/determiners in Ute (only the subject forms are
shown). Two of the three demonstratives in this table are distinguished by pure
distance while the third term is used for referents out of sight.
Most languages in which visibility is a feature of the deictic system have a single
Table 23. Demonstratives in Ute (Givón 1980: 55)
PROXIMAL DISTAL INVISIBLE
INANIMATE
SG.ANIMATE
PL.ANIMATE
Áíca
Áína
Áímuí
máruí
máa
mámuí
Áúru
Áú
Áúmuí
deictic term to indicate a referent out of sight, but Quileute has three: one for
referents nearby (which may be partly visible), one for referents whose location
is known, and one for referents whose location is unknown (cf. Andrade
1933: 252). Table 24 shows the three forms of the invisible demonstrative adverbs
in this language.
Another deictic dimension that is relatively frequent in my sample is elevation.
Table 24. Invisible demonstrative adverbs in Quileute (Andrade 1933: 252)
NEAR (maybe partly visible)
KNOWN PLACE
UNKNOWN PLACE
x» a'x» e
tci''tc'
xu'xwa'
Nine languages have demonstratives that indicate whether the referent is at a
higher or lower elevation relative to the deictic center. These demonstratives
occur in languages spoken in New Guinea (Usan, Hua, Tauya) in the Himalayan
area (Lahu, Khasi, Byansi), in Australia (Dyirbal, Ngiyambaa), and in the
Caucasus (Lezgian). Table 25 shows the deictic system in Khasi, which is based
on six demonstrative roots: three of them locate the referent on a distance scale,
two others indicate a referent at a higher or lower elevation, and one refers to
objects or locations out of sight. The demonstrative roots are either combined
with personal pronouns or with adpositions.
A similar system is employed in Lahu, which has five demonstrative
adverbs: three of them are distinguished by (pure) distance and the other two
SEMANTICS 43
indicate whether the referent is above or below the deictic center. The demonstra-
Table 25. Demonstratives in Khasi (Nagaraja 1985: 11-12; Rabel 1961: 67)
DEM PROs DEM ADVs
M.SG (u 'he') F.SG (ka 'she') PL (ki 'they') (ša 'to')
PROXIMAL
MEDIAL (NEAR H)
DISTAL
UP
DOWN
INVISIBLE
u-ne
u-to
u-tay
u-tey
u-thie
u-ta
ka-ne
ka-to
ka-tay
ka-tey
ka-thie
ka-ta
ki-ne
ki-to
ki-tay
ki-tey
ki-thie
ki-ta
ša-ne
ša-to
ša-tay
ša-tey
ša-thie
ša-ta
tives in Lahu are monomorphemic, they are uninflected and do not combine with
any other morpheme.
While Khasi and Lahu employ particular demonstrative roots in order to refer to
Table 26. Demonstrative adverbs in Lahu (Matisoff 1973: 110-1)
PROXIMAL
MEDIAL
DISTAL
UP
DOWN
chò
ô
cô
nô
mô
entities at different elevations, Tauya expresses the same deictic contrast through
prefixes that are attached to a demonstrative root: pise- refers to a location above
the deictic center, and tofe- indicates a referent at a lower elevation.
Like Tauya, Dyirbal indicates the horizontal location of a referent through bound
Table 27. Elevation in Tauya (MacDonald 1990: 102)
ABOVE BELOW
PROXIMAL
DISTAL
pise-me
pise-e
tofe-me
tofe-e
morphemes: -gali 'down' and -gala 'up' (cf. Dixon 1972: 48). In addition,
44 DEMONSTRATIVES
Dyirbal has also a series of bound forms that indicate whether the referent is
uphill or downhill from the perspective of the speaker:
Dixon treats the forms in Table 28 as single morphemes, but Anderson and
Table 28. Downhill and uphill in Dyirbal (Dixon 1972: 48)
SHORT DISTANCE DOWNHILL
MEDIUM DISTANCE DOWNHILL
LONG DISTANCE DOWNHILL
SHORT DISTANCE UPHILL
MEDIUM DISTANCE UPHILL
LONG DISTANCE UPHILL
-baydí i
-baydí a
-baydí u
-dayi
-daya
-dayu
Keenan (1985: 292) break them down into two elements: -baydí and -day, which
encode the geographical features 'downhill' and 'uphill', and the word final
vowels, -i, -a and -u, which indicate the distance features: 'short', 'medium' and
'long' (i.e. 'proximal', 'medial' and 'distal'). The forms in Table 28 are part of
a more complex system in which the feature 'hill' contrasts with the feature
'river', encoded by the following forms:
According to Dixon (1972: 48), "'river' is the marked feature in the system 'river
Table 29. Downriver and upriver in Dyirbal (Dixon 1972: 48)
MEDIUM DISTANCE DOWNRIVER
LONG DISTANCE DOWNRIVER
MEDIUM DISTANCE UPRIVER
LONG DISTANCE UPRIVER
ACROSS THE RIVER
-balbala
-balbulu
-dawala
-dawulu
-guya
versus hill'". Like the 'hill' suffixes, the 'river' suffixes consist of two morphemes,
-balb 'downriver' and -daw 'upriver', and the distance markers, -ala
'medial' and -ulu 'distal'.10
Geographical features such as uphill and downhill or upriver and downriver
are crosslinguistically uncommon. Apart from Dyirbal, there are only two other
languages in my sample in which they occur: Hua and West Greenlandic. Like
Dyirbal, Hua has demonstratives which indicate whether the referent is uphill or
downhill. They distinguish two degrees of distance: buga refers to a location a
short distance uphill; biga indicates a location a long distance uphill; muna refers
SEMANTICS 45
to a location a short distance downhill; and mina indicates a referent a long
distance downhill.
West Greenlandic has two demonstratives that refer to objects or locations along
Table 30. Uphill and downhill in Hua (Haiman 1980: 258)
UPHILL DOWNHILL
SHORT DISTANCE
LONG DISTANCE
buga
biga
muna
mina
the coastline: anna indicates a referent to the north along the coastline from the
perspective of the speaker, and qanna refers southwards relative to the location
of the speaker. The deictic system of West Greenlandic is especially complex.
Apart from anna and qanna, it includes three demonstrative roots that are marked
for pure distance: ma(t)- 'proximal', u(a)- 'medial', and ik- 'distal'; four demonstrative
roots that indicate distance and elevation: kat-/kan- 'down a medial
distance', sam- 'down a long distance', pik- 'up a medial distance', and pav- 'up
a long distance'; and two demonstrative roots that refer to an object or location
that is either "outside beyond a wall" or "on the other side (interior or exterior)
of some intervening surface, usually a wall or door": qam- 'interior/exterior' and
kig- 'exterior' (Fortescue 1984: 260). In addition, there is an archaic form used
for referents out of sight: im- 'invisible'. Table 31 shows the twelve demonstrative
roots employed in West Greenlandic combined with an absolutive case
marker.
All demonstratives that we have seen thus far in this section indicate a
stationary referent. In some languages, demonstratives are also used to indicate
that the referent is moving in a certain direction relative to the deictic center.
Movement (or direction) is often expressed by bound morphemes that attach to
a demonstrative stem. For instance, Nunggubuyu has three "kinetic suffixes"
(Heath 1980: 152) that indicate whether the referent is moving (i) toward the
speaker, (ii) away from the speaker, or (iii) across the speaker's line of vision.
The examples in (2a-b) illustrate the use of the directional markers. They are
suffixed to demonstrative identifiers in these examples, but they also occur with
demonstrative pronouns/determiners and adverbs (Heath 1984: 281-291).
46 DEMONSTRATIVES
(2) Nunggubuyu (Heath 1980: 152, 152, 152)
Table 31. Demonstrative roots in West Greenlandic (Fortescue 1984: 259-262)
DEM.ABS ROOT
DISTANCE
ELEVATION
IN⁄EXTERIOR
COASTLINE
VISIBILITY
PROXIMAL
MEDIAL
DISTAL
DOWN.MEDIAL
DOWN.DISTAL
UP.MEDIAL
UP.DISTAL
IN⁄OUT
OUTSIDE
IN.THE.NORTH
IN.THE.SOUTH
INVISIBLE
manna
una
innga
kanna
sanna
pinnga
panna
qanna
kinnga
anna
qanna
inna
ma(t)-
u(a)-
ikkat-/
kansampikpavqamkigavqavima.
yuwa-gi-Ála
DISTAL-NC-TOWARD.S
'There he/she comes.'
b. yuwa-gi-Áli
DISTAL-NC-AWAY.FROM.S
'There he/she goes away.'
c. yuwa-gi-yaj
DISTAL-NC-ACROSS
'There he/she goes across.'
Similar demonstratives occur in several other languages in my sample. Kiowa, for
instance, has three directional markers that have the same meaning as the kinetic
affixes in Nunggubuyu: -dé indicates a referent moving toward the deictic center;
-p marks referents that are moving away from the deictic center; and -pé attaches
to demonstratives that refer to entities moving across the visual field of the speaker.
(3) Kiowa (Watkins 1984: 189, 189, 190)
a. f'-dé Ø-df'-+áî
there-toward 3SG-move-come
'He was coming here (toward me from far away).'
b. f'-p Carnegie-kù Ø-df'-+hf'
there-away Carnegie-to 3SG-move-go
'He went off there toward Carnegie.'
SEMANTICS 47
c. f'-pé Ø-áî
there-along 3SG-come.PAST
'She came along there.'
Finally, Inuktitut has a deictic prefix which indicates that the deictic center has
been shifted from the speaker to another person in the speech situation. Demonstratives
that are marked by this prefix refer to a location relative to the person to which
the deictic center has been shifted. Compare the following two examples:
(4) Inuktitut (Denny 1982: 362, 362)
a. pik-unga
up.there-to
'up there from my perspective' (speaker's perspective)
b. ta-ik-unga
SHIFT-up.there-to
'up there from your/his/her/their perspective(s)'
The demonstrative in (4a) consists of a deictic root and a locative marker; it
refers to a location that is 'up there' from the perspective of the speaker. The demonstrative
in (4b) occurs with the same locative marker, but in addition it is
marked by the "field shifting prefix" ta- (Denny 1982: 362). Ta- indicates that the
deictic center has been transferred from the speaker to another person so that the
referent of the demonstrative is 'up there' from the perspective of the person to
which the deictic center has been shifted.
3.1.2 Qualitative features
In addition to deictic information, demonstratives usually provide some qualitative
information about the referent. They may indicate, for instance, whether the
referent is animate or inanimate or whether it is a single entity or a set. I have
divided the qualitative features into six categories: (i) ontology, (ii) animacy, (iii)
humanness, (iv) sex, (v) number, and (vi) boundedness. I will discuss these
categories in turn.
The category of ontology subsumes two semantic features which indicate
whether a demonstrative refers to a location or to an object or person. In most
languages, demonstrative adverbs can only refer to a location while demonstrative
pronouns are used to indicate a person or object. In other words, the semantic
distinction between locational and non-locational referents corresponds rather
closely with the categorial distinction between demonstrative adverbs and
pronouns.11
48 DEMONSTRATIVES
The categories animacy, humanness and sex overlap to some extent: a demonstrative
that indicates a human referent presupposes, for instance, that the
referent is also animate. However, since animacy, humanness and sex are not
synonymous, they must be kept separate.
Animacy distinctions are encoded by the demonstratives in several American
Indian languages in my sample (e.g. Apalai, Barasano, Hixkaryana, Passamaquoddy-
Maliseet, Ute). Table 32 shows the animate and inanimate demonstratives
in Apalai.
One of the few languages in which demonstratives are marked for humanness in
Table 32. (In)animate demonstratives in Apalai (Koehn and Koehn 1986: 95)
ANIMATE INANIMATE
NON-COLL COLL NON-COLL COLL
PROXIMAL
MEDIAL
DISTAL
mokyro
moky
moxiamo
mokaro
mokamo
seny
mony
senohne
morohne
monohne
my sample is Burushaski, an isolate language spoken in Pakistan. Burushaski
indicates humanness and animacy through distinct demonstrative roots. In
addition, it expresses sex as a secondary feature through certain case suffixes.
Table 33 shows the nominative/accusative, genitive and dative forms of the
proximal demonstrative pronouns. The nominative/accusative forms are unmarked
for sex, but the genitive and dative forms indicate the sex features of the referent
indirectly through their case endings.
In addition to animacy, humanness and sex, demonstratives may be marked for
Table 33. Demonstrative pronouns in Burushaski (Lorimer 1935: 141)
HUMAN
ANIMATE
MASC
HUMAN
ANIMATE
FEM
NON-HUMAN
ANIMATE
(+few others)
NON-HUMAN
INANIMATE
NOM/ACC
GEN
DAT
kin7/kin
kin7
kin7r/kin6r
kin7/kin
kin7mo
kin7m~r
g~s7
g~s7
g~s7r
g~t7
g~t7
g~t7r
number. Number is the most frequent non-deictic category of the demonstratives
SEMANTICS 49
in my sample (cf. 2.2.2). Most languages distinguish between singular and plural
forms only, but some languages have also a dual (cf. Ambulas 2.1.2). Table 34
shows that the demonstratives in Wardaman have as many as four number
features: (i) singular, (ii) dual, (iii) plural, and (iv) collective:
The last semantic category to be discussed in this section is boundedness (for a
Table 34. Demonstrative pronouns/determiners in Wardaman (Merlan 1994: 139)
PROXIMAL MEDIAL DISTAL
ABS.SG
ABS.DU
ABS.PL
ABS.COLL
dana
dan-guya
dan-mulu
dan-ganung
nana
nan-guya
nan-mulu
nan-ganung
darni
dang-guya
dang-mulu
dang-ganung
general discussion of the category 'boundedness' see Talmy 1988: 178-80).
Boundedness is a central category of the deictic system in Inuktitut. Inuktitut has
two series of demonstratives: one indicates a restricted referent and the other is
specifically used to indicate an unbound or extended referent. The two series are
shown in Table 35.
The bound forms, which Denny (1982: 360) calls "restricted", refer to an object
Table 35. Demonstrative roots in Inuktitut (Denny 1982: 372)
RESTRICTED EXTENDED
PROXIMAL
DISTAL
UP.THERE
DOWN.THERE
IN.THERE
OUT.THERE
uvikpikkanqavkigmajavpagugqavqagor
location "whose entire extent is comprehensible to the eye in a single glance",
while the unbound forms, which Denny calls "extended", refer to objects and
locations "whose entire extent is not comprehensible in a single glance" (Denny
1982: 360). The choice between these forms is not determined by the objective
size of the referent; crucial is how the speaker conceptualizes the entity to which
s/he refers.
50 DEMONSTRATIVES
3.1.3 Summary: the semantic features of demonstratives
The two previous sections described the semantic features of demonstratives,
which were divided into two categories: (i) deictic features, which indicate the
location of the referent relative to the deictic center, and (ii) qualitative features,
which provide some classificatory information about the referent. The major
results of these two sections are summarized in 1 to 6:
1. All languages have at least two demonstratives that are deictically contrastive:
a proximal demonstrative referring to an entity near the deictic center
and a distal demonstrative indicating a referent that is located in some
distance to the speaker.
2. In some languages, pronominal, adnominal and/or identificational demonstratives
are distance-neutral, but adverbial demonstratives are always deictically
contrastive.
3. Deictic systems that involve more than two deictic terms can be divided into
distance-oriented systems, in which the deictic center is the only point of
reference for the location of the referent, and person-oriented systems, in
which, in addition to the deictic center, the location of the hearer serves as
another reference point.
4. Distance-oriented systems have usually not more than three deictic terms
while person-oriented systems may have up to four.
5. In addition to distance, demonstratives often encode a number of 'special'
deictic features: they may indicate, for instance, whether the referent is
visible or out of sight, at a higher or lower elevation, uphill or downhill,
upriver or downriver, or moving toward or away from the deictic center.
6. Apart from deictic information, demonstratives usually provide some
qualitative information about the referent: they may indicate whether the
referent is a location, object or person, whether it is animate or inanimate,
human or non-human, female or male, a single entity or set, or conceptualized
as a restricted or extended entity.
3.2 The features of demonstratives: a systematic overview
In addition to semantic information, demonstratives often provide some information
concerning aspects of their pragmatic use and syntactic function. There are,
for instance, languages in which exophoric, anaphoric, discourse deictic, and
recognitional demonstratives have different forms, and there are also languages
SEMANTICS 51
in which pronominal, adnominal, adverbial, and identificational demonstratives
are formally distinguished. Thus, there are three kinds of features that demonstratives
encode: (i) semantic features, which indicate the kind of referent and its
location; (ii) pragmatic features, which indicate how demonstratives are used; and
(iii) syntactic features, which indicate their syntactic functions. Table 36 provides
an overview of all features - semantic, pragmatic and syntactic - that are
encoded by demonstratives in my sample.
Table 36. An overview of the features encoded by demonstratives
Semantics
(i) Deixis
Distance Visibility Elevation Geography Movement
neutral
proximal
medial
etc.
visible
invisible
up
down
uphill
downhill
upriver
downriver
toward S
away from S
across the visual
field of S
(ii) Quality
Ontology Animacy Humanness Sex Number Boundedness
location
object/person
animate
inanimate
human
nonhuman
female
male
singular
plural
etc.
bound
unbound
Syntax
Category Case Agreement
pronoun
determiner
adverb
identifier
acc
etc.
(i) Gender
masc
fem
etc.
(ii) Number
singular
plural
etc.
(iii) Case
acc
etc.
Pragmatics
Use Reference
exophoric
anaphoric
discourse deictic
recognitional
(i) Emphasis
emphatic
non-emphatic
(ii) Contrast
contrastive
non-contrastive
(iii) Precision
precise
vague
52 DEMONSTRATIVES
The features in Table 36 are organized into three main categories: semantics,
syntax and pragmatics (given in bold italics). The three main categories are
divided into several subcategories (in italics), which subsume the features that are
directly encoded by demonstratives (in roman type). The features listed under
each subcategory are in contrastive distribution, but they may cooccur with
features of other subcategories. For instance, a demonstrative cannot have two
case features, but it may have a case feature, a category feature, up to three
agreement features, and so on. Let me emphasize that these features are meant
to characterize the information that is directly encoded in the morphological form
of a demonstrative. A demonstrative may be used, for example, as an anaphoric
pronoun referring to a human being, but if this is not reflected in its form it will
not have the features 'anaphoric', 'pronoun' and 'human'. Each feature is either
expressed by a demonstrative root or by one of the morphemes with which it
combines. In addition, there are a few forms in which a feature is associated with
the combination of a demonstrative root and a particular affix. For instance, some
languages have demonstratives that are interpreted as pronouns when they are
case-marked due to the fact that case markers only occur with pronominal demonstratives
in these languages (cf. 2.2.2 and 4.1). In such a case, the feature
'pronoun' is associated with the entire form rather than with one of its components.
The semantic features are divided into two categories: Deixis and Quality.
The category Deixis has five subcategories: (i) Distance, (ii) Visibility, (iii)
Elevation, (iv) Geography, and (v) Movement (or Direction). The category
Quality is divided into six subcategories: (i) Ontology, (ii) Animacy, (iii) Humanness,
(iv) Sex, (v) Number, and (vi) Boundedness. The features of these categories
were discussed at length in the previous two sections.
The syntactic features are divided into three categories: (i) Category, (ii)
Case, and (iii) Agreement. Category subsumes four features which indicate the
categorial status of a demonstrative: (i) 'pronoun', (ii) 'determiner', (iii) 'adverb',
and (iv) 'identifier'. The categorial status of demonstratives will be examined in
the following chapter. The agreement features are subsumed by three subcategories:
(i) Gender, (ii), Number and (iii) Case. Case is listed twice because case is
not only an agreement feature; its primary function is to indicate grammatical
relations.
The pragmatic features are divided into two categories: Use and Reference.
The category Use has four features: (i) 'exophoric', (ii) 'anaphoric', (iii) 'discourse
deictic', and (iv) 'recognitional'. Chapter 5 will show that many languages
employ distinct demonstrative forms for these four uses. The category Reference
is further divided into (i) Emphasis, (ii) Contrast, and (iii) Precision. The features
of these categories indicate the kind of reference that is expressed by a demonSEMANTICS
53
strative. More specifically, they indicate whether a demonstrative is (i) emphatic
or non-emphatic, (ii) contrastive or non-contrastive, and (iii) whether it is used
with vague or precise reference. Since these features are not discussed anywhere
else in this study, I use the remainder of this section to illustrate their form and
function.
Emphasis is usually expressed by emphatic suffixes that attach to a demonstrative
root, as in the following example from Ngiti.
(5) Ngiti (Kutsch 1994: 374)
w'f-r'I nf'ny ' } tsìtsì
DEM-EMPH eat.PF banana(s)
'That one has eaten bananas.'
Similar emphatic affixes augment demonstratives in Basque (Saltarelli 1988: 215-6),
Dyirbal (Dixon 1972: 48), Logbara (Crazzolara 1960: 55), Ewondo (Redden
1980: 70), and Ponapean (Rehg 1981: 143-154).12 Table 37 shows the emphatic
and non-emphatic demonstrative determiners in Ponapean (allomorphs are
omitted). The emphatic forms consist of the non-emphatic singular forms and a
numeral classifier. The classifier shown in this table is men, which indicates an
animate referent. Inanimate demonstrative pronouns involve a different classifier.
The emphatic plural forms are formed by combining the non-emphatic demonstratives
with the morpheme pwu-, which does not occur in any other context.
Like Emphasis, Contrast is usually expressed through a particular affix. Woleaian,
Table 37. Emphatic demonstrative determiners in Ponapean (Rehg 1981: 144, 149)
NON-EMPHATIC EMPHATIC
SG PL SG PL
NEAR S
NEAR H
AWAY FROM S+H
-e(t)
-en
-o
-ka(t)
-kan
-kau
pwuka(t)
pwukan
pwukau
for instance, has a suffix that indicates a contrastive referent "as when pointing
out one member of a group" (Anderson and Keenan 1985: 289):
(6) Woleaian (Anderson and Keenan 1985: 289)
mwu(u)-l
that.NEAR.H-CONTRAST
'that one near you'
54 DEMONSTRATIVES
Like Woleaian, Manam marks contrastiveness by a particular suffix. Demonstratives
being marked by this suffix indicate that the speaker selects the referent
"out of a set" (Lichtenberk 1983: 334).
(7) Manam (Lichtenberk 1983: 334)
tomóata ]áe-ni-Ø y-ún-a
man this-SELECT-3SG 3SG-hit-1SG.OBJ
'This man (out of several) hit me.'
Finally, there are demonstratives that indicate either vague or precise reference.
Ewondo, for instance, has two series of adverbial demonstratives that convey this
kind of information: vá 'near speaker', vála¯ 'near hearer', válí 'away from
speaker and hearer, and álí 'far away from speaker and hearer' are used with
precise reference, while mú 'around the location of the speaker', múlu¯ 'around the
location of the hearer', wóé 'away from the location of speaker and hearer' and
múlí 'far away from the location of speaker and hearer' indicate a location
somewhere around (or in the vicinity of) a certain point of reference (cf. 2.3).
Similarly, in Daga, a language spoken in New Guinea, demonstratives occur with
the suffix -na if the location of the referent is vague, and they are unmarked if
they refer to a precise location (cf. Anderson and Keenan 1985: 291).13
Concluding this chapter, I describe some demonstratives using the features
in Table 36. The number of features that are encoded by a demonstrative varies
with the size of the demonstrative system: demonstratives of complex systems
provide more information than demonstratives of small systems. Acehnese, for
instance, has only three invariable demonstratives which carry one semantic
feature (cf. 8), while the demonstratives in Tümpisa Shoshone, which has more
than a hundred different forms, indicate seven features (cf. 9).
(8) Acehnese nyoe
Syntax -
Semantics proximal
Pragmatics -
(9) Tümp. Shoshone s-u-tungku 'ANA-INVIS-NOM.DU'
Syntax pronoun/determiner
dual
Semantics invisible
object/person
dual
Pragmatics anaphoric
SEMANTICS 55
(10) English there
Syntax adverbial
Semantics distal
location
Pragmatics exophoric/anaphoric
(11) Ngiyambaa ]a-ni-la: 'DIST-LOC-GIVEN'
Syntax pronoun/determiner/(adverb/identifier)
locative
singular
Semantics distal
location
Pragmatics anaphoric
(12) Korean i
Syntax determiner
Semantics proximal
Pragmatics exophoric/anaphoric
(13) German dessen
Syntax pronoun/determiner
singular
masculin/neuter
genitive
Semantics neutral
object/person
singular
(male)
Pragmatics exophoric/anaphoric
CHAPTER 4
Syntax
Having described the morphological and semantic properties of demonstratives,
I now examine their syntactic features. As pointed out in the introduction, I
distinguish between the use of a demonstrative in a specific syntactic context and
its categorial status. Demonstratives occur in four different syntactic contexts: (i)
they are used as independent pronouns in argument position of verbs and
adpositions, (ii) they may cooccur with a noun in a noun phrase, (iii) they may
function as verb modifiers, and (iv) they occur in copular and nonverbal clauses.
I refer to demonstratives being used in one of these four contexts as (i) pronominal,
(ii) adnominal, (iii) adverbial, and (iv) identificational demonstratives,
respectively. Some languages have only one series of demonstratives that they use
in all four contexts, but most languages employ distinct demonstrative forms in
some or all of these positions. If adnominal, pronominal, adverbial, and identificational
demonstratives are formally distinguished, I assume that they belong to
different grammatical categories, which I refer to as (i) demonstrative pronouns,
(ii) demonstrative determiners, (iii) demonstrative adverbs, and (iv) demonstrative
identifiers, respectively.
The term demonstrative pronoun is probably the most transparent term of
these four notions. Demonstrative pronouns are pro-nominals; they are used in
lieu of a noun (phrase) and have the usual morphological features of nominals
(i.e. gender, number and case), if the nominals of a particular language are
marked for these features (cf. 2.2.2).
The term demonstrative determiner applies to adnominal demonstratives that
are formally distinguished from demonstratives in other syntactic contexts.
Traditional grammar assumes that demonstrative determiners are noun modifiers
(e.g. Bloomfield 1933), but Hudson (1984), Abney (1987) and others have argued
that the determiner is head of NP (or DP). Below I argue against the "Determiner-
as-head hypothesis". More specifically, I show that some of the head
features are shared by a demonstrative determiner and a cooccurring noun.
58 DEMONSTRATIVES
The term demonstrative adverb is adopted from Fillmore (1982: 47), who
uses this notion for locational deictics such as English here and there (and also
for manner demonstratives, cf. 4.2). The category adverb applies to a variety of
items that are semantically quite diverse and morphologically often not consistently
marked as a particular word class (cf. Schachter 1985: 20). Syntactically,
adverbs are used as modifiers of verbs, adjectives and other adverbs. Since
locational deictics are primarily used to indicate the location of the event or
situation denoted by a cooccurring verb they may be classified as adverbs.
The term demonstrative identifier is used for demonstratives in copular and
nonverbal clauses that are formally distinguished from demonstratives in other
sentence types. The term has the connotation of a semantic or pragmatic notion,
but I use it as a label for a grammatical category on a par with demonstrative
pronouns, determiners and adverbs. In two previous studies (Diessel 1997a,
forthcoming) I referred to demonstrative identifiers as "predicative demonstratives".
I adopted this notion from studies by Denny (1982: 365) and Heath
(1984: 269-336), where it is used to refer to a particular class of demonstratives
in Inuktitut and Nunggubuyu. Other notions that I have found in the literature that
seem to correspond to the notion of demonstrative identifier are "demonstrative
predicator" (Schuh 1977), "predicative pronoun" (Marconnès 1931: 110), "copulative
demonstrative" (Ziervogel 1952: 47-8), "existential demonstrative" (Benton
1971: 90), "pointing demonstrative" (Rehg 1981: 143), and "deictic identifier
pronoun" (Carlson 1994: 160). Since demonstrative identifiers often occur in
nonverbal clauses, they are sometimes considered to be functionally equivalent
to a demonstrative plus copula, which many languages require in this construction
(Hengeveld 1992). In fact, demonstrative identifiers are often glossed as
'this/that.is' or 'here/there.is' (e.g. Carlson 1994: 241; Dayley 1989: 145). This
explains why some studies use the attribute 'predicative' in order to characterize
demonstrative identifiers. The occurrence of demonstrative identifiers is, however,
not restricted to nonverbal clauses. Demonstratives in copular sentences are also
often distinguished from (pronominal) demonstratives in other sentence types.
Since the demonstratives in copular clauses are certainly not predicative, I
decided to replace the notion predicative demonstrative by demonstrative identifier.
In the following three sections, I discuss the evidence for the distinction
between demonstrative pronouns, determiners, adverbs, and identifiers, and I take
a closer look at languages in which these categories are not distinguished. I begin
by examining the distinction between demonstrative pronouns and determiners,
then I discuss demonstrative adverbs, and finally I consider the evidence for my
hypothesis that many languages have a class of demonstrative identifiers.
SYNTAX 59
4.1 Demonstrative pronouns and demonstrative determiners
The majority of languages uses the same demonstrative forms as independent
pronouns and as modifiers of a cooccurring noun. In my sample, there are only
twenty-four languages in which pronominal and adnominal demonstratives are
formally distinguished. In some of these languages they have different stems, as
in the following examples from Mulao and Japanese:
Mulao uses the demonstratives ni5 and hui5 as independent pronouns and nai6
Table 38. Demonstratives in Mulao (Wang and Guoqiao 1993: 52)
DEM PROs DEM DETs
PROXIMAL
DISTAL
ni5
hui5
nai6
ka6
Table 39. Demonstratives in Japanese (Kuno 1973: 27)
DEM PROs DEM DETs
NEAR S
NEAR H
AWAY FROM S+H
kore
are
kono
ano
and ka6 as modifiers of a cooccurring noun (the superscript numbers indicate
tone). The Japanese demonstratives consist of a distance and a category marker:
ko-, so- and a- indicate the relative distance between the referent and the deictic
center, and -re and -no indicate whether the demonstrative functions as a pronoun
or determiner.
In other languages, pronominal and adnominal demonstratives have the same
stems but differ in their inflection (cf. 2.2.2). Two examples from Turkish and
Lezgian are shown in Table 40 and 41 respectively.
Turkish has three demonstrative roots: bu 'proximal', s¸u 'medial', and o
'distal' (Lewis 1967 glosses s¸u 'the following'). The demonstrative determiners
are uninflected, but the demonstrative pronouns occur with number and case
suffixes, which are joined to the demonstrative root by an alveolar nasal. Like
Turkish, Lezgian has three demonstrative roots, i 'proximal', a 'distal' and at'a
'yonder'. They are marked for gender and number when they are used as
60 DEMONSTRATIVES
independent pronouns, but they are uninflected when they cooccur with an
Table 40. Demonstratives in Turkish (Kornfilt 1997: 106, 311)
DEM PROs DEM DETs
PROX MED DIST PROX MED DIST
SG ABS
ACC
GEN
DAT
LOC
bu
bun-u
bun-un
bun-a
bun-da
s¸u
s¸un-u
s¸un-un
s¸un-a
s¸un-da
on-u
on-un
on-a
on-da
SG/PL bu s¸u o
PL ABS bun-lar s¸un-lar on-lar
Table 41. Demonstratives in Lezgian (Haspelmath 1993: 111)
DEM PROs DEM DETs
PROX DIST YONDER PROX DIST YONDER
SG
PL
ABS
ERG
GEN
ABS
i-m
i-da
i-da-n
i-bur
a-m
a-da
a-da-n
a-bur
at'a-m
at'a-da
at'a-da-n
at'a-bur
SG/PL i a at'a
(inflected) noun (cf. 2.2.2).
If pronominal and adnominal demonstratives have different stems as in
Mulao and Japanese or if they differ in their inflectional behavior as in Turkish
and Lezgian, I assume that they belong to different grammatical categories,
which I refer to as demonstrative pronouns and demonstrative determiners,
respectively.
4.1.1 Adnominal demonstrative pronouns
Unlike Mulao, Japanese, Turkish, and Lezgian, most languages use the same demonstrative
forms in the position of independent pronouns and adjacent to a
cooccuring noun. In my sample, there are sixty-one languages in which
adnominal and pronominal demonstratives have the same stems and the same
inflectional features. In most of these languages there is no evidence that
pronominal and adnominal demonstratives belong to different categories. Both
SYNTAX 61
demonstratives are often independent pronouns, which are either used as arguments
of verbs and adpositions or adjoined to a coreferential noun in apposition
(cf. Hale 1983; Heath 1986; Baker 1996: chap4). Tuscarora, for instance, has two
demonstratives, hèník6˜ 'this/these' and kyèník6˜ 'that/those', which are either
used as independent pronouns or with a cooccurring noun (Mithun 1987). When
hèník6˜ and kyèník6˜ are used adnominally they are only loosely combined with
the juxtaposed noun: (i) both noun and demonstrative can represent the entire NP
without the other element, (ii) their position with respect to each other is flexible
(cf. 1a-b), and (iii) they are often separated by an intonational break (cf. 1c):
(1) Tuscarora (Mithun 1987: 184, 184, 186)
a. hèník6˜ áhaq
that horse
'that horse'
b. unéwak hèník6˜
ghost that
'that ghost'
c. watkaháhiq hèník6˜, ... ruyakwáhehr
it.met.it that he.body.carries
'It met that dinosaur.'
Based on these examples, Mithun (1987) argues that adnominal demonstratives
in Tuscarora are free nominals that cooccur with a coreferential noun in apposition.
There are several other languages in my sample in which adnominal demonstratives
behave in the same way as in Tuscarora and have been analyzed as
independent pronouns that are juxtaposed to a coreferential noun (e.g. Dyirbal,
Nunggubuyu, Wardaman, Oneida, West Greenlandic, Karanga). In some of these
languages, adnominal demonstratives may even be separated from the noun by
an intervening constituent. Such discontinuous noun phrases are quite common
in Australian languages (cf. Dixon 1972; Hale 1983; Heath 1986). An example
from Wardaman is shown in (2).
(2) Wardaman (Merlan 1994: 143)
dang-nyi wunggun-bu-ndi yibiyan-yi
yonder-ERG 3SG:3NON.SG-hit-PAST man-ERG
'That man hit them.'
In Tuscarora and Wardaman, adnominal demonstratives are not categorially
distinguished from demonstrative pronouns. These languages do not have a
62 DEMONSTRATIVES
particular class of demonstrative determiners. Adnominal demonstratives are
demonstrative pronouns that are adjoined to a neighboring noun in some kind of
appositional structure.14
4.1.2 Adnominal demonstratives in English
At this point one might ask whether adnominal and pronominal demonstratives
generally belong to the same category if they are phonologically and morphologically
indistinguishable. Consider, for instance, the adnominal demonstratives in
English. They have the same form as pronominal demonstratives, but are they independent
pronouns that are joined to an appositive noun, or do they function as
determiners? The categorial status of demonstratives in English is controversial.
To begin with, Van Valin and LaPolla (1997: 62) argue that "demonstratives
are pronominal in nature". In their view, this and that are pronouns no matter
where they occur. If they are used adnominally they occupy a particular slot in
the noun phrase. Van Valin and LaPolla use tree diagrams with two "projections"
in order to describe the syntactic and semantic relationships between the elements
of a noun phrase: a "constituent projection", which represents the major constituents
of the noun phrase, and an "operator projection", which represents the
operators of the noun and the entire NP. Both the constituent and operator
projection have several layers: nucleus, core, periphery, and the whole noun
phrase. For Van Valin and LaPolla's treatment of adnominal demonstratives, it
is essential to distinguish between elements that occur inside of the core layer and
elements outside of the core. To be precise, there is only one slot outside of the
core, called the "NP-initial position", which is the place where possessor NPs
usually occur. Articles, adjectives and numerals are treated as operators of the
core; they are only represented in the operator projection:
SYNTAX 63
NP
CORE
NUC
REF
N
book-s
N
NUC
CORE
CORE
NP
N
N
N
N
N
ADJ
three
QNT
the
DEF
NUM operator projection
constituent projection
Figure 2. Layered structure of NP (Van Valin and LaPolla 1997: 59)
Adnominal demonstratives are also treated as operators by Van Valin and
LaPolla. But unlike articles, demonstratives are not 'pure' operators in their view.
They are also represented in the constituent projection where they occupy the
NP-initial position (NPIP) outside of the core.
64 DEMONSTRATIVES
NP
CORE
NUC
REF
N
book
N
NP
NP
N
N
that
DEIC
DEF
operator projection
constituent projection
NPIP
NP
PRODEM
Figure 3. Layered structure of NP with adnominal DEM (Van Valin and LaPolla 1997: 62)
In this analysis, adnominal demonstratives are pronouns that occur in a specific
syntactic slot in which independent nominals (i.e. pronouns or nouns) function as
operators of a juxtaposed noun (phrase). Van Valin and LaPolla (1997: 62) point
out that in their analysis the house has a different syntactic structure from this
house because articles and adnominal demonstratives are treated as members of
two distinct categories, which are associated with different slots in the noun
phrase: articles are pure operators inside of the core, while (adnominal) demonstratives
are treated as independent pronouns that may "occur as NP modifiers"
outside of the core (Van Valin and LaPolla 1997: 62). In defense of their analysis,
Van Valin and LaPolla argue that although this "might seem odd from an English
perspective" there is good crosslinguistic evidence that articles and adnominal demonstratives
are categorially distinguished. They point out that in some languages
articles and adnominal demonstratives may cooccur, and they cite Dryer's work
on word order correlations (Dryer 1989a, 1992a), which shows that articles and
demonstratives follow different word order patterns across languages: while the
order of article and noun correlates with the order of verb and object, the order
of demonstrative and noun is in principle independent of the basic word order
pattern that characterizes a particular language (cf. Dryer 1989a: 89, 1992a: 103-4).
Crosslinguistically, demonstratives tend to precede the noun regardless of the
SYNTAX 65
order of verb and object (cf. Dryer 1989a: 95, 1992a: 96). This might indicate, as
Dryer points out, that articles and demonstratives do not always belong to the
same category.
Though articles and demonstratives are not always members of the same
category, there is good evidence that English the and this/that have the same
categorial status. English is one of the few languages in my sample in which
adnominal demonstratives are in paradigmatic relationship with articles, possessives
and other noun operators (e.g. every), which share a number of syntactic
features. This is a clear indication that adnominal this and that belong to a
category of determiners subsuming (adnominal) demonstratives, articles, possessives
and several other items such as every. Van Valin and LaPolla's argument
is based on the assumption that adnominal demonstratives have the same categorial
status across all languages, but that is not the case. As shown above,
there are languages in which adnominal demonstratives are independent pronouns
that are adjoined to a coreferential noun in apposition, and there are other
languages in which adnominal demonstratives are determiners, which require a
cooccurring noun.
To summarize the discussion thus far, I consider adnominal demonstratives
determiners if they are formally distinguished from demonstratives in other
contexts. They might differ from demonstrative pronouns, adverbs and identifiers
in three ways. First, they might have a particular phonological form, as in Mulao
and Japanese (cf. 4.1). Second, they might differ in their inflectional behavior,
as in Turkish and Lezgian (cf. 4.1). And third, they might have specific syntactic
properties, as in English, where adnominal demonstratives belong to the same
paradigm as articles and possessives. Adnominal this and that cannot be treated
as pro-nominals, unless one wants to claim that all determiners are independent
pronouns in English.
Assuming that the adnominal demonstratives in English are determiners, one
could argue, as an alternative to Van Valin and LaPolla's analysis, that pronominal
this and that are determiners that accompany an empty head instead of a
cooccurring noun. This view is captured in the tree diagram in Figure 4, which
I adopted from Abney (1987: 279-280). Abney discusses this structure as an
alternative to the analysis that he eventually suggests (see below).
66 DEMONSTRATIVES
NP
DP
DET
that
N¢
N
Figure 4. Demonstrative determiner with empty head noun (Abney 1987: 280)
Figure 4 provides an analysis that solves the problem in theory: if pronominal demonstratives
in English generally cooccur with an empty head, they are indeed
determiners. But as far as I can see, there is no empirical evidence for this claim.
Pronominal this and that do not constitute empty-headed NPs. There is no
structural and no psycholinguistic evidence that would support such a claim. The
assumption that adnominal demonstratives accompany an empty head is motivated
by theory internal considerations, which are irrelevant to the current investigation.
After considering the structure in Figure 4, Abney (1987) presents an
alternative analysis. He maintains that pronouns and determiners are in principle
not distinguished. That is, Abney claims that pronouns such as I, he and someone
belong to the same category as determiners such as a, the and every (cf. Postal
1969; Vennemann and Harlow 1977; Hudson 1984: 90-2). Abney refers to the
members of this category as determiners, but they should be called 'pro-terminers'
to be neutral.
In order to understand Abney's hypothesis, one has to consider the context
in which it is proposed. Abney is not interested in the categorial status of this and
that; he is primarily concerned with the structure of the noun phrase in English
and constituency in universal grammar. His ultimate goal is to show that NP has
basically the same constituent structure as IP. Abney assumes, for conceptual
reasons, that the constituent structure of all phrases is essentially the same and
should be reduced to a single X-bar schema. IP is a maximal projection of a
functional category, namely of INFL. Abney suggests that determiners are the
equivalent of INFL in NP. That is, he claims that determiners project a phrase in
which the determiner functions as head and the noun as a complement of the
former.
SYNTAX 67
DP
DET
that
NP
N
Figure 5. Determiner phrase (Abney 1987: 279)
In order to account for the different behavior of traditional pronouns and
traditional determiners, Abney distinguishes between different subtypes of
determiners. More specifically, he claims that the category determiner is subcategorized
in the same way as the category verb. Like verbs, determiners can be
transitive or intransitive. Traditional determiners such as the definite article the
are transitive; they take a noun as their complement. Traditional pronouns, on the
other hand, are intransitive determiners that do not take a complement. In this
approach, demonstratives are treated as determiners with variable valency: they
may be used intransitively without a nominal complement, or they are used
transitively in a determiner phrase. Demonstratives correspond to verbs such as
eat or burn, which can also be both transitive and intransitive.15
Abney's analysis is largely motivated by conceptual and theory internal
considerations, as he himself points out (Abney 1987: 351). He abandons the
distinction between determiners and pronouns primarily because this strengthens
the "Det-as-head analysis", which, in turn, allows him to maintain that the
structure of NP/DP is basically the same as the structure of IP.16 If determiners
belong to the same category as pronouns it appears to be more plausible to
analyze determiners as the head of NP/DP because pronouns are commonly
treated as the head of NP. There is, however, only little empirical support for this
claim. The only piece of evidence that might suggest that pronouns and determiners
are not distinguished comes from items such as this and that that can be
both independent pronouns and determiners. But since most pronouns cannot
function as determiners and most determiners cannot be used as independent
pronouns this argument alone is not convincing. Abney seeks to support his
analysis by two other empirical observations. First, he argues that in certain
contexts personal pronouns such as I and you are also used as determiners, as in
I Claudius and you idiot(s) (Abney 1987: 282). This argument is, however, based
on false assumptions. The personal pronouns that are used in these examples are
not restrictive. They do not determine the referential scope of the associated
nominal; rather, they are used as independent pronouns that cooccur with an
appositive noun.17
68 DEMONSTRATIVES
Second, Abney (1987: 283) claims that determiners are also similar to pronouns
in that they are the basic site of the grammatical features that characterize
a noun phrase. Chapter 2 has shown, however, that pronominal and adnominal
demonstratives do not always behave in the same way in this respect. There are
many languages in which adnominal demonstratives are not the site of grammatical
features even though pronominal demonstratives inflect. Abney's hypothesis that
determiners and pronouns form a single category is thus empirically unmotivated
and therefore I suggest that pronominal and adnominal this and that are to be
categorially distinguished.
To summarize, we have considered three analyses of this and that providing
three different answers to the question: what is the categorial status of adnominal
demonstratives in English? Van Valin and LaPolla argue that both adnominal and
pronominal this and that are independent pronouns. Alternatively, it has been
argued that they function as determiners, which cooccur either with a noun or an
empty head. Finally, Abney suggests that one should abandon the distinction
between pronouns and determiners altogether. All three analyses treat pronominal
and adnominal demonstratives as members of the same category, but ultimately
they are not convincing. I follow therefore the traditional view, which holds that
pronominal and adnominal this and that belong to different categories despite the
fact that they are phonologically and morphologically not distinguished (cf.
Bloomfield 1933). Adnominal this and that occur in a specific syntactic slot
where they cannot be analyzed as pronouns, and there is no evidence that
pronominal this and that would function as determiners. The English demonstratives
have the syntactic properties of two distinct word classes, and it seems that
one can only account for this if one assumes that adnominal and pronominal demonstratives
are categorially distinguished. Like many other languages, English
has both demonstrative pronouns and demonstrative determiners.
It is, of course, no accident that pronominal and adnominal this and that
share the same phonological and morphological features. Historically, the
determiners can be traced back to independent pronouns. Old English did not
have demonstrative determiners. Unlike adnominal demonstratives in Modern
English, adnominal demonstratives in Old English were not obligatory to form a
noun phrase and could cooccur with other elements that determine a noun
semantically such as possessives (cf. Traugott 1992: 173):
SYNTAX 69
(3) Old English (Traugott 1992: 173)
þa com þær gan in to me heofoncund Wisdom, &
then came there going in to me heavenly Wisdom and
þæt min murnede mod mid his wordum gegrette
that my sad spirit with his words greeted
'Then heavenly Wisdom came to me there and greeted my sad spirit
with his words.'
In example (3), the demonstrative þæt cooccurs with a possessive pronoun. In
such a case, the demonstrative usually precedes the possessor as in (3), but
Traugott (1992: 173) points out that there are also examples in Old English in
which the demonstrative follows the possessor. Both the fact that demonstratives
could cooccur with a possessive (pro)noun and that their position vis-à-vis the
possessor was not entirely fixed shows that the noun phrase in Old English was
not as tightly constrained as the noun phrase in Modern English. Traugott
(1992: 173) argues that there are examples in the Old English data in which
adnominal demonstratives can only be analyzed as independent pronouns.
Consider, for instance, the demonstrative in (4):
(4) Old English (Traugott 1992: 173)
se heora cyning ongan ða singan
that their king began then to.sing
'He (that one), their king, then began to sing.'
Old English is a so-called 'verb-second language' in which the finite verb occurs
after the first constituent of a clause, which is typically an adverb (cf. Traugott
1992: 275). Sentence (4) includes the temporal adverb ða 'then', which was
strongly preferred in sentence initial position. Traugott (1992: 173) argues that ða
does not occur before the verb in this example because the demonstrative se "is
probably a pronoun in a topicalized construction", which motivates the unusual
occurrence of the subject NP before the verb, where it is joined to the coreferential
demonstrative in some kind of appositional structure. If it is correct that the
adnominal demonstratives of Old English were independent pronouns, as the
analysis of (4) suggests, one has to assume that the demonstrative determiners of
Modern English developed from a pronominal source.
Himmelmann (1997) argues that the evolution of determiners from independent
pronouns is due to a common historical process whereby pragmatic discourse
structures develop into tightly organized phrase structure configurations (cf. Givón
1979; Hopper 1987; Lehmann 1995a). In his view, syntactic structure is the result
of a grammaticalization process by which elements in apposition (or juxtaposition)
70 DEMONSTRATIVES
are reanalyzed as hierarchically organized phrases (cf. Givón 1979: chap5).
Himmelmann points out that this process has an effect on the categorial status
of the elements that are involved in an emergent phrase structure configuration.
An appositional noun phrase, for instance, consists of two nominals, one of which
might be a pronoun. When such a structure is reanalyzed as a hierarchically
organized NP the categorial statuses of the items involved in this structure change
and new grammatical categories emerge. In the case of the English noun phrase,
pronominal elements such as demonstratives and possessives were reanalyzed as
determiners. They are distinguished from independent pronouns in that they fill
a specific slot in the NP, which can only be occupied by an element that belongs
to a particular class of linguistic items. The categorial change that led from
pronouns to determiners is not yet reflected in the morphology of this and that,
but as the grammaticalization process continues pronouns and determiners might
eventually assume different forms.
The grammaticalization process that gave rise to the structure of the noun
phrase in Modern English did not only affect the demonstratives and other
adnominal elements, it also changed the status of the cooccurring noun. In Old
English, nouns were free nominals which could form a noun phrase without a
cooccurring article, demonstrative or possessive. In Modern English, the occurrence
of bare nouns is restricted to certain types of nouns (e.g. proper names,
mass nouns) and certain uses (e.g. indefinite plural). In all other contexts, a noun
requires a cooccurring determiner as much as a determiner requires a cooccurring
noun. The noun phrase of Modern English is a grammatical construction in which
neither the noun nor the determiner can be the sole representation of the whole
phrase.18 Himmelmann (1997: 144-157) argues that the head and dependent
properties of the English NP is split between its components. Following Zwicky
(1985, 1993), he assumes that the distinction between heads and dependents is
based on a set of semantic, syntactic and morphological features, as shown in
Table 42.
In the unmarked case, the head and dependent features are divided among
the constituents of the phrase, but both Zwicky and Himmelmann argue that in
some constructions the head and dependent features are split among their
components. The noun phrase of Modern English provides an excellent example.
Semantically, the noun is the head of the construction, and since many determiners
do not inflect (e.g. a, the, every) one might argue that the noun is also the
morphosyntactic locus. However, the syntactic head features are represented by
both the noun and the determiner: (i) both constituents are usually required, (ii)
both the demonstrative and the noun have word rank (i.e. there can only be one
element of each in a noun phrase), and (iii) together they determine the categorial
SYNTAX 71
status of the phrase and its external syntax (Lyons 1977: 392). The adnominal de-
Table 42. Head and dependent features (Zwicky 1993: 298)
Head Dependent
Semantics
Syntax
Morphology
characterizing
required
word rank
category determinant
external representative
morphosyntactic locus
contributory
accessory
phrase rank
non-determinant
externally transparent
morphosyntactically irrelevant
monstratives in English have therefore both head and dependent features. They
must be distinguished from pure modifiers such as adjectives, which have only
dependent features. Unlike demonstratives, adjectives (i) are not obligatory, (ii)
they do not have word rank (i.e. they can be iterated and modified by an adverb),
and (iii) they are irrelevant to the external syntactic behavior of a noun phrase.
In order to indicate that determiners are distinguished from ordinary noun
modifiers, they are sometimes given a special status. In the generative literature,
demonstratives are often classified as specifiers (Chomsky 1981), which are
distinguished from modifiers such adjectives due to the fact that they have both
head and dependent features (cf. Zwicky 1993; Pollard and Sag 1994; Borsley
1996).
4.1.3 Pronominal demonstrative determiners
In the previous two sections, we have seen that some languages distinguish
between demonstrative pronouns and determiners while other languages have only
demonstrative pronouns. In this section, I show that there are also languages that
have only demonstrative determiners and lack a class of demonstrative pronouns.
We have already seen an example of such a language in Chapter 2, where I have
argued that Korean uses a demonstrative determiner and a defective noun in lieu
of a demonstrative pronoun. Korean does not have demonstratives that are used
as independent pronouns; (pronominal) demonstratives are always accompanied
by a nominal constituent, as in the following example:
72 DEMONSTRATIVES
(5) Korean (Sohn 1994: 295)
[ce il-ul] nwu-ka mak-keyss-ni
[that thing-ACC who-NOM block-will-Q
'Who would be able to block that?'
There are several other languages in my sample in which demonstratives are
generally embedded in a noun phrase. In some of these languages demonstratives
are accompanied by a classifier, as in the following example from Nùng:
(6) Nùng (Saul and Freiberger Wilson 1980: 61)
[tú té] non cá mu'n
[CLASS that sleep all night
'That one slept all night.'
In other languages, demonstratives cooccur with a third person pronoun, as in (7)
from Kusaiean:
(7) Kusaiean (Lee 1975: 101)
el uh
3SG this
'this (one)'
In Chapter 2, we saw that some languages form demonstrative pronouns from a
demonstrative root and a classifier (e.g. Barasano) or a third person pronoun (e.g.
Ao). The examples from Nùng and Kusaiean suggest that these demonstratives
are historically derived from a noun phrase in which the two elements were
independent.
Although some languages do not have independent demonstrative pronouns,
there are two contexts in which demonstratives usually do not cooccur with a
nominal constituent: (i) in copular and nonverbal clauses, and (ii) in contexts
where a demonstrative refers to an adjacent proposition. Consider the following
two examples:
(8) Kusaiean (Lee 1975: 109)
Sohn pa nge
John (COP) this
'This is John.'
(9) Nùng (Saul and Freiberger Wilson 1980: 21)
da» lão tê va» tê
grandmother that say that
'That grandmother said that.'
SYNTAX 73
Example (8) shows a demonstrative in a copular clause; and example (9) shows
a demonstrative that refers to an adjacent proposition or speech act. Although demonstratives
are usually accompanied by a third person pronoun or a classifier
in Kusaiean and Nùng, the demonstratives in (8) and (9) are used without a
cooccurring nominal. The missing nominal reflects the fact that the demonstratives
in these examples do not have the status of a common pronoun. The demonstrative
in (8) is an identificational demonstrative, which many languages
distinguish from demonstrative pronouns (cf. 4.3), and the demonstrative in (9)
is a discourse deictic demonstrative referring to a chunk of the surrounding
discourse. Discourse deictic demonstratives can be seen as pro-forms of the
proposition(s) to which they refer, but unlike ordinary pronouns they do not
substitute for a common noun or noun phrase denoting a person or object (cf.
5.3). In other words, discourse deictic demonstratives are not prototypical pronouns.
If one takes the notion of pronoun in its narrow sense as a pro-nominal
that replaces a common noun (i.e. a noun denoting an object or person) in an
ordinary main clause (i.e. a clause with a common main verb rather than a
copula), the demonstratives in (8) and (9) do not qualify as pronouns. I maintain
therefore my hypothesis that Kusaiean and Nùng do not have a class of demonstrative
pronouns.
4.1.4 Pronominal and adnominal demonstratives: an overview
Table 43 summarizes what I have said in this section about the categorial status
Table 43. Demonstrative pronouns and demonstrative determiners
DEM PROs DEM DETs
Mulao
Japanese
Turkish
Lezgian
English
Tuscarora
Wardaman
Korean
Kusaiean
Nùng
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
of adnominal and pronominal demonstratives. It distinguishes three types of
languages: (i) languages in which adnominal and pronominal demonstratives
74 DEMONSTRATIVES
belong to different categories; (ii) languages that have only demonstrative
pronouns; and (iii) languages that have only demonstrative determiners.
4.2 Demonstrative adverbs
Having examined the categorial status of pronominal and adnominal demonstratives,
I turn now to demonstrative adverbs. The term demonstrative adverb applies
to locational deictics such as here and there in English. Locational deictics are
adverbial in that they are primarily used to indicate the location of the event or
situation that is expressed by a cooccurring verb; that is, locational deictics
function as some sort of verb modifiers. In many languages, locational deictics
can also be used adnominally as in English this guy here or German das Haus da
'this/that house there'. In this use, they usually cooccur with a demonstrative
determiner that they intensify. That is, if a demonstrative adverb is used adnominally
it usually does not function as an operator of the noun; rather, it is
used to reinforce a cooccurring demonstrative determiner. In some languages, this
use has been grammaticalized and has given rise to new demonstrative forms that
consist of an old demonstrative determiner and a locational adverb. Afrikaans, for
instance, has two demonstratives, hierdie 'this' and daardie 'that', which are
historically derived from the Dutch demonstrative/article die and the demonstrative
adverbs hier 'here' and daar 'there' (cf. Raidt 1993). When these forms first
appeared they were only used adnominally, but now they are also increasingly
used as independent pronouns. Similar demonstratives occur in Swedish (e.g. det
här hus-et 'the/this here house-the') and French (e.g. cette maison-là 'this/that
house-there').19
Apart from locational deictics, manner demonstratives are usually classified
as demonstrative adverbs (cf. Fillmore 1982: 48).20 Manner demonstratives have
been largely ignored in the literature on deixis. Most of the sources that I
consulted list only their forms and gloss them as 'in this/that way' or 'like
this/that', but they do not explain their meaning or function. Manner demonstratives
seem to involve some sort of comparison and they are often used as
discourse deictics (cf. 5.3). Some examples are given in Table 44. Since the use
and function of manner demonstratives is not sufficiently explained in most of
my sources, the following discussion concentrates on locational deictics.
Most languages distinguish locational deictics from pronominal and
adnominal demonstratives, but there are a few languages in my sample in which
they have the same form. In Chapter 2 we saw, for instance, that Guugu Yimidhirr
SYNTAX 75
uses demonstratives with certain (locative) case endings as locational deictics.
Table 44. Manner demonstratives
Japanese
Ainu
Ambulas
Finnish
T. Shoshone
Pangasinan
koo 'in this way', soo 'in that way', aa 'in that way'
taa 'in this way', too 'in that way'
kéga 'like this', aga 'like that', waga 'like that'
näin 'in this way', noin 'in that way', niin 'so' etc.
inni 'this way'/'like that', enni 'this way'/'like that' etc.
onyá 'like this', ontán 'like that (near H)', onmán 'like that'
They belong to the same morphological paradigm as demonstratives that correspond
to this and that in English. There is no evidence for the existence of an
independent class of demonstrative adverbs in Guugu Yimidhirr.
My sample includes only a few other languages in which adverbial demonstratives
are members of the same category as pronominal and adnominal demonstratives.
Ponapean is one of them. Unlike Guugu Yimidhirr, Ponapean distinguishes
demonstrative pronouns from demonstrative determiners, but adverbial
demonstratives have the same form as the singular demonstrative pronouns. That
is, adverbial and pronominal demonstratives belong to the same category in
Ponapean:
Finnish is particularly interesting in this regard. In Finnish, demonstratives can
Table 45. Demonstratives in Ponapean (Rehg 1981: 143-154)
DEM DETs DEM PROs DEM ADVs
SG PL SG PL
NEAR S
NEAR H
AWAY FROM S+H
-e(t)
-en
-o
-ka(t)
-kan
-kau
mwo
metakan
menakan
mwohkan
mwo
be arranged on a cline of adverbiality ranging from forms that are clearly
pronominal to forms that are primarily adverbial (cf. Laury 1997: 128-146).
Finnish has three demonstrative roots: tämä, tuo and se. Tämä refers to entities
near the deictic center, tuo indicates a referent at some distance to the deictic
center, and se is mostly used anaphorically and is perhaps best analyzed as a
definite article in contemporary Finnish (cf. Laury 1995, 1997: 147-264). The
three demonstratives are always case-marked. Finnish has a complex case system
including six locational case forms, which are divided into two groups: three
76 DEMONSTRATIVES
"internal case forms" - inessive, elative and illative - and "three external case
forms" - adessive, ablative and allative. The internal and external case forms of
tämä, tuo and se are shown in Table 46.
The external demonstratives are primarily used to indicate a possessor,
Table 46. Locational case forms of the Finnish demonstratives (Laury 1997: 129)
(a) Internal case forms = Internal demonstratives
INESSIVE ELATIVE ILLATIVE
SG PL SG PL SG PL
PROXIMAL
DISTAL
ANAPHORIC
tässä
tuossa
siinä
näissä
noissa
niissä
tästä
tuosta
siitä
näistä
noista
niistä
tähän
tuohon
siihen
näihin
noihin
niihin
(b) External case forms = External demonstratives
ADESSIVE ABLATIVE ALLATIVE
SG PL SG PL SG PL
PROXIMAL
DISTAL
ANAPHORIC
tällä
tuolla
sillä
näillä
noilla
niillä
tältä
tuolta
siltä
näiltä
noilta
niiltä
tälle
tuolle
sille
näille
noille
niille
recipient or instrument; they are only occasionally used to indicate a location. The
internal demonstratives, on the other hand, are primarily used with reference to
places.
In addition to the forms shown in Table 46, Finnish has a set of "locational
demonstratives", which occur in three different case forms: adessive, ablative and
lative. The locational demonstratives are morphologically similar to the singular
forms of the external demonstratives, but they do not have plural forms.
Table 47. Locational demonstratives in Finnish (Laury 1997: 129)
ADESSIVE ABLATIVE LATIVE
PROXIMAL
DISTAL
ANAPHORIC
täälla
tuolla
siellä
täältä
tuolta
sieltä
tänne
tuonne
SYNTAX 77
The adessive and ablative forms of tämä and se have a long or diphthongized
vowel in the first syllable, which distinguishes these forms from the
corresponding external forms. The adessive and ablative forms of tuo are
indistinguishable from the external demonstratives in written Finnish, but Laury
points out that they are differently pronounced. Finally, the lative forms of all
three locational demonstratives - tänne, tuonne, and sinne - are clearly
distinguished from the corresponding allative forms, tälle, tuolle, and sille.
The external demonstratives are usually considered demonstrative pronouns
by Finnish linguists, but the status of the internal and locational demonstratives
is controversial. Traditionally, they have been considered adverbs (cf. Laury
1997: 134), but Laury points out that there is no morphological evidence in
support of this view. Although the internal demonstratives are commonly used to
refer to a location they behave morphologically like ordinary demonstrative
pronouns rather than adverbs. Moreover, Laury shows that even the locational demonstratives
are not prototypical adverbs. Unlike demonstrative adverbs such as
English here and there, the locational demonstratives in Finnish are case-marked
and they are frequently used adnominally. These are typical properties of demonstrative
determiners (or pronouns). However, when the locational demonstratives
are used adnominally they do not always agree with the cooccurring noun as a
demonstrative determiner (or pronoun). The following example shows a locative
demonstrative in lative case preceding a noun with an illative case marker.
(10) Finnish (Laury 1997: 135)
katotaas sinne kurkkuu
look.PASS DEM.LOC.LAT throat.ILL
'Let's look at that throat.'
The lack of agreement between a locational demonstrative and a cooccurring
noun suggests that the case endings of the locational demonstratives do not have
a syntactic function: they do not indicate grammatical relations, rather they are
used to specify the location of the referent. In other words, the case markers have
primarily a semantic function (similar to a locational adposition). The locational
demonstratives can therefore be seen as demonstrative adverbs despite the fact
that they are case-marked.
The internal demonstratives are integrated into the morphological paradigm
of demonstrative pronouns, but like locational demonstratives they do not
generally agree with a noun when they are used adnominally. The following
example shows an internal demonstrative in inessive case followed by a noun
with an adessive case marker.
78 DEMONSTRATIVES
(11) Finnish (Laury 1997: 136-7)
ja siin puuhellalla,... kerran ni,... mää
and DEM.INE wood.stove.ADE once so 1SG
illalla paistoin
evening.ADE fry.PAST.1SG
'And on the wood stove, one time, I was frying.'
According to Laury, the internal demonstratives have features of both demonstrative
pronouns and demonstrative adverbs. Morphologically, they behave like
ordinary demonstrative pronouns, but syntactically they are often used like demonstrative
adverbs, and semantically they are equivalent to locational deictics
(i.e. adverbial demonstratives). Laury (1997: 138) argues that the Finnish demonstratives
can be arranged on a cline of adverbiality ranging from forms that are
primarily used as pronouns to forms that behave more like typical adverbs:
Up to this point, I have argued that demonstratives can be divided into four
Table 48. Locational demonstratives in Finnish arranged on a cline of adverbiality (cf.
Laury 1997: 138)
<< LESS ADVERBIAL <<
EXTERNAL DEMS INTERNAL DEMS
>> MORE ADVERBIAL >>
LOCATIONAL DEMS
• they are case- and number-
marked like DEM PROs
• they are case- and number-
marked like DEM PROs
• they are case-marked like
DEM PROs, but they are
unmarked for number
• they do not always agree
with a cooccurring noun
• they do not always agree
with a cooccurring noun
• they behave syntactically
like DEM ADVs, but they
may cooccur with a noun
• they are primarily used to
indicate a possessor, recipient,
or instrument
• they are semantically
equivalent to locational
deictics
• they are semantically
equivalent to locational
deictics
syntactic categories, but the Finnish demonstratives show that the boundaries
between these categories are not always clear-cut.
4.3 Demonstrative identifiers
While demonstrative pronouns, determiners, and adverbs are well established
categories of linguistic analysis, the last category to be discussed in this chapter
SYNTAX 79
is widely unknown. Demonstrative identifiers have been described under various
names in reference grammars, but they have never been recognized in the
typological and theoretical literature on demonstratives (but see Himmelmann
1997: 126).
Demonstrative identifiers occur in copular and nonverbal clauses. Like other
demonstratives, they are used to focus the hearer's attention on entities in the
surrounding situation or in the universe of discourse. Most studies consider the
demonstratives in copular and nonverbal clauses demonstrative pronouns, but as
pointed out in the introduction, the demonstratives being used in these constructions
are often formally distinguished from pronominal demonstratives in other
contexts: they may have a different phonological form or may differ in their
inflection. If the demonstratives in copular and nonverbal clauses are phonologically
or morphologically distinguished from pronominal demonstratives in other
clause types, I assume that they form a class of demonstrative identifiers independent
of demonstrative pronouns. If, on the other hand, the demonstratives in
copular and nonverbal clauses are formally indistinguishable from demonstrative
pronouns, I assume that they belong to the same grammatical category. English,
for instance, does not distinguish between demonstrative pronouns and demonstrative
identifiers. The demonstratives in copular clauses have the same phonological
and morphological features as pronominal demonstratives in other contexts and
hence they are considered demonstrative pronouns.21
Demonstrative identifiers are similar to deictic presentatives such as French
voilà, Latin ecce, and Russian vot. Fillmore (1982: 47) calls such presentatives
"sentential demonstratives". Both demonstrative identifiers and sentential demonstratives
are commonly used to introduce new discourse topics, but they have
different syntactic properties. Demonstrative identifiers are embedded in a
specific grammatical construction, a copular or nonverbal clause, while sentential
demonstratives are syntactically more independent. Although they might occur in
sentences that are functionally equivalent to copular and nonverbal clauses (e.g.
Voilà un taxi. 'Here is a taxi.'), they are more commonly used as one word
utterances, which may be loosely adjoined to a neighboring constituent. I assume
therefore that demonstrative identifiers are distinguished from sentential demonstratives,
but the distinction is not clear-cut (see the discussion of Nunggubuyu
below).
In Chapter 6 I will show that demonstrative identifiers are a common
historical source for nonverbal copulas. The distinction between demonstrative
identifiers and copulas is not always immediately obvious: a demonstrative
identifier is easily confused with a copula that appears in a sentence with no overt
subject. Consider, for instance, the following example from Ambulas.
80 DEMONSTRATIVES
(12) Ambulas (Wilson 1980: 454)
kén bakna walkamu taalé
this just little place
'This is just a little place.'
Example (12) shows a nonverbal clause consisting of the demonstrative identifier
kén and a predicate nominal (DEM Ø NP). Without further evidence, example
(12) could be taken as a copular clause with no overt subject, in which kén would
function as a copula (Ø COP NP). However, since kén refers to a location and
is deictically contrastive (kén 'proximal' vs. wan 'distal'), it would be mistaken
to analyze it as a copula. If kén were a copula it would be non-referential and
non-contrastive. Thus, although (12) is syntactically ambiguous, kén can only be
interpreted as a demonstrative because of its meaning. There is, in other words,
a clear semantic contrast between a nonverbal copula and an identificational demonstrative
in a nonverbal clause.
The following two sections present evidence for my hypothesis that many
languages have a distinct class of demonstrative identifiers. In Section 4.3.1, I
consider demonstrative identifiers whose stems are phonologically distinguished
from the stems of demonstrative pronouns, and in Section 4.3.2, I examine demonstrative
identifiers that differ from demonstrative pronouns in their inflection.
4.3.1 Phonological evidence
The strongest evidence for my hypothesis that many languages have a particular
class of demonstrative identifiers comes from languages in which the stems of
identificational demonstratives are phonologically distinguished from the stems
of demonstrative pronouns. In this section, I discuss examples from Supyire,
Karanga, Ponapean, Western Bade, and Kilba. My sample includes further
examples from Izi, Swazi, Margi, and Pangasinan, which will not be considered
(the Pangasinan demonstratives are discussed in Section 4.4).
Table 49 shows the demonstrative pronouns and demonstrative identifiers in
Supyire. The demonstrative pronouns have an initial nasal consonant, which does
not occur with the demonstrative identifiers. Both demonstrative pronouns and demonstrative
identifiers are inflected for gender (noun class) and number. Note that
Supyire does not distinguish between proximal and distal forms (cf. 3.1.1).
Carlson (1994: 240) points out that the demonstrative identifiers might have
developed from a pronominal demonstrative and a copula. Their use is restricted
to affirmative nonverbal clauses. In negative contexts, Supyire uses a particular
negative identifier; and in copular clauses demonstrative identifiers are replaced
SYNTAX 81
by demonstrative pronouns. Example (13a) shows a demonstrative pronoun, and
Table 49. Demonstrative pronouns - identifiers in Supyire (Carlson 1994: 159-61)
DEM PROs (DETs) DEM IDENTs
SG PL SG PL
NC1 ]'gé m' píí we pii
NC2 ' ]ké ' \jé ke ye
NC3 'ndé ' \cíí le cii
NC4 'nté te
NC5 m' pé pe
example (13b) shows one of the demonstrative identifiers.
(13) Supyire (Carlson 1994: 190, 241)
a. mu à pyi a ']gé cè la
you PERF PAST PERF DEM.G1SG know Q
'Did you know this/that one?'
b. ku kè
it.G2SG here.is.G2SG
'Here/there it is.'
Supyire has also a set of "simple identifier pronouns", which occur in the same
syntactic context as the demonstrative identifiers in Table 49. The identifier
pronouns are, however, non-deictic; they are usually glossed as 'it.is' (cf. Carlson
1994: 160).
Table 50 shows the demonstrative pronouns and demonstrative identifiers of
three gender classes in Karanga, which are inflected for number and noun class;
proximal and distal demonstratives are formally distinguished. The demonstrative
pronouns begin with a vowel, while the demonstrative identifiers occur with an
initial h-. Both the demonstrative pronouns and identifiers have allomorphs that
are omitted in Table 50. Marconnès (1931: 110) argues that the demonstrative
identifiers, which he calls "predicative pronouns", form "a complete predicate, i.e.
include the verb 'to be' or some other verb, and correspond to English 'here is',
'here are', 'there is', 'there are', 'there goes' etc." Examples of demonstrative
pronouns and demonstrative identifiers are given in (14a-b) respectively.
82 DEMONSTRATIVES
(14) Karanga (Marconnès 1931: 102, 111)
Table 50. Demonstrative pronouns - identifiers in Karanga (Marconnès 1931: 101-10)
DEM PROs (DETs) DEM IDENTs
SG PL SG PL
PROX NC1
NC2
NC3
etc.
iyi
uyu
ichi
idzi
ava
izvi
heyi
hoyu
hechi
hedzi
hava
hezvi
DIST NC1
NC2
NC3
etc.
iyo
uyo
icho
idzo
avo
izvo
heyo
hoyo
hecho
hedzo
havo
hezvo
a. ndi no da uyu
I like this
'I like this one.'
b. hero sadza
there.is porridge
'There is the porridge.'
Western Bade has three demonstrative roots: one referring to an object or location
near the speaker, one referring to an object or location away from the speaker,
and another one which Schuh (1977: 19-20) glosses as "particular". The distal demonstrative
pronouns have the same form as the distal demonstrative identifiers,
but the proximal and particular forms are different: the demonstrative pronouns
end in a rounded mid back vowel, which is replaced by a long low vowel in the
corresponding forms of the demonstrative identifiers. Moreover, the proximal and
particular forms of the demonstrative identifiers occur optionally with the suffix
-ni which, according to Schuh, does not change their meaning. Both demonstrative
pronouns and demonstrative identifiers are inflected for gender and number.22
The demonstrative identifiers, which Schuh calls "deictic predicators", occur only
in nonverbal clauses:
(15) Western Bade (Schuh 1977: 20)
m' sàa wúnáajàa]íi
this/here your.dog
'Here's your dog.'
SYNTAX 83
Table 52 shows the demonstrative pronouns and demonstrative identifiers in
Table 51. Demonstrative pronouns - identifiers in Western Bade (Schuh 1977: 19-20)
DEM PROs (DETs) DEM IDENTs
SG.M SG.F PL SG.M SG.F PL
PROXIMAL
DISTAL
PARTICULAR
m' só
m' síi
m' s6'nò
m' có
m' cíi
m' c6'nò
m' dó
m' díi
m' d6'nò
m' sàa(ní)
m' sîi
m' s6'náa(ní)
m' càa(ní)
m' cîi
m' c6'náa(ní)
m' dàa(ní)
m' dîi
m' d6'náa(ní)
Ponapean. The demonstrative pronouns begin with a bilabial nasal and the demonstrative
identifiers have an initial high front vowel.
Example (16a-b) illustrate the use of these forms: (16a) shows a demonstrative
Table 52. Demonstrative pronouns - identifiers in Ponapean (Rehg 1981: 150-53)
DEM PROs DEM IDENTs
SG PL SG PL
NEAR S
NEAR H
AWAY FROM S+H
mwo
metakan
menakan
mwohkan
ietakan
ienakan
iohkan
pronoun functioning as the subject of the verb mengila 'wither', and example
(16b) shows a demonstrative identifier in a nonverbal clause.
(16) Ponapean (Rehg 1981: 143, 150)
a. met pahn mengila
this will wither
'This will wither.'
b. iet noumw naipen
this/here your knife
'Here is your knife.'
Finally, in Kilba demonstrative identifiers are monosyllabic enclitics while demonstrative
pronouns are free forms consisting of two or more syllables. It is
unclear whether the demonstratives in Kilba are marked for number. Schuh
(1983b) does not discuss their inflectional features; he only provides the forms
in Table 53.
84 DEMONSTRATIVES
(17) Kilba (Schuh 1983b: 318)
Table 53. Demonstrative pronouns - identifiers in Kilba (Schuh 1983b: 315-317)
DEM PROs (DETs) DEM IDENTs
PROXIMAL
DISTAL
REMOVED
(n ' 6)n ' 6nnà
(nà)ndándà
(]g'6)]g'6]gà
=ná
=ndá
=]gá
k'6n'6]=ná
sheep=DEM
'It's a sheep.'
Schuh (1983b: 317) classifies the enclitics =ná, =ndá and =]gá as demonstratives,
but he commonly translates them by a third person pronoun or an expletive.
The demonstrative enclitics are usually unstressed like third person pronouns, but
since they indicate a deictic contrast I consider them demonstratives. Note that
all demonstrative identifiers cited in this section are marked for distance. Demonstrative
identifiers are genuine deictic expressions; they are not expletives such
as English it in It is Friday. In fact, in some languages demonstrative identifiers
are primarily used exophorically. This is reflected in the terminology that some
of my sources use for demonstrative identifiers: Rehg (1981: 15) calls them
"pointing demonstratives" and Carlson (1994: 160) uses the notion "deictic
identifier pronoun".
4.3.2 Morphological evidence
The previous section has shown that demonstratives in copular and nonverbal
clauses are often phonologically distinguished from pronominal demonstratives
in other sentence types. This section shows that they may also differ in their
inflection. As pointed out in Chapter 2, in the majority of languages pronominal
demonstratives have the same grammatical features as other nominals: if nouns
are inflected for gender, number and/or case, pronominal demonstratives usually
have the same inflectional features. Unlike pronominal demonstratives, identificational
demonstratives do not always occur with the same inflectional endings as
other nominals. There is a substantial number of languages in my sample in
which pronominal demonstratives are inflected for gender, number and/or case
while identificational demonstratives are morphologically invariable. If pronominal
and identificational demonstratives differ in their inflection, I assume that
they belong to distinct categories. That is, if identificational demonstratives are
SYNTAX 85
not inflected in a language in which pronominal demonstratives are marked for
gender, number and/or case, I assume that they belong to a particular class of demonstrative
identifiers. This section discusses examples from Duwai, Nunggubuyu,
Tümpisa Shoshone, Inuktitut, French, and German. Other examples from
Ambulas (cf. 2.1.2), Pangasinan (cf. 4.4), and Modern Hebrew (cf. 6.6.1) are
discussed elsewhere in this study.
Duwai has two demonstrative identifiers, n'6mù 'proximal' and náamù
'distal', which are uninflected. They are distinguished from the demonstrative
pronouns shown on the left hand side of Table 54. The pronouns have a different
stem form and are differentiated for number.23
In Nunggubuyu, demonstrative pronouns occur with two noun class markers, a
Table 54. Demonstrative pronouns - identifiers in Duwai (Schuh 1977: 25-26)
DEM PROs (DETs) DEM IDENTs
SG PL
PROXIMAL
DISTAL
' ]gàannó
' ]gàanàwó
'ndìiwnó
'ndìiwnàwó
n'6mù
náamù
prefix and a suffix, while demonstrative identifiers take only the suffix. The
noun class affixes are also used to indicate number distinctions. Both demonstratives
occur optionally with a case marker. Table 55 shows only the masculine
singular forms; the demonstratives of other noun classes have the same structure.
In Nunggubuyu, demonstrative identifiers are often used without a cooccurring
Table 55. Demonstrative pronouns - identifiers in Nunggubuyu (Heath 1984: 272-4)
DEM PROs (DETs) DEM IDENTs
PROXIMAL
MEDIAL
DISTAL
ANAPHORIC
na-'-gi
na-da-gi
nu-'wa-gi
nu-'ba-gi
ya-gi
da-gi
yuwa-gi
ba-gi
nominal so that one might argue that they are better analyzed as sentential demonstratives
(see above). However, since Nunggubuyu is a non-configurational
language, in which all constituents are syntactically more independent than in
languages with rigid phrase structure configurations (cf. Heath 1984, 1986), I
86 DEMONSTRATIVES
assume that the ability to use demonstrative identifiers without a cooccurring
noun is not a property of the demonstratives but rather a consequence of general
typological characteristics. Example (18) shows a demonstrative identifier that is
accompanied by a coreferential noun.
(18) Nunggubuyu (Heath 1984: 278)
ya-gi na-walyi-nyung
this-M.SG M.SG-male-HUMAN.SG
'Here is the man.'
In Tümpisa Shoshone, demonstrative pronouns are inflected for number and case
and may take the prefix s-, which Dayley (1989: 136) calls an "obviative marker".
The demonstrative identifiers are unmarked for number, they take the suffix
-sü(n) instead of a regular case ending, and they never occur in the obviative
form. Table 56 shows only the proximal and medial forms; there are parallel
forms built on three other demonstrative roots.
The use of demonstrative pronouns and identifiers is exemplified in (19a-b)
Table 56. Demonstrative pronouns - identifiers in T. Shoshone (Dayley 1989: 137-43)
DEM PROs (DETs) DEM IDENTs
SUBJ OBJ
PROXIMAL SG
DU
PL
(s)-i-tü
(s)-i-tungku
(s)-i-tümmü
(s)-i-kka
(s)-i-tuhi
(s)-i-tümmi
i-sü(n)
MEDIAL SG
DU
PL
(s)-e-tü
(s)-e-tungku
(s)-e-tümmü
(s)-e-kka
(s)-e-tuhi
(s)-e-tümmi
e-sü(n)
respectively.
(19) Tümpisa Shoshone (Dayley 1989: 141, 145)
a. u punikka setü
it see that
'This one saw it.'
b. esü nahim pungku
this.is our.DU pet
'This is our pet.'
SYNTAX 87
Inuktitut has demonstrative identifiers that behave syntactically like demonstrative
identifiers in Nunggubuyu: since they are frequently used as one word utterances
they could be classified as sentential demonstratives rather than demonstrative
identifiers. However, since Inuktitut is a non-configurational language like
Nunggubuyu, I assume that the specific properties of the identificational demonstratives
are due to the particular structure of this language.
Table 57 shows the demonstrative identifiers in Inuktitut in comparison to
the demonstrative pronouns. The demonstrative pronouns consist of three
morphemes: a deictic root, a case suffix, and the nominalizer -sum-. The demonstrative
identifiers, on the other hand, are formed by a morphological process
which doubles the final consonant and adds a low back vowel:
The following sentences illustrate the use of demonstrative pronouns and demon-
Table 57. Demonstrative pronouns - identifiers in Inuktitut (Denny 1982: 364-5)
DEM PROs (DETs) DEM IDENTs
PROXIMAL uv-sum-ing
'PROX-the.one-ACC'
uvva
'here'
DISTAL ik-sum-ing
'DIST-the.one-ACC'
ikka
'there'
UP.DISTAL pik-sum-ing
'UP.DIST-the.one-ACC'
pikka
'up there'
DOWN.DISTAL kan-sum-ing
'DOWN.DIST-the.one-ACC'
kanna
'down there'
strative identifiers respectively:
(20) Inuktitut (Denny 1982: 365)
a. pik-sum-inga takujuq
UP.DIST-the.one-ACC he.sees
'He sees the (one) up there.'
b. Piita uvva
Peter DEM.PROX
'Here is Peter.'
French has a demonstrative particle, ce 'this/that/there/it', which is used only in
copular clauses.24 Ce does not generally function as a demonstrative. In fact, in
most instances ce is non-deictic and functions either as a third person pronoun or
88 DEMONSTRATIVES
as an expletive (cf. Reed 1994). However, in those cases in which ce has a
deictic interpretation, it can be analyzed as a demonstrative identifier. The demonstrative
pronouns, celui and celle, are morphologically and phonologically
distinguished from ce. They are marked for gender and number, and they are
usually reinforced by the deictic particles ci and là (cf. 3.1.1). Ce does not inflect
for gender and number and is replaced by celui, celle, ceci, or cela if the speaker
seeks to indicate the relative distance between the referent and the deictic center.
In other words, French has a demonstrative identifier, but its use is not obligatory
in copular clauses.
Finally, in German, where pronominal demonstratives are inflected for gender,
Table 58. Demonstrative pronouns - identifiers in French (Calvez 1993: 33, 62)
DEM PROs DEM IDENTs
SG.M SG.F PL.M PL.F
PROXIMAL
DISTAL
celui-(ci)
celui-(là)
celle-(ci)
celle-(là)
ceux-(ci)
ceux-(là)
celles-(ci)
celles-(là)
ce
ce
number, and case, identificational demonstratives are uninflected. Consider the
examples in (21a-b).
(21) German
a. Das ist meine Schwester.
DEM.NOM/ACC.SG.N is my sister.SG.F
'This is my sister.'
b. Das sind meine Freunde.
dem.nom/ACC.SG.N are my friend.PL
'These are my friends.'
Example (21a-b) show two copular clauses including the demonstrative das. The
demonstrative has the same form as the nominative/accusative, singular, neuter
form of the demonstrative pronouns, but unlike pronominal demonstratives it is
uninflected. Note that the coreferential predicate nominal in (21a) has feminine
gender and that the predicate nominal in (21b) occurs in plural. Since the demonstratives
in these examples are uninflected and do not agree with the coreferential
noun, they must be distinguished from demonstrative pronouns. They are demonstrative
identifiers, which occur only in copular clauses.
SYNTAX 89
4.4 Summary
In this chapter, I have examined the syntactic properties of demonstratives. I have
argued that one has to distinguish between the use of a demonstrative in a
specific syntactic context and its categorial status. Demonstratives occur in four
different syntactic contexts: (i) they are used as independent pronouns in
argument positions of verbs and adpositions; (ii) they occur together with a noun
in a noun phrase; (iii) they may function as locational adverbs modifying a
cooccurring verb; and (iv) they are used in copular and nonverbal clauses. I have
shown that the demonstratives being used in these four contexts are often
formally distinguished from one another. They might have different stem forms,
they might differ in their inflection, or they might have different syntactic
properties. If they are distinguished by any of these criteria, they belong to
different grammatical categories, for which I suggested the terms (i) demonstrative
pronoun, (ii) demonstrative determiner, (iii) demonstrative adverb, and (iv)
demonstrative identifier. I have shown that languages differ as to how they
exploit these categories. Many languages distinguish between demonstrative
pronouns and demonstrative determiners, but some languages have only one of
these two categories. Languages that do not have demonstrative determiners use
instead demonstrative pronouns with a coreferential noun in apposition, while
languages that do not have demonstrative pronouns use demonstrative determiners
together with a classifier or a third person pronoun in lieu of a demonstrative
pronoun. Unlike pronominal and adnominal demonstratives, adverbial demonstratives
usually have a special form. There are only a few languages in my sample
that do not have a distinct class of demonstrative adverbs. Demonstrative
identifiers occur in copular and nonverbal clauses. They may differ from demonstrative
pronouns in two ways: they may have a particular phonological form or
they may have other inflectional features. The category demonstrative identifier
is crosslinguistically not as common as the category demonstrative adverb, but the
distinction between demonstrative pronouns and demonstrative identifiers is at
least as frequent as the distinction between demonstrative pronouns and determiners.
I will conclude this chapter with a short discussion of the demonstratives
from two languages that exemplify the extent of variation in this domain: one in
which pronominal, adnominal, adverbial, and identificational demonstratives are
formally distinguished, and one in which they belong to the same category.
Acehnese represents the latter. The Acehnese demonstratives are shown in
Table 59.
Acehnese has three demonstrative particles that indicate three degrees of
90 DEMONSTRATIVES
distance: nyoe 'proximal', nyan 'medial' and jêh 'distal'. The three demonstra-
Table 59. Demonstratives in Acehnese (Durie 1985: 130)
FREE FORMS BOUND ALLOMORPHS
PROXIMAL
MEDIAL
DISTAL
nyoe
nyan
jêh
=noe
=nan
=dêh
tives have three bound allomorphs: =noe 'proximal', =nan 'medial' and =dêh
'distal'. All six demonstratives may occur in every possible syntactic context (cf.
the discussion of demonstratives in Guugu Yimidhirr in 2.1.1). That is, they may
be used adnominally as in (22a), they may function as independent pronouns or
as locational adverbs (cf. 22b-c), and they are also used as identificational
markers in nonverbal clauses (cf. 22d).
(22) Acehnese (Durie 1985: 191, 268, 256, 132)
a. ureueng=nyan
person=that
'that person.'
b. neu=peusom nyan bek ji=teu-peu lê=gop
2=hide that NEG 3=know-what by=other.person
'Hide that so that no one else will know.'
c. nyan ji=pura-pura teungeut jih
there 3=pretend-pretend sleep he
'There he goes pretending to be asleep.'
d. nyan aneuk=lông
that child=1SG
'That is my child.'
All four examples include the medial demonstrative nyan. In (22a), nyan is used
adnominally. Adnominal demonstratives usually cliticize to a preceding noun, but
they are not generally bound.25 In (22b), nyan is used as an independent pronoun
functioning as the object of the verb peusom 'to hide'. The demonstrative in (22c)
is ambiguous: it is either used to indicate a location or it functions as a presentational
marker. Durie (1985: 132) points out that the Acehnese demonstratives may
serve as locational adverbs (which he calls "locative pronouns"), but apart from
(22c) I did not find any example in Durie's grammar in which nyan might be
interpreted as an adverbial demonstrative. There are, however, several examples
in which some of the other demonstratives are used adverbially (often after a
SYNTAX 91
preposition). In the final example, nyan serves as an identificational marker in a
nonverbal clause. The sentences in (22a-d) show that adnominal, pronominal,
adverbial, and identificational demonstratives are formally indistinguishable in
Acehnese. They belong to the same grammatical category, which may occur in
four different syntactic contexts.
The demonstratives in Pangasinan represent the other end of the spectrum.
Pangasinan uses particular demonstrative forms in each of the four contexts in
which demonstratives occur. The Pangasinan demonstratives are shown in
Table 60.
The Pangasinan demonstratives are divided into four grammatical categories:
Table 60. Demonstratives in Pangasinan (Benton 1971: 51-52, 88-91)
DEM DETs DEM PROs DEM ADVs DEM IDENTs
PROXIMAL sá-ta-y (SG/PL)
sa-rá-ta-y (PL)
(i)yá (SG)
(i)rá-ya (PL)
diá nía
NEAR H -
-
(i)tán (SG)
(i)rá-tan (PL)
ditán nítan
DISTAL sá-ma-y (SG/PL)
sa-rá-ma-y (PL)
(i)mán (SG)
(i)rá-man (PL)
dimán níman
determiners, pronouns, adverbs, and identifiers.26 The demonstrative determiners
are formed from the article sa, the deictic roots ta 'proximal' and ma 'distal', and
the suffix -y, which Benton calls a topic marker, but which is probably a linker
(Nikolaus Himmelmann p.c.). In the plural, demonstrative determiners are marked
by -ra-, which precedes the deictic root and the linker. Benton (1971: 51-2)
classifies these forms as articles, but since they indicate a deictic contrast I
consider them demonstratives. The demonstrative pronouns consist of the demonstrative
roots ya 'proximal', tan 'nearer hearer' and man 'distal', which optionally
occur with an initial high front vowel; the plural forms are also marked by -ra-.
The demonstrative adverbs occur with an initial stop and do not have plural
forms, while the demonstrative identifiers take an initial nasal and are also
unmarked for number. The sentences in (23a-d) exemplify the use of these forms.
92 DEMONSTRATIVES
(23) Pangasinan (Benton 1971: 53, 89, 90, 91)
a. sá-ma-y apók
ART-DEM-LK grandchild.my
'My grandchild' (i.e 'that grandchild of mine')
b. sikató so analíw imán
he TOPIC bought that
'He (is the one who) bought that.'
c. sikató-y inmogíp ditán
he-LK slept here/there
'He (was the one who) slept here/there.'
d. nía so kánen mo
here.is TOPIC food your
'Here's your food.'
The demonstrative in (23a) is a demonstrative determiner; the demonstrative in
(23b) is a pronoun; the one in (23c) is a demonstrative adverb; and the final
example shows a demonstrative identifier.
Acehnese and Pangasinan represent the two ends of a spectrum ranging from
languages in which all demonstratives belong to the same category to languages
in which demonstratives are divided into four distinct classes. Most languages fall
somewhere in between these two extremes. English, for instance, distinguishes
three demonstrative categories: demonstrative adverbs, demonstrative determiners,
and demonstrative pronouns. It does not have a separate class of demonstrative
identifiers; the demonstratives being used in copular constructions are ordinary
demonstrative pronouns. Korean has two demonstrative categories: demonstrative
determiners and demonstrative adverbs. It does not have demonstrative pronouns
and demonstrative identifiers. The functional equivalent of the latter two are noun
phrases that consist of a demonstrative determiner and a defective noun (cf.
2.1.4). To cite one other example, Nunggubuyu has demonstrative pronouns, demonstrative
adverbs and demonstrative identifiers. Adnominal demonstratives are
demonstrative pronouns that cooccur with an appositive noun (Heath 1986). The
three demonstrative categories are formally distinguished through noun class
markers: the demonstrative pronouns take two noun class markers, a prefix and
a suffix, the demonstrative identifiers do not take noun class prefixes, and the demonstrative
adverbs do not occur with noun class suffixes (Heath 1984: 274-318).
CHAPTER 5
Pragmatic use
Demonstratives serve important pragmatic functions in the communicative
interaction between the interlocutors. They are primarily used to orient the hearer
in the speech situation, focusing his or her attention on objects, locations, or
persons, but they also serve a variety of other pragmatic functions. Following
Halliday and Hasan (1976: 57-76), I use the notion exophoric for demonstratives
that are used with reference to entities in the speech situation, and I use the term
endophoric for all other uses. The endophoric use is further subdivided into the
anaphoric, discourse deictic and recognitional uses. Anaphoric and discourse
deictic demonstratives refer to elements of the ongoing discourse (cf. Fillmore
1997; Lyons 1977; Levinson 1983; Himmelmann 1996, 1997). Anaphoric demonstratives
are coreferential with a prior NP; they keep track of discourse participants.
Discourse deictic demonstratives refer to propositions; they link the clause
in which they are embedded to the proposition to which they refer. Recognitional
demonstratives do not refer to elements of the surrounding discourse; rather, they
are used to indicate that the hearer is able to identify the referent based on
specific shared knowledge. The recognitional use is restricted to adnominal demonstratives,
while the demonstratives of all other uses may occur in any
possible syntactic context. That is, recognitional demonstratives are always used
with a cooccurring noun, while exophoric, anaphoric and discourse deictic demonstratives
may also be used pronominally, adverbially and in copular and
nonverbal clauses.
This chapter examines the various pragmatic uses of demonstratives from a
crosslinguistic perspective. It follows rather closely the work by Himmelmann
(1996, 1997), which I have cited above. My analysis is consistent with Himmelmann's
investigation except for one claim that he makes regarding the status of
the exophoric use. Himmelmann argues, in disagreement with much previous
work, that all four uses have equal status. Challenging this hypothesis, I maintain
that the exophoric use is indeed the basic use from which all other uses derive.
I support my hypothesis with evidence from language acquisition, markedness
theory, and grammaticalization. My analysis proceeds as follows: Sections 5.1 to
94 DEMONSTRATIVES
5.4 examine the exophoric, anaphoric, discourse deictic, and recognitional uses
respectively, and Section 5.5 argues that the exophoric use is basic.
5.1 The exophoric use
Exophoric demonstratives focus the hearer's attention on entities in the situation
surrounding the interlocutors. They have three distinctive features: first, they
involve the speaker (or some other person) as the deictic center; second, they
indicate a deictic contrast on a distance scale (unless they belong to the small
minority of demonstratives that are distance-neutral; cf. 3.1.1); and third, they are
often accompanied by a pointing gesture. None of these features is shared by the
three endophoric uses.
Fillmore (1997: 63) distinguishes between two uses that are exophoric from
my perspective: the gestural and the symbolic use (cf. Levinson 1983: 65-66).
The gestural use requires monitoring the speech event in order to identify the
referent, whereas the symbolic use involves activating knowledge about the
communicative situation and the referent. The two uses are exemplified by the
following examples, which Levinson (1983) provides in order to illustrate the
difference.
(1) English (Levinson 1983: 66, 66)
a. This finger hurts.
b. This city stinks.
The demonstratives in both sentences involve the speaker (or some other person)
as the deictic center. They are anchored in the speech situation, which indicates
that they are exophoric. However, only the demonstrative in (1a) can be accompanied
by a pointing gesture. This example illustrates the gestural use. The demonstrative
in (1b), which does not involve a pointing gesture, draws on knowledge
about the larger situational context, which involves more than what is immediately
visible in the surrounding situation. This example illustrates the symbolic use.
The symbolic use shows that the exophoric use is not limited to concrete
referents that are present in the surrounding situation. Exophoric demonstratives
are sometimes described as 'pointers' which simply locate an object in the
physical world, but this view is too simplistic (for a critique of this view see
Hanks 1990, De Mulder 1996, and Himmelmann 1996). Exophoric demonstratives
may also refer to entities that are not immediately visible in the speech situation,
as in (1b) (where the city as a whole is not visible) and in the following example.27
PRAGMATIC USE 95
(2) English (Levinson 1983: 66)
Hello, is Peter there? (on the telephone)
Moreover, exophoric demonstratives are also commonly used with reference to
entities that do not have a physical existence, as in (3):
(3) English
This is a nice feeling.
Even more abstract than the demonstratives in (2) and (3) is the use that Bühler
calls "Deixis am Phantasma" (Bühler 1934: 121-140). This use involves shifting
the deictic center from the speaker in the current speech situation to a person in
a different situation that is evoked by the ongoing discourse. This phenomenon,
which Lyons (1977: 579) calls "deictic projection" (cf. Jakobson 1957; Ehlich
1979; Sitta 1991), is characteristic of narratives and descriptions. Himmelmann
(1996) cites the following example from the Pear Stories, in which the proximal
demonstrative this refers to a location that only exists in the imagination of the
interlocutors.
(4) English (Himmelmann 1996: 222)
And he's... you see a scene where he's... coming on his bicycle this
way.
In (4), the deictic center has been shifted from the speaker to an imaginary
observer in the story world. The demonstrative is deictically anchored in the
situation evoked by the ongoing discourse (cf. Linde and Labov 1975; Ullmer-
Ehrich 1979).
McNeill, Cassell and Levy (1993) show that demonstratives 'am Phantasma'
can also be accompanied by a pointing gesture, just like demonstratives that are
anchored in the immediate speech situation. The referents are physically absent,
but they do exist in the universe of discourse, and speakers point to them as if
they were there. The use of deictic gestures provides strong evidence for my
hypothesis that the use of demonstratives 'am Phantasma' is a subtype of the
exophoric usage.
5.2 The anaphoric use
Anaphoric demonstratives are coreferential with a noun or noun phrase in the
previous discourse. They refer to the same referent as their antecedent (cf. Lyons
1977: 660). Unlike exophoric demonstratives, which are primarily used to orient
96 DEMONSTRATIVES
the hearer in the outside world, anaphoric demonstratives serve a languageinternal
function: they are used to track participants of the preceding discourse.28
Anaphoric demonstratives interact with other tracking devices such as
personal pronouns, definite articles, zero anaphors, and pronominal affixes on the
verb. There are a number of studies that examine the specific properties of
anaphoric demonstratives in comparison to other tracking means (cf. Linde 1979;
Ehlich 1979, 1982; Givón 1983; Sidner 1983; Ariel 1988; Gundel et al. 1993;
Lichtenberk 1988, 1996; Himmelmann 1996; Comrie forthcoming). These studies
show that anaphoric demonstratives are often used to indicate a referent that is
somewhat unexpected and not currently in the focus of attention. Comrie
(forthcoming) observes, for instance, that anaphoric demonstratives in Dutch,
German and Russian exclude the topic of the preceding discourse as a possible
antecedent. Continuing topics are tracked by third person pronouns or definite
noun phrases in these languages, while anaphoric demonstratives are coreferential
with non-topical antecedents, which are usually less expected. Consider the
following example from German.
(5) German
Der Anwalti sprach mit einem Klientenj. Da eri/derj
the lawyer talked with a client since he/this.one
nicht viel Zeit hatte, vereinbarten sie ein weiteres
not much time had agreed.on they a further
Gespräch nächste Woche.
conversation next week
'The lawyer talked to a client. Since he didn't have much time, they
agreed to have another meeting next week.'
The referent of the third person pronoun er is the subject NP der Anwalt 'the
lawyer' of the preceding sentence. The pronoun continues the topic of the
previous discourse. By contrast, the demonstrative der can only be coreferential
with the non-topical NP at the end of the first sentence, einen Klienten 'a client'.
Anaphoric demonstratives in German do not track continuing topics; rather, they
indicate a topic shift. Very often, they occur after the first mention of a thematically
prominent referent that persists in the subsequent discourse (cf. Himmelmann
1997: 229).
Similar conditions license the use of anaphoric demonstratives in many other
languages. For instance, Lichtenberk (1996) points out that in To'aba'ita demonstratives
are especially frequent after a new referent has been mentioned for the
first time. The following example is characteristic in this respect.
PRAGMATIC USE 97
(6) To'aba'ita (Lichtenberk 1996: 387-8)
Si u'unu 'eri 'e lae suli-a te'e wane bia
CLASS story that it:FACT go about-them one man and
kwai-na bia 'a-daro'a te'e wela, wela wane.
spouse-his and BEN-their.DU one child child man
Wela 'eri kali wela fa'ekwa ni bana. 'e a'i
child that little child small PART only it NEG
si tala 'a-na kai lae 'a-si
NEG be.possible BEN-his he:NONFACT go to-CLASS
kula n-e nii daa.
place REL-it:FACT be.located far
'This story is about a man, his wife, and their child, a boy. The child
was very little. He wasn't able to go faraway places.'
Example (6) shows the first paragraph of a narrative about a man, his wife, and
their child. The three major participants are mentioned for the first time in the
initial sentence. The subsequent discourse concentrates on wela 'the child', which
is the main topic of the sentences that follow. When the child is mentioned for
the second time (at the beginning of the second sentence), it is marked by the
anaphoric demonstrative 'eri 'that'. Similar to the demonstrative in (5), the demonstrative
in (6) functions to establish a new discourse topic. Note that wela is
tracked by a third person pronoun once it is in the focus of attention (from the
third sentence onwards). One might object to this analysis by arguing that the use
of the demonstrative at the beginning of the second sentence is motivated by the
fact that the child is selected out of a group of several potential topics. However,
the same strategy is used in contexts in which anaphoric demonstratives are not
selective. Consider, for instance, the following example.
(7) To'aba'ita (Lichtenberk 1996: 385-7)
... keka soeto'o ta ai ura wela 'eri, ma
they ask some woman for child that and
imole 'e-ki keka sore'e: "Kamili'a 'e a'i
person that-PL they say we.EXCL it:FACT NEG
si thaito'oma-na."
NEG know-it
'... they asked some women about the child, and those people said:
"We don't know.'"
According to Lichtenberk (1996: 387), the noun phrase imole 'eki 'those people'
is coreferential with ta ai 'some women' in the preceding clause, which intro98
DEMONSTRATIVES
duced a new discourse referent. As in all previous examples, the referent is
marked by an anaphoric demonstrative when it is mentioned for the second time.
The use of anaphoric demonstratives after first mention is a common strategy
to establish major discourse participants in the universe of discourse. Cyr (1993a,
1993b, 1996) notes the use of this strategy in Montagnais (Algonquian), and
Himmelmann (1996) provides the following example from Tagalog (Austronesian).
(8) Tagalog (Himmelmann 1996: 229)
May kasaysayan sa isang manlalakbay; (0.7 sec)
EXIST statement LOC one traveler (0.7 sec)
ang manlalakbay na ito ay si Pepito.
SPEC traveler LK DEM PRED PROPER.NAME Pepito
'(One incident) is told about a traveler; this traveler (his name) was
Pepito.'
Example (8) includes the anaphoric demonstrative na ito after the noun manlalakbay
'traveler'. Similar to the demonstratives in previous examples, the demonstrative
in (8) indicates that the focus of attention has been shifted to a new
participant that was mentioned for the first time in the preceding sentence.
Himmelmann (1996: 229) points out that the use of anaphoric demonstratives
after the first mention of a new discourse participant is especially common in
languages that do not have a definite article; however, even in languages that
employ a definite article, demonstratives are often preferred in this position (cf.
Christophersen 1939: 29). Once a new discourse participant has been established
as topic, it is usually tracked by third person pronouns, zero anaphors, definite
articles, or pronominal affixes on the verb; but when a referent is mentioned for
the second time, demonstratives are often the most common tracking device. The
three steps that are involved in this strategy are summarized in Table 61.
Anaphoric demonstratives are not only used to establish new discourse topics.
Table 61. The use of anaphoric demonstratives after first mention
• 1st mention
• (indefinite) NP
• new referent
• 2nd mention
• anaphoric DEM
• referent established as topic
• subsequent mentions
• 3.PRO, definite ART etc.
• (topical) referent continued
Lichtenberk (1988) observes, for instance, that in To'aba'ita demonstratives are
also commonly used to reactivate a referent that occurred at some distance in the
preceding discourse, while the referent of an immediately preceding clause is
PRAGMATIC USE 99
usually tracked by a third person pronoun or a pronominal affix on the verb.29
What all anaphoric demonstratives have in common is that they do not just
continue the focus of attention; rather, they indicate that the antecedent is not the
referent that the hearer would expect in this context (i.e. the most topical NP).
Anaphoric demonstratives are used when reference tracking is somewhat problematic,
or as Himmelmann puts it: "demonstratives are used for tracking only if
other tracking devices fail" (Himmelmann 1996: 227).30
Many languages have particular demonstratives that are specialized for the
anaphoric use. Japanese, for instance, has three demonstrative roots: ko- 'near
speaker', so- 'near hearer' and a- 'away from speaker and hearer', which
combine with a category marker (e.g. -ro 'pronoun', -no 'determiner' etc.). All
three demonstrative roots can be used exophorically, but only the so- demonstratives
are commonly used as anaphors (cf. Imai 1996; see also Anderson and
Keenan 1985: 285-6). Consider the following example.31
(9) Japanese (Kuno 1973: 284)
Kinoo Yamada to yuu hito ni aimasita.
yesterday Yamada as named person met
Sono (*kono *ano) hito, miti ni mayotte
that person way in lose
komatte.ita node, tasukete agemasita.
was.in.trouble because helping gave (the favor of)
'Yesterday, I met a man by the name of Yamada. Since he lost his
way and was having difficulties, I helped him.'
The expression sono hito 'that person' is coreferential with a noun phrase in the
preceding sentence. The only demonstrative that is acceptable in this context is
sono.
Like Japanese, Wardaman and Finnish have three demonstratives, but only
one of them is commonly used to track prior discourse participants. Wardaman
uses the medial demonstrative nana for anaphoric reference (Merlan 1994: 138),
and Finnish uses the distal demonstrative se (cf. Laury 1995, 1997). In both
languages, anaphoric demonstratives can be used as independent pronouns and
adjacent to a coreferential noun.
The anaphoric demonstratives in Japanese, Wardaman and Finnish can also
be used exophorically, but some languages employ demonstratives that are
exclusively used as anaphors. In Latin, for instance, the demonstrative is
'this/that' can only be used anaphorically, whereas hic 'near speaker', iste 'near
hearer' and ille 'away from speaker and hearer' are primarily used with reference
to entities in the surrounding situation. Note that is is not a third person pronoun:
100 DEMONSTRATIVES
continuing topics are tracked by zero anaphors in Latin; is is predominantly used
to track topics that are discontinuitive, contrastive or emphatic.
Like Latin, Lezgian has three exophoric demonstratives and a special form
for anaphoric reference, which Haspelmath (1993) glosses as 'the aforementioned'.
Pronominal demonstratives are generally marked for number and case
in Lezgian, while adnominal demonstratives are uninflected (cf. 4.1.1). In the
following example the anaphoric demonstrative ha is used to indicate that the
noun universitet is coreferential with the St. Petersburg University mentioned in
the preceding sentence. Note that ha can also function as an independent pronoun.
(10) Lezgian (Haspelmath 1993: 191)
Zun sifte Sankt=Peterburg.di-n universitet.di-z,
I.ABS first St.=Petersburg-GEN university-DAT
fizika.di-n fakul'tet.di-z, haˆx-na-j... Axpa zun ha
physics-GEN faculty-DAT enter-AOR-PAST then I.ABS that
universitet.di-n juridioeˇ skij fakul'et.di-z haˆx-na.
university-GEN juridical faculty-DAT enter-AOR
'I first entered St. Petersburg University, faculty of physics. Then I
entered the law faculty of that university.'
There are several other languages in my sample that have special anaphoric demonstratives:
Hixkaryana (Derbyshire 1985: 7, 130), Ngiti (Kutsch 1994: 372-5),
Urubu-Kaapor (Kakumasu 1986: 381), Maricopa (Gordon 1986: 55), and Koyra
Chiini (Heath 1999: 62).
Finally, there are languages in which anaphoric demonstratives are marked
by an additional affix. Anaphoric demonstratives in Tümpisa Shoshone, for
instance, take the prefix s-, which is "used to signal given or definite information"
(Dayley 1989: 136):
(11) Tümpisa Shoshone (Dayley 1989: 142)
S-a-tü s-a-kka u tukummahanningkünna.
ANA-DEM-NOM ANA-DEM-OBJ him cook.for
'She cooked that for him.'
Similar anaphoric markers occur in Usan (Reesink 1987: 80), West Greenlandic
(Fortescue 1984: 259), and Ngiyambaa (Donaldson 1980: 137).
5.3 The discourse deictic use
Like anaphoric demonstratives, discourse deictic demonstratives refer to elements
PRAGMATIC USE 101
of the surrounding discourse. Discourse deictic demonstratives are, however, not
coreferential with a prior NP; rather, they refer to propositions (cf. Lyons 1977;
Webber 1991; Canisius and Sitta 1991; Grenoble 1994; Herring 1994; Himmelmann
1996: 224-229; Fillmore 1997: 103-106).32 More specifically, discourse
deictic demonstratives focus the hearer's attention on aspects of meaning,
expressed by a clause, a sentence, a paragraph, or an entire story. Consider the
following example:
(12) English (Webber 1991: 111-2)
A: Hey, management has reconsidered its position. They've
promoted Fred to second vice president.
B: a. That's false. (reference to proposition)
b. That's a lie. (reference to illocution)
The demonstratives in (12a-b) refer to an entire sentence. More precisely, the demonstrative
in (12a) refers to the propositional content of the preceding utterance,
while the demonstrative in (12b) focuses the hearer's attention on its illocutionary
force (cf. Webber 1991: 112).
Discourse deictic demonstratives, like the ones in this example, must be
distinguished from what Lyons calls "pure text deixis" (Lyons 1977: 668).33 Pure
text deictic demonstratives refer to the material side of language, as exemplified
in the following sentences.
(13) English (Webber 1991: 108)
I'm sorry. I didn't hear you. Could you repeat that?
The demonstrative in (13) does not refer to a preceding proposition or speech act;
rather, it refers to a string of speech sounds that the speaker could not interpret.
Pure text deixis is a particular instance of the exophoric use. It refers to linguistic
entities treated as an object of the surrounding situation. By contrast, discourse
deictic demonstratives focus the hearer's attention on aspects of meaning evoked
by the ongoing discourse. The referent of a discourse deictic demonstrative has
no existence outside of the universe of discourse in the physical world.
Like anaphoric demonstratives, discourse deictic demonstratives serve a
language-internal function. There are, however, a number of significant differences
between the two uses. To begin with, anaphoric demonstratives are used to
track prior discourse participants, while discourse deictic demonstratives function
to establish an overt link between two propositions: the one in which they are
embedded and the one to which they refer (cf. Grenoble 1994). Consider the
following example:
102 DEMONSTRATIVES
(14) English (USA Today page 25C, Dec. 12, 1997)
"The object is to make fun", said Jon Butler, executive director of Pop
Warner. "Teams have been working together since August to get here
and we want them to have a good time."
That's why PopWarner moved to the Disney complex three years
ago. With more than 5,000 players, coaches and parents attending in
1994, it was growing.
The demonstrative at the beginning of the second paragraph summarizes the
information expressed in the preceding discourse, providing a thematic ground for
the sentences that follow. Being thematically associated with both the paragraph
that precedes and the paragraph that follows, the demonstrative creates an overt
link between two discourse units. It is similar to a sentence connective in that it
functions to combine two chunks of discourse. In the following chapter, I will
argue that discourse deictic demonstratives provide a common historical source
for the development of conjunctions and complementizers (cf. 6.3.3 and 6.3.4).
Another difference between anaphoric and discourse deictic demonstratives
concerns the likelihood with which the referents of these demonstratives persist
in the ongoing discourse. Himmelmann (1996: 225) points out that the referents
of anaphoric demonstratives usually continue in the subsequent discourse, while
the referents of discourse deictic demonstratives normally do not persist. This is
a consequence of their discourse pragmatic function: anaphoric demonstratives are
commonly used to introduce a major discourse participant, which is often the
main topic of the sentences that follow, whereas discourse deictic demonstratives
function to provide a thematic link between two propositions (or speech acts) at
one particular point in the progressing discourse.
Finally, discourse deictic demonstratives may refer to the discourse that
precedes as well as to the discourse that follows, while anaphoric demonstratives
can only be coreferential with a noun phrase in the previous discourse. That is,
anaphoric demonstratives are always anaphoric, while discourse deictic demonstratives
can be both anaphoric and cataphoric.34 This is illustrated in (15a-b).
(15) English
a. A: I've heard you will move to Hawaii?
B: Who told you that (*this)?
b. A: Listen to this (*that): John will move to Hawaii.
The distal demonstrative in (15a) refers back to the preceding proposition,
whereas the proximal demonstrative in (15b) anticipates upcoming information
expressed in the subsequent clause. Note that this and that are not interchangeable
with one another in these examples. The use of both demonstratives is limited to
PRAGMATIC USE 103
certain contexts. Discourse deictic that is used only with anaphoric reference; that
is, that may not refer to portions of the following discourse (cf. Fillmore
1997: 104-5; see also Halliday and Hasan 1976: 68). Discourse deictic this, on the
other hand, can be both anaphoric and cataphoric, but it refers only to utterances
produced by the same speaker; that is, this cannot refer to propositions or speech
acts across speaker boundaries (cf. Chen 1990: 144; Gundel et al. 1993: 288).
Table 62 summarizes the major differences between the discourse deictic
and anaphoric (tracking) uses.
Like anaphoric demonstratives, discourse deictic demonstratives are often
Table 62. Anaphoric and discourse deictic demonstratives
Anaphoric demonstratives Discourse deictic demonstratives
• they are coreferential with a prior NP • they refer to propositions/speech
acts
• they keep track of discourse participants
• they link two discourse units
• the referent commonly persists in the
subsequent discourse
• the referent usually does not persist
in the subsequent discourse
• only anaphoric • anaphoric and cataphoric
formally distinguished from demonstratives serving other pragmatic functions.
Usan, for instance, has two demonstratives, ende and ete, that are only used as
discourse deictics (cf. Reesink 1987: 81; see also Himmelmann 1997: 128). Ende
refers to propositions of the previous discourse; it consists of the demonstrative
root e 'proximal', the anaphoric marker -ng 'given', the adposition -t 'locative',
and a second demonstrative root at the end of the word. Ete is used to refer to
subsequent propositions; it is formed from e 'proximal', -t 'locative' and e
'proximal' (cf. Reesink 1987: 81). In addition to ende and ete, Usan has four
exophoric demonstratives and a special form for anaphoric reference: eng. The
latter consists of the proximal demonstrative e and the anaphoric marker -ng.
Table 63 shows the exophoric, anaphoric and discourse deictic demonstratives
that occur in Usan. The use of ende and ete is exemplified in (16a-b).
104 DEMONSTRATIVES
(16) Usan (Himmelmann 1997: 303, 309-10)
Table 63. Demonstratives in Usan (Reesink 1987: 76-81)
PROXIMAL
UP.THERE
DOWN.THERE
ACROSS.THERE
ANAPHORIC
DISCOURSE.DEICTIC (backward)
DISCOURSE.DEICTIC (forward)
ité
úmo
iré
ende
a. Âgin amug igo namaibâ, ...
sister.in.law bush be 2SG.OBJ.angry.SG.FUT.3SG ...
wai qeru me netinei.
animal blood not 2SG.OBJ.give.1SG.FUT(UNCERTAIN)
Ende qâmâra bo is âin erobon
DEM say.3SG again descend bamboo under.cover
igoai.
be.3SG.PAST
'"My sister-in-law is in the bush and will be angry with you...
I can't give the animal blood to you." When he said thus, (the
moon) went down again and stayed inside the bamboo.'
b. End ete qâmb igurei: ... "See inaun
DEM DEM say be.3PL.PAST "now moon
ag et igâma ende qi eng in
on.the.ground DEM be.3SG DEM 'or' DEM 1PL
munangit wai igumune ...".
human animal be.1PL ..."
'Therefore they used to say thus: "If the moon were still here
on the ground, we humans would have been animals,..."'
In (16a), ende summarizes the information of the preceding sentence, and in
(16b), ete is used to anticipate the information expressed in the subsequent quote.
As pointed out in Chapter 4, some languages have a particular class of
manner demonstratives, which are usually glossed as 'in this way', 'like this', or
'thus'. Manner demonstratives are frequently used as discourse deictics, as in the
following example from Ainu.
PRAGMATIC USE 105
(17) Ainu (Refsing 1986: 98)
Ikoytupa an hi kamuy nukar wa, pon cep
suffer.from.lack.of we NLZ gods see and be.small fish
poronno an eimekkar wa, taa e se ruwe ne.
a.lot PASS bestow and in.this.way you bring ASS
'The gods saw our suffering and a lot of small fish were bestowed
upon us, and in this way it is that you have brought them.'
The manner demonstrative in (17), taa 'in this way', is derived from the root of
the proximal demonstrative ta. It functions as an anaphoric discourse deictic
referring back to the preceding propositions. There are several other languages
in my sample that use manner demonstratives as discourse deictics. In Chapter
2 we saw, for instance, that Ambulas uses the manner demonstratives kéga 'like
this', aga 'like that', and waga 'like that' in order to refer to propositions: kéga
and aga are cataphoric discourse deictics, while waga refers (anaphorically) to
elements of the preceding discourse.
5.4 The recognitional use
The recognitional use has received much less attention in the literature than any
of the other uses. Although this use is recognized in a number of studies (e.g.
Lakoff 1974; Auer 1981, 1984; Chen 1990; Gundel et al. 1993), it has never
been described in detail until recently. Himmelmann (1996, 1997) is the first to
provide a systematic account of this use.
The recognitional use has two properties that distinguish it from all other
uses. First, recognitional demonstratives are only used adnominally.35 Second,
recognitional demonstratives do not have a referent in the preceding discourse or
the surrounding situation; rather, they are used to activate specific shared
knowledge.36 Consider the following example:
(18) English (Himmelmann 1996: 230)
... it was filmed in California, those dusty kind of hills that they have
out here in Stockton and all, ... so ...
In (18) the dusty hills are mentioned for the first time. Although first mentions
are usually marked by an indefinite article in English, the dusty hills occur with
the distal demonstrative those. The demonstrative indicates that the following
noun expresses information that is familiar to the hearer due to shared experience.
Example (19) is similar in this respect.
106 DEMONSTRATIVES
(19) English (Gundel et al. 1993: 278)
I couldn't sleep last night. That dog (next door) kept me awake.
As in the previous example, example (19) includes a noun that occurs with a demonstrative
at its first mention. The demonstrative does not refer to an entity in
the surrounding discourse or speech situation; rather, it indicates that the speaker
believes that the hearer knows the referent.
Recognitional demonstratives mark information that is discourse new and
hearer old. Prince (1992) introduces the terms "discourse new / discourse old"
and "hearer new / hearer old" in order to distinguish information that has been
evoked by the preceding discourse from information that is already in the hearer's
knowledge store (i.e. old with respect to the speaker's beliefs). Discourse old
information is also hearer old information, but hearer old information might be
discourse new: the hearer might know something although it was previously not
mentioned. Such information is unactivated (cf. Chafe 1987; 1994), but pragmatically
presupposed (cf. Dryer 1996). Recognitional demonstratives are specifically
used to mark information that is discourse new (i.e. unactivated) and hearer old
(i.e. pragmatically presupposed). More precisely, recognitional demonstratives
mark information that is (i) discourse new, (ii) hearer old, and (iii) 'private'
(Himmelmann uses the term "specific" rather than private). Private information
is information that speaker and hearer share due to common experience in the
past. It is distinguished from general cultural information shared by all members
of the speech community. General cultural information is also hearer old at its
first mention, but unlike private hearer old information it is marked by a definite
article in English. Hawkins (1978: 115-122) refers to the use of definite articles
with nouns expressing general cultural information as the "larger situational use
of the definite article". Two examples are given in (20a-b).
(20) English
a. Last night I met the President.
b. I joined the navy for two years.
The President and the navy express general cultural information which is familiar
to all speakers of the speech community including the current speaker and hearer.
Since the President and the navy do not convey private information, recognitional
demonstratives are not allowed in (20a-b). These examples show that the
recognitional use is restricted to nouns that encode private information.
The recognitional use often implies that speaker and hearer share the same
view or that they sympathize with one another. Consider the following examples
adapted from Lakoff (1974).
PRAGMATIC USE 107
(21) English (Lakoff 1974: 351, 352)
a. How's that throat?
b. That Henry Kissinger sure knows his way around in Hollywood.
The demonstrative in (21a) indicates that the speaker shares the hearer's concern
about his or her throat, and the demonstrative in (21b) suggests that the interlocutors
share the same view about Henry Kissinger. As in these examples, recognitional
demonstratives are often used to indicate emotional closeness, sympathy,
and shared beliefs and therefore Lakoff (1974) calls this use "emotional deixis".
Since the referent of a recognitional noun phrase was not previously
activated, the speaker cannot be certain as to whether the hearer will indeed
identify the referent. The information expressed by the noun following a recognitional
demonstrative may not be sufficient for the hearer to find the referent
in his or her knowledge store. In order to facilitate the identification task, the
speaker may provide additional information about the referent in a relative clause,
as in following example from German.
(22) German (Auer 1984: 637)
Was isn eigentlich mit diesem Haustelephon, was
what is MD with that house.telephone what
mir (wir) immer khabt ham;...
we used.to had have...
'What happened to that house telephone that we used to have?'
The Haustelephon 'house telephone' is a new discourse entity which is mentioned
for the first time in this sentence. It is marked by an adnominal demonstrative in
order to indicate that the hearer is familiar with the referent. However, since the
Haustelephon was not previously activated the speaker cannot be certain as to
whether the hearer will indeed identify the referent, and for that reason s/he
provides additional information about the Haustelephon in a relative clause.
Relative clauses are so frequently used after a recognitional mention that Himmelmann
(1996: 230) considers the occurrence of relative clauses and other noun
modifiers a secondary feature of this use. Furthermore, he notes that the noun
marked by a recognitional demonstrative is often followed by a pause, providing
a chance for the hearer to ask for clarification if s/he could not identify the
referent (cf. Himmelmann 1997: 72-3). This is exemplified in (23).
108 DEMONSTRATIVES
(23) German (Himmelmann 1997: 58)
A: Was hast n (dann) gelesen? (0.2sec)
what have you (then) read (0.2sec)
B: (Ja) diesen Aufsatz von dem Olson. (1.5sec)
(Yeah) that essay by the/that Olson (1.5sec)
A: Was is n des für einer? (0.4sec) Ach so! (0.2sec)
what is PART that P one (0.4sec) INTJEC (0.2sec)
Von dem hab ich immer noch nix mitgekriegt.
About him have I still yet anything heard
A: 'What did you read next?' (0.2sec)
B: 'That essay by Olson.' (1.5sec)
A: 'What kind of person is he?' (0.4sec) 'Oh wait! Right!'(0.2 sec)
'I still haven't heard anything about him.'
The 'essay by Olson', which speaker B mentions in line two, is a new discourse
topic. Since speaker B believes that speaker A is familiar with the referent,
Aufsatz 'essay' is marked by a recognitional demonstrative. Following this speech
act, there is a pause of 1.5 seconds which allows the hearer to ask for clarification.
The recognitional use must be distinguished from the use of the demonstratives
in the following examples:
(24) English (Himmelmann 1997: 78)
a. Similar payroll tax boosts would be imposed on those under the
railroad retirement system.
b. The true artist is like one of those scientists who, from a single
bone, can reconstruct an animal's entire body.
Similar to the recognitional use, the demonstratives in (24a-b) occur with a
subsequent relative clause. However, they do not appeal to private shared
knowledge, nor do they focus the hearer's attention on entities in the speech
situation or discourse. These demonstratives do not fit any of the uses discussed
thus far. Himmelmann (1997: 78) characterizes them as semantically empty
proforms that provide an anchorage point for the relative clauses that follow. In
chapter 6, I will argue that the head of a relative clause is often marked by a
grammatical item that developed from a recognitional demonstrative. I will refer
to such grammatical items as determinatives. The demonstratives in (24a-b) can
be seen as determinatives at an early stage of the grammaticalization process.
They are formally indistinguishable from adnominal demonstratives, but they have
lost the discourse pragmatic properties of a (recognitional) demonstrative and
PRAGMATIC USE 109
serve a purely grammatical function. In other languages, determinatives are
further grammaticalized and formally distinguished from their historical source
(cf. 6.4.4).
Apart from the recognitional use, there is one other context in which nonexophoric
demonstratives are commonly used without a referent in the preceding
discourse. In colloquial English, unstressed this is frequently used to introduce
new discourse topics (cf. Prince 1981; Wald 1983; Wright and Givón 1987;
Gernsbacher and Shroyer 1989). Consider the following example:
(25) English (Prince 1981: 233)
A few years ago, there was this hippie, long-haired, slovenly. He
confronted me...
Like recognitional that, unstressed this precedes a noun that is mentioned for the
first time, but it serves a very different discourse pragmatic function: recognitional
that introduces hearer old information, while unstressed this marks
information that is hearer new. Furthermore, while the referent of recognitional
that is usually of low topicality and does not recur in the subsequent discourse
(Himmelmann 1996: 230; 1997: 83), unstressed this marks important new topics
that usually persists in the subsequent discourse (cf. Prince 1981; Wald 1983;
Wright and Givón 1987; Givón 1990). Following Wright and Givón (1987), I
assume that unstressed this is not just a different use of the proximal demonstrative;
rather, it functions as an article used to mark specific indefinite information
(cf. Gundel et al. 1993: 275). I will discuss specific indefinite this together with
similar articles from other languages in the following chapter on grammaticalization
(cf. 6.4.6).
Like anaphoric and discourse deictic demonstratives, recognitional demonstratives
may have a particular form. Himmelmann (1996: 231-234, 1997: 62-71) reports
that there are several Australian languages that employ a particular demonstrative
in order to activate private hearer old knowledge. Yankunytjatjara, for instance,
uses the demonstratives panya in order to "call the listener's attention to the fact
that he or she is already familiar with the referent" (Goddard 1985: 107).
Himmelmann (1996: 232) argues that panya is a recognitional demonstrative and
he discusses further examples from Nunggubuyu and Mparntwe Arrernte.
5.5 The special status of exophoric demonstratives
The previous sections discussed four different pragmatic uses of demonstratives:
the exophoric, anaphoric, discourse deictic, and recognitional uses. In the
110 DEMONSTRATIVES
literature, it is often assumed that the exophoric use is the basic use from which
all other uses derive (cf. Brugmann 1904: 7-8; Bühler 1934: 390; Lyons
1977: 671). This view has recently been questioned by Himmelmann (1996; see
also Hanks 1990; Fuchs 1993; Laury 1997), who maintains that there is no
evidence for the common assumption that the exophoric use is basic. Moreover,
he argues that the four pragmatic uses discussed in the preceding sections must
be of equal status because all four uses are universally attested. If the three
endophoric uses were derived from the exophoric use, one would expect,
according to Himmelmann (1996: 242), that the transpositions (or extensions)
from the exophoric use to the three other uses would be less pervasive and less
regular.
Himmelmann proposes this hypothesis based on very little data. In my view,
it is still an open question whether all four uses are universally attested. And even
if it turns out that all four uses are universal, it would not rule out that the
exophoric use is basic and that the three other uses are derived from this use by
regular transpositions. In the remainder of this chapter, I will argue, following
previous work by Brugmann (1904), Bühler (1934), Lyons (1977), and many
others, that the exophoric use is indeed the basic use from which all other uses
derive. I present three arguments in support of this view. First, the exophoric use
is prior in language acquisition. Second, exophoric demonstratives are morphologically
and distributionally unmarked. And third, the grammaticalization of demonstratives
originates from the anaphoric, discourse deictic and recognitional uses;
that is, exophoric demonstratives are never immediately reanalyzed as grammatical
markers. All three arguments suggest that the exophoric use has a special
status. It is the prototypical use from which all other uses derive.
In a study on the acquisition of deictic words in English, Eve Clark (1978)
has shown that gestures are crucial for young children to learn the use of deictic
words such as this and that and here and there. She argues that the acquisition
of deictic expressions occurs in four steps. At first, children use a pointing
gesture without any words to focus the hearer's attention on entities in the
surrounding situation. Then they begin to use an isolated demonstrative accompanied
by a gesture. Next, they combine demonstratives with other linguistic
expressions, producing utterances such as this shoe or that mine. And finally, they
learn how to use demonstratives without a deictic gesture if the identification of
the referent is sufficiently determined by situational clues (Clark 1978: 96-97; cf.
Weissenborn 1988). The development is summarized in Table 64, which I
adopted from Clark's paper (1978: 97).
Although Clark does not distinguish between different pragmatic uses, it is
clear from her argumentation that the exophoric use is learned prior to the other
PRAGMATIC USE 111
uses because the exophoric use is the only use that is commonly accompanied by
Table 64. From deictic gesture to deictic word (Clark 1978: 97)
Stage Gesture Utterance Example
1
2
3
4
+
+
'da' (= 'that')
'that shoe'
'that coat is mine'
a pointing gesture. Children learn the use of exophoric demonstratives based on
the directive force of deictic gestures, which in turn provides the ground for the
acquisition of those uses that do not involve a gesture. The exophoric use of demonstratives
is thus of central significance for the development of deictic words
from deictic gestures, which suggests that the exophoric use plays a central role
within the deictic system.37
Additional support for this view comes from markedness theory (cf. Greenberg
1966; Croft 1990: chap4). As shown in the preceding sections, many
languages employ distinct demonstratives for different uses. In particular,
anaphoric demonstratives are often marked by an affix that is added to a demonstrative
root. I have given an example from Tümpisa Shoshone in Section 5.2.
Three further examples from Usan, Ngiyambaa, and West Greenlandic are shown
in (26) to (28).
(26) Usan (Reesink 1987: 80)
Exophoric e 'this'
Anaphoric e-ng 'this-GIVEN'
(27) Ngiyambaa (Donaldson 1980: 139)
Exophoirc ]ilu 'this.ERG'
Anaphoric ]ilu-la 'this.ERG-GIVEN'
(28) West Greenlandic (Fortescue 1984: 261)
Exophoric manna 'this'
Anaphoric ta-manna 'ANA-this'
Like anaphoric demonstratives in Tümpisa Shoshone, anaphoric demonstratives
in Usan, Ngiyambaa, and West Greenlandic are formed by adding an extra
morpheme to a demonstrative root. The anaphoric demonstratives are thus morphologically
more complex than their exophoric counterparts. Morphological complexity
is one of the criteria that typologists use in order to determine the unmarked
value of a grammatical category. Croft (1990) defines this criterion as follows:
112 DEMONSTRATIVES
Structure: the marked value of a grammatical category will be expressed by at
least as many morphemes as is the unmarked value of that category. (Croft
1990: 73)
Assuming this criterion, anaphoric demonstratives are structurally marked: in all
languages included in my sample, anaphoric demonstratives involve at least as
many morphemes as exophoric demonstratives, and in some languages they are
distinguished from exophoric demonstratives by an additional morpheme. That is,
if anaphoric and exophoric demonstratives do not have the same number of
morphemes, anaphoric demonstratives are always more complex.
Like anaphoric demonstratives, discourse deictic and recognitional demonstratives
may have a special form, but they are usually not more complex than
exophoric demonstratives. Since they do not occur with an additional morpheme
they are structurally unmarked. However, recognitional demonstratives have a
marked distribution. As pointed out above, recognitional demonstratives are only
used adnominally, while all other demonstratives occur in various syntactic
contexts. Lexical items that occur in fewer syntactic contexts than other members
of the same category are distributionally marked according to the following
criterion:
Behavior (distributional): if the marked value occurs in a certain number of
distinct grammatical contexts (construction types), then the unmarked value will
also occur in at least those contexts that the marked value occurs in. (Croft
1990: 82)
Thus, anaphoric demonstratives are morphologically (or structurally) marked
vis-à-vis demonstratives of the exophoric use, and recognitional demonstratives
have a marked distribution relative to the demonstratives of all other uses. Only
discourse deictic demonstratives do not seem to be marked by any of the
markedness criteria that typologists assume.
Finally, the exophoric use must be considered basic because the grammaticalization
of demonstratives always originates from one of the three endophoric
uses. Anaphoric demonstratives are, for instance, frequently reanalyzed as third
person pronouns (cf. 6.3.1); discourse deictic demonstratives provide a common
historical source for sentence connectives (cf. 6.3.4); and recognitional demonstratives
may develop into determinatives (cf. 6.4.4). Anaphoric, discourse deictic and
recognitional demonstratives serve language-internal functions; they are already
to some extent grammaticalized in that they function to organize the information
that is encoded in the ongoing discourse. Since exophoric demonstratives serve
a language-external function, they cannot be immediately reanalyzed as grammatical
markers. A demonstrative will always first go through a stage at which it is used
PRAGMATIC USE 113
with reference to linguistic entities in the surrounding discourse before it assumes
a specific grammatical function. One can think of the grammaticalization of demonstratives
as a continuum ranging from items that are used to orient the hearer
in the outside world to items that are routinely used to organize the lexical
material within the ongoing discourse. Exophoric demonstratives mark one end
of this cline. The other is represented by grammatical items such as third person
pronouns, sentence connectives, and determinatives. The three endophoric uses
are somewhere in between the two ends of this cline, referring to linguistic
entities within the universe of discourse. Figure 6 shows the cline of grammaticalization
that I propose.
Exophoric Endophoric Grammatical
anaphoric
discourse deictic
recognitional
e.g. 3rd person pronoun
e.g. sentence connective
e.g. determinative
exophoric
Figure 6. The grammaticalization cline of demonstratives
5.6 Summary
In this chapter, I have examined the pragmatic uses of demonstratives in discourse
and face-to-face conversations. Following Himmelmann (1996, 1997), I
have distinguished four different uses: the exophoric, anaphoric, discourse deictic,
and recognitional uses. Exophoric demonstratives are primarily used to orient the
hearer in the speech situation. They focus his or her attention on entities in the
world outside of discourse and they are often accompanied by a pointing gesture.
I have shown that the exophoric use is not restricted to concrete referents that are
visible in the speech situation. Exophoric demonstratives are also commonly used
with reference to abstract and removed objects, and they may indicate a referent
in a fictive situation evoked by the ongoing discourse (Deixis 'am Phantasma').
Anaphoric demonstratives are coreferential with a noun or noun phrase in the
previous discourse; they keep track of discourse participants that are contrastive,
emphatic and somewhat unexpected. Very often, they occur after the first mention
of a new referent in order to establish a new discourse topic. Discourse deictic
demonstratives refer to an adjacent chunk of discourse. More specifically, they
114 DEMONSTRATIVES
refer to aspects of meaning: the propositional content or illocutionary force of an
utterance. Unlike anaphoric demonstratives, discourse deictic demonstratives do
not track continuing topics; rather, they express an overt link between two
propositions or speech acts. Finally, recognitional demonstratives activate private
hearer old knowledge. They mark new discourse referents that speaker and hearer
know from common experience in the past. Very often, recognitional demonstratives
are accompanied by a relative clause (or a prepositional phrase) providing
additional information about the referent. In the final section, I argued that the
exophoric use of demonstratives is basic and unmarked. I presented three
arguments in support of this view: first, the exophoric use is crucial for the
acquisition of demonstratives by young children; second, exophoric demonstratives
are morphologically unmarked relative to anaphoric demonstratives and
distributionally unmarked vis-à-vis recognitional demonstratives; and third, the
grammaticalization of demonstratives originates from the three endophoric uses.
CHAPTER 6
Grammaticalization
The previous chapters dealt with synchronic aspects of demonstratives: their
morphological structures, semantic features, syntactic functions, and pragmatic
uses. This chapter is concerned with the diachrony of demonstratives. More
specifically, it investigates the development of demonstratives into grammatical
markers. Crosslinguistically, demonstratives provide a common historical source
for a wide variety of grammatical items such as definite articles, relative and third
person pronouns, copulas, sentence connectives, complementizers, number
markers, and possessives. The development of multiple grammatical markers from
a single source item has been called polygrammaticalization (Craig 1991). It
occurs when a single lexeme undergoes grammaticalization in several syntactic
contexts (cf. Lehmann 1995b: 1258). Grammaticalization is often viewed as a
process that involves isolated linguistic items, but it is the entire grammatical
construction, rather than an isolated item, that is subject to grammaticalization (cf.
Bybee, Perkins and Pagliuca 1994: 11). This chapter shows that the path-ofevolution
that a demonstrative might take is crucially determined by the syntactic
context in which it occurs. Specifically, it is argued that pronominal, adnominal,
adverbial, and identificational demonstratives give rise to four different sets of
grammatical markers which usually retain some of the syntactic properties that
the demonstrative had in the source construction. Pronominal demonstratives
develop into grammatical items that are either still used as pronouns (or have at
least some of the properties of a pro-nominal). Adnominal demonstratives give
rise to grammatical items that function as operators of nominal constituents.
Adverbial demonstratives evolve into operators of verbs or verb phrases. And
identificational demonstratives develop into grammatical markers that interact
with nominal constituents derived from predicate nominals. There is thus a rather
close correspondence between the syntactic function of the demonstrative in the
source construction and the grammatical function of the target.
This chapter is organized as follows. Section 6.1 discusses some general
principles of grammaticalization, providing the theoretical background for the
116 DEMONSTRATIVES
following investigation. Section 6.2 defines the criteria that I will use in order to
determine if and to what extent a demonstrative has undergone grammaticalization.
Sections 6.3 to 6.6 describe the grammaticalization channels that originate from
pronominal, adnominal, adverbial, and identificational demonstratives, respectively.
Finally, Section 6.7 addresses the question: where do demonstratives come
from - what is their historical source?
6.1 Some general principles of grammaticalization
Grammaticalization is usually defined as the process whereby lexical items
develop into grammatical items and items that are already grammaticalized
assume new grammatical functions (cf. Meillet 1921: 131-133; KurySowicz
1965: 52; Lehmann 1985: 303; Heine, Claudi and Hünnemeyer 1991a: 1-5;
Hopper and Traugott 1993:xv; Bybee, Perkins and Pagliuca 1994: 4-5). Lexical
items are content words, which encode the main semantic concepts of an
utterance. Grammatical items, on the other hand, are function words, which
organize the lexical material within a sentence. Lexical items comprise the major
word classes-nouns, verbs, and adjectives-while grammatical items subsume
such elements as prepositions, conjunctions, and auxiliaries. The former are
always open classes (except for adjectives, which may be open or closed class;
see Dixon 1982), while the latter are closed class items (cf. Talmy 1988).
Lexical and grammatical items form a cline of grammaticality ranging from
free content words to bound grammatical morphemes (Hopper and Traugott
1993: 7). Between the two ends of this cline there is a wide variety of items that
are more or less grammaticalized. Grammaticalization can be seen as the process
by which an item moves towards the grammatical end of this cline. This process
is unidirectional (cf. Hopper and Thompson 1993: 94-129). That is, lexical items
develop into grammatical items and grammatical items may further grammaticalize,
but grammatical items do not develop into lexical items or items that are less
grammaticalized. A few counterexamples have been cited in the literature (cf.
Jefferson and Zwicky 1980; Matsumoto 1988; Harris and Campbell 1995: 336-339),
but most of them are controversial (cf. Bybee, Pagliuca and Perkins 1994: 13-14),
and even if there are cases of "degrammaticalization" (Heine, Claudi and
Hünnemeyer 1991b: 149) they would be so rare that they hardly undermine the
claim that grammaticalization is basically a unidirectional process.
The development of grammatical items follows certain pathways, called
grammaticalization channels (cf. Heine, Claudi and Hünnemeyer 1991a: 221).
Grammaticalization channels are patterns of historical change leading time and
GRAMMATICALIZATION 117
again from the same source to the same target. Crosslinguistically, they are very
similar and they tend to be stable over time. That is, new grammatical items may
evolve through a certain grammaticalization channel at different times, giving rise
to grammatical categories with different layers. Many Indo-European languages
have, for instance, several layers of adpositions that represent different stages of
a grammaticalization process, leading from constructions including a relational
noun (e.g. in front of) to adpositions with highly abstract meanings (e.g. of) (cf.
Lehmann 1985).
Grammaticalization may affect all aspects of a linguistic sign: its phonological
form, its morphosyntactic features, and its meaning or function (cf. Heine and
Reh 1984; Hopper and Traugott 1993; Lehmann 1993, 1995a, 1995b). At the
phonological level, grammaticalization often involves a process of phonological
reduction and coalescence (cf. Bybee, Pagliuca and Perkins 1994: 4-9). Items
tend to shorten and to fuse with other elements in their environment when they
grammaticalize; free forms lose some of their phonological substance, turn into
clitics and then into affixes before they eventually disappear (cf. Heine and Reh
1984: 17-28). Phonological reduction and coalescence are perhaps the most
obvious signs that grammaticalization has occurred, but they are not restricted to
grammaticalization. Lehmann (1989) points out that lexicalization may also
involve phonological reduction and coalescence. He defines lexicalization as a
word formation process by which formerly independent items are combined into
complex words with idiosyncratic semantic properties. Unlike lexicalization,
grammaticalization gives rise to forms that are morphologically and semantically
regular and transparent (cf. Lehmann 1995b: 1263-1264).38
At the morphosyntactic level, grammaticalization often restricts the distributional
freedom of an item: more grammaticalized items tend to occur in a
specific slot in a grammatical construction. Very often they are arranged in
paradigms and their occurrence is obligatory in certain contexts (cf. Lehmann
1995a: 137-143). Moreover, lexical items may lose the ability to inflect when they
grammaticalize. Although this is frequently accompanied by phonological
reduction, it is in principle an independent process. Grammatical items may lose
the ability to inflect even if they keep all of their phonological substance: they
may occur with a frozen affix that has lost its meaning or function.
At the semantic level, grammaticalization involves a process of semantic
bleaching or fading (Sweetser 1988, 1990): lexical items become semantically
less concrete and pragmatically less significant (cf. Heine and Reh 1984: 15). At
the same time, they gain new grammatical functions or meanings (cf. Sweetser
1988; Heine, Claudi and Hünnemeyer 1991a). According to Traugott (1989),
linguistic signs become more subjective when they grammaticalize. She argues
118 DEMONSTRATIVES
that the meaning of a grammatical item is often situated in the speaker's subjective
belief toward the situation.
Heine, Claudi and Hünnemeyer (1991a) have shown that grammaticalization
processes are commonly motivated by metaphorical extensions, metonymic
transfers, and pragmatic inferences (cf. Claudi and Heine 1986; Traugott and
König 1991). In their view, and in the view of many others, the initial stage of
a grammaticalization process usually involves the metaphorical (or metonymic)
use of a lexical expression or a conversational implicature that is triggered by a
specific item.
6.2 Criteria for the grammaticalization of demonstratives
Having described the general principles of grammaticalization, I turn now to the
features that characterize the grammaticalization of demonstratives. As pointed
out at the end of Chapter 5, the grammaticalization of demonstratives is a
continuous process leading from exophoric demonstratives used to orient the
hearer in the outside world to grammatical items serving specific syntactic
functions. This process involves a number of changes that one might summarize
as follows:
Functional changes
1. Grammatical items that developed from demonstratives are no longer used
to focus the hearer's attention on entities in the outside world.
2. They are deictically non-contrastive.
Syntactic changes
3. Their occurrence is often restricted to a particular syntactic context.
4. They are often obligatory to form a certain grammatical construction.
Morphological changes
5. They are usually restricted to the distal or, less frequently, the proximal
form.
6. They may have lost their ability to inflect.
Phonological changes
7. They may have undergone a process of phonological reduction.
8. They may have coalesced with other free forms.
These are eight criteria that one might use in order to determine if and to what
extent a demonstrative has grammaticalized. The two functional criteria apply to
(almost) all demonstratives that have undergone grammaticalization: grammatical
GRAMMATICALIZATION 119
items that developed from demonstratives do not function to orient the hearer in
the outside world and they are always non-contrastive (but see Section 6.5.2 for
discussion of a notable exception). The other criteria may or may not apply,
depending on the grammaticalization channel, the properties of the source item,
and the stage that an emergent grammatical marker has reached. More grammaticalized
demonstratives are more likely to have undergone any of the formal
changes than less grammaticalized demonstratives. At the initial stage of a
grammaticalization process, grammatical markers often have the same morphosyntactic
and phonological properties as the source items (Hopper 1991).
Given the criteria I have suggested, one has to assume that anaphoric,
discourse deictic, and recognitional demonstratives are already to some extent
grammaticalized (as I have argued in Section 5.5). All three endophoric demonstratives
serve language-internal functions and they are non-contrastive.39
Moreover, anaphoric and recognitional demonstratives are usually restricted to the
distal form, which suggests that they have undergone a morphological process,
and recognitional demonstratives occur only in adnominal position, which indicates
that they have changed syntactically. The division between grammatical items and
endophoric demonstratives is strictly speaking an idealization. There is no clearcut
borderline that separates endophoric demonstratives from grammatical
markers. There are only demonstratives that are more or less grammaticalized.
In the following four sections, I describe eighteen grammaticalization
channels that commonly originate from a demonstrative. For each channel, I
provide at least one example of a grammatical item that has undergone some
morphosyntactic and/or phonological changes so that source and target are
formally distinguished. Some of the grammaticalization processes that I examine
have been discussed extensively in the literature on grammaticalization, but other
changes have only been described in reference grammars or other special sources.
I begin my investigation with the reanalysis of pronominal demonstratives,
followed by the grammaticalization of adnominal and adverbial demonstratives,
and I conclude with grammatical items that developed from identificational demonstratives
in copular and nonverbal clauses.
6.3 The grammaticalization of pronominal demonstratives
6.3.1 Third person pronouns
In many languages, third person pronouns are historically derived from pronominal
demonstratives.40 Givón (1984: 353-360) has shown that the emergence of
120 DEMONSTRATIVES
third person pronouns from demonstratives is part of a diachronic cline that one
might describe as follows:
(1) DEM PRO > third person PRO > clitic PRO > verb agreement
At the beginning of this cline we find anaphoric pronominal demonstratives
tracking emphatic, contrastive and unexpected discourse topics. Anaphoric demonstratives
that develop into third person pronouns become de-stressed and their
use is gradually extended to all persisting topics. When third person pronouns
continue to grammaticalize they may become clitics, which may eventually turn
into agreement markers (cf. Givón 1984: 353; Lehmann 1995a: 39-42).
The entire cline is attested in the history of the French language. Modern
standard French uses pronominal clitics to track continuing topics, but in certain
nonstandard varieties the clitics are essentially used as agreement markers, which
are commonly accompanied by a coreferential (pro)noun (cf. Lambrecht 1981).
Historically, the clitics go back to free third person pronouns, which in turn
developed from the demonstrative ille in Vulgar Latin (e.g. Harris
1978: 100-101).
6.3.2 Relative pronouns
Relative pronouns are often marked by a relativizer, which might be an (uninflected)
particle or an (inflected) pronoun (cf. Keenan 1985). Though relative
pronouns are widely used in European languages (both Indo-European and non-
Indo-European such as Hungarian, Finnish and Georgian), they seem to be very
uncommon in other parts of the world (cf. Comrie 1998). In my sample, there is
only one non-European language that has a relativizer which one might consider
a relative pronoun: Tümpisa Shoshone (Dayley 1989: 357-374). The relative
pronoun in Tümpisa Shoshone is, however, rather different from the kind of
relative pronoun that is commonly found in European languages. In European
languages, relative pronouns tend to occur in the initial position of the relative
clause and their case features are determined by their syntactic function within
the embedded clause (cf. Comrie 1998). By contrast, relative pronouns in
Tümpisa Shoshone may occur in any position in the relative clause and they have
the same case features as the coreferential noun in the superordinate clause.
Relative pronouns have two common historical sources: (anaphoric) pronominal
demonstratives and wh-question words (cf. Lehmann 1984). Consider the
following examples from German:
GRAMMATICALIZATION 121
(2) German
a. Er hat einen neuen Vorschlag gemacht, der
he has a new suggestion made REL
mir besser gefallen hat.
me better pleased has
'He made a new suggestion, which I liked better.'
b. Er hat einen neuen Vorschlag gemacht; der hat
he has a new suggestion made REL has
mir besser gefallen.
me better pleased
'He made a new suggestion; this one I liked better.'
The relative pronoun in (2a) occurs in a dependent clause, marked by the position
of the finite verb at the end of the sentence. It has the same form as the pronominal
demonstrative in (2b), which is embedded in an independent main clause
(marked by a finite verb in second position). Relative pronouns and pronominal
demonstratives are morphologically indistinguishable in German, but since their
syntactic features are different they are commonly distinguished (cf. Eisenberg
1994: 200).41 Relative pronouns are always fronted (i.e. they are always the first
element in the relative clause), whereas pronominal demonstratives occur in the
usual position of a (pro-)noun. That is, the relative pronoun in (2a) can only
occur clause-initially, while the pronominal demonstrative in (2b) may also occur
after the dative pronoun mir and the finite verb hat (cf. relative pronoun: *Mir
der besser gefallen hat vs. demonstrative pronoun: Mir hat der besser gefallen).
Moreover, while relative pronouns are generally unstressed, anaphoric demonstratives
may bear a contrastive accent.
The development of relative pronouns in German is controversial. Behaghel
(1923-1932, III: 766) argues that they derive from an (anaphoric) pronominal demonstrative
that continued a noun of the preceding sentence. However, Paul
(1916-20, IV: 189-191) and Lockwood (1968: 242-244) contend that they
evolved from a pronominal demonstrative that was originally governed by the
verb in the main clause, as in (3):
(3) Old High German (Lockwood 1968: 243)
enti aer antwurta demo za imo sprah
but he answered DEM.DAT to him spoke
'But he answered the one (who) spoke to him.'
In this example, demo is syntactically part of the main clause (i.e. it is governed
by the main verb antwurta 'answer'), but semantically it belongs to both main
122 DEMONSTRATIVES
and subordinate clause: it is an undergoer in the main clause and an actor in the
subordinate clause. If main and subordinate clause require the same case, it is
possible to interpret the demonstrative as a syntactic element of the subordinate
clause:
(4) Old High German (Lockwood 1968: 243)
thô liefun sâr thie nan minnôtun meist
then ran at.once DEM.NOM him loved most
'Then ran at once they who loved him most.'
The demonstrative in (4) is syntactically ambiguous: it occurs in nominative case
and could be interpreted as the syntactic subject of the verb in either the main
clause or the subordinate clause. According to Lockwood, the relative pronoun
in German evolved from pronominal demonstratives that were ambiguous as in
this example. The conversion was completed once the pivotal pronoun was represented
twice, in main and subordinate clause, as in the following example from
Early New High German:
(5) Early New High German (Lockwood 1968: 244)
Wer ist die, die aufgehet aus der Wüste wie
who is the.one who rises from the desert like
ein gerader Rauch?
a straight smoke
'Who is the one who rises from the desert like smoke?'
Lehmann (1984: 378-383) suggests yet another developmental pathway along
which the relative pronoun in German might have emerged. He argues that the
German relative pronoun evolved from an adnominal demonstrative that preceded
a postnominal attribute as in (6):
(6) Old High German (Lehmann 1984: 378)
kuningin thia richun
queen DEM mighty
'the mighty queen'
Lehmann maintains that relative clauses in German are the result of a process
whereby attributive adjectives like the one in (6) are expanded to attributive
clauses in which the adnominal demonstrative is reinterpreted as a relative
pronoun. A crucial stage of this process is marked by participial constructions
like the one in (7), which are commonly found in Old High German:
GRAMMATICALIZATION 123
(7) Old High German (Lehmann 1984: 379)
ich bim Gabriel thie azstantu fora gote
I am Gabriel DEM standing in.front.of god
'I am Gabriel standing in front of God.'
Lehmann argues that the adnominal demonstrative in (7) establishes an overt link
between the preceding noun and the participial construction, which it nominalizes.
This construction has properties of both the attributive adjective in (6) and a fullfledged
relative clause: like the attributive adjective in (6), the participial
construction in (7) is a nominalized constituent that is linked to the preceding
noun by an adnominal demonstrative; and like full-fledged relative clauses in
Modern German, it is a clausal attribute that includes a verb, though the verb is
nonfinite. Lehmann (1984: 379) argues that participial constructions of this sort
developed into finite relative clauses by analogy to an older relative construction
that included a finite verb and later disappeared.
Though Lehmann rejects the common assumption that the relative pronoun
in German evolved from a pronominal demonstrative, he does not deny that there
are languages in which relative pronouns can be traced back to pronominal demonstratives.
On the contrary, he argues that the relative pronoun in Ancient
Greek evolved from an anaphoric demonstrative pronoun. If Lehmann's analysis
is accurate, relative pronouns may arise from two types of demonstratives: (i)
pronominal demonstratives (which are either used as anaphors or as the pivot of
a bi-clausal construction) and (ii) adnominal demonstratives that link a nominal
attribute to the head noun.
6.3.3 Complementizers
Complementizers are frequently based on pronominal demonstratives that are used
as discourse deictics (cf. Frajzyngier 1991). The complementizers of North and
West Germanic languages, for instance, arose from a demonstrative that originally
occurred in the main clause referring forward to the subsequent proposition.
Harris and Campbell (1995: 287) provide the following example from Middle
High German, which exemplifies the source construction.
(8) Middle High German (Harris and Campbell 1995: 287)
joh gizalta in sâr tha, thiu sâlida untar in uuas
and told them immediately that the luck among them was
'And he told them immediately that good fortune was among them.'
The initial clause in (8) includes the pronominal demonstrative tha, which
anticipates the information expressed in the following clause. When the cataphoric
124 DEMONSTRATIVES
demonstrative was reanalyzed as a complementizer it became associated with the
following subordinate clause where it occurs in Modern German.
The same process gave rise to the complementizer that in English (cf.
Traugott 1992: 230-238). Hopper and Traugott (1993: 185-189) argue that the
that-complementizer started out as a copy of a cataphoric pronominal demonstrative
that occurred in the preceding main clause. Consider the following example
from Old English:
(9) Old English (Hopper and Traugott 1993: 186)
þæt gefremede Diulius hiora consul, þæt þæt
DEM arranged Diulius their consul COMP DEM
angin wearð tidlice þurhtogen
beginning was in.time achieved
'Their consul Diulius arranged (it) that it was started on time.'
The initial þæt in (9) is a fronted object pronoun used to anticipate the complement
clause, which is introduced by a copy of the cataphoric demonstrative.
Hopper and Traugott (1993: 186) point out that one could analyze the complement
clause as an appositive of the object pronoun, rather than an argument of the verb
in the preceding clause. The appositive clause turned into a complement clause
when the cataphoric demonstrative was no longer used to anticipate its occurrence.
The complementizers of North Germanic languages such as Swedish att
and Icelandic að developed along the same pathway. Lockwood (1968: 222-223)
provides the following two examples from Faroese, which show that the grammaticalization
channel is still productive.
(10) Faroese (Lockwood 1968: 223)
a. eg sigi, at hann kemur
I say that he comes
'I say that he comes.'
b. eg sigi tað, hann kemur
I say that he comes
'I say that: he comes.'
The complement clause in (10a) is marked by the complementizer at, which
introduces the embedded clause. Like the complementizers in English and
German, the complementizer in Faroese can be traced back to a cataphoric demonstrative
in the main clause. Example (10b) has the same structure as the
construction from which the complement clause in (10a) developed. It includes
a pronominal demonstrative that is used to anticipate the subsequent clause. This
GRAMMATICALIZATION 125
sentence exemplifies the source construction from which the complementizers in
North and West Germanic languages evolved.
6.3.4 Sentence connectives
Sentence connectives are frequently formed from a pronominal demonstrative and
some other element (e.g. an adverb or adposition) that indicates the semantic
relationship between the conjoined propositions. Consider, for instance, the
following example from Hixkaryana:
(11) Hixkaryana (Derbyshire 1985b: 157)
nomokyaknano tuna heno. Gre ke romararGn
it.was.coming rain QNT DEM because.of my.field
hokohra wehxaknano
NEG I.was
'It was raining heavily. Therefore I did not work on my field.'
Example (11) shows two clauses linked by the pronominal demonstrative Gre and
the causal postposition ke. Derbyshire (1985a: 57, 1985b: 157) treats Gre ke as a
sentence connective, consisting of two words (cf. English so that), which are
routinely used in combination to express a causal link between two propositions.
Similar sentence connectives occur in many other languages in my sample. Epena
Pedee, for instance, has a number of 'temporal relators' (Harms 1994: 144)
formed from the distal demonstrative ma and a suffix that indicates the semantic
relationship between the linked propositions. The most common temporal relator
is mapái 'and/so then', consisting of ma and -pái, which Harms (1994: 144)
glosses as 'only'. Mapái is used to express a link between two closely related
events, as in (12):
(12) Epena Pedee (Harms 1994: 144)
tá-cˇi náwe te cˇáa peecˇiadai-pá-ri ma-pái cˇi
our-REF mother house each dizzy-HAB-PRES that-only REF
khari-pa-rí-pa khari-pá-ri kh½i ra pa-ru-má-a
sing-HAB-PRES-ERG sing-HAB-PRES face arrive-PRES-LOC-DAT
'The shaman gets dizzy and loses her balance in each house. And
then the singer sings until she (the shaman) revives.'
Sentence connectives are either based on pronominal demonstratives that are used
as discourse deictics, or they involve a manner (adverbial) demonstrative. Epena
Pedee, for instance, has a number of sentence connectives that consist of the
manner demonstrative maa 'like that' and a morpheme that specifies the semantic
126 DEMONSTRATIVES
relationship between the conjoined propositions. Example (13) includes the
sentence connective maaphéda 'after that' formed from maa 'like that' and -phéda
meaning 'after':
(13) Epena Pedee (Harms 1994: 145)
perõrá-pa imáma wárra pee-thaa-hí maa-phéda
spotted.cavy-ERG tiger son kill-OBJ-PAST like.that-after
unu-hi-dá ewári ába mée
find-PAST-PL day one jungle
'A spotted cavy killed a tiger's child. After that, one day they met
in the jungle.'
Like Epena Pedee, Khasi has a set of sentence connectives that are formed from
a distal demonstrative and a bound morpheme. Example (14) exhibits a complex
sentence consisting of two clauses linked by na]ta 'then', which is formed from
the adpositional marker na]- and the demonstrative root -ta:
(14) Khasi (Nagaraja 1985: 100)
u khla u la bam na]-ta u la thyú
ART tiger ART PAST ate P-DEM ART PAST slept
'The tiger ate then he slept.'
Finally, German employs a large number of pronominal adverbs that developed
from an old oblique form of the pronominal demonstrative das (i.e. dara, dar)
and an adposition:
The pronominal adverbs in Table 65 are either used to substitute for a preposi-
Table 65. Pronominal adverbs in German (Paul 1992)
Modern form Source Gloss
damit
darüber
darum
dabei
darin
darauf
dazu
dafür
dagegen
da:r-mit(i)
dara ubiri
da:r-umbi
da:r-bi:
da:r-inne
da:r-u:f
da:ra-zuo
dara-fure
dara-gegene
DEM.OBL-with
DEM.OBL-above
DEM.OBL-because.of
DEM.OBL-by
DEM.OBL-in
DEM.OBL-on.top.of
DEM.OBL-to
DEM.OBL-for
DEM.OBL-against
tional phrase or they function to join two neighboring clauses. These forms are
GRAMMATICALIZATION 127
only weakly grammaticalized: the prepositions are still governed by the verb and
the demonstratives establish an anaphoric link like discourse deictics but unlike
fully grammaticalized conjunctions. In the following example, the pronominal
adverb darüber 'about that/it' is used to indicate a link between the preceding
question and the following clause.
(15) German
A: Was machst du, wenn du fertig bist?
what do you when you finished are
'What are you going to do when you are finished?'
B: Darüber habe ich noch nicht nachgedacht.
DEM.OBL.about have I yet not thought.about
'I haven't thought about it yet.'
Darüber is composed of the old oblique form of the pronominal demonstrative
das, which is no longer used in Modern German, and the preposition über
'about', which is governed by the verb nachdenken at the end of the clause.
6.3.5 Possessives
Many languages do not have particular possessive markers, instead using personal
or demonstrative pronouns (often in genitive case) in order to indicate the
possessor. For instance, in Supyire a possessor may be realized by a pronominal
demonstrative preceding the noun denoting the possessee. Possessive demonstratives
have the same form as adnominal demonstratives, but unlike the latter,
possessive demonstratives agree with their antecedent rather than with the head
noun. The different agreement behavior of adnominal and possessive demonstratives
is shown in (16a-b): the adnominal demonstrative in (16a) has the same
noun class features as the following noun, while the noun class of the possessive
demonstrative in (16b) corresponds to the noun class of its referent or antecedent.
(16) Supyire (Carlson 1994: 200, 201)
a. ']ké ba-gé
DEM(G2S) house-DEF(G2S)
'this/that house'
b. ']gé ba-gé
DEM(G1S) house-DEF(G2S)
'that/this one'e house'
The possessive demonstratives in Supyire are ordinary pronominal demonstratives
serving a particular semantic function in this construction. They may, however,
128 DEMONSTRATIVES
turn into possessive markers if they become disassociated from pronominal demonstratives
that are used in other contexts. The development of possessive
markers from pronominal demonstratives is well attested. For instance, the French
possessive leur 'their' developed from the genitive masculine plural form of the
pronominal demonstrative ille in Vulgar Latin. Harris (1978: 87-95) describes the
development as follows. Classical Latin used the possessive pronoun suus in
order to indicate a possessor within the same sentence, and it used the genitive
forms of the anaphoric demonstrative is in order to indicate a possessor that is not
mentioned in the same clause. In other words, Classical Latin distinguished
reflexive from non-reflexive possessives: suus was used as a reflexive possessive,
and eius SG.M, eorum PL.M and earum PL.F functioned as non-reflexive possessives.
The latter were later replaced by illius SG.M⁄F, illorum PL.M and illarum
PL.F when the demonstrative ille took over the function of is. Old French lost the
distinction between reflexive and non-reflexive possessives and restricted the use
of suus to singular while the former masculine plural form of the non-reflexive
possessives, illorum, was adopted as the corresponding plural form. The singular
and plural feminine forms, illius and illarum, disappeared. As the grammaticalization
process continued, illorum was shortened to leur (il-lor-um) and by
analogy it developed a new plural form, leurs, used to indicate multiple possessees.
6.4 The grammaticalization of adnominal demonstratives
6.4.1 Definite articles and noun class markers
Adnominal demonstratives provide a common historical source for definite
articles. The development has been described in numerous studies including
Christophersen (1939), Heinrichs (1956), Krámský (1972), Ultan (1978a), Harris
(1978, 1980), Greenberg (1978, 1991), Lüdtke (1991), Vogel (1993) Cyr (1993a,
1993b, 1996), Leiss (1994), Epstein (1994, 1995), Lehmann (1995a), Laury
(1995, 1997), and Himmelmann (1997, 1998). This section summarizes the
central findings discussed in these works.
Most of the studies that I have cited assume that definite articles arise from
anaphoric adnominal demonstratives (e.g. Greenberg 1978: 69; see Himmelmann
1997 for a different view). The use of anaphoric demonstratives is usually
confined to non-topical antecedents that tend to be somewhat unexpected,
contrastive or emphatic (cf. 5.2). When anaphoric demonstratives develop into
definite articles their use is gradually extended from non-topical antecedents to
GRAMMATICALIZATION 129
all kinds of referents in the preceding discourse. In the course of this development,
demonstratives lose their deictic function and turn into formal markers of
definiteness. An example of such a definite marker is the article the in English.
The semantic reanalysis of adnominal demonstratives as markers of definiteness
is usually accompanied by certain formal changes. Since articles are
generally unstressed they often lose some of their phonological substance and
cliticize to an element in their environment. The definite articles in Swedish,
Rumanian and Basque, for instance, are bound morphemes (i.e. enclitics or
suffixes) that evolved from free forms.
Furthermore, as we have seen in Chapter 4, adnominal demonstratives are
often independent pronouns that are only loosely adjoined to a coreferential noun
in apposition. Since articles are in general syntactically dependent, one has to
assume that adnominal demonstratives often lose their status as free nominals
when they become reanalyzed as definite markers (cf. Himmelmann
1997: 144-157).
Finally, Plank and Moravcsik (1996) report (based on unpublished work by
Siewierska and Bakker 1992) that, at least in European languages, demonstratives
are significantly more often inflected than articles, which suggests that adnominal
demonstratives often lose the ability to inflect when they grammaticalize as
definite markers.
Once demonstratives have turned into definite markers, their use may spread
from definite nouns to nouns expressing specific indefinite information (cf.
Greenberg 1978). When this happens, articles occur with (almost) every noun,
definite and indefinite, unless the noun is (i) non-specific (i.e. generic), (ii)
inherently definite (e.g. proper names), or (iii) otherwise marked for definiteness
(e.g. by a demonstrative). Greenberg (1978) mentions several Bantu languages
having articles of this type (Bemba, Zulu, Xhosa). When such articles continue
to grammaticalize they often turn into gender or noun class markers before they
eventually disappear (cf. Schuh 1990). Table 66 shows three nouns from Turkana
marked by three different gender prefixes whose historical relationship to the
distal demonstratives is still reflected in their phonological form:
Table 66. Noun class markers in Turkana (Dimmendaal 1983: 307)
NC-NOUN DEM
(\)a-b7r-~'
(\)e-kìle
(\)i-ì]ok
'woman'
'man'
'dog'
ya'
ye'
yi'
'that.F'
'that.M'
'that.N'
130 DEMONSTRATIVES
6.4.2 Linkers
Definite articles are used to indicate the information status of a nominal expression.
They must be distinguished from linking articles, or linkers, which function
to establish an overt link between the elements of a complex noun phrase. Unlike
definite articles, which tend to occur at the margin of an NP, linkers are usually
placed between the head noun and the associated modifiers. Consider, for
instance, the following examples from Tagalog, in which the noun modifiers (i.e.
an adjective in (17a), a relative clause in (17b), a numeral in (17c), a demonstrative
in (17d), and an interrogative pronoun in (17e)) are linked to the following
noun by the linking article na (allomorph -ng).
(17) Tagalog (Himmelmann 1997: 160, 161, 161, 162; Foley 1980: 181)
a. ang maliít na langgám
SPEC small LK ant
'the little ant'
b. ang paa ng mama... na bàbaríl sa kanyá
SPEC foot GEN man LK gun LOC 3SG.DAT
'the feet of the man who was going to shoot at him'
c. sa isa-ng mànlalakbáy
LOC one-LK traveler
'about a traveler'
d. ay yuu-ng mama
PRED DEM-LK man
'when that man'
e. ano-ng puno
what-LK tree
'what tree'
Similar constructions occur in many other Austronesian languages, in which the
linker has often the same or a very similar form as the linker in Tagalog (cf.
Foley 1980). Examples from Toba Batak, Tolai, Wolio, and Ilokano are given in
(18) to (21).
(18) Toba Batak (Foley 1980: 186)
huta na leban
village LK another
'another village'
GRAMMATICALIZATION 131
(19) Tolai (Foley 1980: 189)
a mamat na vat
ART heavy LK stone
'a/the heavy stone'
(20) Wolio (Foley 1980: 192)
heqgane na be-a-umba-mo
time LK INTENTION-3SG-COME-DEF
'the time he will come'
(21) Ilokano (Foley 1980: 185)
ti kuarta nga in-gatang-mo
ART money LK PERF-buy-2SG
'the money with which you bought'
Himmelmann (1997: 172-188) argues that the linkers in Tagalog and other
Austronesian languages developed from adnominal demonstratives. He points out
that na (and nga) is not only a common form of the Austronesian linker; the
same form is also frequently used as a medial or distal demonstrative (e.g. na
'that' Sichule, nana 'that' Kambera) (Himmelmann 1997: 164). Table 67 shows
the Proto-Oceanic demonstratives as reconstructed by Ross (1988; adopted from
Himmelmann 1997: 164). Note that the medial demonstrative na has the same
form as the linkers in Tagalog, Toba Batak, Tolai, and Wolio, which probably
emerged from this form.
Linkers are not only found in the Austronesian language family. There are also
Table 67. Proto-Oceanic demonstratives (Ross 1988: 100)
PROXIMAL
MEDIAL
DISTAL
*e/ne
*a/na
*o/no
linkers in Chadic (Schuh 1983a, 1990), Cushitic (Hetzron 1995), and Albanian
(Sasse 1991). In Chadic and Cushitic, the linker is a genitive suffix that evolved
from a (demonstrative) determiner (cf. Schuh 1990: 601-607). Schuh (1983a,
1990) presents two possible scenarios for the development of the "genitive-linking
morphemes" in Bade, Hausa, Kera, and other Chadic languages, without committing
himself to either one. Either they evolved from a determiner of the possessed
noun, as in (22a), or they developed from a pronoun that functioned as the head
of the possessor, as in (22b).
132 DEMONSTRATIVES
(22) The evolution of genitive-linking morphemes in Chadic (cf. Schuh
1990: 605-6)
a. [[possessee DET] possessor ] > possessee-GEN possessor
[[horse that] of John's]
b. possessee [PRO possessor] > possessee-GEN possessor
[horse [that of John's]]
If (22b) is the right scenario the genitive linkers would not derive from determiners
(i.e. adnominal demonstratives) but rather from (demonstrative) pronouns
that occurred with a possessor NP in apposition.
The Albanian linker is a free morpheme, which is obligatory with adjectives
(23a) and genitive nouns (23b), but does not occur with demonstratives (23c) and
other noun operators.
(23) Albanian (Himmelmann 1997: 167, 171, 171)
a. (një) shok i mirë
one friend.INDEF.M LK.NOM.SG.M good
'a good friend'
b. nën-a e vajz-ë
mother.DEF.NOM.SG.F LK.NOM.SG.F girl-INDEF.GEN.SG.F
'a girl's mother'
c. ky libër
this.NOM.SG.M book.INDEF.M
'this book'
Himmelmann (1997: 172-183) shows that linkers occur more frequently with
lexical attributes such as adjectives, genitives and relative clauses than with noun
operators such as demonstratives and interrogative pronouns (cf. Foley 1980).
Based on this finding, he argues that the grammaticalization of linkers originates
in constructions that involve a noun and a lexical attribute and that the use of
linkers with noun operators is due to later extensions. If Himmelmann's hypothesis
is correct the linker in Tagalog is further grammaticalized than the linker in
Albanian. The former is used with both attributes and operators, while the use of
the latter is restricted to certain types of lexical attributes and does not occur with
noun operators.
6.4.3 Boundary markers of postnominal relative clauses/attributes
Relative pronouns are only one of several relative markers that may arise from
a demonstrative. Many African languages have relative clauses in which the head
of the relative construction is marked by an adnominal demonstrative that is
GRAMMATICALIZATION 133
repeated at the end of the relative clause.42 Consider, for instance, the following
example from Izi:
(24) Izi (Meier, Meier and Bendor-Samuel 1975: 165)
kèbé ndú» ònó !nwé né ngú ònó ré
keep people DEM own mother your DEM well
'Keep your mother's relatives well.'
The relative construction in (24) includes two adnominal demonstratives: one that
occurs after the head noun, and one that occurs at the end of the relative clause
(which is a nonverbal clause in this case). The initial demonstrative can be
analyzed as a modifier of the head noun, but the final demonstrative does not
have an obvious function. Meier, Meier and Bendor-Samuel (1975: 165) argue
that the final instance of ònó serves as a boundary marker of the relative clause:
it "gives cohesion to what otherwise would be a rather loose construction". In
their perspective, the final ònó is not an ordinary demonstrative, but rather a
grammatical marker used to form relative clauses. The same type of relative
construction occurs in Sango:
(25) Sango (Samarin 1967: 73)
á.famille só ahé mbi só, mbi yí ála p7p7
relatives that laugh 1SG that 1SG like them NEG
'Relatives who make fun of me, I don't like them.'
Like relative clauses in Izi, relative clauses in Sango are often marked by a copy
of the demonstrative that accompanies the head noun at the end of the relative
clause. According to Samarin (1967: 73), the final demonstrative is used to tie the
whole construction together; it is a specific grammatical marker used to indicate
the final boundary of the relative clause.
Sankoff and Brown (1976) describe the emergence of a similar relative
construction in Tok Pisin. Like relative clauses in Izi and Sango, relative clauses
in Tok Pisin occur with an adnominal demonstrative at the final boundary of the
relative clause.
(26) Tok Pisin (Sankoff and Brown 1976: 632)
meri ia, em i yangpela meri, draipela meri ia,
girl REL 3SG PRED young girl big girl REL
em harim istap
3SG listen ASP
'This girl, who was a young girl, big girl, was listening.'
134 DEMONSTRATIVES
The relative clause in (26) is marked by the particle ia, which is etymologically
related to the adverbial demonstrative here in English. Ia was first reanalyzed as
an adnominal demonstrative before it assumed the function of a boundary marker
in relative clauses (e.g. man ia 'this guy'; Sankoff and Brown 1976: 639-641).
Sankoff and Brown (1976: 657) point out that ia is often omitted when the
relative clause occurs at the end of a sentence, where the final boundary of the
relative clause is sufficiently marked by intonation. They characterize the two
instances of ia in (26) as a "bracketing device" used to mark relative clauses and
other postnominal attributes (Sankoff and Brown 1976: 631); that is, both
instances of ia serve a grammatical function in this construction. Example (27)
shows that the ia...ia construction is not only used to mark relative clauses; it
also occurs with nominal attributes that follow a preceding noun.
(27) Tok Pisin (Sankoff and Brown 1976: 642)
... na em, man ia, lapun man ia, stap autsait ia
man this old man this stayed outside here
'... and this man, this old man, stayed outside.'
In this example, ia "brackets" the noun phrase in apposition to man. Sankoff and
Brown maintain that ia...ia has basically the same function in this construction
as in relative clauses. In both contexts, it is used to mark lexical material that
provides necessary information for the identification of the preceding noun
(Sankoff and Brown 1976: 640).43
Finally, Ewe has relative clauses that are marked by two relative particles,
si and lá, which seem to have the same function as the two ias in Tok Pisin.
(28) Ewe (Heine and Reh 1984: 251)
nyf'nu si vá étsf lá mé-ga-le o
woman REL come yesterday REL NEG-yet-be NEG
'The woman who came yesterday is no longer here.'
Sentence (28) includes a relative construction that is marked by two relative
particles: si and lá. Si follows the head noun and lá occurs at the end of the
relative clause. Historically, si is related to the proximal demonstrative sia, which
is composed of a demonstrative root and a definite marker, and lá can be traced
back to a definite article, which in turn may have developed from an adnominal
demonstrative. Heine and Reh (1984: 251) consider si and lá a "discontinuous
morpheme" used to mark relative clauses. According to their analysis, si...lá has
basically the same function as ia...ia in Tok Pisin. In both languages, relative
clauses are framed by two relative markers that derive from adnominal demonstratives
(cf. Benveniste 1966).44 Heine and Reh (1984: 251) argue that relative
GRAMMATICALIZATION 135
clauses in Ewe might have developed from an afterthought construction that was
used to clarify the meaning of the preceding noun, which has now become the
head of the relative construction.
6.4.4 Determinatives
In Chapter 5 it was pointed out that a recognitional demonstrative may turn into
a relative marker that indicates the nominal head of a relative clause. Himmelmann
(1997: 77-78) argues that English has such a relative marker at an early
stage of the grammaticalization process.
(29) English (Himmelmann 1997: 78)
...provision was made for payment for unemployment relief by nationwide
taxation rather than by a levy only on those states afflicted with
manpower surplus.
The distal demonstrative in this example does not indicate a referent in the
preceding discourse or speech situation, nor is it used to activate private hearer
old knowledge; rather, it functions to mark the nominal head of a relative clause.
Demonstratives of this sort are sometimes called determinatives (cf. Quirk et al.
1972: 217; Himmelmann 1997: 77-80). They are not only used adnominally, but
also as independent pronouns functioning as the head of the subsequent relative
clause:
(30) English (Himmelmann 1997: 77)
Those who backed a similar plan last year hailed the message.
Like English, Swedish has determinatives, which serve the same grammatical
function. They are morphologically indistinguishable from adnominal demonstratives,
but their syntactic properties are different. While adnominal demonstratives
always cooccur with a definite article that is attached as a suffix to the following
noun, adnominal determinatives precede a noun that is not marked by a definite
article. Consider the following examples:
(31) Swedish (Holmes and Hinchliffe 1994: 167, 168; Viberg et al.
1991: 144)
a. De turist-er-na fick mycket sol.
DEM tourist-PL-DEF got lot.of sun
'Those tourists got a lot of sun.'
136 DEMONSTRATIVES
b. De turist-er som åkte till Island fick mycket sol,
DTM tourist-PL REL went to Island got lot.of sun
medan de turist-er som åkte till Italien fick
while DTM tourist-PL REL went to Italy got
regn varje dag.
rain every day
'Those tourists who went to Iceland got a lot of sun, while
those tourists who went to Italy had rain every day.'
c. Jag vill ha tillbaka bok-en som du lånade
I want have back book-DEF REL you borrowed
i förra veckan.
in last week
'I'd like the book back that you borrowed last week.'
Example (31a) shows an adnominal demonstrative modifying a noun that is
marked for definiteness by the suffix -na. Sentence (31b), on the other hand,
includes two adnominal determinatives that precede a noun without a definite
article (i.e. without the suffix -na). Note that the nominal head of a relative
clause is not generally used without a definite article. The definite article is only
omitted if the head noun is accompanied by a determinative as in (31b). Example
(31c) shows that the head noun of a relative construction does occur with a
definite article if it is not marked by a determinative.
Like determinatives in English, determinatives in Swedish can be used
pronominally, as in the following example:
(32) Swedish (Holmes and Hinchliffe 1994: 169)
Island har mycket att bjuda dem som gillar äventyr.
Iceland has a.lot to offer DTM REL like adventure
'Iceland has a lot to offer those who like adventures.'
German has a determinative that developed from the definite article der and the
demonstrative jener (Lockwood 1968: 73). Derjenige and its inflected forms are
primarily used as the head of a relative clause as in the following example (cf.
Drosdowski 1995: 336):
(33) German
Derjenige, der das gemacht hat, wird bestraft.
the.one who this/it did has will be.punished
'The one who has done this will be punished.'
In colloquial German, derjenige may also occur with a subsequent prepositional
phrase (e.g. Wir nehmen denjenigen mit dem besten Angebot. 'We will take the
GRAMMATICALIZATION 137
one with the best offer.'), and occasionally it is used as a plain pronoun (e.g.
Derjenige soll kommen. 'That one is supposed to come.'). However, according to
Lockwood (1968: 73), these are extensions of the use of derjenige with a following
relative clause; originally, derjenige was used only as a determinative
pronoun.45
6.4.5 Number markers
In a recent study, Frajzyngier (1997) has shown that demonstratives may be the
historical source for plural markers. He discusses data from several Chadic
languages in which plural markers and demonstratives are morphologically
related.46 Consider the following examples from Mupun, Hona and Podoko:
(34) Mupun (Frajzyngier 1997: 201)
saar 'hand'
saar mo 'hands'
mo 'these/they'
(35) Hona (Frajzyngier 1997: 204)
kwàlàmbá 'bottle'
kwàlàmbá-yà 'bottles'
dí-yà 'this'
(36) Podoko (Frajzyngier 1997: 207-8)
16ya 'bird'
16ya-kaki 'birds'
yma-ká 'that'
In all three examples, the plural marker has the same form as a demonstrative or
one of its components. Based on these and parallel data from several other
Chadic languages, Frajzyngier maintains that the plural markers in Chadic
developed from former demonstratives. His analysis appears to be straightforward
in the case of plural markers that evolved from plural demonstratives: the latter
are readily reinterpreted as plural markers if they lose their deictic function.
However, Frajzyngier maintains that plural markers also arose from singular demonstratives.
He discusses several factors that may have contributed to the
grammaticalization of singular demonstratives as plural markers. Most importantly,
he points out that plural marking in Chadic is often confined to definite nouns
marked by an adnominal demonstrative or a related noun modifier. Due to the
cooccurrence of definiteness and plural marking, adnominal demonstratives may
become associated with the semantic feature of plurality and then they are
138 DEMONSTRATIVES
immediately reanalyzed as plural markers if they lose their deictic function.
Plural marking is not confined to nouns in Chadic. Verbs are also commonly
marked by a plural affix. As in many other languages, the plural affixes of verbs
are often similar to the plural markers of nouns in Chadic. Frajzyngier attributes
the morphological resemblance of verbal and nominal plural markers to a
common historical origin. He claims that both plural markers developed from demonstratives.
I suspect, however, that nominal and verbal plural markers originate
from demonstratives in two different source constructions: nominal plural
markers are probably derived from adnominal demonstratives that accompany a
juxtaposed noun, while verbal plural markers develop from pronominal demonstratives
that cliticize to a verb stem.
Frajzyngier's study is primarily concerned with Chadic languages, but he
points out that there are also many other languages in which plural markers and
demonstratives are morphologically related (cf. Dryer 1989b). It is thus conceivable
that the development of plural markers from demonstratives is a widespread
phenomenon and by no means restricted to Chadic.
6.4.6 Specific indefinite articles
Wright and Givón (1987) have shown that many languages distinguish between
two different indefinite nouns: indefinite nouns having a specific referent and
indefinite nouns denoting a non-specific entity. Specific indefinites are often used
to introduce a major discourse participant that will persist in the subsequent
discourse, whereas non-specific indefinites do not usually recur in the sentences
that follow. Many languages mark specific indefinites by an article based on the
numeral 'one' and non-specific indefinites by zero (cf. Wright and Givón 1987;
Givón 1995). Standard English does not distinguish between the two indefinites;
both specific and non-specific indefinites occur with the article a. However, in
colloquial English, unstressed this and these are commonly used to mark specific
indefinite information that will persist in the subsequent discourse (cf. Wright and
Givón 1987: 15-28; cf. also Prince 1981; Wald 1983; Gernsbacher and Shroyer
1989). A typical example is shown in (37):
(37) English (Givón 1990: 921)
...So next he passes this bum and boy, the guy was real ragged, run
down and all, was not even begging, just sitting there; so he stops and
gives him a dollar and the next thing you know the guy is screaming...
The noun phrase this bum introduces a new discourse participant, which is one
of the main topics in the sentences that follow. Following Wright and Givón
GRAMMATICALIZATION 139
(1987), I assume that unstressed this is a particular indefinite article, strictly
distinct from the adnominal demonstrative from which it descended. Unlike the
demonstrative, indefinite this is generally non-deictic; that is, indefinite this does
not function to orient the hearer in the speech situation or in the universe of
discourse; rather, it provides particular processing instructions. As Givón
(1990: 921) puts it, indefinite this is a "grammatical signal" that "instructs the
hearer to open and activate a file for the referent".47
There is, at least, one other language in my sample that seem to have a
specific indefinite article derived from an adnominal demonstrative. Like English,
Urim uses a former demonstrative to introduce new discourse topics. In this
function, demonstratives are often accompanied by the indefinite article ur, as in
the following example.
(38) Urim (Hemmilä 1989: 46)
kin ur pa ekg naren ampen tukgwan
woman a that two gather breadfruit ripe
'Two women were gathering ripe breadfruits.'
6.5 The grammaticalization of adverbial demonstratives
6.5.1 Temporal adverbs
Time is an abstract concept that is often metaphorically structured in spatial terms
(e.g. Lakoff and Johnson 1980). Mapping spatial expressions onto the temporal
dimension provides a common historical source for the development of temporal
markers (cf. Bybee and Dal 1989; Bybee, Pagliuca and Perkins 1991; Lichtenberk
1991; Haspelmath 1997). Since temporal expressions are semantically more
abstract and subjective than spatial terms, it is commonly assumed that the
development of temporal markers from spatial expressions is a case of grammaticalization
(cf. Heine, Claudi and Hünnemeyer 1991a: 156-157; Haspelmath 1997).
Temporal adverbs such as now and then in English are frequently derived
from adverbial demonstratives (cf. Anderson and Keenan 1985: 297-299). Very
often, adverbial demonstratives are directly imported into the temporal domain.
Anderson and Keenan (1985: 298) cite, for instance, the examples in Table 68
from Wik-Munkan (Pama-Nyungan), which are used both as locational and
temporal deictics. They consist of three deictic stems, in- 'proximal', an- 'medial'
and nan- 'distal', and two suffixes, -pal, which indicates movement from a
certain direction, and -man, which denotes a stationary referent.48
140 DEMONSTRATIVES
Anderson and Keenan (1985: 298) point out that some of the terms in Table 68
Table 68. Spatial-temporal deictics in Wik-Munkan (Anderson and Keenan 1985: 298)
Form Spatial sense Temporal sense
inman
inpal
anpal
anman
nanpal
nanman
'right here'
'from here'
'from there (distant)'
'around there'
'from there (near)'
'there (close), that place'
'right now, today'
'from now'
'from then (on)'
'around now'
'from then (recent)'
'now (general), any near time'
have acquired special meanings so that the temporal senses are not always
predictable from the corresponding spatial terms. This is a clear indication that
the temporal expressions have become independent of the spatial demonstratives
from which they derive.
According to Anderson and Keenan (1985: 298), it is fairly uncommon for
a language to employ temporal deictics that are completely independent of the demonstrative
system. However, temporal and spatial deictics do not always have
the same morphological form as in Wik-Munkan. Consider, for instance, the
following forms from Kannada:
The temporal adverbs, iga 'now' and aga 'then', involve the same deictic roots
Table 69. Demonstratives and temporal adverbs in Kannada (Schiffmann 1983: 38-9)
DEM DETs DEM PROs DEM ADVs TEMP ADVs
PROXIMAL
DISTAL
i
a
ivanu
avanu
illa
alla
iga
aga
as all other elements of the deictic system, but they are clearly distinguished from
adnominal, pronominal and identificational demonstratives. There are many other
languages in my sample in which temporal adverbs are formally distinguished
from demonstratives, but involve the same deictic roots.49
6.5.2 Directional/locational preverbs
Preverbs are elements such as con-, re- and dis- in Latin that are affixed to the
verb stem. According to Lehmann (1995a: 97-104), preverbs are commonly
GRAMMATICALIZATION 141
derived from relational adverbs that indicate the semantic relationship between
a verb and a noun. Lehmann shows that a language may have several layers of
preverbs whose syntactic and semantic properties can be quite different. For
instance, the oldest layer of preverbs in German includes inseparable prefixes of
the verb (e.g. be-, er- and ver-). Their semantic contribution to the verb is vague
and the meaning of the resulting forms is usually quite idiosyncratic. Preverbs
that developed more recently (e.g. aus-, auf- and ab-) tend to be semantically
more transparent and they are separable from the verb stem in certain contexts.
Directional preverbs are often derived from adverbial demonstratives.50
German, for instance, has two directional preverbs, hin 'hither' and her 'thither',
which developed from an old demonstrative root, hi, which only survived in a
few forms such as hin and her and hier 'here' and heute 'today' (cf. Lockwood
1968: 36, 72). Hin and her are still sometimes used as independent adverbs, but
more frequently they function as preverbs. Lehmann maintains that preverbation
in German and other Indo-European languages is not an instance of grammaticalization
but rather of lexicalization.51 He argues that the use of preverbs in
these languages is usually not fully productive and that most verbs including a
preverb are semantically irregular. These are typical properties of a word
formation process rather than grammaticalization (cf. Lehmann 1989).
Although I would not dispute Lehmann's general conclusion, it seems to me
that the formation of complex verbs including a directional preverb are usually
more regular and productive than other instances of preverbation. The two
directional preverbs in German, for instance, combine fully productively with all
verbs expressing a directional process and the resulting forms are semantically
regular and transparent. Some examples are shown in Table 70.
The verbs shown in this table express a process or activity that is directed
toward a specific location. The meanings of these verbs are completely regular
and their formation is fully productive. Verbs that do not fit this pattern such as
hindeuten 'to indicate', hinrichten 'to execute', or hinweisen 'to point out'
developed from verbs that were at one point semantically regular. It is thus
essential to distinguish between the development by which the demonstrative
adverbs hin and her turned into directional preverbs and subsequent changes that
affected the entire verb form. The former is an instance of grammaticalization,
giving rise to verbs that are semantically regular and transparent. Only the latter
is a lexicalization process whereby a verb including hin and her may assume a
new meaning that diverges from the original pattern.
There are several other languages in my sample in which directional
preverbs developed from demonstratives. Papago, for instance, has two directional
142 DEMONSTRATIVES
preverbs, i- 'toward' and a(m)- 'away', which are based on the adverbial de-
Table 70. Directional preverbs in German
hin-/her-kommen
hin-/her-fahren
hin-/her-laufen
hin-/her-rennen
hin-/her-schwimmen
hin-/her-kriechen
hin-/her-fliegen
hin-/her-jagen
hin-/her-bringen
hin-/her-holen
hin-/her-ziehen
hin-/her-tragen
hin-/her-hören
hin-/her-sehen
'to come hither / thither'
'to go by vehicle hither / thither'
'to run hither / thither'
'to run hither / thither'
'to swim hither / thither'
'to crawl hither / thither'
'to fly hither / thither'
'to chase hither / thither'
'to take hither / thither'
'to get hither / thither'
'to drag hither / thither'
'to carry hither / thither'
'to listen to s.th. or s.o. / to me'
'to look at s.th. or s.o. / at me'
monstratives ia 'here' and am 'there' (Mason 1950: 42, 65).
(39) Papago (Mason 1950: 42)
a. i-gebeı
TOWARD-take.it
'Take it there.'
b. am-hihi
AWAY-they.went
'They went there.'
Two further examples from Mojave and Inuktitut are shown in (40) and (41)
respectively. In both languages, preverbs have the same form as some of the demonstratives
that are used as independent words in other contexts. Note that the
preverb in (40) does not indicate the direction of a motion verb; rather, it is used
to mark the location expressed by a positional verb. Apart from motion verbs,
verbs denoting a location or existence provide a common context for the occurrence
of a directional/locational preverb.
(40) Mojave (Munro 1976: 35)
v--uwa-k
DEM-1-be.at-TNS
'I am around.' (or 'Here I am.')
GRAMMATICALIZATION 143
(41) Inuktitut (Denny 1982: 373)
pa-una-ar-puq
up.there.RESTRICTED-via-move-3SG
'He is going via around up there.'
6.6 The grammaticalization of identificational demonstratives
6.6.1 Nonverbal copulas
In a frequently cited paper, Li and Thompson (1977) have shown that copulas
often arise from demonstratives and third person pronouns. More specifically,
they argue that nonverbal copulas derive from anaphoric pronouns, either from
anaphoric personal pronouns or from anaphoric demonstrative pronouns. Subsequent
studies by Schuh (1983b), Eid (1983), Gildea (1993), and Devitt (1994)
supported their finding (see Stassen 1997: 76-91 for a crosslinguistic overview
of nonverbal copula constructions).
In this section, I argue that Li and Thompson's analysis is only partially
correct. Though I agree with their hypothesis that nonverbal copulas often
develop from anaphoric third person pronouns, I disagree with their claim that
copulas may develop along the same path from anaphoric pronominal demonstratives.
Challenging their view, I maintain that the development of nonverbal
copulas from third person pronouns and from demonstratives follow two different
pathways. Before I discuss the demonstrative-to-copula path-of-evolution I will
briefly describe the mechanism whereby personal pronouns develop into copulas.
The mechanism is shown in (42), which I adapted from Devitt (1994: 144):
(42) [NP NP] / [NPi [PROi NP] ] Þ [NPi COPi NP]
SUBJ PRED TOP SUBJ PRED SUBJ PRED
nonverbal clause topicalization copular clause
Li and Thompson maintain that the reanalysis of anaphoric pronouns as nonverbal
copulas originates from a topic-comment construction in which the topical NP is
resumed by an anaphoric subject pronoun. Since the topic and the pronominal
subject are coreferential they will agree if there is any agreement marking in the
language. When such a topic-comment construction is routinely used to express
an identity relation between the topic and the predicate nominal, the topicalized
NP is eventually reanalyzed as the subject of an equational sentence in which the
anaphoric pronoun assumes the function of a copula. Li and Thompson support
144 DEMONSTRATIVES
their analysis by data from several languages including Modern Hebrew, where
the reanalysis of third person pronouns as copulas is due to a very recent
development; so recent, indeed, that their status as copulas is not immediately
obvious. Consider the following examples:
(43) Modern Hebrew (Glinert 1989: 189, 188)
a. ha-sha'oni hui matana
the-clock.M.SG is/he.M.SG present.F.SG
'The clock is a present.'
b. hevrati bóing hii taagid anaki
company.F.SG Boeing is/she corporation.M.SG giant
'The Boeing company is a giant corporation.'
In both sentences hu 3SG.M and hi 3SG.F agree with the sentence-initial NP, which
one might interpret either as the topicalized noun phrase of a nonverbal clause or
as the subject of a copular construction. If the initial NP is the subject of a
copular clause, hu and hi would be nonverbal copulas; but if it is the topic of a
topic-comment construction, hu and hi would function as anaphoric pronouns.
Following Berman and Grosu (1976), Li and Thompson argue that hu and hi are
nonverbal copulas in this context. They present three arguments in support of
their view. First, they point out that the NP preceding hu and hi can be a first or
second person pronoun, as in (44).
(44) Modern Hebrew (Berman and Grosu 1976: 271)
ani/ata/hu hu hašoter
I/you/he 3SG.M the.policeman
'I am / you are / he is the policeman.'
If hu were a pronoun in this example, the sentence would be ungrammatical
because pronominal hu and hi have to agree with their antecedent.
Second, a topicalized noun phrase is usually separated from the following
clause by an intonational break. Since hu and hi follow the sentence-initial
constituent without a pause, the latter must be the subject of a copular clause,
rather than a topicalized noun phrase of a nonverbal clause.
And finally, while the predicate nominal of an identificational sentence can
be questioned (45a-b), it is impossible to question the postverbal NP of a topiccomment
construction (46a-b), as illustrated by the following examples adopted
from Berman and Grosu (1976):
GRAMMATICALIZATION 145
(45) Modern Hebrew (Berman and Grosu 1976: 277, 277)
a. moše hu xayal
Moshe he/is soldier
'Moshe is a soldier.'
b. ma hu moše
what he/is Moshe
'What is Moshe?'
(46) Modern Hebrew (Berman and Grosu 1976: 277, 277)
a. moše, hu ohev et rivka
Moshe he/is loves ACC Rivka
'Moshe, he loves Rivka.'
b. *et mi moše, hu ohev
ACC whom Moshe he loves
*'Who is such that Moshe, he loves her?'
Thus far, I agree with Li and Thompson's analysis. I challenge, however, their
claim that the development of nonverbal copulas from demonstratives involves
the same mechanism as the development of copulas from third person pronouns.
More precisely, I disagree with their hypothesis that nonverbal copulas derive
from anaphoric pronominal demonstratives that resume a topicalized noun phrase.
Questioning this part of their analysis, I maintain that nonverbal copulas that are
based on demonstratives develop from identificational demonstratives in nonverbal
clauses. Crucial evidence for my hypothesis also comes from Modern
Hebrew.
Modern Hebrew has not only copulas that are derived from personal
pronouns, but also a set of nonverbal copulas that developed from the demonstratives
ze M.SG, zot F.SG and éle PL. Like hu and hi, the demonstratives are still
used with their original function; that is, apart from their use as copulas, they are
still used as demonstratives. When ze, zot and éle are used as demonstratives they
may function as pronominal demonstratives or as identificational demonstratives
in nonverbal clauses. Pronominal and identificational demonstratives have the
same form, but they differ in their agreement behavior: anaphoric pronominal demonstratives
agree in gender and number with their antecedent, while identificational
demonstratives agree with the predicate nominal that follows. This is
exemplified in (47a-b).
146 DEMONSTRATIVES
(47) Modern Hebrew (Glinert 1989: 100, Informant)
a. ten li kasdai aHéret, ani sone et zoti
give me helmet.F.SG other I hate ACC DEM.F.SG
'Give me another helmet, I hate this (one).'
b. zei abai sheli
DEM.M.SG father.M.SG mine
'This is my father.'
The demonstrative in (47a) is an anaphoric pronominal demonstrative. It agrees
in gender and number with the noun kasda 'helmet' in the preceding clause (both
demonstrative and noun are feminine singular). Example (47b), on the other hand,
shows an identificational demonstrative in a nonverbal clause. In this sentence,
the demonstrative agrees in gender and number with the following predicate
nominal aba (sheli) '(my) father' (both demonstrative and noun are masculine
singular). In order to determine whether the nonverbal copula developed from a
pronominal or identificational demonstrative, one has to examine the agreement
properties of the copula:
(48) Modern Hebrew (Glinert 1989: 189)
ha-báyit shelHa zoti dugmai tova
the-house.M.SG your COP/DEM.F.SG example.F.SG good
'Your house is a good example.'
Example (48) includes two noun phrases of different genders and the feminine
singular demonstrative zot, which Glinert (1989: 189) analyzes as a copula in this
example. The copula agrees in gender and number with the predicate nominal at
the end of the sentence, rather than with the sentence initial NP. The plural
copula in (49) exhibits the same kind of agreement.
(49) Modern Hebrew (Glinert 1989: 190)
ha-músika she-baHárti élei (ze) ktaimi
the-music.M.SG SUB-I.picked COP/DEM.PL pieces.PL
she-kulam ohavim
SUB-love everyone
'The music I picked is pieces everyone loves.'
Like zot in (48), the plural copula éle agrees with the predicate nominal rather
than the preceding subject. Note that in casual speech, the copual ze may occur
instead of éle (or zot) regardless of the gender and number features of the
coreferential noun. Since the copulas in (48) and (49) show the same agreement
behavior as the demonstratives in nonverbal clauses, I assume that they derive
from identificational demonstratives rather than anaphoric demonstrative
GRAMMATICALIZATION 147
pronouns, as Li and Thompson have argued. (50) shows the mechanism that I
suggest for the developmental pathway from demonstratives to copulas.
(50) [NP] [DEMi NPi] Þ [NP COPi NPi]
There are two crucial differences between the grammaticalization path shown in
(50) and the mechanism described by Li and Thompson (which I believe is only
appropriate for nonverbal copulas derived from third person pronouns). First, the
agreement features of copulas derived from demonstratives are determined by the
predicate nominal rather than the subject; and second, copular clauses including
a former demonstrative develop from two intonation units, a topical NP and a
nonverbal clause, that merge into a single construction, whereas copular clauses
that include a former third person pronoun emerge from a construction that is
expanded by left-dislocation.52
Additional support for my analysis comes from Kilba. Kilba has three
nonverbal copulas which have the same form as identificational demonstratives
in nonverbal clauses, while they differ from demonstrative pronouns. The demonstrative
pronouns are complex free forms whereas the identificational demonstratives
and nonverbal copulas are monosyllabic enclitics. Consider the forms in
Table 71, which is repeated from Section 4.3.1.
Schuh (1983b) shows that the identificational demonstratives have turned into
Table 71. Demonstrative pronouns - identifiers in Kilba (Schuh 1983b: 315-317)
DEM PROs/DETs DEM IDENTs
PROXIMAL
DISTAL
REMOVED
(n ' 6)n ' 6nnà
(nà)ndándà
(]g'6)]g'6]gà
=ná
=ndá
=]gá
copulas that mark three different tenses: =ná has been reanalyzed as a present
tense copula, =ndá indicates past tense, and =]gá is used in copular clauses
whose subject is out of sight (a similar development occurred in Panare; cf.
Gildea 1993):
(51) Kilba (Schuh 1983b: 321)
a. ùsmân h'6bà ná
Usman Kilba is
'Usman is a Kilba.' (Usman is present)
148 DEMONSTRATIVES
b. ùsmân h'6bà ndá
Usman Kilba was
'Usman was a Kilba.'
c. ùsmân h'6bà ]gá
Usman Kilba was
'Usman is a Kilba.' (Usman is not present)
6.6.2 Focus markers
In a recent study, Luo (1997) has shown that focus markers often have the same
or a very similar morphological form as copulas and/or demonstratives. The
morphological resemblance among these categories suggests that in such cases
they are historically related. Givón (1979: 246-248) and Heine and Reh
(1984: 147-182) have shown that in many African languages focus markers
emerged from former copulas (see McWhorter 1994 for an alternative analysis
of the evolution of the focus marker ni in Swahili). More specifically, they have
argued that focus markers frequently arise from copulas in cleft constructions.
The mechanism is shown in (52):
(52) [[Ø COP NP]S [REL CL]S]S Þ [[FOC NP]NP ...]S
The source construction consists of two clauses: a copula clause (with zero
subject) providing focal information and a relative clause providing presupposed
information. In the target construction, the copula clause has been reanalyzed as
a focal NP of the former relative clause. Given that demonstratives are a common
historical source for nonverbal copulas (see above), one might posit that the
morphological relationship between demonstratives and focus markers is due to
the development of focus markers from nonverbal copulas that in turn evolved
from identificational demonstratives in nonverbal clauses:
(53) IDENT DEM > COPULA > FOCUS MARKER
Though (53) appears to be a likely pathway for the development of focus markers
(cf. Luo 1997), it is conceivable that focus markers may also develop directly
from identificational demonstratives in nonverbal clauses. (54) shows the
mechanism by which they would thus emerge.
(54) [[DEM Ø NP]S [REL CL]S]S Þ [[FOC NP]NP ]S
GRAMMATICALIZATION 149
Note that the source construction in (54) has the same structure as the one in (52)
except that the focal part of the cleft construction includes an identificational demonstrative
in a nonverbal clause rather than a copula with zero subject. The
mechanism in (54) may account for the development of focus markers in
languages such as Ambulas and Mokilese, where demonstratives and focus
markers are morphologically related but show no obvious relationship to a copula.
As shown in 2.1.2, Ambulas has two demonstrative identifiers, kén 'proximal'
and wan 'distal', which are frequently used as focus markers:
(55) Ambulas (Wilson 1980: 157)
véte dé wak a [wan méné] kaapuk yéménén
see.and he said ah FOC you not you.went
'He saw him and said, "Ah, so you did not go.'"
Similarly, Mokilese (Austronesian) employs a focus marker (ioar) which,
according to Harrison (1976: 309), can be traced back to an old deictic form that
is cognate to a demonstrative identifier in Ponapean (see 4.3.1).
(56) Mokilese (Harrison 1976: 311)
ioar Wilson ma pwehng ih mehu
FOCUS Wilson REL told him that
'It was Wilson who told him that.'
Since there is no evidence that the focus markers in Ambulas and Mokilese
developed from a copula, it is at least a plausible hypothesis that they emerged
directly from a demonstrative identifier in a nonverbal clause.
6.6.3 Expletives
Expletives are semantically empty pro-forms that some languages require to form
certain syntactic constructions. Two examples from French and Modern Hebrew
are given in (57) and (58) respectively.
(57) French (Calvez 1993: 332)
C'est toi que je connais le mieux.
it.is you that I know the best
'You are the one I know best.'
150 DEMONSTRATIVES
(58) Modern Hebrew (Glinert 1989: 63)
(ze) tov [she-bat]
it good SUB-you.came
'It's good (that) you came.'
In these examples, ce and ze function as dummy pro-forms (the use of ze is
optional): they do not have a referent and serve a purely grammatical function.
Historically, ce and ze are derived from identificational demonstratives in copular
clauses. In other syntactic contexts, expletives developed from demonstratives that
were originally used as pronouns or adverbs (cf. Traugott 1992: 216-219). An
example of the former is English it in It rained; and an example of the latter is
the existential there in sentences like There was an old man who lived in Western
New York.
6.7 The diachronic origin of demonstratives
Having described the grammaticalization of demonstratives, I now address the
question: where do demonstratives come from - what is their historical source?
Demonstratives are usually considered grammatical items (for a different
view see Woodworth 1991: 285). Grammaticalization theory holds that all
grammatical items are eventually derived from lexical expressions (cf. Hopper
and Traugott 1993: 104), but there is no evidence from any language that demonstratives
developed from a lexical source or any other source, for that matter, that
is non-deictic. Demonstratives are sometimes reinforced by lexical items such as
ecce 'behold' in Vulgar Latin, which strengthened the weakened demonstrative
ille (VL ecce ille > OFr cest cel > Fr ce; Harris 1978: 70-77). In such a case, the
lexical item may become part of the demonstrative, and if the original demonstrative
subsequently disappears the lexical item may assume a deictic function (cf.
Brugmann 1904).53 This might be, however, the only mechanism by which a nondeictic
term may evolve into a demonstrative. Frajzyngier (1987, 1996) claims
that the demonstratives in Mupun and several other Chadic languages developed
from motion verbs and verbs of saying, but his analysis is very speculative and
the suggested development appears to be unmotivated.54 As Himmelmann
(1997: 20) notes, apart from those cases where a demonstrative developed from
a lexical item that functioned to reinforce a weakened demonstrative, there is no
convincing evidence from any language that demonstratives may have evolved
from a lexical source (cf. Traugott 1982: 245; Hopper 1991: 31; Hopper and
Traugott 1993: 129). A number of studies have therefore suggested that demonGRAMMATICALIZATION
151
stratives might present an exception to the hypothesis that all grammatical expressions
are eventually derived from lexical items (cf. Plank 1979a; Traugott
1982).55 Demonstratives, or the deictic elements on which they are based, might
belong to the basic vocabulary of every language. This is not only suggested by
the absence of any positive evidence for a lexical source, it is also supported by
the fact that demonstratives belong to the very few items that display a nonarbitrary
relationship between phonetic form and meaning. Based on a representative
sample of 26 languages, Woodworth (1991) has shown "that there is a
systematic relationship between vowel quality and distance such that the vowel
quality of the form with proximal meaning has a higher pitch than that of the
form with distal meaning." This confirms earlier studies by Sapir (1949) and
Ultan (1978b) who cast the same finding in articulatory terms: the vowels of
proximal demonstratives tend to be higher and more advanced than the vowels
of the corresponding medial and distal forms. Plank (1979a) argues that the iconic
relationship between phonetic shape and meaning might indicate that demonstratives
are newly created words (cf. also Plank 1979b). Questioning the hypothesis
that all grammatical items are either derived from lexical items or from grammatical
items that previously developed from a lexical source, he argues:
doch möchte ich bei meiner Konzeption von lexikalisch und grammatisch nicht
a priori ausschliessen, dass bestimmte grammatische Mittel ihren historischen
Ursprung nicht in lexikalischen oder ehemals funktionsverschiedenen grammatischen
Ausdrucksmitteln haben. Zumindest bei grammatischen Kategorien aus
Bereichen, in denen Form-Bedeutungszuordnungen tendentiell auf ikonischer
statt auf rein symbolischer Basis erfolgen, sollte auch das erste Verfahren der
"Urschöpfung" von Ausdrücken in unmittelbar grammatischer Funktion im
Auge behalten werden. Zu denken wäre dabei an zwei- oder mehrstufige
Systeme der Ortsdeixis (hier - dort), deren Vokalmuster in vielen Sprachen
frappante Ähnlichkeiten aufweisen, ... (Plank 1979a: 331-332)
[but for my conception of lexical and grammatical I do not want to exclude a
priori that certain grammatical items do not originate from lexical items or
items that previously served another grammatical function. At least for those
grammatical categories for which the pairing of form and meaning is partially
iconic rather than purely symbolic, the first strategy of creating new terms that
are immediately used to serve a grammatical function should also be considered.
Place deictic systems including two or more terms show, for instance, a
pattern of vowel alternation that is strikingly similar in many languages, ...]
The iconic relationship between sound shape and meaning could, of course, also
be the result of a phonological process that is involved in the grammaticalization
of demonstratives. But given that sound symbolism is usually associated with
152 DEMONSTRATIVES
newly created words, the systematic relationship between vowel quality and
distance seems to support the hypothesis that demonstratives are new creations.
This would also be in accordance with the fact that demonstratives are not
ordinary grammatical items. Grammatical items function to organize the lexical
material in discourse, while demonstratives serve a language-external function (at
least in their most basic use). A number of scholars, including Peirce (1955) and
Bühler (1934), have argued that demonstratives and other deictics form a
particular class of items that is distinct from all other linguistic expressions (see
also Ehlich 1979, 1982, 1983, 1987). Demonstratives are used to orient the hearer
in the speech situation, focusing his or her attention on objects of interest. This
is one of the most basic functions of human communication for which there
might be a particular class of linguistic expressions that emerged very early in the
evolution of language. In this connection, it is interesting to note that demonstratives
are also among the very first items in language acquisition. According to
Eve Clark (1978: 95), demonstratives often appear in the first ten words of
English-speaking children, and they are always among the first fifty. Apart from
demonstratives there are hardly any other closed class items that English-speaking
children learn before they begin to construct their first simple sentences (Brown
1973). All this suggests that demonstratives are not ordinary grammatical
markers, and hence it would make sense if they do not derive from lexical items
as all other grammatical markers.
If demonstratives are not derived from lexical items, there would be two
different sources from which grammatical markers may emerge: lexical expressions
and demonstratives. The grammaticalization of items from both domains
would be unidirectional. That is, grammatical items develop from lexical expressions
and demonstratives but never vice versa. Furthermore, there would be no
transitions between the two source domains - demonstratives do not develop
into lexical items, nor is there evidence that lexical items have ever been
reanalyzed as demonstratives. These considerations are summarized in Figure 7.
GRAMMATICALIZATION 153
LEXICAL
MORPHEMES
DEMONSTRATIVES
(DEICTIC MORPHEMES)
GRAMMATICAL
MORPHEMES
(NEW) GRAMMATICAL
MORPHEMES
Figure 7. The two source domains of grammatical items
According to Figure 7, grammatical markers evolve from two different sources:
lexical items and demonstratives. The grammaticalization processes that originate
from these two sources differ significantly. The grammaticalization of lexical
items involves such mechanisms as metaphorical extension, metonymic transfer
and conversational implicature (cf. Sweetser 1990; Heine, Claudi and Hünnemeyer
1991a; Traugott and König 1991; Hopper and Traugott 1993; Bybee,
Perkins and Pagliuca 1994). These mechanisms are irrelevant to the grammaticalization
of demonstratives, which is based on the extension of the exophoric use
to the three endophoric uses (cf. 5.5). Endophoric demonstratives serve a
language-internal function, providing a natural starting point for the development
of grammatical markers. It seems that grammatical morphemes always emerge
from a secondary use of the source expression: lexical items that are used in a
metaphorical (or metonymic) sense or that trigger a conversational implicature
and demonstratives that function to indicate a referent in the universe of discourse
or to activate private shared knowledge.
6.8 Summary
In this chapter, I have shown that demonstratives provide a common historical
154 DEMONSTRATIVES
source for a wide variety of grammatical items. I have examined the grammaticalization
path of eighteen grammatical markers that frequently arise from demonstratives.
I have shown that the path-of-evolution that a demonstrative takes is
largely determined by the syntactic context in which it undergoes grammaticalization.
Pronominal, adnominal, adverbial, and identificational demonstratives
provide input into different grammaticalization channels giving rise to grammatical
items that usually retain some of the syntactic properties of their historical
source. There are, however, a few grammatical items that may arise from more
than one source. Sentence connectives, for instance, derive from pronominal demonstratives
that are used as discourse deictics and from manner demonstratives,
which are commonly classified as adverbs. Also, expletives may originate from
demonstratives in three different contexts: (i) identificational demonstratives in
copular and nonverbal clauses, (ii) pronominal demonstratives functioning as
subjects of an impersonal verb, and (iii) adverbial demonstratives in existential
constructions. Table 72 provides an overview of the grammaticalization channels
that I have examined in this chapter.
The list of grammatical items hown in Table 72 is by no means exhaustive.
Other grammaticalization processes starting from a demonstrative have been
suggested in the literature. For instance, Vries (1995) has argued that a number
of Papuan languages have topic markers that are based on pronominal demonstratives,
and Ehlich (1986, 1987) has shown that interjections and discourse markers
may evolve from deictics (i.e. demonstratives).56 Finally, it is well known that in
many languages demonstratives are commonly used as hesitation signals (e.g.
Korean, Japanese, Mandarin Chinese, Finnish, Passamaquoddy-Maliseet).
In the final section of this chapter, I have considered the hypothesis that demonstratives
(or the deictic roots on which demonstratives are based) belong to
the basic vocabulary of every language. I presented three arguments that seem to
support this view. First, there is no positive evidence that demonstratives derive
from lexical items. Second, demonstratives exhibit an iconic relationship between
phonetic shape and meaning, which is usually associated with newly created
words. And third, demonstratives do not serve a language-internal function like
ordinary grammatical markers; rather, they are used for one of the most basic
functions of human communication, for which there might be a class of linguistic
items that is distinct from both lexical and grammatical expressions.
GRAMMATICALIZATION 155
Table 72. The grammaticalization of demonstratives
Source Target
Pronominal demonstratives: third person pronouns
relative pronouns
complementizers
sentence connectives
possessives
adnominal determinatives
verbal number markers
expletives
(linkers)
Adnominal demonstratives: nominal number markers
definite articles/noun class markers
linkers
boundary markers of attributes
pronominal determinatives
specific indefinite articles
(relative pronouns)
Adverbial demonstratives: directional preverbs
temporal adverbs
expletives
Identificational demonstratives: nonverbal copulas
focus markers
expletives
CHAPTER 7
Conclusion
It has been the purpose of this study to provide a comprehensive overview of the
form, function and grammaticalization of demonstratives in the world's languages.
This final chapter summarizes the major results of my investigation and discusses
some of the areas that merit further examination.
7.1 Major findings
Chapter 2 examined the morphological features of demonstratives: the occurrence
of demonstrative clitics, the inflectional properties of demonstratives in different
syntactic contexts, and the formation of demonstrative stems. It was shown that
the occurrence of demonstrative clitics is largely restricted to adnominal demonstratives;
pronominal, adverbial and identificational demonstratives are almost
always unbound. The inflectional properties of demonstratives vary with their
syntactic function. Pronominal demonstratives are more often inflected than
adnominal and identificational demonstratives, which in turn are more likely to
inflect than adverbial demonstratives; the latter are usually uninflected unless they
occur with a set of locative case markers. The stem of most demonstratives is
monomorphomemic. If it consists of more than one morpheme, pronominal demonstratives
are often formed from a demonstrative root and a nominalizing
suffix, a third person pronoun or a classifier, and adverbial demonstratives may
be composed of a demonstrative root and a locative marker or a locational noun.
Adnominal and identificational demonstratives have either the same stems as
pronominal demonstratives or they consist solely of a deictic root.
Chapter 3 examined the semantic features of demonstratives. It was argued
that demonstratives have two kinds of semantic features: deictic features, which
indicate the location of the referent in the speech situation, and qualitative
features, which characterize the referent. The deictic features indicate whether the
referent is near or away from the deictic center, whether it is at a higher or lower
elevation, whether the referent is visible or out of sight, uphill or downhill,
upriver or downriver, or whether it is moving toward or away from the deictic
158 DEMONSTRATIVES
center or across the speaker's line of vision. I argued that although pronominal
and adnominal demonstratives are not always deictically contrastive, all languages
have at least two (adverbial) demonstratives that indicate a contrast on a distance
scale: a proximal demonstrative, which refers to a location near the deictic center,
and a distal demonstrative, which indicates a referent at greater distance. The
qualitative features of demonstratives are subsumed under six semantic categories:
(i) ontological status, (ii) animacy, (iii) humanness, (iv) sex, (v) number, and (vi)
boundedness. The features of these categories classify the referent, which
provides important clues for its identification by the hearer.
Chapter 4 investigated the distribution and the categorial status of demonstratives.
I argued that demonstratives occur in four syntactic contexts: (i) they are
used as independent pronouns in argument position of verbs and adpositions, (ii)
they are used together with a cooccurring noun or (iii) a cooccurring verb, and
(iv) they may occur in copular and nonverbal clauses. Some languages employ
the same demonstratives in all four contexts, but most languages have several
series of demonstratives that serve specific syntactic functions in particular
environments. If the demonstratives being used in one of these four contexts are
phonologically distinguished or if they have different morphosyntactic properties,
they belong to different grammatical categories, which I call (i) demonstrative
pronouns, (ii) demonstrative determiners, (iii) demonstrative adverbs, and (iv) demonstrative
identifiers. I used the corresponding adjectives - pronominal,
adnominal, adverbial, and identificational - in order to refer to the use of a demonstrative
in a specific syntactic context regardless of its categorial status. It
was shown that some languages do not have a particular class of demonstrative
determiners, instead, using a demonstrative pronoun with a coreferential noun in
apposition. Other languages lack a particular class of demonstrative pronouns and
use demonstrative determiners together with a classifier, a third person pronoun,
or a nominal particle where most languages use a demonstrative pronoun.
Adnominal demonstratives in English are phonologically and morphologically
indistinguishable from demonstrative pronouns, but since they are associated with
two different word classes one must assume that pronominal and adnominal this
and that belong to distinct categories. In Old English, adnominal demonstratives
had the status of independent pronouns that were adjoined to a coreferential noun
in apposition, but in Modern English, adnominal demonstratives are determiners
and categorially distinguished from demonstrative pronouns. Adverbial demonstratives
are usually distinct from demonstrative pronouns and determiners. There are
only a few languages in my sample in which adverbial demonstratives belong to
the same category as pronominal and adnominal demonstratives. Finally, identificational
demonstratives are usually considered demonstrative pronouns, but my
CONCLUSION 159
investigation showed that the demonstratives in copular and nonverbal clauses are
often formally (i.e. categorially) distinguished from pronominal demonstratives
in other contexts: they may have a particular phonological form or they may
differ in their inflection.
Chapter 5 was concerned with the pragmatic uses of demonstratives.
Following a recent study by Himmelmann (1996), I distinguished four different
uses: (i) the exophoric, (ii) anaphoric, (iii) discourse deictic, and (iv) recognitional
uses. Exophoric demonstratives function to orient the hearer in the speech
situation; they serve a language-external function. The other three uses function
to refer to linguistic elements in the surrounding discourse or to activate private
shared knowledge; they serve a language-internal function enhancing discourse
coherence. I used the notion of endophoric demonstrative in order to distinguish
these three uses from the exophoric use. Anaphoric demonstratives are coreferential
with a prior NP; they keep track of discourse topics. Discourse deictic
demonstratives refer to a chunk of the surrounding discourse; they express an
overt link between two discourse units. Finally, recognitional demonstratives
activate information that is already in the hearer's knowledge store. The exophoric
use is the most basic use from which all other uses derive: exophoric demonstratives
are prior in language acquisition, they are morphologically and
distributionally unmarked (vis-à-vis anaphoric and recognitional demonstratives),
and the grammaticalization of demonstratives originates from one of the three
endophoric uses. One can think of the grammaticalization of demonstratives as
a cline ranging from exophoric demonstratives that are used to orient the hearer
in the outside world to grammatical markers that serve a specific syntactic
function. Anaphoric, discourse deictic, and recognitional demonstratives are
somewhere in between the two ends of this cline.
In Chapter 6 I showed that the grammaticalization of demonstratives is
largely determined by the syntactic context in which a demonstrative occurs.
Pronominal, adnominal, adverbial, and identificational demonstratives are
associated with grammatical items that usually retain some of the syntactic
properties of their historical source. Pronominal demonstratives are frequently
reanalyzed as third person pronouns, relative pronouns, complementizers, sentence
connectives, possessives, expletives, and verbal number markers; adnominal demonstratives
may develop into definite articles (noun class markers), linkers,
adnominal determinatives, nominal number markers, boundary markers of
postnominal attributes, and, perhaps, relative pronouns; adverbial demonstratives
are often reanalyzed as temporal adverbs and in some languages they developed
into directional preverbs, sentence connectives, and expletives; and identificational
demonstratives are the source for nonverbal copulas, focus markers, and expletives.
160 DEMONSTRATIVES
Finally, I considered the hypothesis that demonstratives might not derive from
lexical items, as commonly assumed, but rather from a class of genuine deictics
that belong to the basic vocabulary of every language. I presented three arguments
in support of this view. First, there is no convincing evidence from any
language that would indicate that demonstratives evolve from a lexical source.
Second, demonstratives are among the very few items that exhibit an iconic
relationship between form and meaning, which is characteristic of newly created
words. And third, demonstratives serve one of the most fundamental functions of
human communication for which all languages use a set of deictics that might
have developed very early in the evolution of language.
7.2 Future research
My findings raise a number of questions that deserve further investigation. In this
final section, I outline some areas of future research that would thematically
continue the current work.
In Chapter 4, I discussed the semantic values of demonstratives as described
in reference grammars. The vast majority of grammars that I consulted use
semantic labels such as 'proximal' or 'near speaker' in order to characterize the
meanings of demonstratives. These labels are, however, only rough approximations.
The meaning of a demonstrative is often more complex. It would be a very
interesting project to study the semantic values of demonstratives in greater detail.
One issue that requires special attention is the distinction between distanceoriented
and person-oriented systems, as suggested by Anderson and Keenan. Are
these really two different systems, or is a person-oriented system just a variant
of a distance-based system with a special deictic term for entities near the hearer,
as I have argued in my discussion of the deictic system in Quileute?
Another interesting question that could not be addressed in this work is why
languages differ as to the number and kind of deictic terms that they employ. It
has been repeatedly argued that the cultural environment of a speech community
determines, at least to some extent, the size of a deictic system (cf. Frei 1944;
Denny 1978; Perkins 1992). Specifically, it has been claimed that the size of a
deictic system decreases "as the degree to which the spatial environment is manmade
increases" (e.g. Denny 1978: 80). This claim has been put forward based
on crosslinguistic comparison: the deictic systems of many aboriginal languages
are much more complex than the deictic systems of languages that are spoken by
people in modern societies. This hypothesis presupposes that the deictic systems
of languages such as English, French or Japanese were more complex at an
CONCLUSION 161
earlier stage in their history. In order to test the hypothesized correlation between
the size of a deictic system and the cultural environment of the speech community
one has to study the historical development of deictic systems. In particular,
it is necessary to examine the deictic systems of languages spoken in societies
that underwent significant socioeconomic changes.57
Another topic that would be worthwhile to investigate is whether the
distinction between different demonstrative categories correlates with other
syntactic features of a language. What motivates, for instance, the categorial
distinction between demonstrative pronouns and demonstrative determiners, which
occurs in some languages but not in others? And why do most languages employ
a spearate class of demonstrative adverbs? I suspect that the existence of different
demonstrative categories correlates with the existence of more general word
classes. For instance, if a language has a class of determiners, I would expect that
adnominal demonstrative will be distinguished from pronominal, adverbial and
identificational demonstratives. If, on the other hand, a language does not have
a class of determiners (like Old English), it is quite likely that adnominal demonstratives
belong to the same category as demonstratives that occur in other
contexts. In other words, my hypothesis is that the distinction between demonstrative
pronouns, determiners, adverbs, and identifiers is motivated by the devision
between more general word class that occur in a particular language.
In the chapter on grammaticalization, I did not consider the question of
whether the distance features of demonstratives might have an effect on their
path-of-evolution. Greenberg (1978: 61) and Givón (1984: 226) argued that
definite articles are almost always derived from distal demonstratives. There
might be similar correlations between other grammatical markers and certain
distance categories. My impression is that most grammatical markers derive from
distal demonstratives, but this needs thorough investigation.
Finally, it is left for future research to decide whether demonstratives are
derived from lexical expressions, as most previous studies assumed, or whether
they are based on a set of genuine deictics, as I suggested. Although it is, of
course, impossible to prove that demonstratives form a genuine class of deictics
that did not evolve from lexical expressions, this hypothesis will become increasingly
convincing if future studies fail to find evidence for a non-deictic source
from which demonstratives might have emerged.
Notes
1. An overview of some of the philosophical literature on deixis and demonstratives is given in
Levinson (1983: 55-61).
2. Rijkhoff and Bakker (1998) provide an overview of various sampling methodologies. See also
Bell (1978), Dryer (1989c), Nichols (1992), Rijkhoff, Bakker, Hengeveld, and Kahrel (1993),
and Bybee, Perkins and Pagliuca (1994).
3. Note that some of the forms in Table 8 end in a long vowel (e.g. mı¯líí). The final vowel is only
lengthened when the demonstratives are used pronominally; in adnominal position, the final
vowel is always short.
4. In "extremely limited non-colloquial situations" the demonstratives ku and i can occur without
a defective noun, ku meaning 'he' and i 'this fact' (cf. Sohn 1994: 294).
5. Note that adverbial demonstratives may attach to a noun when they are used adnominally, as
in French cette maison-là 'that house there' (cf. 4.2).
6. In some studies, clitics are defined in phonological terms: clitics are usually unstressed and they
are not affected by phonological rules that are sensitive to word boundaries. However, Anderson
(1992: 200) shows that phonological criteria are insufficient to define clitics. He argues that
clitics can only be defined in syntactic terms. Clitics are "phrasal affixes" (Klavans 1985: 95),
which may attach to various elements of a phrase: (i) the initial word, (ii) the final word, (iii)
the head etc. (cf. Anderson 1992: 202).
7. The case markers of demonstratives have probably a different origin. As pointed out above, the
case endings of demonstratives are usually very similar to the case endings of nouns and
therefore it is reasonable to assume that they developed from the same source. As Lehmann
(1985, 1995a), Lord (1993) and others have shown, case markers are frequently derived from
adpositions or serial verbs.
8. Himmelmann does not consider the categorial status of das. While some German linguists
consider das a demonstrative when it is stressed (e.g. Engel 1988), others assume that stressed
das is not distinguished from the (unstressed) definite article of the same form (e.g. Bisle-Müller
1991).
9. The same argument applies, of course, to person-oriented systems with three deictic terms,
which can be seen as variants of two-term systems.
10. Dixon (1972: 48) points out that the suffixes of the 'river' series refer to 'water' features in
general, and that the 'hill' suffixes may also refer to 'cliff', 'tree' etc.
11. The use of demonstrative determiners and demonstrative identifiers is usually not restricted to
a certain kind of referent: they refer to whatever is denoted by the noun or predicate nominal
with which they cooccur.
164 NOTES
12. Some of my sources use the notion intensive instead of emphatic (cf. Redden 1980: 70; Dixon
1972: 48). I assume that emphatic and intensive markers have basically the same pragmatic
function.
13. Note that 'precision' is different from 'boundedness'. The features of the category 'precision'
indicate whether the speaker draws the hearer's attention to a specific location or whether the
location of the referent is not made precise, while the features of the category 'boundedness'
indicate how the referent is conceptualized.
14. See Moravcsik (1997) for an interesting discussion of adnominal demonstratives in Hungarian,
which behave in some ways like independent pronouns and in others like "satellites of the noun"
(cf. 6.4.1).
15. Note that under Abney's analysis determiners are the only functional elements that are
subcategorized like verbs (and other lexical elements). All other functional elements require a
complement, i.e. they cannot be intransitive (cf. Abney 1987: 285).
16. Further theory internal support for the Det-as-head analysis comes from the following consideration.
Jackendoff (1977) has shown that there can be two specifiers in the English noun phrase:
[The]1 [many]2 good men. The traditional NP analysis cannot accommodate such structures
because X-bar theory does not allow for more than one specifier per phrase. The Det-as-head
analysis provides room for an additional specifier by treating one of them as the head of DP and
the other as its specifier (cf. Abney 1987: 287-295). This argument relies, of course, crucially
on the assumption that many and other adnominal elements that may occur in the 'second
determiner slot' are in fact determiners.
17. Abney points out that this is a possible interpretation, but his discussion suggests that there is
also a non-restrictive reading (which I don't get).
18. I use the term 'construction' in the sense of Construction Grammar, in which constructions are
defined as conventional pairings of form and meaning (cf. Fillmore, Kay and O'Connor 1988;
Fillmore and Kay 1993; Goldberg 1995; Lakoff 1987; Diessel 1997b).
19. The French forms are less grammaticalized in that the adverbial intensifiers ci and là are not
obligatory like their counterparts in Afrikaans and Swedish.
20. Temporal deictics such as English now and then are also sometimes classified as demonstrative
adverbs (cf. Anderson and Keenan 1985: 297), but I will keep them separate from demonstratives.
In Section 6.5.1, I argue that temporal deictics are grammatical markers that frequently
develop from adverbial demonstratives.
21. There is, however, a semantic difference between pronominal and identificational demonstratives
in English: in copular clauses, this and that may refer to a person as in This is my friend. In all
other syntactic contexts, pronominal this and that cannot denote a human referent (cf. Halliday
and Hasan 1976: 62-63).
22. Schuh (1977) does not decompose the demonstrative in Table 51, but they seem to consist of
a deictic root (m-), a gender/number affix (-s MASC, -c FEM, -d PL), and a distance/category
marker (-o 'proximal pronoun', -aa(ni) 'proximal identifier'; ii 'distal pronoun/identifier', -6no
'particular pronoun', -6naa(ni) 'particular identifier').
23. Schuh (1977) does not indicate morpheme boundaries, but the plural forms of the demonstrative
pronouns are apparently formed from a demonstrative root and the plural marker -aw-, which
is inserted into the root before the final mid back vowel.
24. The ce in copular clauses must be kept separate from the masculine singular form of the demonNOTES
165
strative determiners; cf. C'est Pascal 'This is Pascal' vs. ce cadeau 'this gift'.
25. In Durie's grammar (1985), there are only examples of adnominal demonstratives that cliticize
to a preceding element, but his description suggests that adnominal demonstratives can also be
free forms.
26. In addition to the forms in Table 60, Pangasinan has two further series of demonstratives, which
Benton (1971: 91-93) calls "demonstratives of similarity" and "independent demonstratives".
The demonstratives of similarity are manner demonstratives; they belong to the category of demonstrative
adverbs. The independent demonstratives occur in a variety of contexts: (i) in
equational sentences, (ii) after the marker ed (which seems to function as an adposition), and
(iii) linked to a noun phrase by the particle ya. They appear to be similar to sentential demonstratives
such as French voilà, but this needs further investigation.
27. Since the out-of-sight demonstratives that we saw in Chapter 3 are anchored in the speech
situation and are deictically contrastive, they can be viewed as exophoric demonstratives despite
the fact that they refer to entities that are not present in the speech situation.
28. Himmelmann (1996) refers to the anaphoric use as the "tracking use", emphasizing the discourse
pragmatic function of demonstratives that are coreferential with a prior NP.
29. The average referential distance of third person pronouns is 1.7 in Lichtenberk's data. The
referential distance of demonstratives is much higher: it is 3.4 for the proximal demonstrative
'eri, and 8.6 for the distal demonstrative baa. These are the figures for adnominal demonstratives.
Lichtenberk does not consider the pronominal use of 'eri and baa.
30. An exception to this might occur in languages that do not have third person pronouns and/or
definite articles (or languages in which the use of third person pronouns is restricted to
human/animate referents). These languages frequently employ anaphoric demonstratives to track
major discourse participants (Nikolaus Himmelmann p.c.).
31. Kuno (1973: 282-290) argues that ano can also be used anaphorically. He points out, however,
that unlike sono, ano only occurs when the antecedent denotes an entity that speaker and hearer
know personally from previous experience. This suggests that ano functions as a recognitional
demonstrative rather than an anaphor.
32. Demonstratives referring to an event or situation are often subsumed under the discourse deictic
use (cf. Webber 1991; Himmelmann 1996).
33. Lyons (1977: 668) refers to discourse deictic demonstratives as "impure text deixis". The notion
of text deixis is sometimes used as a synonym of the notion discourse deixis or as a cover term
subsuming both text and discourse deixis.
34. In the literature, the notion anaphoric is used in two different ways: on the one hand it refers
to the tracking use of demonstratives, and on the other hand it indicates that a pronoun, noun
(phrase) or adverb refers to an element of the preceding discourse. In the former sense,
'anaphoric' contrasts with the terms 'exophoric', 'discourse deictic' and 'recognitional'; in the
latter sense, it contrasts with the term 'cataphoric'.
35. According to Nikolaus Himmelmann (p.c.), recognitional demonstratives in Nunggubuyu can
also be used pronominally, but in all languages with which I am familiar recognitional demonstratives
are used only adnominally.
36. Himmelmann (1996: 236-9) points out that a recognitional demonstrative may be coreferential
with a distant NP that is no longer activated.
37. In accordance with the data from language acquisition, Brugmann (1904) and Bühler (1934)
166 NOTES
argue that the exophoric use of demonstratives is prior in the evolution of language. That is,
they claim that the more abstract uses of demonstratives developed phylogenetically from demonstratives
that were used to focus the hearer's attention on concrete objects in the speech
situation. If this is correct, it would strengthen my hypothesis that the exophoric use is basic.
38. Note that the notion of lexicalization is also used to denote the (uncommon) process by which
grammatical items develop into lexical forms (cf. Hopper and Traugott 1993: 104).
39. Discourse deictic demonstratives may indicate a contrast between anaphoric and cataphoric
reference, but they do not indicate two different locations on a distance scale.
40. First and second person pronouns are usually not derived from demonstratives. But see
Humboldt (1832) for some examples.
41. There is one minor difference in genitive plural. The genitive plural of the relative pronoun is
deren. The pronominal demonstrative, on the other hand, has two forms: deren and derer
(Drosdowski 1995: 335-336). Deren is used to indicate the possessor in a possessive noun
phrase (e.g. die Schüler und deren Eltern 'the pupils and their parents'), while derer is used
either to refer to a subsequent relative clause (e.g. das Schicksal derer, die... 'the fate of those
who...'), or as a free standing pronoun functioning as the object of verbs that take an argument
in genitive case (e.g. Wir gedenken derer nicht mehr. 'We don't commemorate those (people)
any more') (cf. Drosdowski 1995: 334).
42. I use the following terminology: a 'relative clause' is an embedded clause that is used to modify
a noun phrase in the main clause, which I refer to as the 'head of the relative construction'. The
'relative construction' comprises the embedded clause and the head noun.
43. According to Verhaar (1995: 215-6), the ia...ia construction occurs only in one particular dialect
and is almost entirely absent in "standard or received Tok Pisin". It is one of several relative
constructions used "by speakers when they do not know immediately how to continue what they
are saying...". (Verhaar 1995: 216).
44. Like Heine and Reh, Benveniste (1966) argues that si and lá "frame" the relative clause in Ewe.
However, in contrast to Heine and Reh, Benveniste does not consider si ... lá a discontinuous
morpheme; rather, he contends that si and lá serve two separate functions in this construction.
45. To be precise, Lockwood only considers derjenige plus prepositional phrase. That is, he does
not mention the use of derjenige as a plain pronoun and he also ignores the occasional use of
derjenige with a subsequent noun (e.g. Diejenigen Leute, die das gesagt haben,...' 'Those people
who said that...').
46. Frajzyngier uses the notion of demonstrative as a cover term for demonstratives, definite articles
and anaphoric pronouns. I only cite examples that Frajzyngier glosses as demonstratives.
47. Himmelmann (1996: 222) argues that unstressed this is an extension of the use of Deixis am
Phantasma (cf. 5.1) and therefore should not be considered a grammatical marker. If indefinite
this was just another use of the proximal demonstrative, one would expect that it occurred in
contrast to the distal form that and that it is stressed in certain contexts. Since indefinite this is
always unstressed and non-contrastive, I assume, with Wright and Givón, that unstressed this
does not represent a particular use but rather a specific grammatical marker.
48. The latter contrasts with a third suffix (not shown in this table) which indicates a "location in
a time-dependent fashion: that is, location at some fixed point in time" (Anderson and Keenan
1985: 298).
49. Wari' has three demonstratives that combine spatial and temporal reference. They indicate how
NOTES 167
long the referent, which is either a person or object, has been absent from the current speech
situation: paca' 'that just occurred (always heard but never seen)', cara ne 'that recently absent',
and cara pane 'that long absent' (these are the adnominal forms). While ne and pane are
sentence-final temporal deictics, it is unclear whether paca' and cara can be traced back to any
other form; in particular, it is unclear if they derived from locational deictics.
50. Turkish has two locational preverbs, bu 'proximal' and o 'distal', that have the same morphological
form as the adnominal demonstratives while they differ from demonstrative adverbs and
pronouns; the latter two are morphologically more complex. At first glance, these data seem to
suggest that the preverbs in Turkish originated from adnominal demonstratives. It is, however,
difficult to conceive how a preverb might have developed from an adnominal demonstrative,
given that preverbs and adnominal demonstratives occur in very different syntactic contexts.
Following Himmelmann (1997: 21), I assume that demonstratives are historically based on
deictic particles with no specific syntactic function. Given that all demonstratives in Turkish
include the deictic roots, bu and o, it is quite likely that bu and o are the direct descendants of
such particles. I suspect that the preverbs of Modern Turkish developed from these particles.
More specifically, I make the following claim (which is subject to future investigation): the
preverbs bu and o evolved from deictic particles that were used adverbially before the demonstratives
of the current system emerged.
51. Lehmann (1995a: 101-3) examines another type of preverbation in Totonac and Abkhaz which,
in his view, might be an instance of grammaticalization.
52. Li and Thompson do not use the term left-dislocation and thus one cannot be certain as to
whether they mean that the topical NP is a left-dislocated constituent of a nonverbal clause. This
is, however, how subsequent studies have interpreted their analysis (cf. Gildea 1993; Devitt 1994).
53. Brugmann (1904: 123) discusses a similar case in Swiss German, where the identity pronoun
selb 'ipse', which many German dialects use to strengthen the weakened demonstrative der,
assumed a deictic function.
54. Williams (1976: 33) claims that Tuscarora has a demonstrative (kyé:nv: 'this') that is derived
from the verb: k-yenv: 'I am holding it'. She does not explain, however, how this might have
happened. It seems that her hypothesis is solely based on the morphological similarity between
the two forms, which could simply be an accident.
55. Traugott (1982: 245) says, for instance: "However, it is dubious whether we can trace all
grammatical markers derived by processes of grammaticalization to lexical items rather than to
certain seemingly fundamental grammatical items, such as demonstrative pronouns and
interrogatives. The Indo-European t-demonstrative and kU-interrogative, for example, have been
remarkably resistant to change over several thousand years, and no lexical source seems
reconstructable for them."
56. Wilkins (1992) argues that interjections and deictics have many features in common, which
might suggest that there is a common diachronic relationship between interjections and deictics
(i.e. demonstratives).
57. Fuchs (1996) shows that standard Croatian (spoken in the city of Zagreb) has (almost) lost the
middle term of a three-term deictic system, which most rural dialects of Croatian have preserved.
Her discussion implies that the change in the Zagreb dialect has been caused by changes in the
socioeconomic environment of the speakers. According to Frei (1944: 123) two-term deictic
systems are in general derived from three-term systems by reduction of the middle term as in
the Zagreb dialect of Croatian.
APPENDIX A
Data sources
Acehnese Durie (1985)
Ainu Refsing (1986), Dettmar (1989)
Alamblak Bruce (1984)
Ambulas Wilson (1980)
Ao Gowda (1975)
Apalai Koehn and Koehn (1986)
Barasano Jones and Jones (1991)
Basque Saltarelli (1988)
Burushaski Lorimer (1935), Berger (1998)
Byansi Trivedi (1991)
Canela-Krahô Popjes and Popjes (1986)
Czech Harkins (1953)
Duwai Schuh (1977)
Dyirbal Dixon (1972)
Epena Pedee Harms (1994)
Ewondo Redden (1980)
Finnish Sulkala and Karjalainen (1992), Laury (1995, 1997)
French Calvez (1993)
Georgian Hewitt (1995)
German Eisenberg (1994)
Gulf Arabic Holes (1990, 1995)
Guugu Yimidhirr Haviland (1979)
Halkomelem Galloway (1993)
Hixkaryana Derbyshire (1985a, 1985b)
Hua Haiman (1980)
Izi Meier, Meier and Bendor-Samuel (1975)
Japanese Kuno (1973), Imai (1996)
Kannada Schiffmann (1983), Sridhar (1990)
Karanga Marconnès (1931)
Khasi Rabel (1961), Nagaraja (1985)
Kiowa Watkins (1984)
Korean Lee (1989), Sohn (1994)
Koyra Chiini Heath (1999)
Kunuz Nubian Abdel-Hafiz (1988)
Kusaiean Lee (1975)
Lahu Matisoff (1973)
Lango Noonan (1992)
Lealao Chinantec Rupp (1989)
170 APPENDIX A
Lezgian Haspelmath (1993)
Logbara Crazzolara (1960)
Mam England (1983)
Manam Lichtenberk (1983)
Mandarin Chinese Chao (1968), Lin (1981), Li and Thompson (1981)
Margi Hoffmann (1963)
Modern Hebrew Glinert (1989)
Mojave Munro (1976)
Mulao Wang and Guoqiao (1993)
Nama Hagman (1977)
Nandi Creider and Tapsubei Creider (1989)
Ngiti Kutsch (1994)
Ngiyambaa Donaldson (1980)
Nùng Saul and Freiberger Wilson (1980)
Nunggubuyu Heath (1980, 1984)
Oneida Michelson (1996), Diessel (1999b)
Pangasinan Benton (1971)
Passamaquoddy-Maliseet Leavitt (1996), Ng (1999)
Picurís Zaharlick (1977)
Ponapean Rehg (1981)
Punjabi Bhatia (1993)
Quileute Andrade (1933)
Santali Bodding (1929)
Slave Rice (1989)
Supyire Carlson (1994)
Swazi Ziervogel (1952)
Swedish Viberg et al. (1995), Holmes and Hinchliffe (1994)
Tauya MacDonald (1990)
Tok Pisin Verhaar (1995)
Tümpisa Shoshone Dayley (1989)
Turkana Dimmendaal (1983)
Turkish Lewis (1967), Underhill (1976), Kornfilt (1997)
Tuscarora Mithun (1987)
Tzutujil Dayley (1985)
Urim Hemmilä (1989)
Urubu-Kaapor Kakumasu (1986)
Usan Reesink (1987)
Ute Givón (1980)
Vietnamese Thompson (1965), Phú Phong (1992)
Wardaman Merlan (1994)
Wari' Everett and Kern (1997)
Western Bade Schuh (1977)
West Futuna-Aniwa Dougherty (1983)
West Greenlandic Fortescue (1984)
Yagua Payne and Payne (1990)
Yankunytjatjara Goddard (1985)
Yimas Foley (1991)
APPENDIX B
The inflectional features of
pronominal demonstratives1
Number Gender Case None2
Acehnese +
Ainu +
Alamblak + +
Ambulas + +
Ao + +
Apalai + +
Barasano + +
Basque + + +
Burushaski + + +
Byansi + +
Canela-Krahô +
Czech + + +
Duwai +
Dyirbal + +
Epena Pedee + + +
Ewondo + +
Finnish + +
French + +
Georgian + +
German + + +
Gulf Arabic + +
Guugu Yimidhirr + +
Halkomelem + +
Hixkaryana + +
Hua +
Izi +
Japanese +
Kannada + + +
Karanga + +
Khasi + +
Kiowa + +
Korean *
Koyra Chiini +
Kunuz Nubian +
Kusaiean *
172 APPENDIX B
Lahu *
Lango +
Lealao Chinantec (+)
Lezgian + +
Logbara +
Mam +
Manam +
Mandarin Chinese *
Margi +
Modern Hebrew + +
Mojave + +
Mulao +
Nama + +
Nandi +
Ngiti +
Ngiyambaa + +
Nùng *
Nunggubuyu + + +
Oneida +
Pangasinan +
Passamaquoddy-Maliseet + +
Picurís + +
Ponapean +
Punjabi + + +
Quileute + +
Santali + + +
Slave +
Supyire + +
Swazi (+) +
Swedish + +
Tauya +
Tok Pisin +
Tümpisa Shoshone + +
Turkana + +
Turkish + +
Tuscarora +
Tzutujil +
Urim +
Urubu-Kaapor +
Usan +
Ute + + +
Vietnamese *
Wardaman + +
Wari' + +
Western Bade + +
West Futuna-Aniwa *
West Greenlandic + +
APPENDIX B 173
Yagua + +
Yankunytjatjara + +
Yimas + +
Total 64 38 25 17
1. Some of the sources that I consulted do not explicitly discuss the inflectional features of
pronominal demonstratives. In particular, in grammars of languages in which demonstratives
are not marked for gender, number and/or case, it is often not stated that pronominal demonstratives
are uninflected. Moreover, there are several grammars in my sample that do not distinguish
between (inflectional) affixes and clitics (cf. 2.2.1). The reader would therefore be
advised to check the original sources before citing information from this appendix.
2. The star * indicates that the language does not have pronominal demonstratives.
References
Abdel-Hafiz, A. S. 1988. A Reference Grammar of Kunuz Nubian. Buffalo, New
York: SUNY Buffalo dissertation.
Abney, S. P. 1987. The English Noun Phrase in its Sentential Aspect. Cambridge,
Massachusetts, MIT dissertation.
Andersen, H. (ed.) 1995. Historical Linguistics 1993. Amsterdam: John
Benjamins.
Anderson, S. R. 1992. A-Morphous Morphology. Cambridge: Cambridge University
Press.
Anderson, S. R. and Keenan, E. L. 1985. "Deixis." In Shopen (ed.) 3, 259-308.
Andrade, M. J. 1933. "Quileute." In Handbook of American Indian Languages, F.
Boas (ed.), 149-292. New York: Augustin.
Ariel, M. 1990. Accessing Noun-Phrase Antecedents. London: Routledge.
Auer, J. C. P. 1981. "Zur indexikalitätsmarkierenden Funktion der demonstrativen
Artikelform in deutschen Konversationen." In Sprache: Verstehen und
Handeln, G. Hindelang and W. Zillig (eds.), 301-311. Tübingen: Niemeyer.
Auer, J. C. P. 1984. "Referential Problems in Conversation." Journal of Pragmatics
8: 627-648.
Baker, M. C. 1996. The Polysynthesis Parameter. New York: Oxford University
Press.
Bhatia, T. K. 1993. Punjabi: A Cognitive-Descriptive Grammar. London:
Routledge.
Behaghel, O. 1923-32. Deutsche Syntax: Eine geschichtliche Darstellung. 4 vols.
Heidelberg: Winter.
Bell, A. 1978. "Language Samples." In Greenberg et al. (eds.) 1, 123-156.
Benton, R. A. 1971. Pangasinan Reference Grammar. Honolulu: University of
Hawaii Press.
Benveniste, E. 1966. "La phrase relative: Problème de syntaxe générale." In E.
Benveniste, Problèmes de linguistique générale, 208-222. Paris: Gallimard.
Berger, H. 1998. Die Burushaski-Sprache von Hunza und Nager. Teil 1:
Grammatik. Wiesbaden: Harrassowitz.
176 REFERENCES
Berman, R. A. and Grosu, A. 1976. "Aspects of the Copula in Modern Hebrew."
In Studies in Modern Hebrew Syntax and Semantics: The transformationalgenerative
approach, P. Cole (ed.), 265-285. Amsterdam: North-Holland.
Bisle-Müller, H. 1991. Artikelwörter im Deutschen. Tübingen: Niemeyer.
Bloomfield, L. 1933. Language. New York: Holt, Rinehart, and Winston.
Bodding, P. O. 1929. Materials for a Santali Grammar. Dunka: Santal Mission of
the Northern Churches.
Borsley, R. D. 1996. Modern Phrase Structure Grammar. Oxford: Blackwell.
Brown, C. 1985. "Polysemy, Overt Marking, and Function Words." Language
Sciences 7: 283-332.
Brown, R. 1973. A First Language. Harvard: Harvard University Press.
Bruce, L. 1984. The Alamblak Language of Papua New Guinea (East Sepik).
Canberra: Australian National University.
Brugmann, K. 1904. Demonstrativpronomina der indogermanischen Sprachen.
Leipzig: Teubner.
Brugmann, K. and Delbrück, B. 1911. Grundriss der vergleichenden Grammatik
der indogermanischen Sprachen. 2 vols. Strassburg: Teubner.
Bühler, K. 1934. Sprachtheorie: Die Darstellungsfunktion der Sprache. Jena:
Fischer.
Bybee, J. L. and Dahl, Ö. 1989. "The Creation of Tense and Aspect Systems in
the Languages of the World." Studies in Language 13: 51-103.
Bybee, J. L., Pagliuca, W. and Perkins, R. D. 1991. "Back to the Future." In
Traugott and Heine (eds.) 2, 17-58.
Bybee, J. L., Perkins, R. D. and Pagliuca, W. 1994. The Evolution of Grammar:
Tense, aspect, and modality in the languages of the world. Chicago: University
of Chicago Press.
Calvez, D. J. 1993. French Reference Grammar. Lincolnwood: National Textbook.
Canisius, P. and Sitta, G. 1991. "Textdeixis: Zum Verhältnis von Deixis, Substitution
und Anaphora." In Betriebslinguistik und Linguistikbetrieb: Akten des
24. Linguistischen Kolloquiums Bremen 1989, E. Klein, F. Pouradier Duteil
and K. H. Wagner (eds.), 143-152. Tübingen: Niemeyer.
Carlson, R. 1994. A Grammar of Supyire. Berlin: Mouton de Gruyter.
Chafe, W. L. 1987. "Cognitive Constraints on Information Flow." In Coherence
and Grounding in Discourse, R. S. Tomlin (ed.), 21-52. Amsterdam: John
Benjamins.
Chafe, W. L. 1994. Discourse, Consciousness, and Time: The flow and displacement
of conscious experience in speaking and writing. Chicago: University of
Chicago Press.
REFERENCES 177
Channon, R. 1985. "Anaphoric that: A friend in need." Chicago Linguistic Society
11: 98-109.
Chao, Y. R. 1968. A Grammar of Spoken Chinese. Berkeley: University of California
Press.
Chen, R. 1990. "English Demonstratives: A case study of semantic expansion."
Language Sciences 12: 139-153.
Chomsky, N. 1981. Lectures on Government and Binding. Dordrecht: Foris.
Christophersen, P. 1939. The Articles: A study of their theory and use in English.
Copenhagen: Munksgaard.
Clark, E.V. 1978. "From Gesture to Word: On the natural history of deixis in
language acquisition." In Human Growth and Development, J. S. Bruner and
A. Garton (eds.), 85-120. Oxford: Oxford University Press.
Claudi, U. and Heine, B. 1986. "On the Metaphorical Base of Grammar." Studies
in Language 10: 297-335.
Comrie, B. 1998. "Rethinking the Typology of Relative Clauses." Language
Design 1: 59-86.
Comrie, B. Forthcoming. "Pragmatic Binding: Demonstratives as anaphors in
Dutch." Berkeley Linguistics Society 23.
Craig, C. G. 1991. "Ways to Go in Rama: A case study in poly-grammaticalization."
In Traugott and Heine (eds.) 2, 455-492.
Crazzolara, J. P. 1960. A Study of the Logbara (Ma'di) Language. London: Oxford
University Press.
Creider, C. A. and Tapsubei Creider, J. 1989. A Grammar of Nandi. Hamburg:
Buske.
Croft, W. 1990. Typology and Universals. Cambridge: Cambridge University
Press.
Cyr, D. E. 1993a. "Cross-linguistic Quantification: Definite articles vs. demonstratives."
Language Sciences 15: 195-229.
Cyr, D. E. 1993b. "Definite Articles and Demonstratives in Plains Cree." In
Papers from the Twenty-Fourth Algonquian Conference, W. Cowan (ed.),
64-80. Ottawa: Carleton University.
Cyr, D. E. 1996. "Nikotwâsik iskwâhtêm, pâskihtêpayih!" In Studies in Honour
of H. C. Wolfart, J. D. Nichols and A. C. Ogg (eds.), 77-111. Winnipeg:
Memoir.
Dayley, J. D. 1985. Tzutujil Grammar. Berkeley: University of California Press.
Dayley, J. D. 1989. Tümpisa (Panamint) Shoshone Grammar. Berkeley: University
of California Press.
De Mulder, W. 1996. "Demonstratives as Locating Expressions." In Pütz and
Dirven (eds.), 29-47.
178 REFERENCES
Denny, J. P. 1978. "Locating the Universals in Lexical Systems for Spatial
Deixis." Chicago Linguistic Society: Papers from the Parasession on the
Lexicon: 71-84.
Denny, J. P. 1982. "Semantics of the Inuktitut (Eskimo) Spatial Deictics."
International Journal of American Linguistics 48: 359-384.
Derbyshire, D. C. 1985a. Hixkaryana. Amsterdam: North-Holland.
Derbyshire, D. C. 1985b. Hixkaryana and Linguistic Typology. Dallas: Summer
Institute of Linguistics and University of Texas at Arlington Press.
Derbyshire, D. C. and Pullum, G. K. (eds.) 1986-1998. Handbook of Amazonian
Languages. 4 vols. Berlin: Mouton de Gruyter.
Dettmar, H. A. 1989. Ainu-Grammatik. 2 vols. Wiesbaden: Harrassowitz.
Devitt, D. 1994. Copula Constructions in Crosslinguistic and Diachronic Perspective.
Buffalo, New York: SUNY Buffalo dissertation.
Diessel, H. 1997a. "The Diachronic Reanalysis of Demonstratives in Crosslinguistic
Perspective." Chicago Linguistic Society 33: 83-98.
Diessel, H. 1997b. "Verb-first Constructions in German." In Lexical and Syntactical
Constructions and the Construction of Meaning, M. Verspoor, K. D. Lee
and E. Sweetser (eds.), 51-68. Amsterdam: John Benjamins.
Diessel, H. 1998. Demonstratives in Crosslinguistic and Diachronic Perspective.
Buffalo, New York: SUNY Buffalo dissertation.
Diessel, H. 1999a. "The Morphosyntax of Demonstratives in Synchrony and
Diachrony." Linguistic Typology 3: 1-49.
Diessel, H. 1999b. "The Use of Demonstratives in Oneida Narratives." In Buffalo
Papers in Linguistics, K. Michelson (ed.). Buffalo, New York: SUNY
Buffalo.
Diessel, H. Forthcoming. "Predicative Demonstratives." Berkeley Linguistics
Society 23.
Dimmendaal, G. J. 1983. The Turkana Language. Dordrecht: Foris.
Dixon, R. M.W. 1972. The Dyirbal Language of North Queensland. Cambridge:
Cambridge University Press.
Dixon, R. M.W. 1982. Where Have All the Adjectives Gone. Berlin: Mouton.
Donaldson, T. 1980. Ngiyambaa: The language of the Wangaaybuwan. Cambridge:
Cambridge University Press.
Dougherty, J.W. D. 1983. West Futuna-Aniwa: An introduction to a Polynesian
outlier language. Berkeley: University of California Press.
Drosdowski, G. (ed.) 1995. Duden. Grammatik der deutschen Gegenwartssprache.
Mannheim: Dudenverlag.
Dryer, M. S. 1989a. "Article-Noun Order." Chicago Linguistic Society 25: 82-97.
Dryer, M. S. 1989b. "Plural Words." Linguistics 27: 865-895.
REFERENCES 179
Dryer, M. S. 1989c. "Large Linguistic Areas and Language Sampling." Studies in
Language 13: 257-292.
Dryer, M. S. 1992a. "The Greenbergian Word Order Correlations." Language 68:
81-138.
Dryer, M. S. 1992b. "SVO Languages and the OV: VO Typology." Linguistics
27: 443-482.
Dryer, M. S. 1996. "Focus, Pragmatic Presupposition, and Activated Propositions."
Journal of Pragmatics 26: 475-523.
Durie, M. 1985. A Grammar of Acehnese on the Basis of a Dialect of North Aceh.
Dordrecht: Foris.
Ehlich, K. 1979. Verwendungen der Deixis beim sprachlichen Handeln. 2 vols.
Bern: Lang.
Ehlich, K. 1982. "Anaphora and Deixis: Same, similar, or different?" In Jarvella
and Klein (eds.), 315-338.
Ehlich, K. 1983. "Deixis und Anapher." In Rauh (ed.), 79-98.
Ehlich, K. 1986. Interjektionen. Tübingen: Narr.
Ehlich, K. 1987. "so-Überlegungen zum Verhältnis sprachlicher Formen und
sprachlichen Handelns, allgemein und an einem widerspenstigen Beispiel."
In Sprache und Pragmatik. Lunder Symposium 1986, I. Rosengren (ed.),
279-298. Stockholm: Almqvist and Wiksell.
Ehrich, V. 1992. Hier und Jetzt: Studien zur lokalen und temporalen Deixis im
Deutschen. Tübingen: Niemeyer.
Eid, M. 1983. "The Copula Function of Pronouns." Lingua 59: 197-207.
Eisenberg, P. 1994. Grundriss der deutschen Grammatik. [Third edition]. Stuttgart:
Metzler.
Engel, U. 1988. Deutsche Grammatik. Heidelberg: Groos.
England, N. C. 1983. A Grammar of Mam: A Mayan language. Austin: University
of Texas Press.
Epstein, R. 1994. "The Development of the Definite Article in French." In
Perspectives on Grammaticalization, W. Pagliuca (ed.), 63-80. Amsterdam:
John Benjamins.
Epstein, R. 1995. "The Later Stages in the Development of the Definite Article:
Evidence from French." In Andersen (ed.), 159-175.
Everett, D. L. and Kern, B. 1997. Wari': The Pacaas Novos language of Western
Brazil. London: Routledge.
Fillmore, C. J. 1982. "Towards a Descriptive Framework for Spatial Deixis." In
Jarvella and Klein (eds.), 31-59.
Fillmore, C. J. 1997. Lectures on Deixis. Stanford: CSLI Publications.
180 REFERENCES
Fillmore, C. J. and Kay, P. 1993. Construction Grammar. Berkeley, University of
California, Ms.
Fillmore, C. J., Kay, P. and O'Connor, C. 1988. "Regularity and Idiomaticity in
Grammatical Constructions: The case of let alone." Language 64: 501-538.
Foley, W. A. 1980. "Toward a Universal Typology of the Noun Phrase." Studies
in Language 4: 171-199.
Foley, W. A. 1991. The Yimas Language of New Guinea. Stanford: Stanford
University Press.
Fortescue, M. 1984. West Greenlandic. London: Croom Helm.
Fox, B. A. (ed.) 1996. Studies in Anaphora. Amsterdam: John Benjamins.
Frajzyngier, Z. 1987. "From Verb to Anaphora." Lingua 72: 155-168.
Frajzyngier, Z. 1991. "The de dicto Domain in Language." In Traugott and Heine
(eds.) 1, 219-251.
Frajzyngier, Z. 1996. "On Sources of Demonstratives and Anaphors." In Fox
(ed.), 169-203.
Frajzyngier, Z. 1997. "Grammaticalization of Number: From demonstratives to
nominal and verbal plural." Linguistic Typology 1: 193-242.
Frei, H. 1944. "Systèmes de déictiques." Acta Linguistics 4: 111-129.
Fuchs, A. 1993. Remarks on Deixis. Heidelberg: Groos.
Fuchs, Ž. M. 1996. "Here and There in Croatian: A case study of an urban
standard variety." In Pütz and Dirven (eds.), 49-62.
Galloway, B. D. 1993. A Grammar of Upriver Halkomelem. Berkeley: University
of California Press.
Gernsbacher, M. A. and Shroyer, S. 1989. "The Cataphoric Use of the Indefinite
this in Spoken Narratives." Memory and Cognition 17: 536-540.
Gildea, S. 1993. "The Development of Tense Markers from Demonstrative
Pronouns in Panare (Cariban)." Studies in Language 17: 53-73.
Givón, T. 1979. On Understanding Grammar. New York: Academic Press.
Givón, T. 1980. Ute Reference Grammar. Ignacio: Ute Press.
Givón, T. (ed.) 1983. Topic Continuity in Discourse: Quantified Cross-Language
Studies. Amsterdam: John Benjamins.
Givón, T. 1984. Syntax. A Functional-Typological Introduction, vol. 1. Amsterdam:
John Benjamins.
Givón, T. 1990. Syntax. A Functional-Typological Introduction, vol. 2. Amsterdam:
John Benjamins.
Givón, T. 1995. Functionalism and Grammar. Amsterdam: John Benjamins.
Glinert, L. 1989. The Grammar of Modern Hebrew. Cambridge: Cambridge
University Press.
REFERENCES 181
Goddard, C. 1985. A Grammar of Yankunytjatjara. Alice Springs: Institute of
Aboriginal Development.
Goldberg, A. E. 1995. A Construction Grammar Approach to Argument Structure.
Chicago: University of Chicago Press.
Gordon, L. 1986. Maricopa Morphology and Syntax. Berkeley: University of
California Press.
Gowda, K. S. G. 1975. Ao Grammar. Mysore: Central Institute of Indian Languages.
Greenberg, J. H. 1963. "Some Universals of Language, with Particular Reference
to the Order of Meaningful Elements." In Universals of Grammar, J.H.
Greenberg (ed.), 73-113. Cambridge: MIT Press.
Greenberg, J. H. 1966. Language Universals, With Special Reference to Feature
Hierarchies. The Hague: Mouton.
Greenberg, J. H. 1978. "How Does a Language Acquire Gender Markers." In
Greenberg et al. (eds.) 3, 47-82.
Greenberg, J. H. 1985. "Some Iconic Relationships among Place, Time, and
Disourse Deixis." In Iconicity in Syntax, J. Haiman (ed.), 271-287. Amsterdam:
John Benjamins.
Greenberg, J. H. 1991. "The Last Stages of Grammatical Elements: Contractive
and expansive desemanticization." In Traugott and Heine (eds.) 1, 301-314.
Greenberg, J. H., Ferguson, C. A. and Moravcsik, E. A. (eds.) 1978. Universals of
Human Language. 4 vols. Stanford: Stanford University Press.
Grenoble, L. 1994. "Discourse Deixis and Information Tracking." Berkeley
Linguistics Society 20: 208-219.
Grimes, B. F. (ed.) 1997. Ethnologue: Languages of the world. [Thirteenth
edition]. Dallas: Summer Institute of Linguistics.
Gundel, J. K., Hedberg, N. and Zacharski, R. 1993. "Cognitive Status and the
Form of Referring Expressions in Discourse." Language 69: 274-307.
Hagman, R. S. 1977. Nama Hottentot Grammar. Bloomington: Indiana University
Press.
Haiman, J. 1980. Hua: A Papuan Language of the Eastern Highlands of New
Guinea. Amsterdam: John Benjamins.
Haiman, J. and Thompson, S. A. (eds.) 1988. Clause Combining in Grammar and
Discourse. Amsterdam: John Benjamins.
Hale, K. 1983. "Walpiri and the Grammar of Nonconfigurational Languages."
Natural Language and Linguistic Theory 1: 5-47.
Halliday, M. A. K. and Hasan, R. 1976. Cohesion in English. London: Longman.
Hanks, W. F. 1989. "The Indexical Ground of Deictic Reference." Chicago
Linguistic Society: Parasession on Language in Context: 104-122.
Hanks, W. F. 1990. Referential Practice. Chicago: University of Chicago Press.
182 REFERENCES
Harkins, W. E. 1953. A Modern Czech Grammar. New York: King's Crown Press.
Harms, P. L. 1994. Epena Pedee Syntax. Dallas: Summer Institute of Linguistics
and University of Texas at Arlington Press.
Harris, A. C. and Campbell, L. 1995. Historical Syntax in Cross-Linguistic
Perspective. Cambridge: Cambridge University Press.
Harris, M. 1978. The Evolution of French Syntax: A comparative approach.
London: Longman.
Harris, M. 1980. "The Marking of Definiteness in Romance." In Historical
Morphology, J. Fisiak (ed.), 141-156. The Hague: Mouton.
Harrison, S. P. 1976. Mokilese Reference Grammar. Honolulu: University of
Hawaii Press.
Haspelmath, M. 1993. A Grammar of Lezgian. Berlin: Mouton de Gruyter.
Haspelmath, M. 1997. From Space to Time: Temporal adverbials in the world's
languages. Unterschleissheim: Lincom Europa.
Haviland, John 1979. "Guugu Yimidhirr." In Handbook of Australian Languages,
R. M.W. Dixon and B. J. Blake (eds.) 1, 27-180.
Hawkins, J. A. 1978. Definiteness and Indefiniteness. London: Croom Helm.
Heath, J. 1980. "Nunggubuyu Deixis, Anaphora, and Culture." Chicago Linguistic
Society: Parasession on Pronouns and Anaphora: 151-165.
Heath, J. 1984. Functional Grammar of Nunggubuyu. Canberra: Australian
Institute of Aboriginal Studies.
Heath, J. 1986. "Syntactic and Lexical Aspects of Nonconfigurationality in
Nunggubuyu (Australia)." Natural Language and Linguistic Theory 4:
375-408.
Heath, J. 1999. A Grammar of Koyra Chiini: The Songhay of Timbuktu. Berlin:
Mouton de Gruyter.
Heine, B. 1997. Cognitive Foundations of Grammar. New York: Oxford University
Press.
Heine, B. and Reh, M. 1984. Grammaticalization and Reanalysis in African
Languages. Hamburg: Buske.
Heine, B., Claudi, U. and Hünnemeyer, F. 1991a. Grammaticalization: A conceptual
framework. Chicago: University of Chicago Press.
Heine, B., Claudi, U. and Hünnemeyer, F. 1991b. "From Cognition to Grammar:
Evidence from African languages." In Traugott and Heine (eds.) 1, 149-187.
Heinrichs, H. M. 1954. Studien zum bestimmten Artikel in den germanischen
Sprachen. Giessen: Schmitz.
Hemmilä, R. 1989. "The Demonstrative Pronouns pa and ti in Urim Discourse."
Language and Linguistics in Melanesia 20: 41-63.
REFERENCES 183
Hengeveld, K. 1990. "A Functional Analysis of Copula Constructions in Mandarin
Chinese." Studies in Language 14: 291-323.
Hengeveld, K. 1992. Non-Verbal Predication: theory, typology, diachrony. Berlin:
Mouton de Gruyter.
Herring, S. C. 1994. "Discourse Functions of Demonstrative Deixis in Tamil."
Berkeley Linguistics Society 20: 246-259.
Hetzron, R. 1995. "Genitival Agreement in Awngi: Variation on an Afroasiatic
theme." In Double Case: Agreement by Suffixaufnahme, F. Plank (ed.),
325-335. New York: Oxford University Press.
Hewitt, B. G. 1995. Georgian: A structural reference grammar. Amsterdam: John
Benjamins.
Himmelmann, N. 1996. "Demonstratives in Narrative Discourse: A taxonomy of
universal uses." In Fox (ed.), 205-254.
Himmelmann, N. 1997. Deiktikon, Artikel, Nominalphrase: Zur Emergenz syntaktischer
Struktur. Tübingen: Niemeyer.
Himmelmann, N. 1998. "Regularity in Irregularity: Article use in adpositional
phrases." Linguistic Typology 2: 315-353.
Hoffmann, C. 1963. A Grammar of the Margi Language. London: Oxford
University Press.
Holes, C. 1990. Gulf Arabic. London: Routledge.
Holes, C. 1995. Modern Arabic: Structures, functions and varieties. London:
Longman.
Holmes, P. and Hinchliffe, I. 1994. Swedish: A comprehensive grammar. London:
Routledge.
Hopper, P. J. 1987. "Emergent Grammar." Berkeley Linguistics Society 13:
139-157.
Hopper, P. J. 1991. "On Some Principles of Grammaticalization." In Traugott and
Heine (eds.) 1, 17-35.
Hopper, P. J. and Traugott, E. C. 1993. Grammaticalization. Cambridge: Cambridge
University Press.
Hudson, R. 1984. Word Grammar. Oxford: Blackwell.
Humboldt, W.von 1832 (1807). "Ueber die Verwandtschaft der Ortsadverbien mit
dem Pronomen in einigen Sprachen." In Abhandlungen der historischphilologischen
Klasse der Königlichen Akademie derWissenschaften zu Berlin
aus dem Jahre 1829, 1-26. Reprinted in Wilhelm von Humboldts Gesammelte
Schriften, 1. Abteilung, Band 6, 304-330. Berlin: Behr.
Imai, S. 1996. Space Divisions in Terms of Japanese Deictic Demonstratives.
Buffalo, SUNY Buffalo, Ms.
Jackendoff, R. S. 1977. X-Bar Syntax. Cambridge: MIT Press.
184 REFERENCES
Jakobson, R. 1971 (1957). "Shifters, Verbal Categories, and the Russian Verb."
In Roman Jakobson, SelectedWritings, vol. 2, 130-147. The Hague: Mouton.
Jarvella, R. J. and Klein, W. (eds.) 1982. Speech, Place, and Action. Chichester:
John Wiley.
Jefferson, R. J. and Zwicky, A. M. 1980. "The Evolution of Clitics." In Papers
from the Fourth International Conference on Historical Linguistics, E. C.
Traugott, R. La Brum and S. Shepherd (eds.), 221-231. Amsterdam: John
Benjamins.
Jones, W. and Jones, P. 1991. Barasano Syntax. Dallas: Summer Institute of
Linguistics and University of Texas at Arlington Press.
Kakumasu, J. 1986. "Urubu-Kaapor." In Derbyshire and Pullum (eds.) 1, 326-403.
Keenan, E. L. 1985. "Relative Clauses." In Shopen (ed.) 2, 147-170.
Klavans, J. L. 1985. "The Independence of Syntax and Phonology in Cliticization."
Language 61: 95-120.
Koehn, E. and Koehn, S. 1986. "Apalai." In Derbyshire and Pullum (eds.) 1,
33-127.
Kornfilt, J. 1997. Turkish. London: Routledge.
Koshal, S. 1979. Ladakhi Grammar. Delhi: Motilal Banarsidass.
Krámský, J. 1972. The Article and the Concept of Definiteness in Language. The
Hague: Mouton.
Kuno, S. 1973. The Structure of the Japanese Language. Cambridge: MIT Press.
KurySowicz, J. 1965. "The Evolution of Grammatical Categories." Reprinted in
J. KurySowicz, 1976, Esquisses Linguistiques, vol. 2, 38-54. Munich: Fink.
Kutsch Lojenga, C. 1994. Ngiti: A Central-Sudanic language of Zaire. Köln:
Köppe.
Lakoff, R. 1974. "Remarks on this and that." Chicago Linguistic Society 10:
345-356.
Lakoff, G. 1987. Women, Fire, and Dangerous Things. Chicago: University of
Chicago Press.
Lakoff, G. and Johnson, M. 1980. Metaphors We Live By. Chicago: University
of Chicago Press.
Lambrecht, K. 1981. Topic, Antitopic and Verb Agreement in Non-Standard
French. Amsterdam: John Benjamins.
Laury, R. 1995. "On the Grammaticalization of the Definite Article SE in Spoken
Finnish." In Andersen (ed.), 239-250.
Laury, R. 1997. Demonstratives in Interaction: The emergence of a definite article
in Finnish. Amsterdam: John Benjamins.
Leavitt, R. M. 1996. Passamaquoddy-Maliseet. Munich: Lincom Europa.
Lee, K. D.1975. Kusaiean Grammar. Honolulu: University of Hawaii Press.
REFERENCES 185
Lee, H. H. B. 1989. Korean Grammar. Oxford: Oxford University Press.
Lehmann, C. 1984. Der Relativsatz: Typologie seiner Strukturen, Theorie seiner
Funktionen, Kompendium seiner Grammatik. Tübingen: Narr.
Lehmann, C. 1985. "Grammaticalization: Synchronic variation and diachronic
change." Lingua e Stile 20: 303-318.
Lehmann, C. 1989. "Grammatikalisierung und Lexikalisierung." Zeitschrift für
Phonetik, Sprachwissenschaft und Kommunikationsforschung 42: 11-19.
Lehmann, C. 1993. "Theoretical Implications of Grammaticalization Phenomena."
In The Role of Theory in Language Description, W. A. Foley (ed.), 315-340.
Berlin: Mouton de Gruyter.
Lehmann, C. 1995a. Thoughts on Grammaticalization. Munich: Lincom Europa.
Lehmann, C. 1995b. "Synsemantika." In Syntax: Ein internationales Handbuch,
J. Jacobs, A. von Stechow, W. Sternefeld, and T. Vennemann (eds.),
1251-1266. Berlin: Mouton de Gruyter.
Leiss, E. 1994. "Die Entstehung des Artikels im Deutschen." Sprachwissenschaft
19: 307-319.
Levinson, S. C. 1983. Pragmatics. Cambridge: Cambridge University Press.
Lewis, G. L. 1967. Turkish Grammar. Oxford: Oxford University Press.
Li, C. N. and Thompson, S. A. 1977. "A Mechanism for the Development of
Copula Morphemes." In Mechanisms of Syntactic Change, C. N. Li (ed.),
419-444. Austin: University of Texas Press.
Li, C. N. and Thompson, S. A. 1981. Mandarin Chinese: A functional reference
grammar. Berkeley: University of California Press.
Lichtenberk, F. 1983. A Grammar of Manam. Honolulu: University of Hawaii
Press.
Lichtenberk, F. 1988. "The Pragmatic Nature of Nominal Anaphora in
To'aba'ita." Studies in Language 12: 299-344.
Lichtenberk, F. 1991. "On the Gradualness of Grammaticalization." In Traugott
and Heine (eds.) 1, 37-80.
Lichtenberk, F. 1996. "Patterns of Anaphora in To'aba'ita Narrative Discourse."
In Fox (ed.) 1996, 379-411.
Lin, H. T. 1981. Essential Grammar for Modern Chinese. Boston: Cheng and
Tsui.
Linde, C. 1979. "Focus of Attention and the Choice of Pronouns in Discourse."
In Discourse and Syntax [Syntax and Semantics 12], T. Givón (ed.),
337-354. New York: Academic Press.
Linde, C. and Labov, W. 1975. "Spatial Networks as a Site for the Study of
Language and Thought." Language 51: 924-939.
Lockwood, W. B. 1968. Historical German Syntax. Oxford: Clarendon.
186 REFERENCES
Lord, C. 1993. Historical Change in Serial Verb Constructions. Amsterdam: John
Benjamins.
Lorimer, D. L. R. 1935. The Burushaski Language. Vol. 1: Introduction and
Grammar. Oslo: Instituttet for Sammenligende Kulturforskning.
Lüdtke, H. 1991. "Überlegungen zur Entstehung des bestimmten Artikels im
Romanischen." Linguistica 31: 81-97.
Luo, C. 1997. "Iconicity or Economy? Polysemy between demonstratives, copulas
and contrastive focus markers." Chicago Linguistic Society 33: 273-286.
Lyons, J. 1977. Semantics. 2 vols. Cambridge: Cambridge University Press.
Lyons, J. 1979. "Deixis and Anaphora." In The Development of Conversation and
Discourse, T. Myers (ed.), 88-103. Edinburgh: Edinburgh University Press.
MacDonald, L. 1990. A Grammar of Tauya. Berlin: Mouton de Gruyter.
MacNeill, D., Cassell, J. and Levy, E. T. 1993. "Abstract Deixis." Semiotica 95:
5-19.
Marconnès, F. S. J. 1931. A Grammar of Central Karanga: The language of Old
Monomotapa as at present spoken in Central Mashonaland, Southern Rhodesia
[Supplement to Bantu Sudies, 5]. Johannesburg:Witwatersrand University
Press.
Mason, J. A. 1950. The Language of the Papago of Arizona. Philadelphia: University
Museum, University of Pennsylvania.
Matisoff, J. A. 1973. The Grammar of Lahu. Berkeley: University of California
Press.
Matsumoto, Y. 1988. "From Bound Grammatical Markers to Free Discourse
Markers: History of some Japanese connectives." Berkeley Linguistics Society
14: 340-351.
McWhorter, J. 1994. "From Focus Marker to Copula in Swahili." Berkeley
Linguistics Society: Special Session on Historical Issues in African Languages:
57-66.
Meier, P., Meier, I. and Bendor-Samuel, J. 1975. A Grammar of Izi: An Igbo
language. Norman: Summer Institute of Linguistics and University of
Oklahoma Press.
Meillet, A. 1921 (1912). "L'évolution des formes grammaticales." In A. Meillet,
Linguistique historique et linguistique générale, vol. 1, 130-148. Paris:
Klincksieck.
Merlan, F. C. 1994. A Grammar ofWardaman: A language of the Northern Territory
of Australia. Berlin: Mouton de Gruyter.
Michelson, K. 1996. Oneida Grammar. Buffalo, SUNY Buffalo, Ms.
Mithun, M. 1987. "The Grammatical Nature and Discourse Power of Demonstratives."
Berkeley Linguistics Society 13: 184-194.
REFERENCES 187
Moravcsik, E. 1997. "Parts and Wholes in the Hungarian Noun Phrase - a
Typological Study." In Proceedings of LP '96. Typology - Prototypes, Item
Orderings and Universals, B. Palek (ed.), 307-324. Prague: Charles University
Press.
Munro, P. 1976. Mojave Syntax. New York: Garland.
Nagaraja, K. S. 1985. Khasi: A descriptive analysis. Pune: Deccan College.
Ng, E. 1999. When Words with the Same Forms Have Different Functions: Demonstratives
and their derivatives in Maliseet-Passamaquoddy. Buffalo,
SUNY Buffalo, Ms.
Nichols, J. 1992. Linguistic Diversity in Space and Time. Chicago: University of
Chicago Press.
Noonan, M. 1992. A Grammar of Lango. Berlin: Mouton de Gruyter.
Paul, H. 1916-1920. Deutsche Grammatik. 5 vols. Halle an der Saale: Niemeyer.
Paul, H. 1920. Prinzipien der Sprachgeschichte. Tübingen: Niemeyer.
Paul, H. 1992. Deutsches Wörterbuch, H. Henne and G. Objartel (eds.) [Ninth,
revised edition], Tübingen: Niemeyer.
Payne, D. L. and Payne, T. E. 1990. "Yagua." In Derbyshire and Pullum (eds) 2,
249-474.
Peirce, C. S. 1955. Philosophical Writings of Peirce J. Buchler (ed.). New York:
Dover.
Perkins, R. D. 1992. Deixis, Grammar, and Culture. Amsterdam: John Benjamins.
Phú Phong, N. 1992. "Vietnamese Demonstratives Revisited." Mon-Khmer
Studies 20: 127-136.
Plank, F. 1979a. "Ikonisierung und De-Ikonisierung als Prinzipien des Sprachwandels."
Sprachwissenschaft 4: 121-158.
Plank, F. 1979b. "Exklusivierung, Reflexivierung, Identifizierung, relationale
Auszeichnung. Variationen zu einem semantisch-pragmatischen Thema." In
Sprache und Pragmatik. Lunder Symposium 1978, I. Rosengren (ed.),
330-354. Lund: Gleerup.
Plank, F. and Moravcsik, E.1996. "The Maltese Article: Language-particulars and
universals." Rivista di Linguistica 8: 183-212.
Pollard, C. and Sag, I. A. 1994. Head-Driven Phrase Structure Grammar. Chicago:
University of Chicago Press.
Popjes, J. and Popjes, J. 1986. "Canela-Krahô." In Derbyshire and Pullum (eds.)
1, 128-199.
Postal, P. M. 1969. "On So-called 'Pronouns' in English." In Modern Studies in
English: Readings in transformational grammar, D. A. Reibel and S. A.
Schane (eds.), 201-224. Englewood Cliffs: Prentice-Hall.
188 REFERENCES
Prince, E. F. 1981a. "On the Inferencing of Indefinite-this NPs." In Elements of
Discourse Understanding, A. K. Joshi, B. L. Webber and I. A. Sag (eds.),
231-250. Cambridge: Cambridge University Press.
Prince, E. F. 1981b. "Towards a Taxonomy of Given-new Information." In
Radical Pragmatics, P. Cole (ed.), 223-255. New York: Academic Press.
Prince, E. F. 1992. "The ZPG letter: Subjects, definiteness and informationstatus."
In Discourse Description: Diverse analyses of a fundraising text, S. A.
Thompson and W. Mann (eds.), 295-325. Amsterdam: John Benjamins.
Pushpa, K. 1976. Kokborok Grammar. Mysore: Central Institute of Indian
Languages.
Pütz, M. and Dirven, R. (eds.) 1996. The Construal of Space in Language and
Thought. Berlin: Mouton de Gruyter.
Quirk, R., Greenbaum, S., Leech, G. and Svartvik, J. 1972. A Grammar of
Contemporary English. London: Longman.
Rabel, L. 1961. Khasi: A language of Assam. Baton Rouge: Louisiana State
University Press.
Raidt, E. H. 1993. "Linguistic Variants and Language Change: Deictic variants
in some German and Dutch dialects vis-à-vis Afrikaans." In Historical
Linguistics 1991, J. van Marle (ed.), 281-293. Amsterdam: John Benjamins.
Rauh, G. (ed.) 1983. Essays on Deixis. Tübingen: Narr.
Rauh, G. 1983. "Aspects of Deixis." In Rauh (ed.), 9-60.
Redden, J. E. 1980. A Descriptive Grammar of Ewondo. Carbondale, Illinois:
Department of Linguistics.
Reed, L. 1994. "An Aspectual Analysis of French Demonstrative ce." Berkeley
Linguistics Society 20: 300-312.
Reesink, G. P. 1987. Structures and their Function in Usan: A Papuan language
of Papua New Guinea. Amsterdam: John Benjamins.
Refsing, K. 1986. The Ainu language. Århus: Aarhus University Press.
Rehg, K. L. 1981. Ponapean Reference Grammar. Honolulu: University of Hawaii
Press.
Renck, G. L. 1975. A Grammar of Yagaria. Canberra: Australian National University.
Rice, K. 1989. A Grammar of Slave. Berlin: Mouton de Gruyter.
Rijkhoff, J. and Bakker, D. 1998. "Language Sampling." Linguistic Typology 2:
263-314.
Rijkhoff, J., Bakker, D., Hengeveld, K, and Kahrel, P. 1993. "A Method of
Language Sampling." Studies in Language 17: 169-203.
Ross, M. D. 1988. Proto Oceanic and the Austronesian Languages of Western
Melanesia. Canberra: Australian National University.
REFERENCES 189
Ruhlen, M. 1991. A Guide to the World's Languages. Vol. 1: Classification.
[Second edition]. Stanford: Stanford University Press.
Rupp, J. E. 1989. Lealao Chinantec Syntax. Dallas: Summer Institute of Linguistics
and University of Texas at Arlington Press.
Saltarelli, M. 1988. Basque. London: Croom Helm.
Samarin, W. J. 1967. A Grammar of Sango. The Hague: Mouton.
Sankoff, G. and Brown, P. 1976. "The Origins of Syntax in Discourse: A case
study of Tok Pisin relatives." Language 52: 631-666.
Sapir, E. 1949. "A Study in Phonetic Symbolism." In SelectedWritings of Edward
Sapir, D. G. Mandelbaum (ed.), 61-72. Berkeley: University of California
Press.
Sasse, H-J. 1991. Arvanitika: Die albanischen Sprachreste in Grienchenland,
vol. 1. Wiessbaden: Harrassowitz.
Saul, J. E. and Freiberger Wilson, N. 1980. Nùng Grammar. Dallas: Summer
Institute of Linguistics and University of Texas at Arlington Press.
Schachter, P. 1985. "Parts-of-speech systems." In Shopen (ed.) 1, 3-61.
Schiffmann, H. F. 1983. A Reference Grammar of Spoken Kannada. Seattle:
University of Washington Press.
Schuh, R. G. 1977. "Bade/Ngizim determiner system." Afroasiatic Linguistics
4-3: 1-74.
Schuh, R. G. 1983a. "The Evolution of Determiners in Chadic." In Studies in
Chadic and Afroasiatic Linguistics, E. Wolff and H. Meyer-Bahlburg (eds.),
157-210. Hamburg: Buske.
Schuh, R. G. 1983b. "Kilba Equational Sentences." Studies in African Linguistics
14: 311-326.
Schuh, R. G. 1990. "Re-Employment of Grammatical Morphemes in Chadic:
Implications for language history." In Linguistic Change and Reconstruction
Methodology, P. Baldi (ed.), 599-618. Berlin: Mouton de Gruyter.
Shetter, W. Z. 1994. Dutch: An essential grammar. London: Routledge.
Shopen, T. ed. 1985. Language Typology and Syntactic Description. 3 vols.
Cambridge: Cambridge University Press.
Sidner, C. L. 1983. "Focusing in the Comprehension of Definite Anaphora." In
Computational Models of Discourse, M. Brady and R. C. Berwick (eds.),
267-330. Cambridge: MIT Press.
Siewierska, A. and Bakker, D. 1992. Data from Word Order Database. Handout
at Joint Meeting of NP and Constituent Order Groups of Eurotype, Donostia.
Sitta, G. 1991. Deixis am Phantasma: Versuch einer Neubestimmung. Bochum:
Brockmeyer.
Sohn, H-M. 1994. Korean. London: Routledge.
190 REFERENCES
Sridhar, S. N. 1990. Kannada. London: Routledge.
Stassen, L. 1997. Intransitive Predication. Oxford: Clarendon.
Sulkala, H. and Karjalainen, M. 1992. Finnish. London: Routledge.
Sweetser, E. 1988. "Grammaticalization and Semantic Bleaching." Berkeley
Linguistics Society 14: 389-405.
Sweetser, E. 1990. From Etymology to Pragmatics: Metaphorical and cultural
aspects of semantic structure. Cambridge: Cambridge University Press.
Talmy, L. 1988. "The Relation of Grammar and Cognition." In Topics in Cognitive
Linguistics, B. Rudzka-Ostyn (ed.), 165-205. Amsterdam: John
Benjamins.
Thompson, L. C. 1965. A Vietnamese Grammar. Seattle: University of Washington
Press.
Traugott, E. C. 1982. "From Propositional to Textual Meanings: Some semanticpragmatic
aspects of grammaticalization." In Perspectives on Historical
Linguistics, W. P. Lehmann and Y. Malkiel (eds.), 245-271. Amsterdam:
John Benjamins.
Traugott, E. C. 1989. "On the Rise of Epistemic Meanings in English: An
example of subjectification in semantic change." Language 65: 31-55.
Traugott, E. C. 1992. "Syntax." In The Cambridge History of the English Language.
Vol. 1: The Beginnings to 1066, R. M. Hogg (ed.), 168-289. Cambridge:
Cambridge University Press.
Traugott, E. C. and König, E. 1991. "The Semantics-pragmatics of Grammaticalization
Revisited." In Traugott and Heine (eds.) 1, 189-218.
Traugott, E. C. and Heine, B. (eds.) 1991. Approaches to Grammaticalization. 2
vols. Amsterdam: John Benjamins.
Trivedi, G. M. 1991. Descriptive Grammar of Byansi. A Bhotiya language. Calcutta:
Anthropological Survey of India, Government of India, Department of
Culture.
Ullmer-Ehrich, V. 1978. "Wohnraumbeschreibungen." Zeitschrift für Literaturwissenschaft
und Linguistik 9: 58-84.
Ultan, R. 1978a. "On the Development of a Definite Article." In Language
Universals. Papers from the Conference held at Gummersbach/Cologne,
Germany, October 3-8, 1976, H. Seiler (ed.), 249-265. Tübingen: Narr.
Ultan, R. 1978b. "Size-sound Symbolism." In Greenberg et al. (eds.) 2, 525-568.
Underhill, R. 1976. Turkish Grammar. Cambridge: MIT Press.
Van Valin, R. D. Jr. and LaPolla, R. J. 1997. Syntax: Structure, meaning, and
function. Cambridge: Cambridge University Press.
Vennemann, T. and Harlow, R. 1977. "Categorial Grammar and Consistent Basic
XV Serialization." Theoretical Linguistics 4: 227-254.
REFERENCES 191
Verhaar, J.W. M. 1995. Toward a Reference Grammar of Tok Pisin: An experiment
in corpus linguistics. Honolulu: University of Hawaii Press.
Viberg, Å., Ballardini, K. and Stjärnlöf, S. 1995. Essentials of Swedish Grammar.
Lincolnwood, Illinois: Passport.
Vogel, P. M. 1993. "Über den Zusammenhang von definitem Artikel und
Ferndeixis." Sprachtypologie und Universalienforschung 46: 222-233.
Vries, L. de 1995. "Demonstratives, Referent Identification and Topicality in
Wambon and Some Other Papuan Languages." Journal of Pragmatics 24:
513-533.
Wald, B. 1983. "Referents and Topic Within and Across Discourse Units: Observations
from current vernacular English." In Discourse-Perspectives on
Syntax, F. Klein-Andreu (ed.), 91-116. New York: Academic Press.
Wang, C. and Guoqiao, Z. 1993. An Outline of Mulao. Canberra: Australian
National University.
Watkins, L. J. 1984. A Grammar of Kiowa. Lincoln: University of Nebraska Press.
Webber, B. L. 1991. "Structure and Ostension in the Interpretation of Discourse
Deixis." Language and Cognitive Processes 6: 107-135.
Weissenborn, J. 1988. "Von der demonstratio ad oculos zur Deixis am
Phantasma: Die Entwicklung der lokalen Referenz bei Kindern." In Karl
Bühler's Theory of Language: Proceedings of the Conference held at Kirchberg,
August 26, 1984 and Essen, November 21-24, 1984, A. Eschenbach
(ed.), 257-276. Amsterdam: John Benjamins.
Weissenborn, J. and Klein, W. (eds.) 1982. Here and There: Cross-linguistic
studies on deixis and demonstratives. Amsterdam: John Benjamins.
Wilkins, D. P. 1992. "Interjections as Deictics." Journal of Pragmatics 18:
119-158.
Williams, M. M. 1976. A Grammar of Tuscarora. New York: Garland.
Wilson, P. R. 1980. Ambulas Grammar. Ukarumpa, Papua New Guinea: Summer
Institute of Linguistics.
Woodworth, N. L. 1991. "Sound Symbolism in Proximal and Distal Forms."
Linguistics 29: 273-299.
Wright, S. and Givón, T. 1987. "The Pragmatics of Indefinite Reference: Quantified
text-based studies." Studies in Language 11: 1-33.
Zaharlick, A. M. 1977. Picurís Syntax. Washington D. C.: American University
dissertation.
Ziervogel, D. 1952. A Grammar of Swazi. Johannesburg:Witwatersrand University
Press.
192 REFERENCES
Zigmond, M. L., Booth, C. G. and Munro, P. 1991. Kawaiisu. A grammar and
dictionary with texts, P. Munro (ed.). Berkeley: University of California
Press.
Zwicky, A. M. 1977. On Clitics. Bloomington: Indiana University Linguistics
Club.
Zwicky, A. M. 1985. "Heads." Journal of Linguistics 21: 1-29.
Zwicky, A. M. 1993. "Heads, Bases and Functors." In Heads in Grammatical
Theory, G. G. Corbett, N. M. Fraser and S. McGlashan (eds.), 292-315.
Cambridge: Cambridge University Press.
Language Index
A
Abkhaz 167
Acehnese 11, 22, 25, 29 30, 54,
89-92
Afrikaans 74, 164
Ainu 9, 12, 75, 104, 105
Alamblak 12, 36
Albanian 131-2
Ambulas 12, 15-8, 22, 26, 27, 29,
38, 49, 75, 80, 85, 105, 149
Ancient Greek 8, 123
Ao 11, 29, 30, 72
Apalai 10, 48
B
Bade 131
Barasano 10, 29, 30-1, 48, 72
Basque 9, 10, 53, 129
Bemba 129
Burushaski 9, 10, 48
Byansi 11, 26, 42
C
Canela-Krahô 10
Croatian 167
Czech 12, 27, 38
D
Daga 54
Dutch 74, 96
Duwai 11, 26, 27, 85
Dyirbal 12, 42, 43-44, 53, 61
E
English 2, 7, 25, 36, 38, 55, 58,
62-74, 75, 77, 79, 92, 94-95,
101-2 105-7, 108-9, 110, 124,
129, 135, 138-9, 149, 158, 160,
164
Old English 68-70, 124, 158, 161
Epena Pedee 10, 26, 42, 125-6
Ewe 134-5, 166
Ewondo 11, 18-20, 22, 53, 54
F
Faroese 124
Finnish 12, 75-78, 99, 120, 154
French 4, 5, 12, 27, 37, 38, 74, 79,
85, 87-8, 120, 128, 149, 160,
163, 164, 165
G
Georgian 12, 120
German 2, 8, 12, 27, 38, 55, 74, 85,
88, 96, 107, 108, 120-1, 126,
127, 136-7, 140-2, 166, 167
Swiss German 167
Old High German 8, 121-3
Middle High German 124
Early New High German 122
Gulf Arabic 11
Guugu Yimidhirr 12, 13-15, 22, 27,
74-5, 90
H
Halkomelem 10, 42
194 LANGUAGE INDEX
Hausa 131
Hixkaryana 10, 48, 100, 125
Hona 137
Hua 12, 26, 27-8, 29, 42, 44-5
Hungarian 120, 164
I
Icelandic 124
Ilokano 131
Inuktitut 13, 47, 49, 58, 85, 87,
142-3
Izi 11, 25, 80, 133
J
Japanese 12, 39, 59-60, 65, 73, 75,
99, 154, 160
K
Kambera 131
Kannada 12, 26, 140
Karanga 11, 61, 80, 81-2
Kawaiisu 23
Kera 131
Khasi 11, 29, 42-3, 126
Kilba 23, 80, 83-4, 147-8
Kiowa 10, 31, 45-6
Kokborok 29, 32
Korean 9, 12, 20-2, 25, 31-2, 55,
71-2, 73, 92, 154
Koyra Chiini 11, 25, 38, 100
Kunuz Nubian 11
Kusaiean 11, 25, 30, 72-3
L
Lahu 11, 42-3
Lango 11, 22, 24
Latin 79, 99-100, 120, 128, 140, 150
Lealao Chinantec 10
Lezgian 12, 26, 27, 29, 42, 59-60,
65, 73, 100
Logbara 11, 53
M
Mam 11
Manam 11, 54
Mandarin Chinese 11, 30-1, 154
Margi 11, 22, 27, 29, 80
Maricopa 100
Modern Hebrew 11, 85, 144-6,
149-50
Mojave 10, 142
Mokilese 149
Montagnais 98
Mparntwe Arrernte 109
Mulao 11, 25, 59-60, 65, 73
Mupun 137, 150
N
Nama 11, 26, 40
Nandi 11, 22
Ngiti 11, 53, 100
Ngiyambaa 5, 12, 42, 55, 79, 100,
111
Nùng 11, 25, 61
Nunggubuyu 12, 45-6, 58, 61, 85-6,
87, 92, 109
O
Oneida 10, 25, 61
P
Panare 147
Pangasinan 11, 27, 39, 75, 80, 85,
91-2, 165
Papago 141-2
Passamaquoddy-Maliseet 10, 42, 48,
154
Picurís 10
Podoko 137
Ponapean 5, 11, 22, 53, 75, 80, 83,
149
Punjabi 12
Q
Quileute 10, 42, 160
LANGUAGE INDEX 195
R
Rumanian 129
Russian 79, 96
S
Sango 133
Santali 11, 13
Sichule 131
Slave 10
Spanish 39
Supyire 11, 38, 80-1, 127
Swahili 148
Swazi 11, 80
Swedish 12, 27, 74, 124, 129, 135-6,
164
T
Tagalog 98, 130, 131
Tauya 12, 26, 42-3
To'aba'ita 96-7, 98, 165
Toba Batak 130
Tok Pisin 8, 9, 12, 25, 38, 133-4,
166
Tolai 131
Totonac 167
Tümpisa Shoshone 10, 27, 42, 54,
75, 85-6, 100, 111, 120
Turkana 11, 129
Turkish 12, 26, 59-60, 65, 73, 167
Tuscarora 4, 10, 25, 61, 73, 167
Tzutujil 10
U
Urim 12, 25, 139
Urubu-Kaapor 10, 25, 100
Usan 12, 25, 42, 100, 103, 111
Ute 10, 42, 48
V
Vietnamese 11, 36
W
Wardaman 12, 49, 61, 73, 99
Wari' 10, 166-7
Western Bade 11, 80, 82-3
West Futuna-Aniwa 11
West Greenlandic 10, 42, 44-6, 61,
100, 111
Wik Munkan 139-40
Woleaian 53-4
Wolio 131
Yagaria 22-3
Yagua 10, 30-1
Yankunytjatjara 12, 109
Yimas 12, 31-2, 39
X
Xhosa 129
Z
Zulu 129
Name Index
A
Abney, S. P. 57, 65-68, 164
Anderson, S. R. 2, 3, 24, 28, 38, 39,
40, 44, 53, 54, 99, 139, 140, 160,
163, 166
Andrade, M. J. 42
Ariel, M. 96
Auer, J. C. P. 105, 107
B
Baker, M. C. 61
Bakker, D. 129, 163
Behaghel, O. 8, 121
Bell, A. 163
Bendor-Samuel, J. 133
Benton, R. A. 58, 91-2, 165
Benveniste, E. 134, 166
Berman, R. A. 8, 144
Bisle-Müller, H. 163
Bloomfield, L. 57, 68
Bodding, P. O. 13
Borsley, R. D. 71
Brown, C. 28
Brown, P. 8, 133-4
Brown, R. 152
Bruce, L. 36-7
Brugmann, K. 8, 110, 165, 167
Bühler, K. 7, 9, 35, 95, 110, 152,
165
Bybee, J. L. 115-7, 139, 153, 163
C
Calvez, D. J. 37, 88, 149
Campbell, L. 8, 116, 123
Canisius, P. 101
Carlson, R. 38, 58, 80-1, 127
Cassell, J. 95
Chafe, W. L. 106
Chen, R. 7, 103, 105
Chomsky, N. 71
Christophersen, P. 8, 98, 128
Clark, E.V. 110-11, 152
Claudi, U. 8, 116-8, 139, 153
Comrie, B. 96, 120
Craig, C. G. 115
Crazzolara, J. P. 53
Creider, C. A. 24
Croft, W. 111-2
Cyr, D. E. 8, 98, 128
D
Dahl, Ö. 139
Dayley, J. D. 86, 100, 120
Delbrück, B. 8
De Mulder, W. 94
Denny, J. P. 13, 47, 49, 58, 87, 143,
160
Derbyshire, D. C. 100, 125
Devitt, D. 8, 143, 167
Dimmendaal, G. J. 129
Dixon, R. M.W. 43-4, 53, 61, 116,
163, 164
Donaldson, T. 5, 100, 111
Drosdowski, G. 136, 166
Dryer, M. S. 9, 64-5, 106, 138, 163
Durie, M. 90, 165
198 NAME INDEX
E
Ehlich, K. 3, 9, 95, 96, 152, 154
Ehrich, V. 95
Eid, M. 143
Eisenberg, P. 121
Engel, U. 163
Epstein, R. 8, 128
F
Fillmore, C. J. 3, 6, 35-6, 39, 40, 41,
58, 74, 79, 93, 94, 101, 103, 164
Foley, W. A. 32, 39, 130-2
Fortescue, M. 45-6, 100, 111
Frajzyngier, Z. 8, 137-8, 150
Frei, H. 38, 160, 167
Freiberger Wilson, N. 72
Fuchs, A. 110
Fuchs, Z. M. 167
G
Gernsbacher, M. A. 109, 138
Gildea, S. 8, 143, 147, 167
Givón, T. 42, 69, 70, 96, 109,
119-20, 138-9, 148, 161, 166
Glinert, L. 144-6, 149
Goddard, C. 109
Goldberg, A. E. 164
Gordon, L. 100
Gowda, K. S. G. 30, 32
Greenberg, J. H. 8, 28, 111, 128,
129, 161
Grenoble, L. 101
Grimes, B. F. 9
Grosu, A. 8, 144-5
Gundel, J. K. 7, 96, 103, 105, 109
Guoqiao, Z. 59
H
Haiman, J. 28, 29, 45
Hale, K. 61
Halliday, M. A. K. 6, 93, 103, 164
Hanks, W. F. 35, 94, 110
Harlow, R. 66
Harms, P. L. 125-6
Harris, A. C. 8, 116, 123
Harris, M. 1978. 8, 37, 120, 128,
150
Harrison, S. P. 149
Hasan, R. 6, 93, 103, 164
Haspelmath, M. 27, 29, 60, 100, 139
Haviland, J. 14-5
Hawkins, J. A. 106
Heath, J. 38, 45, 58, 61, 85-6, 92,
100
Heine, B. 8, 116-8, 134, 139, 148,
153, 166
Heinrichs, H. M. 8, 128
Hemmilä, R. 139
Hengeveld, K. 8, 58, 163
Herring, S. C. 101
Hetzron, R. 131
Himmelmann, N. 5, 6, 7, 8, 28, 38,
69, 70, 79, 91, 93-6, 98-9,
101-3, 105-10, 113, 128, 130-2,
135, 150, 159, 163, 165, 166,
167
Hinchliffe, I. 135-6
Hoffmann, C. 29
Holmes, P. 135-6
Hopper, P. J. 8, 69, 115-6, 119, 124,
150, 153, 166
Hudson, R. 57, 66
Humboldt, W.von 166
Hünnemeyer, F. 8, 116-8, 139, 153
I
Imai, S. 39, 99
J
Jackendoff, R.S. 164
Jakobson, R. 95
Jarvella, R. J. 3
Jefferson, R. J. 116
Johnson, M. 139
Jones, W. 31
Jones, P. 31
NAME INDEX 199
K
Kahrel, P. 163
Kakumasu, J. 100
Kay, P. 164
Keenan, E. L. 3, 28, 38, 39, 40, 44,
53, 54, 99, 120, 139-40, 160,
166
Klavans, J. L. 24, 163
Klein, W. 3
Koehn, E. 48
Koehn, S. 48
König, E. 118, 153
Kornfilt, J. 60
Krámský, J. 8, 128
Kuno, S. 39, 59, 99, 165
KurySowicz, J. 116
Kutsch Lojenga, C. 100
L
Labov, W. 95
Lakoff, G. 106-7, 139, 164
Lakoff, R. 105
Lambrecht, K. 120
LaPolla, R. J. 62-5, 68
Laury, R. 8, 75-8, 99, 100, 128
Lee, K. D. 72
Lehmann, C. 8, 69, 115-7, 120,
122-3, 128, 140-1, 163, 167
Leiss, E. 8, 128
Levinson, S. C. 6, 35-6, 93-5, 163
Levy, E. T. 95
Lewis, G. L. 59
Li, C. N. 8, 31, 143-5, 167
Lichtenberk, F. 7, 54, 96-8, 139,
165
Linde, C. 95-6
Lockwood, W. B. 8, 121-2, 141,
136-7, 166
Lord, C. 163
Lorimer, D. L. R. 48
Lüdtke, H. 8, 128
Luo, C. 148
Lyons, J. 3, 6, 7, 35, 71, 93, 95,
101, 110, 165
M
MacDonald, L. 43
Marconnès, F. S. J. 58, 81-2
Mason, J. A. 142
Matisoff, J.A. 43
Matsumoto, Y. 116
McNeill 95
McWhorter, J. 148
Meier, I. 133
Meier, P. 133
Meillet, A. 116
Merlan, F. C. 49, 61, 99
Mithun, M. 4, 61
Moravcsik, E. 129, 164
Munro, P. 142
N
Nagaraja, K. S. 43, 126
Nichols, J. 163
Noonan, M. 24
O
O'Connor 164
P
Pagliuca, W. 115-6, 139, 153
Paul, H. 8, 121, 126
Payne, T. E. 31
Payne, D. L. 31
Peirce, C. S. 9, 152
Perkins, R. D. 115-6, 139, 153, 160,
163
Plank, F. 9, 129, 151
Pollard, C. 71
Postal, P. M. 66
Prince, E. F. 106, 109, 138
Q
Quirk, R. 135
200 NAME INDEX
R
Rabel, L. 43
Raidt, E. H. 74
Rauh, G. 3, 37
Redden, J. E. 18-20, 53, 164
Reed, L. 88
Reesink, G. P. 100, 103-4, 111
Refsing, K. 105
Reh, M. 5, 8, 117, 134, 148, 166
Rehg, K. L. 53, 58, 75, 83
Renck, G. L. 23
Rijkhoff, J. 163
Ross, M. D. 131
S
Sag, I. A. 71
Saltarelli, M. 53
Samarin, W. J. 133
Sankoff, G. 8, 133-4
Sapir, E. 151
Sasse, H-J. 131
Saul, J. E. 72
Schachter, P. 58
Schiffmann, H. F. 140
Schuh, R. G. 8, 58, 82-4, 85, 129,
131-2, 143, 147, 164
Shroyer, S. 109, 138
Siewierska, A. 129
Sitta, G. 95, 101
Sohn, H-M. 20-2, 72, 163
Stassen, L. 143
Sweetser, E. 117, 153
T
Talmy, L. 49, 116
Tapsubei Creider, J. 24
Thompson, L. C. 36
Thompson, S. A. 8, 31, 143-5, 167
Traugott, E. C. 8, 9, 68-9, 116-8,
124, 150, 151, 153, 166, 167
U
Ullmer-Ehrich, V. 95
Ultan, R. 8, 128, 151
V
Van Valin, R. D. Jr. 62-5, 68
Vennemann, T. 66
Verhaar, J.W. M. 166
Viberg, Å. 135
Vogel, P. M. 8, 128
Vries, L. de 154
W
Wald, B. 109, 138
Wang, C. 59
Watkins, L. J. 31, 46
Webber, B. L. 6, 101, 165
Weissenborn, J. 3, 110
Wilkins, D. P. 167
Williams, M. M. 167
Wilson, P. R. 16-8, 80, 149
Woodworth, N. L. 9, 28, 150-1
Wright, S. 109, 138, 166
Z
Ziervogel, D. 58
Zigmond, M. L. 23
Zwicky, A. M. 24, 70-1, 116
Subject Index
A
Adverb
Locational 17, 47, 90, 139-40
Temporal 7, 16, 69, 139-40, 155,
159, 164
Animate/inanimate 16, 47-8, 158
Appositional NP 4, 61-2, 65, 67,
69-70, 92, 158
Article
Definite 7-8, 18, 19, 37-8, 64-5,
67, 75, 91, 96, 98, 106, 115,
128-9, 134, 135-6, 155, 159,
163, 166
Indefinite 7, 105
Specific indefinite 7, 109, 129,
138-9, 155, 159
B
Boundary marker 7, 132-5, 155, 159
Boundedness 47, 49, 50, 158, 164
C
Classifier 5, 25, 29, 30, 33, 53, 72,
73, 157-8
Clitic 22-6, 32, 83-4, 90, 120, 129,
157, 163
Complementizer 7, 8, 102, 123-5,
155, 159
Construction Grammar 70, 164
Copula 7, 8, 58, 73, 79-80, 115,
143-8, 148-9, 155, 159
Copular clause 5-6, 57-8, 72-3,
79-92, 93, 143-8, 154, 164
D
Defective noun 20-1, 25, 31, 33,
71-2, 92
Deixis
Am Phantasma 95, 113, 166
Emotional 107
Person 35-6, 47
Place 36, 47
Shift 47, 95
Social 36
Text 101, 165
Time 36
Deictic system
Distance-oriented 39-41, 50, 160
Person-oriented 39-41, 50, 160
One term 2, 37-9, 50
Two term 36-40, 50, 163, 167
Three term 19, 39-41, 50, 163,
167
Four term 19, 40-1, 50
Five, six, seven terms 40
Deictic features
Across the speaker's line of vision
45-7, 157
Along the coast line 45, 50, 157
Away from the speaker 45-7, 50,
157
Distance features 35-41, 50, 158
Toward the speaker 45-7, 50, 157
Up - down 40, 42-3, 50, 157
Uphill - downhill 41, 44-5, 50,
157
202 SUBJECT INDEX
Upriver - downriver 41, 44, 50,
157
Visible - invisible 40, 41-2, 50,
157
Demonstrative
Acquisition 110-11, 152
Adverb/ adverbial 2, 4-5, 7, 14, 17,
19-25, 27-9, 31-3, 36, 38, 42, 45, 47,
50, 53-4, 57-8, 74-8, 89-92, 93, 115,
134, 139-143, 154-5, 158, 161
Complex forms (stems) 21, 28-32,
111-2, 157
Determiner/adnominal 4-5, 7, 14,
16, 18, 20-6, 28-9, 31-3, 36-9, 42,
45, 50, 57-8, 59-74, 75, 89-92, 93,
100, 107-8, 115, 122-3, 128-139,
140, 154-5, 158, 161, 163
Diachronic origin 8-9, 150-4, 160,
161
Diminutive 22
Identifier / identificational 4-7, 14-5,
19, 22-5, 27-8, 32-3, 45, 50, 57-8,
73, 78-88, 89-92, 93, 115, 140,
143-150, 154-5, 158, 163
Inflection 25-8, 30, 32, 84-8, 157
Manner 17, 58, 74-5, 104-5, 125,
154, 165
Phylogenetic development 165-6
Predicative (see identificational)
Pronoun/pronominal 4-5, 7, 14, 16,
18, 22-33, 36-9, 42, 45, 47-8, 50, 53,
57-8, 59-74, 75, 77-8, 89-92, 93,
100, 115, 119-28, 132, 135, 140,
143-8, 154-5, 158, 161
Sentential 79, 85, 87, 165
Determinative 7, 108, 109, 112-3,
135-7, 155, 159
Determiner-as-head hypothesis 57,
66-8, 164
Directional marker (see preverb)
E
Expletive 7, 84, 88, 149-50, 153,
155, 159
F
Female/male 48, 158
Focus marker 7, 148-50, 155, 159
G
Grammaticalization
General 7-9, 30, 69-70, 74,
108-9, 112-13, 115-55, 159,
160-1, 167
Polygrammaticalization 115
Principles 116-119
H
Hesitation signal 154
Human/nonhuman 47-8, 158
I
Interjection 154, 167
L
Linker 7, 91, 130-2, 155, 159
M
Markedness 110-14
Morphological marking
Case 25-8, 32-3, 163
Gender 25-7, 30, 33
Number 25-7, 30, 33, 48-9, 137-8
N
Nominalizer 28, 29, 33, 87, 157
Non-configurational 85-6, 87
Nonverbal clause 5-6, 14-5, 19,
57-8, 72, 79-92, 93, 143-8,
148-9, 154, 159
Noun class marker 30, 128-9, 155,
159
Number marker 7, 115, 137-8, 155,
159
SUBJECT INDEX 203
P
Possessive 7, 65, 68-70, 115, 127-8,
155, 159
Pragmatic uses of demonstratives
Anaphoric 6, 39, 50, 75, 93,
95-100, 101-3, 109-14, 116,
120, 127-8, 143, 159, 166
Cataphoric 102, 103, 105, 123-4,
165, 166
Discourse deictic 6, 17, 36, 50, 73-4,
93, 100-5, 109-14, 123, 125, 127,
154, 159, 166
Endophoric 6, 7, 93, 109-14, 119,
153, 159
Exophoric 6-7, 50, 84, 93, 94-5, 99,
101, 109-14, 159, 165
Gestural 94
Recognitional 6-7, 50, 93, 105-9,
109-14, 116, 159, 165
Symbolic 94
Preverb 7, 140-3, 155, 159, 167
Pronominal NP 25, 30
R
Reference
Contrastive / non-contrastive 52-4
Emphatic/ non-emphatic 52-3
Precise /vague 19-20, 52-4, 164
Relative pronoun 7, 8, 115, 120-3,
155, 159, 166
S
Sentence connective 7, 102, 112-3,
115, 125-8, 154, 155, 159
Sound symbolism 9, 151-2, 160
T
Third person pronoun 5, 7, 8, 15, 16,
23, 25, 29, 30, 33, 38, 72, 73,
84, 87, 96, 97, 99, 112-3, 115,
119-20, 155, 157, 159, 165
Topic marker 154
In the series TYPOLOGICAL STUDIES IN LANGUAGE (TSL) the following titles have
been published thus far:
1. HOPPER, Paul J. (ed.): Tense-Aspect: Between semantics & pragmatics. 1982.
2. HAIMAN, John & Pamela MUNRO (eds): Switch Reference and Universal Grammar.
Proceedings of a symposium on switch reference and universal grammar, Winnipeg, May
1981. 1983.
3. GIVÓN, T.: Topic Continuity in Discourse. A quantitative cross-language study. 1983.
4. CHISHOLM, William, Louis T. MILIC & John A.C. GREPPIN (eds): Interrogativity: A
colloquium on the grammar, typology and pragmatics of questions in seven diverse languages,
Cleveland, Ohio, October 5th 1981-May 3rd 1982. 1984.
5. RUTHERFORD, William E. (ed.): Language Universals and Second Language Acquisition.
1984 (2nd ed. 1987).
6. HAIMAN, John (Ed.): Iconicity in Syntax. Proceedings of a symposium on iconicity in
syntax, Stanford, June 24-26, 1983. 1985.
7. CRAIG, Colette (ed.): Noun Classes and Categorization. Proceedings of a symposium on
categorization and noun classification, Eugene, Oregon, October 1983. 1986.
8. SLOBIN, Dan I. & Karl ZIMMER (eds): Studies in Turkish Linguistics. 1986.
9. BYBEE, Joan L.: Morphology. A Study of the Relation between Meaning and Form. 1985.
10. RANSOM, Evelyn: Complementation: its Meaning and Forms. 1986.
11. TOMLIN, Russel S.: Coherence and Grounding in Discourse. Outcome of a Symposium,
Eugene, Oregon, June 1984. 1987.
12. NEDJALKOV, Vladimir (ed.): Typology of Resultative Constructions. Translated from the
original Russian edition (1983). English translation edited by Bernard Comrie. 1988.
14. HINDS, John, Shoichi IWASAKI & Senko K. MAYNARD (eds): Perspectives on
Topicalization. The case of Japanese WA. 1987.
15. AUSTIN, Peter (ed.): Complex Sentence Constructions in Australian Languages. 1988.
16. SHIBATANI, Masayoshi (ed.): Passive and Voice. 1988.
17. HAMMOND, Michael, Edith A. MORAVCSIK and Jessica WIRTH (eds): Studies in
Syntactic Typology. 1988.
18. HAIMAN, John & Sandra A. THOMPSON (eds): Clause Combining in Grammar and
Discourse. 1988.
19. TRAUGOTT, Elizabeth C. and Bernd HEINE (eds): Approaches to Grammaticalization,
2 volumes (set) 1991
20. CROFT, William, Suzanne KEMMER and Keith DENNING (eds): Studies in Typology
and Diachrony. Papers presented to Joseph H. Greenberg on his 75th birthday. 1990.
21. DOWNING, Pamela, Susan D. LIMA and Michael NOONAN (eds): The Linguistics of
Literacy. 1992.
22. PAYNE, Doris (ed.): Pragmatics of Word Order Flexibility. 1992.
23. KEMMER, Suzanne: The Middle Voice. 1993.
24. PERKINS, Revere D.: Deixis, Grammar, and Culture. 1992.
25. SVOROU, Soteria: The Grammar of Space. 1994.
26. LORD, Carol: Historical Change in Serial Verb Constructions. 1993.
27. FOX, Barbara and Paul J. Hopper (eds): Voice: Form and Function. 1994.
28. GIVÓN, T. (ed.) : Voice and Inversion. 1994.
29. KAHREL, Peter and René van den BERG (eds): Typological Studies in Negation. 1994.
30. DOWNING, Pamela and Michael NOONAN: Word Order in Discourse. 1995.
31. GERNSBACHER, M. A. and T. GIVÓN (eds): Coherence in Spontaneous Text. 1995.
32. BYBEE, Joan and Suzanne FLEISCHMAN (eds): Modality in Grammar and Discourse.
1995.
33. FOX, Barbara (ed.): Studies in Anaphora. 1996.
34. GIVÓN, T. (ed.): Conversation. Cognitive, communicative and social perspectives. 1997.
35. GIVÓN, T. (ed.): Grammatical Relations. A functionalist perspective. 1997.
36. NEWMAN, John (ed.): The Linguistics of Giving. 1998.
37. RAMAT, Anna Giacalone and Paul J. HOPPER (eds): The Limits of Grammaticalization.
1998.
38. SIEWIERSKA, Anna and Jae Jung SONG (eds): Case, Typology and Grammar. In honor
of Barry J. Blake. 1998.
39. PAYNE, Doris L. and Immanuel BARSHI (eds.): External Possession. 1999.
40. FRAJZYNGIER, Zygmunt and Traci S. CURL (eds.): Reflexives. Forms and functions.
2000.
41. FRAJZYNGIER, Zygmunt and Traci S. CURL (eds): Reciprocals. Forms and functions.
2000.
42. DIESSEL, Holger: Demonstratives. Form, function and grammaticalization. 1999.
43. GILDEA, Spike (ed.): Reconstructing Grammar. Comparative Linguistics and Grammaticalization.
2000.
44. VOELTZ, F.K. Erhard and Christa KILLIAN-HATZ (eds.): Ideophones. n.y.p.
45. BYBEE, Joan and Paul HOPPER (eds.): Frequency and the Emergence of Linguistic
Structure. 2001.
46. AIKHENVALD, Alexandra Y., R.M.W. DIXON and Masayuki ONISHI (eds.): Noncanonical
Marking of Subjects and Objects. 2001.
47. BARON, Irene, Michael HERSLUND and Finn SORENSEN (eds.): Dimensions of Possession.
n.y.p.
48. SHIBATANI, Masayoshi (ed.): The Grammar of Causation and Interpersonal Manipulation.
n.y.p.
49. WISCHER, Ilse and Gabriele DIEWALD (eds.): New Reflections on Grammaticalization.
n.y.p.
Bạn đang đọc truyện trên: Truyen2U.Pro