![Preface
Quantification is one of the most basic notions for human beings, and conse-
quently it is one of the most commonly marked notions in languages. It is indi-
cated with different devices, phonological, morphological, syntactic, and others.
In this book, quantification will be observed and studied from three different
perspectives: synchronic, diachronic and typological. Several topics are studied
in the book, with a focus on plurality. The book will take advantage of recent
results from research on quantification in languages of China to deal with more
concrete and more targeted subjects, such as relationships between plurality
and classifiers, plurality and reduplication, plurality expressions in non-Han
languages, and plurality concepts in child language acquisition.
This book originated from a project awarded by the French ANR, Quantifi-
cation and Plurality ANR-06-BLAN0259, then by the IUF (University Institute of
France). Without their substantial financial support, this book could not have
been completed. The editor expresses her sincere gratitude to these institutions.
The authors are from France, Germany, Norway, China, Hongkong and the United
States. Thanks to collective research and to the contribution of each scholar, the
book progressively took shape and became what it is today. Completion of the
project is actually marked by the publication of two books: one was published in
China, Liàng yǔ fùshù de yánjiū 量与复数的研究 ([Quantification and plurality]
2010), and the second (the present book) is being published in Europe. In the
research, as well as in the editing work, my colleagues have helped me in one
way or another and have given me pertinent remarks and useful advice. The
editor’s thanks also go to Murielle Fabre who did the formatting work for
this volume, which is time-consuming labor. Many thanks also to Marie-Claude
Paris and Jingqi Fu for helping me with checking work. Undoubtedly, all faults
are mine.
Dan Xu
Paris
September 2012](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-11-320.jpg)
![Dan Xu
Introduction
Plurality and Classifiers across languages
of China1
1 Background
Plurality has drawn the attention of scholars (Sanches [1971] 1973; Greenberg
1972, 1988; Lü 1984; Dryer 1989, 1992; Bisang 1993, Heide 1995, Ortmann 2000,
Comrie 2008, Xu 2010, among others) for several decades. Papers on plurality
have most often discussed it in terms of numeral classifiers and the absence of
obligatory plural marking. But numeral classifiers and plural marking are not
the only strategies for grammatical quantification; reduplication is a third device
for quantification. In fact, plural marking, numeral classifiers and reduplication
constitute the main means of quantification marking in the domain of grammar.
There are many more lexical devices for expressing quantification, but this book
is interested in grammatical systems. The present introduction will first con-
centrate on the typological correlation between the three different strategies
for quantification, as well as some general issues. In the second section, it
will address individual devices for quantification in Chinese, concentrating on
plurality and reduplication, which have been studied less frequently than classi-
fiers. The third part will go beyond Chinese and will present a short typology of
how quantification is expressed in the other languages of China. A description
of how the book is organized will conclude this introduction.
Languages in which all three strategies of quantification are marked obliga-
torily are rare, while languages lacking all of them are even less frequent (see
Chapter 2 in this volume). It is important to distinguish two notions here, co-
occurrence and co-existence. I refer to the co-existence of three devices (plural
markers, reduplication and classifiers) instead of their co-occurrence, which is
1 This book is one of the outcomes related to the ANR projects Quantification et Pluralité ANR-
06-BLAN0259 and Space and Quantification (IUF). We express our gratitude for their financial
support. Our thanks also go to the anonymous reviewer for his/her constructive comments on
the Introduction and to all participants who have provided their contribution to the present
volume.
My debts to Walter Bisang are numerous. He has given me not only effective support and
help but also very useful suggestions for this Introduction.
Finally thanks to Craig Baker for his help with proofreading.](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-13-320.jpg)
![impossible in most languages. They are correlated in expressing quantity: if one
is present and obligatory, the other two are often absent or optional. This also
assumes that in human languages, quantification has to be explicitly or
implicitly indicated by different means. If a language does not possess obligatory
plural marking, it tends to exploit reduplication as a grammatical device. In
this sense, plural marking and reduplication are comparable in syntax. More
concretely, one or the other is frequently attested in the nominal structures of
human languages.
Reduplication often co-exists with classifiers in Mainland Southeast Asian
languages. This is one of the characteristics of this linguistic area (Enfield 2005).
In fact, two patterns are found in languages: obligatory plural marking on the one
hand vs. reduplication-classifier marking on the other. Reduplication is rarely
obligatory (even though it is fully productive in many languages, cf. Rubino
2005) and adds specific semantics to plurality. For that reason, it can easily co-
exist with the numeral-classifier systems of Chinese and Mainland Southeast
Asian languages.
Even though the mutual exclusion of numeral classifiers and obligatory plural
marking has often been discussed in typological work since Sanches ([1971] 1973)
and Greenberg (1972), the theory is not as water-tight as is often claimed. Even
Greenberg (1972) does not treat this correlation as having no exceptions. Some
publications, especially a monograph on Japanese by P. Downing (1996), have
pointed out that the correlations between plurality and classifiers must be more
widely investigated and better described, since counterexamples exist. Based
on Japanese data, Downing (1996: 28) indicates that “the language [Japanese]
possesses not only numeral classifiers, but several plural markers which may
appear with nouns which are accompanied by classifiers as well.” Bisang (in
this volume) also shows potential counterexamples with data on Northern Kam
and Weining Ahmao.
Reduplication and the occurrence of numeral classifiers are compared much
less in the literature. In contemporary Chinese, plural marking and reduplica-
tion are not obligatory, while classifiers are when using numerals. One device
cannot co-occur with the others as is illustrated by the following examples:
(1) 人们， 人人， 五个人
rén men rén rén wǔ ge rén
person-PL, person-person, five-CL-person
‘people, everybody/people, five people’
(2) *人人们， *五个人们
rén rén men wǔ ge rén men
person-person-PL, five-CL-person-PL
2 Dan Xu](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-14-320.jpg)
![In fact, Chinese seems to have become a nominal classifier language not as
early as has been believed, and the order “numeral + classifier + noun” won as a
dominant order only after the 10th century AD (Wu 2006). This means that Old
Chinese was not a classifier language for a long time. It is thus interesting to
investigate how plurality marking developed during the periods when classifiers
were not obligatory (see Feng in this volume). Plural markers in contemporary
Chinese are obviously based on word plurality (one of the categories of plurality
denoted by Dryer [1989]), while Old and Middle Chinese used other means (Xu
2010). In Chinese, nouns and verbs can be reduplicated to express plurality,
pluractionality, distributivity, etc. Not all nouns can be reduplicated, and not
all reduplicated nouns indicate plurality. Also, productive reduplication is rela-
tively new in Chinese, and very few examples can be found in Old Chinese and
Middle Chinese. In modern Mandarin Chinese, nouns cannot be freely redupli-
cated. Not as many nouns can be reduplicated (such as 人人 rénrén ‘everyone’,
天天 tiāntiān ‘everyday’) as adjectives and verbs. Classifiers, like nouns, can be
repeated indicating distributivity:
(3) 人人， 年年
rénrén, niánnián
‘everyone, every year’
(4) 个个, 张张
gègè, zhāngzhāng
‘every one, every piece (of paper)’
(5) 家家户户
jiājiāhùhù
‘each family’
Example (3) is often cited to show noun reduplication in Chinese. In fact,
this is far less productive than adjective or verb reduplication. Temporal words
such as 年 nián ‘year’ can be reduplicated. In (4), classifiers are reduplicated to
express distributivity. In general, classifiers can be reduplicated when the con-
text is clear. Example (5) shows that some reduplicated words have been com-
pletely lexicalized. These examples are abundant and we shall pass over them
here. One thing is clear: distribution/distributivity is in fact another way to
express plurality; and without a concept of collectivity in which plurality is
implied, the expression “everyone” is not tenable.
Quantification and many other grammatical phenomena are better described
and studied in Indo-European languages because comparative methods have
Introduction: Plurality and Classifiers across languages of China 3](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-15-320.jpg)
![been established by European scholars since the 19th century. The investigation
of the languages of China seems to lag far behind. The linguistic situation in
China is also more complex, since more languages from more families are
involved. The languages of China consist of the Han languages (Mandarin and
at least nine other Sinitic languages), and the Han are surrounded by different
ethnic groups and languages. It is thus important and necessary to study the
Han languages as well as the non-Han languages of China. Since European lan-
guages tend to have obligatory plural marking and lack numeral classifiers, and
since they use reduplication less productively and less frequently (cf. Rubino
2005) than many languages of China, a better understanding of the quantifica-
tion strategies in the languages of China will certainly enrich our comprehension
of human language and thought. In addition, the study of Chinese and the other
languages spoken in China will have its impact on the study of other language
families. The languages of China deserve more investigation, especially the non-
Han languages, because many of them remain unknown or poorly studied.
Plurality is “an area that has been rather neglected, especially in typological
work” (Dryer 1989: 888); and “A clearer understanding of the diachronic origin
(or origins) of plural words is necessary in order to better understand their
synchronic properties.” (Dryer 1989). This book tries to remedy the general lack
of information on grammatical quantification in the languages of China by pro-
viding papers by scholars who have been working in these fields for a long time
and have made important contributions to the study of the languages in China.
A book edited by Li Jinfang and Hu Suhua (2005) observes non-Han languages
of China and studies classifiers. In a recent book Liàng yǔ fùshù de yánjiū 量与
复数的研究 ([Quantification and plurality], Dan Xu [ed.], 2010) which has been
published as a companion to the present volume, seventeen authors (two papers
are written by co-authors) give their contributions to these problems. The
authors of these two books do not overlap, and the first one was written in
Chinese while the present one is in English. These two books aim to review
Chinese and some non-Han languages in depth to better understand the correla-
tion between plurality and classifiers.
2 Devices for marking quantity in Chinese2
Chinese is one of the languages that does not require plural marking in noun
phrases (see Feng Shengli and Harbsmeier in this volume for Old Chinese). Plu-
rality is expressed in the clause by numerous adverbs, or indicated in nominal
2 Studies on the origin and development of the plurality suffix -men in Chinese can be found in
the sister book Quantification and Plurality (2010) mentioned earlier.
4 Dan Xu](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-16-320.jpg)
![phrases by numerals, adjectives or noun modifiers (Paris in this volume). Plurality
in Chinese is not obligatorily expressed by a morpheme or several morphemes
as in Indo-European languages, but by abundant adjunct words. Thus, Chinese
seems to follow Greenberg’s (1972) universal even though the complementary
distribution of obligatory plural markers and numeral classifiers has been chal-
lenged by more recent data (see Bisang, this volume).
Chinese differs from morphologically rich languages, but might have used
morphology for expressing quantification in its early history. Eventually, plural
words (lexical items) became prominent. Though contemporary Chinese does
not have obligatory marking of every NP for plurality, it possesses some adverbs
to indicate plurality (see Lee in this volume), such as 都 dōu ‘all’, 皆 jiē ‘all’,
全 quán ‘all’, etc. The list was much longer in Middle Chinese: 都 dōu, 皆 jiē,
全 quán, 並 bìng, 盡 jìn, 具/俱 jù, 悉 xī, 咸 xián, 縂 zǒng, 擧 jǔ, etc. (see also
Meisterernst in this volume). It seems that the Chinese language prefers marking
plurality in the VP to marking it in the NP.
In previous studies on plural marking in Chinese, scholars have focused on
the plural marker men which is only applicable to noun phrases with the feature
[+human] and it is not obligatory especially in the object position. In Early
Old Chinese (11th–5th centuries BC), some words like 眾 zhòng, 群 qún, 諸 zhū,
marked plural meaning in the prenominal position. Only 余 yú was a post-
nominal plural word. These prenominal plural words became lexicalized later
(see also Meisterernst in this volume) and some postnominal plural words like
曹 cáo, 輩 bèi, 等 děng took over the function of plural marking. Due to its post-
nominal position, yú survived and continued to express plural meaning. There is
no major debate among scholars about the status of these plural words which
are not full-fledged plural markers. The real plural marker men evolved during
the Song Dynasty (960–1279 AD). From diachronic data, Dan Xu (2009a) suggests
that the emergence of postnominal plural marking is correlated with the develop-
ment of classifiers which moved from the postnominal position to the prenominal
position, while plural marking did the reverse (see also Jang-Ling Lin and Alain
Peyraube in this volume). In this book, authors will observe the relationship
between plural marking and classifiers from different angles to better under-
stand the problem. The investigation will not be limited to Chinese, but will
extend to other languages spoken in China in order to work with different per-
spectives on this topic.
Reduplication is a frequent phenomenon in Asian and African languages
(for example, Malay in Indonesia, Lepcha in India, Lomongo in the Congo,
Amharic in Ethiopia, Malagasy in Madagascar, Yoruba in Western Africa, etc.).
It can indicate a variety of concepts such as plurality, distributivity, iteration,
and intensity (cf. Rubino 2005 for a more exhaustive list). One of its meanings
Introduction: Plurality and Classifiers across languages of China 5](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-17-320.jpg)
![(8) 看看， 想想， 说说， 走走
kànkan, xiǎngxiang, shuōshuo, zǒuzou
‘to take a look, to think a bit, to say something, to take a walk’
In the above examples from contemporary Chinese, the reduplicated verbs
mean “to go to do something a little bit, to try”. The grammatical meaning
“unaccomplished, atelic, imperfective, short duration action” is in fact an inter-
pretation of the new status of the reduplicated part: it indicates the future.
Reduplicated verbs are not compatible with the past tense, adverbs implying the
past, or elements indicating accomplishment. An aspect marker [+accomplish] le
has to be inserted between the two parts if one wishes to express the past, a telic
action. Again, I think that some contemporary non-iconic meanings have to pro-
gressively lose their initial iconic signification. Now we will illustrate this point
of view with some diachronic examples:
(9) 看看似相识 (孟浩然 Mèng Hàorán, 689–740 AD)
kànkàn sì xiāngshí
see see like be acquainted
‘Seeing each other, it seems that we are acquainted.’
(10) 望望不见君 (李白 Lǐ Bái, 701–762 AD)
wàngwàng bú jiàn jūn
look look NEG see you
‘I am looking and looking but I cannot see you’
In these sentences, the reduplicated verbs intensify the continuance of the
action in repeating the same verb. The iconic meaning is clear. According to
Cheng Xiangqing (1992) and Liu Xiaonong’s separate descriptions (1992) and to
Wang Ying’s (1996: 234) studies, reduplicated verbs from before or during the
Tang never indicated diminutive meaning. All of them express high frequency,
augmented quantity, continuance, and so on. It is interesting to note that in
modern Yangzhou (located in Jiangsu Province) dialect, reduplicated verbs have
kept the iconic meaning indicating high frequency of the action (Zhu Jingsong
cited by Wang Ying 1996: 240), in contrast with Mandarin Chinese.
How did these iconically motivated reduplicated verbs lose their initial mean-
ing and become non-iconic, i.e. morphological devices, in Mandarin Chinese?
Again diachronic data and contemporary dialects provide us with comple-
mentary evidence. I believe that the rise and the merger of the patterns V1
(V2) kàn ‘see’ (try to do something) and Vyī ‘one’ V (do something once) caused
Introduction: Plurality and Classifiers across languages of China 7](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-19-320.jpg)
![the use of the plural marker does not depend on the semantic feature [+animate]
and all nouns can be marked by plural suffixes. However, the plural marker is
not always obligatory as in European languages. The Salar, Korean, and Manchu
languages employ classifiers due to influence from the Chinese language. Geo-
graphically they have been surrounded by Chinese languages for centuries and
language contact has been very intense. The evidence for this contact-induced
change is that classifiers are not yet obligatory in Salar, Korean, or Manchu,
and the word order of numeral and noun is not fixed. For these reasons, we
suppose that classifiers do not present a significant feature for this type of
language, even though some of them have begun to develop classifiers due to
long contact with Chinese languages.
It is clear that the Sinitic languages have VO order like type (b) and use
classifiers like (a) and (b). It is important to notice that in groups (a) and (b),
almost all languages (42/44) use reduplication devices and all of them employ
classifiers. This fact confirms our hypothesis that reduplication and classifiers
form one pattern with regard to plural marking. Additionally, in languages using
classifiers, plural markers are often distinct in that they are sensitive to the
feature [animate].
4 Organization of the book
In this volume, quantification will be observed and studied from different per-
spectives. The book studies several topics focusing on plurality. Based on the
present results of research on quantification, different chapters deal with con-
crete and targeted subjects such as relations between plurality and classifiers,
plurality and reduplication, expression of plurality in non-Han languages, and
plurality concepts in child language acquisition.
The papers are divided into three sections. The first one takes a typological
perspective and consists of two papers in which correlations between different
types of quantification are discussed. Bisang deals with the correlation between
the presence of numeral classifiers and the non-obligatoriness of plural mark-
ing as claimed by Greenberg (1972); Xu describes reduplication in its interaction
with plural marking and numeral classifiers. The second section deals with
numeral classifiers and their diachronic development in Mandarin Chinese
(Feng) and Early Southern Min (Lin and Peyraube). The expression of plurality
is the topic of the third section. It starts with the expression of plurality in the
history of Chinese, first in its semantic form in classical Chinese (Harbsmeier),
then in the form of the marker 諸 zhū from Han to Wei Jin Nanbeichao Chinese
(Meisterernst). Both of these papers are looking at the domain of nouns. This
Introduction: Plurality and Classifiers across languages of China 11](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-23-320.jpg)
![also applies to the next two papers on the use of 不同 bùtóng ‘different’ (Paris)
and the expression of plurality in the pronominal system of Bai (Fu). The paper
by Liu and Gu is dedicated to the expression of plurality in the domain of the
verb. Finally, Lee’s paper looks at the acquisition of quantification by three-
year-old children.
4.1 Correlations between different types of quantification
Chapter 1, “Numeral classifiers with plural marking. A challenge to Greenberg”
by Walter Bisang, starts out with data on Northern Kam and Weining Ahmao,
in which numeral classifiers obligatorily mark number. These findings are prob-
lematic to Greenberg’s (1974: 25) universal claim that “[n]umeral classifier lan-
guages generally do not have compulsory expression of nominal plurality, but
at most facultative expression.” Even though plural marking is not fully obliga-
tory in these languages, the fact that there are processes of grammaticalization
which support a classifier system that carries obligatory information on number
shows the weakness of Greenberg’s claim. After a discussion of the diachronic
language-internal processes that produced the classifier systems of Northern
Kam and Weining Ahmao, the paper will offer some general observations con-
cerning the status of implicational universals. It supports Newmeyer’s (2005)
view that typological generalizations must be seen as “holistic” in the sense
that there is only an indirect correlation between implicational universals and
rules, constraints and the properties of constructions in individual languages.
Chapter 2, “Reduplication in languages: A case study of languages of China”
by Dan Xu, deals with reduplication from a typological perspective. Reduplica-
tion is broadly attested in human languages, especially in the Southern Hemi-
sphere. In languages where reduplication and classifiers are found extensively,
plural marking is not well-developed, and is sensitive to the semantic feature
[+human]. The more plural marking is developed, such as in Indo-European
languages and Altaic languages, the less this semantic feature is required. The
more a language possesses developed plural-marking morphology, the less it
needs reduplication and classifiers. Reduplication is iconically motivated. Posi-
tive degree constitutes its core meaning. Reduplication with negative meaning
characterizes the grammaticalization process from icon to symbol.
4.2 Numeral classifiers and their diachronic development
Chapter 3, “The Syntax and Prosody of Classifiers in Classical Chinese,” is written
by Shengli Feng, who scrutinizes what mechanism might have triggered the rise of
12 Dan Xu](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-24-320.jpg)
![the syntactic and semantic constraints of the plural word zhū. The author indi-
cates that semantically the function of zhū is to refer to a well-defined plural set
of items, and that syntactically, zhū has to be analyzed as an adjective rather
than a determiner since it can combine with determiners such as demonstratives
and personal pronouns. Using the Animacy Hierarchy framework, Meisterernst
shows that personal pronouns, which are situated at the top of the Animacy
Hierarchy, are not compatible with zhū. Zhū underwent an important extension
from [+animate] [+human] to [–animate]. According to the data from the Wei Jin
Nanbeichao (220–589 AD) periods, and especially from the early 5th century
Buddhist text Gaoseng Faxian zhuan, the author concludes that pluralization
obviously was not the only function of zhū, at least during that period.
Chapter 7, “Bu-tong ‘different’ and nominal plurality in Mandarin Chinese”
by Marie-Claude Paris, studies a pair of two Mandarin (apparent) synonyms —
不同 bùtóng ‘different’ and 不一样 bù yíyàng ‘different’. Although their mean-
ing seems at first sight to be equivalent, bùtóng and bù yíyàng actually behave
very differently, both at the syntactic and at the semantic levels. While they can
be substituted for each other in some cases, in other cases they cannot; hence
their difference is established. Bù yíyàng can only be used as a scalar relative
predicate or as a predicate modifier, whereas bùtóng can function both as
an absolute predicate and as a noun determiner. This paper concludes that (i)
plurality and number constitute two different markings, whose distribution is
not complementary in Mandarin and that (ii) bùtóng can, in some contexts, be
labeled a ‘redundant’ plurality marker.
Chapter 8, “Plurality in the pronominal paradigms of Bai dialects” by Jingqi
Fu, investigates plurality in Bai, a Sino-Tibetan language whose linguistic affilia-
tion is debatable. Examining how plurality and case are expressed in the prono-
minal paradigms of Bai dialects, she proposes to derive the plural pronouns,
possessive and objective pronouns by a complex form of pronoun base followed
by an ending: plural morpheme, possessive and locative markers respectively,
contrary to the analysis where the inflected plural is independently derived
via vowel inflection. In doing so, she accounts for the different forms of plural
pronouns by the differences in the plural endings across dialects. She also inves-
tigates how plural formation interacts with case (objective, subjective and geni-
tive). This account is consistent with the general characteristic of Bai of being
a non-inflected language and puts Bai on a par with other Tibeto-Burman
languages which developed the objective case later.
Chapter 9, “Frequentative Aspect and Pluractionality in Nuosu Yi” by Hong-
yong Liu and Yang Gu, explores plurality’s relationship with frequentative
aspect. Nuosu Yi belongs to the Tibeto-Burman language group and possesses
three different means to encode a frequentative event: by using a preverbal tem-
14 Dan Xu](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-26-320.jpg)
![poral adverb, a postverbal frequentative aspect marker (like the West Green-
landic suffix “-tar”), or by both of them co-occurring in the same sentence. The
authors propose a unified way to explain these three different means. Following
Van Geenhoven (2004, 2005), the authors argue that the Nuosu Yi frequenta-
tive aspect marker is a pluractional operator, but the preverbal adverb is not.
Because of semantic overlap between the preverbal adverb and the frequenta-
tive aspect marker, the authors study in detail the semantic and syntactic con-
straints on the non/co-occurrence of these two markers in Nuosu Yi.
Chapter 10, “Quantificational structures in three-year-old Chinese-speaking
children” by Thomas Hun-tak Lee, investigates the realizations of existential
quantification, universal quantification, negation and modals in children’s lan-
guage. The experiments are performed on Mandarin-speaking and Cantonese-
speaking children from one and a half to three and a half years of age. Cognitive
structures relative to these quantifications are explored. The standard quantifica-
tional structures have been much expanded to describe restricted quantification,
binding of sets and predicates, questions and various kinds of modality. Based
on rich data on child language, the author makes some remarks on how the
logical operators are realized in Chinese, providing evidence for these operators.
Lee’s data show that existential quantification occurs earlier than universal
quantification.
This book is expected to have an impact on the study of linguistic typology,
language contact, and patterns of the evolution (grammaticalization, lexicaliza-
tion) of languages and linguistic areas. More and more linguists are realizing
that forming conclusions on patterns in general linguistics without a better
understanding of Chinese and languages of China is risky and dubious. This
volume may also provide useful material for cognitive sciences and anthropol-
ogy. Typological studies require a solid description of Chinese and the languages
of China. Investigation of quantification in diachronic Chinese and non-Han lan-
guages will offer general linguists a schema complementary to other languages of
the world.
References
Andler, Daniel. 2004. Reprint. Introduction aux sciences cognitives [Introduction to cognitive
sciences]. Paris: Editions Gallimard, 1992.
Behr, Wolfgang. 2006. “Homosomatic juxtaposition” and the problem of “syssemantic” (huì yì )
characters. In Ecriture chinoise. Données, usages et représentations [Chinese writing,
data, use and representations], Françoise Bottéro and Redouane Djamouri (eds.), 75–114.
Paris: EHESS/CRLAO.
Bisang, Walter. 1993. Classifiers, Quantifiers and Class Nouns in Hmong. Studies in Languages
17: 1–51.
Introduction: Plurality and Classifiers across languages of China 15](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-27-320.jpg)
![Borer, Hagit. 2005. In Name Only. Oxford: Oxford University Press.
Bybee, Joan, Perkins Revere and Pagliuca William. 1994. The Evolution of Grammar. Chicago/
London: University of Chicago Press.
Cai, Jinghao. 1990. Chongtan yuzhuci ‘kan’ de qiyuan [Again on the origin of the particle kan],
Zhongguo Yuwen 1: 75–76.
Cao, Zhiyun (eds.). 2008. Hanyu fangyan ditu ji [Chinese dialects Atlas]. Beijing: Commercial
Press.
Carlson, Robert. 1991. Grammaticalization of Postpositions and Word Order in Senufo
Languages. In Approaches to Grammaticalization, Elizabeth Closs Traugott and Bernd
Heine (eds.), vol. II: 201–223. Amsterdam/Philadelphia: John Benjamins Publishing
Company.
Chao, Yuen Ren. 1968. A Grammar of Spoken Chinese. Berkeley/Los Angeles: University of
California Press.
Chen, Chusheng. 2004. Jinwen changyong zidian [Dictionary of inscriptions on bronze in
common use]. Xi’an: Shanxi renmin chubanshe.
Cheng, Xiangqing. 1992. Bianwen fuyinci yanjiu [dissyllabic words in the Bianwen]. In Sui Tang
Wudai hanyu yanjiu [Studies of Chinese in the Sui, Tang and Wudai], Cheng Xiangqing
(ed.), 1–132. Jinan: Shandong Jiaoyu chubanshe.
Clark, Herbert H. 1973. Space, Time, Semantics and the Child. In Cognitive Development and
the Acquisition of Language, Timothy E. Moore (ed.), 27–63. New York/London: Academic
Press.
Comrie, Bernard. 2008. The Areal Typology of Chinese: Between North and Southeast Asia. In
Chinese Linguistics in Leipzig CLE 2, Redouane Djamouri, Barbara Meisterernst and Rint
Sybesma (eds.), 1–21. Paris: EHESS/CRLAO.
Corbett, Greville. 2000. Number. Cambridge: Cambridge University Press.
Croft, William. 1990. Typology and Universals. Cambridge: Cambridge University Press.
Dai, Qingxia. 1990. Yimianyu de jiegou zhuci [The grammatical markers in Tibeto-Burman
languages]. In Zangmian yuzu yuyan yanjiu [Research on Tibeto-Burman languages],
83–97. Beijing: Minzu chubanshe.
Downing, Pamela. 1996. Numeral Classifier Systems: The case of Japanese. Amsterdam/
Philadelphia: John Benjamins Publishing Company.
Dryer, Matthew. 1989. Plural words. Linguistics 27: 865–895.
Dryer, Matthew. 1992. The Greenbergian word order correlations. Language 68 (1): 81–138.
Dryer, Matthew. 2003. Word order in Sino-Tibetan languages from a typological and
geographical perspective. In The Sino-Tibetan languages, Graham Thurgood and
Randy J. LaPolla (eds), 43–55. London/New York: Routledge.
Enfield, Nick J. 2005. Areal Linguistics and Mainland Southeast Asia. Annual Review of
Anthropology 34: 181–206.
Feng, Chuntian. 1991. Jindai hanyu yufa wenti yanjiu [Grammatical studies in Modern Chinese].
Jinan: Shandong jiaoyu chubanshe.
Greenberg, Joseph. 1972. Numeral classifiers and substantival number: problems in the
genesis of a linguistic type. Language Universals 9: 1–39.
Greenberg, Joseph. 1974. Numeral classifiers and substantival number: problems in the genesis
of a linguistic type. In Proceedings of the 11th International Congress of Linguistics,
Bologna – Florence, Aug–Sept 1972, Bologna 1974: 17–37. (Reprinted in Greenberg 1990:
16–93).
16 Dan Xu](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-28-320.jpg)
![Greenberg, Joseph. 1988. The first person inclusive dual as an ambiguous category. Studies in
Language 12: 1–18.
Greenberg, Joseph. 1990. On language. In Selected writings of Joseph H. Greenberg, Denning,
K. and S. Kemmer (eds.). Stanford: Stanford University Press.
Guo, Pan. 2004. Hanyu sheshu wenti yanjiu [Studies on numbers in Chinese]. Beijing:
Zhonghua shuju.
Guo, Wengao. 1995. ‘Yu’ qianzhi yu shuci de yongfa. [‘Yu’ used before numbers]. Zhongguo
Yuwen 5: 367–368.
Hagège, Claude. 2010. Adposition. Oxford: Oxford University Press.
Haiman, John. 1980. The Iconicity of grammar: Isomorphism and motivation. Language 56 (3):
515–540.
Haiman, John (ed.). 1985. Iconicity in Syntax. Amsterdam/Philadelphia: John Benjamins
Publishing Company.
Heide, Wegener. 1995. The German plural and its acquisition in the light of markedness theory.
In The Development of Morphology Systematicity. A Cross-Linguistic Perspective, Hanna
Pishwa and Karl Maroldt (eds), 247–261. Tübingen: Gunter Narr.
Heine, Bernd and Tania Kuteva. 2002. World lexicon of grammaticalization. Cambridge:
University of Cambridge.
Heine, Bernd, Ulrike Claudi and Friederike Hünnemeyer. 1991. From Cognition to Grammar-
Evidence from African Languages. In Approaches to grammaticalization, Elizabeth Closs
Traugott and Bernd Heine (eds.), Volume 1: 149–187. Amsterdam/Philadelphia: John
Benjamins Publishing Company.
Hopper, Paul J. and Sandra A. Thompson. 1980. Transitivity in grammar and discourse.
Language 56: 251–299.
Iljic, Robert. 2001. The problem of the suffix-men in Chinese grammar. Journal of Chinese
Linguistics 29 (1): 11–68.
Jiang, Lansheng. 1995. Shuo “mo” yu “men” tongyuan [“Mo” and “men” have the same origin]
Zhongguo Yuwen 3: 180–190.
Li, Jinfang and Suhua, Hu. 2005. Han Zang yuxi liangci yanjiu [Studies on classifiers in Sino-
Tibetan]. Beijing: Zhongyang Minzu daxue chubanshe.
Li, Yanhui and Yuzhi, Shi. 2000. Hanyu liangci xitong de jianli yu fushu biaoji “men” de fazhan
[The establishment of the classifier system and the development of the plural marker
“men”]. Dangdai yuyanxue 2: 27–36.
Liu, Xiaonong. 1992. Tang Wudai ci dieyin chutan [Preliminary study of reduplicated sounds in
Tang and Wudai]. In Sui Tang Wudai hanyu yanjiu [Studies of Chinese in Sui, Tang and
Wudai], Cheng Xiangqing (ed.), 440–471. Jinan: Shandong Jiaoyu chubanshe.
Lü, Shuxiang and Jiang Lansheng. 1985. Jindai hanyu zhidaici [Demonstratives and pronouns in
Modern Chinese]. Shanghai: Xuelin chubanshe.
Lü, Shuxiang. 1984. Hanyu yufa lunwenji [Collected papers on Chinese grammar]. Beijing:
Commercial Press.
Ma, Guoquan. 1979. Liang Zhou tongqi mingwen shuci liangci chutan [A preliminary study on
numerals and classifiers in bronze inscriptions during the Eastern and Western Zhou].
Gu wenzi yanjiu 1: 126–136.
Mei, Tsu-lin. 1986. Guanyu jindai hanyu zhidaici [About demonstratives and pronouns in
Modern Chinese]. Zhongguo Yuwen 6: 401–412.
Montaut, Annie. 2009. Réduplication et constructions en écho en hindi/ourdou [Reduplication
and echo constructions in Hindi/Urdu]. Faits de Langue/Les Cahiers 1: 9–44.
Introduction: Plurality and Classifiers across languages of China 17](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-29-320.jpg)
![Newmeyer, Frederick J. 2005. Possible and probable languages: A generative perspective on
linguistic typology. Oxford: Oxford University Press.
Ortmann, Albert. 2000. Where plural refuses to agree: Feature unification and morphological
economy. Acta Linguistica Hungarica, Volume 47: 249–288.
Peyraube, Alain. 1998. On the History of Classifiers in Archaic and Medieval Chinese. In Studia
Linguistica Serica, Benjamin K. Tsou (ed.), 39–68. Hong Kong: City University of Hong Kong.
Peyraube, Alain and Wiebusch Thekla. 1993. Le rôle des classificateurs nominaux en chinois et
leur évolution historique: un cas de changement cyclique [The role of nominal classifiers
in Chinese and their historical development: a case of cyclical change]. Faits de langues
2: 47–57.
Rubino, Carl. 2005. Order of subject, object and verb. In The world atlas of language structures,
Martin Haspelmath, Matthew S. Dryer, David Gil and Bernard Comrie (eds.). Oxford:
Oxford University Press.
Sagart, Laurent. 1999. The Roots of Old Chinese. Amsterdam/Philadelphia: John Benjamins
Publishing Company.
Sanches, Mary. 1973. Numeral classifiers and plural marking : an implicational universal.
Language Universals 11: 1–22.
Singh, Rajendra. 2005. Reduplication in Modern Hindi and the theory of reduplication. In
Grammar, Comparative and general–Reduplication, Bernard Hurch (ed.), 263–281. Berlin/
New York: Mouton de Gruyter.
Sun, Hongkai. 2001. Lun Zangmian yuzu zhong de qiangyuzhi yuyan [On the Qiangic branch of
the Tibeto-Burman languages]. Languages and linguistics 2: 1, 157–181.
Sun, Hongkai, Huang Xing and Hu Zengyi (eds). 2007. Zhongguo de yuyan [Languages in
China]. Beijing: Commercial Press.
Svorou, Soteria. 1993. The Grammar of Space. Amsterdam/Philadelphia: John Benjamins
Publishing Company.
Tsou, Benjamin. 1976. The structure of nominal classifier systems. In Austroasiatic Studies,
Philip N. Jenner, Laurence C. Thompson, and Stanley Starosta (eds.), Part II: 1215–1247.
Honolulu: University of Hawaii Press.
Talmy, Leonard. 2000a. How Language Structures Space. Toward a Cognitive Semantics, Volume
1: 177–254. Cambridge/Massachusetts: Massachusetts Institute of Technology Press.
Talmy, Leonard. 2000b. A Typology of Event Intergration. Toward a Cognitive Semantics, Volume
2: 213–288. Cambridge/Massachusetts: Massachusetts Institute of Technology Press.
Van Geenhoven, Veerle. 2004. For-adverbials, frequentative aspect, and pluractionality. Natural
Language Semantics 12: 135–190.
Van Geenhoven, Veerle. 2005. Atelicity, pluractionality and adverbial quantification. In
Perspectives on Aspect, Henk Verkuyl, Henriëtte de Swart and Angeliek van Hout (eds.),
107–124. Netherlands: Springer.
Wang, Ying. 1996. Tangshi zhong de dongci chongdie [Reduplicated verbs in Tang poetry],
Zhongguo Yuwen 3: 233–234, 240.
Wang, Huayun. (preprint). Fushu ciwei “men” de laiyuan [Origin of the plural suffix -men].
Wu, Fuxiang. 1995. Changshitai zhuci “kan” de lishi kaocha [Diachronic examination of the
tentative aspect particle “kan”]. Yuyan yanjiu 2: 161–166.
Wu, Fuxiang. 2006. Wei Jin Nanbeichao shiqi hanyu mingliangci fanchou de yufahua chengdu
[Degree of grammaticalization of nominal classifiers during the Wei Jin Nanbeichao
periods]. In Linguistic Studies in Chinese and Neighboring Languages: Festschrift in Honor
of Professor Pang-Hsin Ting on His 70th Birthday. Language and Linguistics Monograph
Series Number W-6: 553–571.
18 Dan Xu](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-30-320.jpg)
![Xing, Fuyi. 1960. Lun “men” he “zhuwei” zhilei bingyong [On the coexistence of “men” he
“zhuwei”]. Zhongguo Yuwen 6: 289, 292.
Xu, Dan. 1994. Guanyu hanyu li dongci + X + didianci de juxing [On the V + X + locative pattern in
Chinese]. Zhongguo Yuwen 3: 180–185.
Xu, Dan. 2006. Typological change in Chinese syntax. Oxford: Oxford University Press.
Xu, Dan. 2009a. Cong Zhanguo zonghengjia shu kan Xi Han chuqi fushu gainian de biaoda
[Plurality in Chinese during the early Han as seen in the Zhanguo zonghengjia shu], Lishi
yuyanxue yanjiu, Volume 2. Beijing: Commercial Press.
Xu, Dan. 2009b. Xiayou Shixingyu de mouxie tedian [Some syntactic features of Down-river
Shixing]. Minzu Yuwen, 1: 25–42.
Xu, Dan. 2010. Cong yuyan leixing kan hanyu fushu xingshi de fazhan [The development of
plurality from a typological perspective]. In: Dan Xu (ed.), Liang yu fushu de yanjiu
[Quantification and plurality], 90–112. Beijing: Shangwu yinshuguan.
Xu, Dan (ed.). 2008. Space in Languages of China: Cross-linguistic, synchronic and diachronic
perspectives. Dordrecht: Springer.
Xu, Dan (ed.). 2010. Liang yu fushu de yanjiu [Quantification and plurality]. Beijing: Shangwu
yinshuguan.
Xu, Dan, and Fu Jingqi. 2012. Classifiers and some typological considerations. To appear in
Breaking down the barriers: Interdisciplinary studies in Chinese linguistics and beyond.
(Language and Linguistics Monograph Series) Taipei: Institute of Linguistics, Academia
Sinica.
Xu, Zhongshu (ed.). 1989. Jiaguwen zidian [Dictionary of bone inscriptions]. Chengdu: Sichuan
cishu chubanshe.
Zhang, Meilan. 2001. Jindai Hanyu houzhui xingrongci [Suffixed adjectives in Modern Chinese].
Guiyang: Guizhou jiaoyu chubanshe.
Zhao, Jie. 1989. Xiandai Manyu yanjiu [Study on Contemporary Mandchou], Beijing: Minzu
chubanshe.
Introduction: Plurality and Classifiers across languages of China 19](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-31-320.jpg)
![In addition to these noun classes, Ejagham has five numeral classifiers which
are derived from nouns and have their own noun-class prefixes. One of them is
the classifier ɴ̀-sûm (singular)/à-sûm (plural) of noun class 3/6 for “any fruit or
root which is long” (Watters 1981: 310). As can be seen from (3) on the overall
structure of the noun phrase, the numeral classifier takes the initial position of
the noun phrase:
(3) CL GEN-LINK N NUM
The genitive linker (GEN-LINK) between the classifier and the noun is con-
trolled by the noun class of the classifier (CL). The numeral, which is situated
at the end of the construction1, also agrees with the class of the classifier. In
the following example, the noun -tə̀bɛ̀ ‘okra pod’, which belongs to noun class
9/14 (singular: ɴ̀-tə̀bɛ̀; plural: ɔ̀-tə̀bɛ Watters 1981: 300) is presented in com-
bination with the numerals -d ‘one’ and -bá1ɛ́ ‘two’. As can be seen, the head
noun and the classifier are both marked for number:
(4) The classifier Ǹ-sûm (sg.)/à-sûm in Ejagham
a. ɴ̀-sûm ì ɴ̀-tə̀bɛ́ yə́-d
NCL3:SG-CL:fruit GEN-LINK NCL9:SG-okra.pod NCL3:SG-one
‘one okra pod’
b. à-sûm ´ ɔ́-tə̀bɛ́ á-bá1ɛ́
NCL6:PL-CL:fruit GEN-LINK NCL14:PL-okra.pod NCL6:PL-two
‘two okra pods’ (Watters 1981: 310)
The numeral classifier system of Ejagham is not pervasive. As Watters (1981:
313) points out, “most nouns do not use a classifier when being enumerated, but
for the various types of nouns listed above, e.g. seeds, grains, kernels, nuts,
long fruits and roots, round fruits and roots, clusters of fruits, plants, trees and
vegetables a classifier is generally required”.
1 The structure of the classifier construction in Ejagham additionally contradicts Greenberg’s
(1974: 31) claim that the classifier and the numeral must form a continous constituent. As can
be seen from (3) and (4), word order in the classifier construction of Ejagham is [CL-N-NUM].
While classifiers form a coherent unit with the numeral in East and mainland Southeast Asian
languages, they establish a more coherent relation with the noun to be counted in Ejagham
and other Niger-Congo languages. In Ejagham, this can be seen from the genitive-linker, which
agrees with the classifier and creates a head-modifier relation between the classifier and the
classified noun. The coherence between the numeral and the classifier is maintained by the
agreement relation as it is controlled by the class-membership of the classifier.
Numeral classifiers with plural marking. A challenge to Greenberg 25](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-37-320.jpg)
![There are other Niger-Congo languages in which the numeral-classifier sys-
tem is pervasive, i.e., each count noun needs to have a classifier in the context
of enumeration. One language of this type is Kana, another Cross River language
spoken in the South of Nigeria, which has lost the Bantu noun class system
(Ikoro 1994, 1996):2
(5) Kana
a. bàɛ̀ bēè ŋwíí
two CL:fruit child
‘two children’
b. tāà té ítòbtòb
three CL:tree motorcycle
‘three motorcycles’ (Ikoro 1996: 91–92)
Even if the use of classifiers is limited to certain semantic domains in Ejagham,
the combination of obligatory number marking and numeral classifiers remains
a challenge to Greenberg’s (1974) claim. This challenge may be mitigated to a
certain extent if it turns out that fully pervasive systems of numeral classifiers
are limited to those Niger-Congo languages in which the obligatory opposition
of singular vs. plural reflected in the noun-class system has vanished (see
Section 4).
In spite of a considerable amount of counter-evidence against Greenberg’s
implicational claim from various languages and areas across the globe, there
hasn’t been much counter-evidence from the area of East and mainland South-
east Asia. More recent research on Northern Kam (Gerner 2006) and on Weining
Ahmao (Gerner and Bisang 2008, 2009, 2010) shows that there is evidence
against this universal claim even in the isolating languages of that area. In
fact, the few instances of morphology that developed in these two languages
are concerned with the morphophonological form of the classifier and – what is
even more remarkable – with the emergence of singular/plural-distinctions that
are marked on the classifier itself. As will be shown in the next section, classifier
morphology is the result of fusion with other words or markers which lose their
syntactic independence. Such processes are not limited to the development of
classifier inflection, they are also attested in other domains of grammar. One of
2 As can be seen from (5), Kana does not violate the word-order constraints claimed by
Greenberg (1974: 31) (cf. endnote 1). Its word order is [NUM-CL-N]. In spite of this, the fusion
between CL and N is higher than between NUM and CL. There is a genitive linker in the form
of a floating low tone which takes the final syllable of the head of the construction, i.e., the
classifier (CL) followed by the noun (N) as its dependent. Thus, the form bēè ‘CL:fruit’ in (5a) is
a combination of [bēē plus low tone]. In the case of té ‘CL:tree’ in (5b), the low tone does not
show any effects.
26 Walter Bisang](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-38-320.jpg)
![them is discussed by Fu (in this volume) on the integration of plural markers
into pronominal paradigms of Bai. It is needless to say that the integration of
plural marking into the morphophonology of classifiers or pronominal systems
considerably weakens Greenberg’s universal claim.
The paper is organized as follows: Section 2 presents the classifier systems
of Northern Kam (see Subsection 2.1) and Weining Ahmao (see Subsection 2.2).
Section 3 describes the factors that operate against Greenberg’s (1974) universal
claim. It will turn out that none of these factors are associated with cognitive
motivations that are generally mentioned in linguistic typology. A short con-
clusion in section 4 will take up the discussion of the validity of typological
generalizations where there are various factors that potentially undermine them.
2 The data from Northern Kam and Weining
Ahmao
2.1 Number-sensitive numeral classifiers in Northern Kam
Northern Kam belongs to the Kam-Tai branch of the Kadai language family.3
Like most of the other East and mainland Southeast Asian languages, Northern
Kam is an isolating language. There is only one exception, and that exception
shows up in the numeral-classifier system. Since this is nicely described by
Gerner (2006), I will only summarise his findings in the present subsection.
Numeral classifiers in Northern Kam show a regular morphophonological
contrast between singular and plural forms. The initial consonant of the singular
form is almost always a glide ([w], [j]) or a voiced fricative ([ʑ], [ɤ]), while the
plural form of classifiers starts with a voiceless stop or a nasal. This distinction
is made only in Northern Kam. The numeral classifier system of Southern Kam
does not have such a distinction. Since the single classifier form of Southern
3 There are two different classifications of this language family. The older one is from
Edmondson and Solnit (1988) and distinguishes three subfamilies, i.e., Kam-Tai, Hlai (= Li) and
Geyang. In this classification, which is also called the “Kadai hypothesis”, Kam-Tai is further
divided into Kam-Sui (to which Kam belongs) and Tai. A more recent classification by
Chamberlain (1997), which is sometimes dubbed “Tai-Kadai hypothesis”, splits Kam-Tai into
two independent subbranches and thus distinguishes four subfamilies, i.e., Tai, Kam-Sui, Hlai
and Kadai. This classification is known under the heading of “Zhuang-Dong” in the Chinese
context. Since Gerner’s (2006) comparative analysis of core classifiers in 22 Kam-Tai languages
provides supportive evidence for the Kadai hypothesis of Edmondson and Solnit (1988),
I adopt this classification. My above description of the genetic situation is from Gerner
(2006: 239).
Numeral classifiers with plural marking. A challenge to Greenberg 27](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-39-320.jpg)
![Kam systematically corresponds to the plural form of Northern Kam, the singu-
lar form of Northern Kam must be a special, historically derived form which only
developed in that language (Gerner 2006: 244–247). The derivational rules are
described in detail in Gerner (2006). For that reason, only a few examples will
be presented in this paper:
(6) Singular vs. plural form of Northern Kam (Gerner 2006: 244–245)4
Derivation Class meaning CL:SG CL:PL
a. [w] <– [p] human wəu45 pəu45
b. [w] <– [m] dual body parts waŋ55 maŋ55
c. [w] <– [m] clothes wəi31 məi31
d. [j] <– [ɕ] entities with handle jaŋ45 ɕa ŋ45
e. [j] <– [t
˛] 1-dim entities jiu22 t
˛iu22
f. [ʑ] <– [n] 3-dim entities ʑən45 nən45
g. [ɣ] <– [ʔ] small diverse range ɣa55 ʔa55
As is to be expected from the above description, the singular form of the
classifier is used with the numeral i45 ‘one’ in (7a). For numbers greater than
one, the plural form is employed in (7b):
(7) Northern Kam
(Gerner 2006: 243–244; for the form of the classifier, cf. [5e])
a. i45 jiu22 na45
one CL:SG river
‘one river’
b. ham11 t
˛iu22 na45
three CL:PL river
‘three rivers’
Numeral classifiers also occur in [classifier + noun] constructions. In this
construction, they mark indefiniteness. The singular form stands for singular
indefinite (8a) and (9a), while the plural form expresses plural indefinite in (8b)
and (9b):
4 The tones are represented by number combinations. The numbers represent relative pitch on
a scale from 1 (lowest) to 5 (highest). The first number represents the beginning of the tonal
contour, the second number stands for its end.
28 Walter Bisang](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-40-320.jpg)
![(8) Northern Kam
(Gerner 2006: 249; for the form of the classifier, cf. [6c])
a. wəi31 tu33 b. məi31 tu33
CL:SG garment CL:PL garment
‘a garment’ ‘garments’
(9) Northern Kam
(Gerner 2006: 249; for the form of the classifier, cf. [6g])
a. ɣa55 hoŋ22 b. ʔa55 hoŋ22
CL:SG loom CL:PL loom
‘a loom’ ‘looms’
In other grammatical contexts, only the plural form of the classifier is used.
In the case of the demonstrative construction, the classifier marks singular, even
though the plural form is used:
(10) Northern Kam
(Gerner 2006: 251; for the form of the classifier, cf. [6b])
maŋ55 ta45 ai33
CL:PL eye DEM:this
‘this eye’
The situation as it has been described so far is remarkable in at least two
respects. First, it is typologically rare that the plural form provides the basic
form from which the singular is derived. Of course, singulative forms are
attested in a considerable number of languages but this is usually limited to a
subset of nouns. In Northern Kam, the derivation of the singular from the plural
is pervasive. Second, the form which explicitly marks plural in the [numeral +
classifier]-construction (7b) and in the [classifier + noun]-construction ([8b] and
[9b]) marks singular in the demonstrative construction (10). Both of these prop-
erties can be accounted for if one looks at the historical development of the
number distinction in Northern Kam classifiers.
Based on Shi (1997), Gerner (2006: 244) convincingly argues that the singular/
plural-distinction of numeral classifiers developed “when the ancestor classifiers
collocated with the numeral i45 ‘one’”. The new classifier forms that emerged from
the collocation with that numeral were reanalysed as singular forms in contrast to
the original form which collocated with all other numerals and did not undergo
any morphophonological change. Since this unchanged form occurred with all
numerals higher than one, it was only natural to associate this form with plural
Numeral classifiers with plural marking. A challenge to Greenberg 29](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-41-320.jpg)
![meaning. Thus, the typologically remarkable situation that the singular is
derived from the plural synchronically is due to a morphophonological process
of change. Once the distinction of singular vs. plural was established with
numerals, it moved into the [classifier + noun]-construction with its indefinite
meaning. This scenario is quite plausible if one takes into account that the con-
struction [one + classifier + noun] is used for expressing indefiniteness in a vast
number of languages (Dryer 2005). As soon as the [singular classifier + noun]-
construction was associated with indefiniteness as a short form of the [one +
classifier + noun]-construction, the ancient form of the classifier was reanalysed
as a plural indefinite marker in the [plural classifier+ noun]-construction.
As can be seen from example (10), the number distinction is not fully gram-
maticalized through all constructions which take a numeral classifier. In these
constructions, the plural classifier is used even though it does not necessarily
mean plural. If one takes the plural form as the original form of the classifier it
does not come as a surprise that this form occurs in all grammatical contexts in
which the singular/plural-distinction remains irrelevant. As I will argue in the
remainder of this subsection, the fact that it actually marks singular in the
demonstrative construction can be accounted for by a look at Thai as another
Kam-Tai language. In Thai, numeral classifiers are optional in the demonstrative
construction. If they occur with demonstratives there is a very strong tendency to
interpret them as singulative markers (Hundius and Kölver 1983, Becker 2005).
Thus, (11a) with no classifier can be singular or plural, while (11b) with a classifier
is most likely to be singular:
(11) Thai
a. rót níi b. rót khan níi
car this car CL this
‘this car/these cars’ ‘this car’
If one assumes that the situation in Northern Kam was similar to that in
Thai before the emergence of number-sensitive classifiers, the occurence of a
classifier in the demonstrative construction also triggered a singular interpreta-
tion in Northern Kam. Since the singular/plural-distinction did not go beyond
the numeral construction and the [classifier + noun]-construction, the original
form is still the only option for the demonstrative construction. The seemingly
paradoxical case that the form used with plural function in some constructions
triggers a singular interpretation in the demonstrative construction can thus be
resolved by the argument that there is no other form available for the demon-
strative construction and that that form triggers singular interpretation as it did
before the singular/plural-distinction took shape.
30 Walter Bisang](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-42-320.jpg)
![2.2 Classifier inflection in Weining Ahmao
Weining Ahmao is a Miao language spoken in the Weining County of Western
Guizhou. Its classifier system is described since Wang Fushi (1957 [1972])5 and
Wang Deguang (1986, 1987) and it seems to be unparalleled in the world’s lan-
guages (for more data and the analysis of the inflectional paradigm, see Gerner
and Bisang 2008, 2010). Each of its classifiers is inflected for the categories of
number (singular vs. plural), reference (definite vs. indefinite) and size (augmen-
tative, medial, small). The forms concerning size also express social deixis in
dialogues (not in narratives). Roughly speaking, male speakers use the augmen-
tative form of the classifier, female speakers the medial form and children the
diminutive form (for more details, see Gerner and Bisang 2008, 2010). The
following table shows how the form of Ahmao classifiers varies within the
twelve-slot paradigm determined by these three categories:
Table 1: The inflectional paradigm of classifiers in Weining Ahmao (Gerner and Bisang 2008:
721; Gerner and Bisang 2010: 79)
Gender/Age
Register Size
Singular Plural
Definite Indefinite Definite Indefinite
Male Augmentative CVT C*VT ti55 a11 CVT’ di31 a11 C*VT’
Female Medial Cai55 C*ai213 tiai55 a11 CVT’ diai213 a11 C*VT’
Children Diminutive Ca53 C*a35 tia55 a11 CVT’ dia55 a11 C*VT’
The definite singular augmentative form with its structure of CVT is the
basic form of the classifier. It consists of a consonant (C), a vowel nucleus (V)
and a lexically determined tone (T)6. Each of these elements can be subject to
morphophonological change in the other slots. The symbol C* stands for
changes in the properties of the consonant in some classifiers. Depending on
the individual classifier, the properties involved are voicing and aspiration to
distinguish indefinite from definite classifiers. The vowel quality [ai]7 is charac-
teristic of medial size, while [a] marks diminutive size. Medial-definite forms are
typically expressed with [55] tone, medial-indefinite forms exhibit a [213] tone.
Diminutive-definite forms display the tones [53] (singular) or [55] (plural) and
5 The description of Wang Fushi (1957) slightly differs from the one presented in Gerner and
Bisang (2008). Wang Fushi (1957) does not mention the distinction between definite and
indefinite augmentative/male forms in the singular (cf. Tables 1 and 2).
6 On the notation of tones by the combination of numbers, cf. fn. 2.
7 Wang Fushi (1957) has [ae] instead of [ai].
Numeral classifiers with plural marking. A challenge to Greenberg 31](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-43-320.jpg)
![finally diminutive-indefinite forms are associated with the tones [35] (singular)
and [55] (plural) (Gerner and Bisang 2008: 722). The plural forms are characterized
by the marker ti55, which undergoes similar changes as the singular classifier
forms described in table 1. This marker is followed by a11 plus the classifier in
its form of C*VT’, in which T’ indicates that the tone takes a special form at least
with some classifiers.
There are some 40 to 50 numeral classifiers in Weining Ahmao. To illustrate
how the pattern described above works with a concrete classifier, table 2 shows
the inflectional forms of the classifier lu55, one of the most common classifiers
which is used with nouns denoting inanimates.
Table 2: The inflectional paradigm of the classifier lu55 in Weining Ahmao (Gerner and Bisang
2008: 722)
Gender/Age
Register Size
Singular Plural
Definite Indefinite Definite Indefinite
Male Augmentative lu55 lu33 ti55 a11 lu55 di31 a11 lu55
Female Medial lai55 lai213 tiai55 a11 lu55 diai213 a11 lu55
Children Diminutive la53 la35 tia55 a11 lu55 dia55 a11 lu55
The following example shows how the different classifier forms are used in
a text. The text in (12) is a Weining Ahmao version of the well-know fable of the
fox and the crow. In the first line, the classifier tu44 for animals occurs in its
indefinite (medial singular) form, since the noun a33 dy33 ‘fox’ is introduced
into the text8. Later, it is taken up in the corresponding definite form tae33. The
noun nG‘ae35 ‘meat’, which stands for the cheese in the western version of the
fable, is already known from previous context and is thus marked by the definite
(ts‘ae33: singular, medial) form of the classifier. The crow is also marked by the
definite (ta33: singular, diminutive) form of the classifier because it has been
introduced earlier. Since classifiers also occur in possessive constructions, we
find the classifier lu55 in its definite (singular, diminutive) form la55 on the last
line.9 Possessive constructions have the form [possessor CLi possesseei], in
8 In narratives, the selection of size depends on the real size of the objects described. In (12),
the fox is marked by the medial-size form, not by the augmentative form. In principle, the
selection of the size form depends on the view of the speaker. If she wants to point out the big
size of a given nominal concept she can chose the augmentative form. Since foxes are not
extremely big, the use of the medial form corresponds to the default. Once the size of a given
element of a narrative is determined, this feature contributes to reference tracking.
9 Wang Fushi (1957) reports a different tone for this classifier form. As can be seen from table
2, the medial definite singular form of lu55 is la53 in Gerner and Bisang (2008) instead.
32 Walter Bisang](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-44-320.jpg)
![Apart from its inflectional classifiers, Weining Ahmao is basically an isolat-
ing language as most other East and mainland Southeast Asian languages. This
begs the question of how the inflectional classifier system emerged. Gerner and
Bisang (2009: 19–29; 2010: 86–90) suggest a two-stage scenario. The first stage
induced the development of the three size distinction. It is charachterized by a
rebracketing process followed by further processes of erosion. The second stage,
which was responsible for the distinctions of definite vs. indefinite, started out
from the combination of the numeral i55 ‘one’ with a classifier.
The first stage is based on derivational markers which are used to form aug-
mentatives and diminutives. In the case of Weining Ahmao, these markers are
a55 ɲie53 (augmentatives) and ŋa11 (diminutives). The augmentative marker is
related to the noun ɲie53 ‘mother’, while the diminutive marker is still present
in the noun ŋa11 ʑau11 ‘child’. Both markers can take the initial position of
nominal compounds. If these markers are combined with nouns for animals,
their initial meaning is still visible – the augmentative marker produces a noun
referring to the female gender of animals, while the diminutive marker derives a
noun denoting the juvenile members in adult-young animal pairs (14a). If com-
bined with inanimate nouns the two markers express a vague idea of largeness
(physically or metaphorically) and smallness, respectively (cf. example [14b]):
(14) Weining Ahmao (Gerner and Bisang 2010: 87)
a. mpa44 ‘pig, hog’ → a55 ɲie53 mpa44 ‘sow
[augmentative, female]’
→ ŋa11 mpa44 ‘piglet
[diminutive, child]’
b. tɕa44 ‘wind’ → a55 ɲie53 tɕa44 ‘storm’
[augmentative]’
→ ŋa11 tɕa44 ‘breze of wind
[diminutive]’
The development of different classifier forms for size started out with a pro-
cess of reanalysis in which the augmentative and diminutive markers no longer
formed a unit with the noun that followed them but were rebracketed into a new
unit consisting of the classifier plus the augmentative/diminutive marker:
(15) NUM CL [a55 ɲie53/ŋa11–NOUN] → NUM [CL–a55 ɲie53/ŋa11] NOUN
After this change of word boundaries, the sequences [CL–a55 ɲie53] and [CL–
ŋa11] were further reduced through various stages (loss of the nasals ɲ and ŋ,
etc.) into [ai] and [a], respectively. The association of the initial augmentative
34 Walter Bisang](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-46-320.jpg)
![form in [ai] as a marker of medial size has to do with the use of these markers in
the social setting between men, women and children (for details, cf. Gerner and
Bisang 2009: 13–18). In this context, males used the original classifiers, while
the forms in [ai] and [a] basically remained for women and children, respec-
tively. Through this process, the size values of the various forms were redistrib-
uted. The original form was associated with augmentative and male speakers,
while the forms in [ai] and [a] were downgraded: [ai] was newly associated
with medial size and female speakers and [a] became the marker for diminutive
size and child speakers. The initial meanings of ‘mother’ and ‘child’ might have
contributed additionally to this process of redistribtion.
The second stage is based on the function of numeral classifiers as markers
of definiteness in the bare-classifier construction [CL + N]. This is illustrated
by ts’ae33 nG‘ae35 [CL meat] ‘the meat’ and ta33 li55 a55 [CL crow] ‘the crow’ in
example (12) from Weining Ahmao. Since Weining Ahmao is accessible only in
its contemporary form with its fully-fledged classifier-inflection system, it is
not possible to illustrate this stage without the simultaneous integration of size.
The vast majority of Miao languages do not express size on the classifier but
they use bare-classifier constructions for marking definiteness. One of these
languages is Hmong, which will be used to illustrate classifiers as definiteness
markers in bare-classifier constructions without the interference of size. In (16),
the two nouns niam ‘wife’ and txiv ‘husband’ are introduced in the first sentence
and are then taken up with their classifier tus as definite NPs in the next
sentence:
(16) Hmong
Thaum ub muaj ob tug niam txiv. Tus txiv tuag lawm.
Long.ago there.are two CL wife husband CL husband die PFV
Tus niam quaj quaj, nrhiav nrhiav tsis tau tus txiv.
CL wife cry cry look.for look.for NEG get CL husband
‘Long ago there were a wife and a husband. The husband died.
The wife was crying a lot but no mattter how hard she tried
she was not able to find her husband.’ (Mottin 1980: 200)
Indefinite nouns are marked by the numeral ib ‘one’ as in the Kam language
described in subsection 2.1.
(17) Hmong
Ua ciav nws pom ib lub nkauj npuas.
suddenly he see one CL pigpen
‘Suddenly, he discovered a pigpen.’ (Mottin 1980)
Numeral classifiers with plural marking. A challenge to Greenberg 35](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-47-320.jpg)
![Parallel to the situation in Northern Kam, the classifier changed its form
under the influence of the preceding numeral one, which is i55 in Weining
Ahmao, written ib10 in Hmong, and can be reconstructed with a closed syllable
as *iet which became *iʔ55 at a later stage. As is argued by Gerner and Bisang
(2009, 2010), it was the interaction of the glottal stop with the classifier in
sequences of the type [*iʔ55 CL] which triggered processes such as voicing, aspi-
ration or tonal change in the phonological structure of the classifier.11 The oppo-
sition that was formerly expressed by [CL + N]definite vs. [one + CL + N]indefinite is
now reflected by different forms of the classifier itself and can be represented
as [CLdef + N] vs. [CLindef + N]. Since this opposition was developed in the context
of the singular form of [one + CL + N]indefinite, the formal distinction was limited
to the singular. For expressing the same distinctions (plus the three size distinc-
tions) in the plural, the language followed another strategy.
The opposition between singular and plural is based on some sort of a
plural classifier which must have undergone the same changes as the normal
classifiers described in table 1. The details of this development need more
research. From the look at another Miao language like Hmong, one can clearly
say that there are plural classifiers. In Hmong, there is even a distinction
between the plural classifiers cov and tej, which roughly mark definiteness and
indefiniteness, respectively. In (18), the definite plural classifier cov in line 2
refers to the seven wives (poj niam) mentioned before in line 1. The plural clas-
sifier tej in (19) refers to an unspecified plural number of widows (poj ntsuag):
(18) Hmong
Ces ob tug mus txog chaw uas Yawm Pus
then two CL go to country/place REL Yao Pu
muaj xya tus poj niam. Ces cov poj niam quaj dheev tias: . . .
have seven CL wife then PL:CL wife cry very.much QUOT
‘Then, the two of them arrived at the country in which Yap Pu had seven
wives. The wives cried and said: . . .’ (Mottin 1980)
10 There are no syllables with consonantal codas in Hmong (the nasal ŋ is the only exception).
The Latin-based writing system of Hmong thus uses the post-nuclear position for encoding
tones by various consonant letters. The consonant -b, for instance, represents the high-level
tone [55].
11 This process is known as “change in glottal articulation” (Lehmann 1992: 193; also cf.
Gerner and Bisang 2009, 2010).
36 Walter Bisang](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-48-320.jpg)
![(19) Hmong
Kuv niam tias: kom kuv tuaj nrhiav tej poj ntsuag
I mother QUOT IMP 2.SG come seek CL:PL widow
tsev nyob xwb.
house live only
‘My mother told me: you only frequent houses of widows.’
[During your journey, you shall only ask for hospitality in houses of
widows.] (Mottin 1980)
3 Discussion: Factors that operate against
Greenberg’s (1974) universal claim
Greenberg’s (1974) implicational universal discussed in Subsection 1.1 excludes
the cooccurrence of obligatory plural marking and numeral classifiers. From
such a strict perspective, the classifier systems of Northern Kam and Weining
Ahmao provide no counter-evidence against Greenberg (1974), since the expres-
sion of number is still not compulsory in these two languages, i.e., the bare
noun alone with no classifier is still sufficient in certain contexts and if it is clear
that plurality is meant. This conclusion is certainly correct if one looks at Green-
berg’s claim from a purely synchronic angle that disregards the historical devel-
opment of grammatical systems. However, a solid universal claim should not
only be observable at the synchronic level but also at the diachronic level. If
combinations of properties that are excluded by a universal claim can be tied
together in historical processes of language change this casts serious doubt on
the solidity of that universal.
In the case of Chinese, the findings from diachronic research basically sup-
port the claim that there is a correlation between transnumerality and numeral
classifiers (cf. Meisterernst, this volume; Harbsmeier, this volume). The marker
zhū does not exclusively express plural, it also quantifies over unmarked plural
nouns and it expresses referentiality (Meisterernst, this volume). Even the exis-
tence of unmarked nouns with plural meaning (Harbsmeier, this volume) is no
straightforward argument against the overall transnumeral character of nouns
as long as plural nouns are a clearly defined subset of lexicalized plurals of a
limited size. Personal pronouns seem to be the only instance in which obligatory
plural marking is of some importance (cf. Meisterenst, this volume; also cf. Fu,
in this volume, on number in the pronominal system of Bai). But this only
affects the highest level of the animacy hierarchy as discussed by Corbett
(2000) and thus concerns only a small number of nominals.
Numeral classifiers with plural marking. A challenge to Greenberg 37](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-49-320.jpg)
![No matter how appropriate the motivations for the presence of numeral
classifiers in a language are in general, the present paper shows that there are
various factors that do not depend on transnumerality and the existence of
a mass/count-distinction. Thus, it seems rather implausible to argue that all
nouns in a Bantoid language such as Ejagham are mass nouns if that language
makes a systematic distinction between singular and plural in its noun-class
system. In spite of this, the language is about to develop a numeral classifier
system. The data from Kana with its fully developed classifier system and its
loss of the old noun-class system (cf. example [5]) may support the hypothesis
that there is a correlation between transnumerality and the existence of numeral
classifiers because the singular/plural distinction associated with the noun-
class system has vanished in the course of time. The question is whether the
disappearance of the noun-class system is directly linked to the development
of a fully-fledged numeral-classifier system. An alternative explanation for this
loss comes from phonetics and phonology and is related to the erosion of
phonological substance, a process which may well be completely unrelated to
transnumerality. A lot of additional research will be needed to understand the
emergence of the numeral-classifier systems of Niger-Congo languages.
In East and mainland Southeast Asian languages such as Northern Kam and
Weining Ahmao, the singular/plural-distinction developed even if it is plausible
to argue that they were transnumeral at least before they integrated number dis-
tinctions into their respective classifier systems. From a typological perspective
which emphasizes the relevance of whether there is a count/mass-distinction in
a language, the shift from a simple classifier system based on transnumeral
nouns to a system with number-distinctions which somehow presuppose a
count/mass-distinction looks rather great. However, the phonetic process that
was ultimately responsible for this shift does not depend at all on that func-
tional motivation.
This leads to the ultimate question of the status of typological universals
and their cognitive motivations. The findings of this paper seem to support
Dryer’s (1997) view that exceptionless universals are extremely rare and that
typologists should better spend their time on statistical universals. They also
argue rather for what Newmeyer (2005) calls holistic functionalism than for
atomistic functionalism. Atomistic functionalism assumes that “[t]here is direct
linkage between properties of particular grammars and functional motivations
for those properties” (Newmeyer 2005: 174). In contrast, holistic functionalism
sees a far more indirect relation between grammars and motivations. As
Newmeyer states, “grammars as wholes reflect the ‘interests’ of language users,
but there is no question of parceling out rules, constraints, constructions, and so
on of individual grammars and assigning to them a functional motivation”
40 Walter Bisang](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-52-320.jpg)
![Feng, Shengli. 2012. The syntax and prosody of classifiers in classical Chinese. This volume.
Fu, Jingqi. 2012. Plurality in the pronominal paradigms of Bai dialects. This volume.
Gerner, Matthias. 2006. Noun classes in Kam and Chinese Kam-Tai languages: their morpho-
syntax, semantics and history. Journal of Chinese Linguistics 34: 237–305.
Gerner, Matthias, and Walter Bisang. 2008. Inflectional speaker-role classifiers in Weining
Ahmao. Journal of Pragmatics 40: 719–731.
Gerner, Matthias, and Walter Bisang. 2009. Inflectional classifiers in Weining Ahmao: mirror of
the history of a people. Folia Linguistica Historica 30: 1–36.
Gerner, Matthias, and Walter Bisang. 2010. Social-deixis classifiers in Weining Ahmao. In Rara
and Rarissima. Documenting the Fringes of Linguistic Diversity, Jan Wohlgemuth and
Michael Cysouw (eds.), 75–94. Berlin/New York: Mouton de Gruyter.
Greenberg, Joseph H. 1974. Numeral classifiers and substantival number: Problems in the
genesis of a linguistic type. In Proceedings of the 11th International Congress of
Linguistics, Bologna – Florence, Aug–Sept 1972. Bologna. 17–37. Reprinted in Greenberg,
1990: 16–93.
Greenberg, Joseph H. 1990. On Language. Selected Writings of Joseph H. Greenberg, ed. by
Denning, K. and S. Kemmer. Stanford: Stanford University Press.
Harbsmeier, Christoph. 2012. Plurality and the subclassification of nouns in classical Chinese.
This volume.
Hundius, Harald, and Ulrike Kölver. 1983. Syntax and semantics of numeral classifiers in Thai.
Studies in Language 7: 165–214.
Ikoro, Suanu M. 1994. Numeral classifiers in Kana. Journal of African Languages and Linguistics
15: 7–28.
Ikoro, Suanu M. 1996. The Kana language. Leiden: Research School CNWS.
Lehmann, Winfred P. 1992. Historical Linguistics. An Introduction. London: Routledge.
Li, Xuping, and Walter Bisang. 2012. Classifiers in Sinitic languages: From individuation to
definiteness marking. Lingua 122: 335–355.
Meisterernst, Barbara. 2012. Number in Chinese: a diachronic study of the case of zhu from
Han to Wei Jin Nanbeichao Chinese. This volume.
Mottin, Jean. 1980. Contes et légendes Hmong Blanc. Bangkok: Don Bosco Press.
Newmeyer, Frederick J. 2005. Possible and Probable Languages. A Generative Perspective on
Linguistic Typology. Oxford: Oxford University Press.
Sanches, Mary, and Linda Slobin. 1973. Numeral classifiers and plural marking: an implica-
tional universal. Working Papers in Language Universals 11: 1–22.
Shi, Lin. 1997. 侗语汉语语法比较研究 Dongyu Hanyu yufa bijiao yanjiu [Dong-Han Com-
parative Grammar Research]. Beijing: Central University of Nationalities Press.
Wang, Deguang. 1986. 威宁苗语话语材料 Weining Miaoyu huayu cailiao [Language material
in the Weining dialect of the Miao language]. Minzu Yuwen 民族语文 3: 69–80.
Wang, Deguang. 1987. 贵州威宁苗语量词拾遗 Guizhou Weining Miaoyu liangci shiyi [Find-
ings on numeral classifiers in the Weining dialect of the Miao language in Guizhou]. Minzu
Yuwen 民族语文 5: 36–38.
Wang, Fushi. 1957. 贵州威宁苗语量词 Guizhou Weining Miaoyu liangci [The classifier in the
Weining dialect of the Miao language in Guizhou]. Yuyan Yanjiu 语言研究 1957: 75–121.
Translated 1972 as: The classifier in the Weining dialect of the Miao language in Kwei-
chou. In Miao and Yao Linguistic Studies: Selected Articles in Chinese, Herbert Purnell
(ed.), 111–185. Ithaca, NY: Cornell University.
Watters, John Roberts. 1981. A phonology and morphology of Ejagham – with notes on dialect
variation. Ph.D Diss., University of California, Los Angeles.
42 Walter Bisang](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-54-320.jpg)
![our every act, and crowned his impertinences by selling us some two
dozen real bad eggs, at eight cash apiece.
Sometimes a traveller is apt to find himself in a very awkward
position, if his carters and an offensive innkeeper happen to be in
league. It may happen that the carters owe the innkeeper money,
and the latter threatens to get them punished unless they help him
in extorting money from the unfortunate foreigner; the innkeeper
then demands a perfectly unreasonable sum from his victim, which
the latter refuses; upon this the carters come up and say that they
cannot possibly leave till the money is paid, and the helpless
traveller finds himself between the devil and the deep sea. Luckily
for us, our agreement effectually stopped any nonsense of this kind,
for if the carters said they could not go on, we at once replied, "All
right, please yourselves—but if you do not arrive by the 27th, you
lose half your hire."
Every day, now, villages became more numerous, and the
country was very thickly cultivated. We stopped a long time at our
midday halt at the city of Shuen-Hua-Fu,[15] as the carters had to
change some silver and transact some business before proceeding.
After leaving, we passed a very large number of gravestones and
memorial stones, with old inscriptions of the rounded characters; the
hillsides on the left of the road were studded with them for several
miles, and in many places the coffins were sticking out of the
ground, where the earth had been washed away from over them. On
the road we met a convoy of mules carrying boxes containing tins of
oil from Batoum, and another with silver; the latter had an escort
armed with spears decorated with red tassels, each mule, too, was
decorated with a red flag, the whole making quite an imposing
show.
Towards evening the road got very bad, and in one place we all
had to get out and steady the carts as they passed down a steep,](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-66-320.jpg)
![We were now very anxious about him, so divided ourselves into
two parties, the carts and servants going one road, while we went
another, which passed through several small towns and ran parallel
to the road taken by the carts. Both parties made inquiries as they
went along, and for some thirty li we still heard of him; he seemed
to have walked straight ahead, never stopping except to buy some
food. His clothes and gait, as well as his dark face, made him very
conspicuous, so that the various guards on the town gates always
noticed him; but after going about ten miles we lost all trace of our
faithful Duffadar, and all we could do was to give notice of his
disappearance in different places, and ask the people to show him
which way we had gone. The most unlucky part of the business was
that he could not speak a solitary word of Chinese, and might get
himself into trouble, but fortunately we were in northern, and not
central or southern China, so we felt sure that sooner or later he
would turn up.[16]
In some of the towns we passed through we saw quantities of
"spirit's paper" lying about the streets, or hanging up in the shops
for sale. This spirit's paper is simply round pieces, cut to represent
cash, each sheet of paper representing any number, or gilt to
represent silver or gold. These are bought by the people and
scattered about the houses, or put in holes in walls and tree trunks,
where they are found by restless, wandering spirits, who, poor
creatures, are easily deceived into thinking them offerings of great
value, and consequently refrain from injuring the pious, but
economical offerer. They also burn large quantities of this paper
money, and effigies of carts, horses, and other signs of worldly
wealth, all of which go to improve the social status of their ancestors
in the nether world.[17] The Chinaman, though decidedly
superstitious and pious, according to his lights, rarely allows his
regard for his ancestors to overcome that for his own pocket.](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-69-320.jpg)
![This was to be another very long day, so we halted for an hour,
about 4 o'clock in the afternoon, and fed the animals, we ourselves
buying some splendid grapes for forty cash a pound. These are
grown south of the town of Shuen-Hua-Fu; also quantities of millet
(huang-mi[18]), which is ground up and eaten like rice by the poor
classes.
While wandering round, telling people of Shahzad Mir's
disappearance and looking at the various shops, we saw hung up in
a cage on the city gate the head of a man who had been found
guilty of theft, and next to this gruesome relic were the shoes of
several officials, whose ability or zeal in discharge of their duties had
merited this special distinction; thus are the best and the worst of
men's deeds kept fresh in the minds of the Chinese people.
We were just about to make our third start for the day when two
Russian merchants entered our inn. We were much struck by the
promptitude with which they extracted their revolvers from the
sedan chairs in which they were travelling, and placed them in
handy and conspicuous positions on top of their luggage. Curiously
enough the only language in which we could converse with them
was Chinese, which one man spoke fluently; and it was a great relief
when we learnt that we need have no fear of the Gulf of Pe Chili
being frozen over, for we were still in plenty of time.
We still had sixty li to do before halting for the night at Huai-Lai-
Hsien, so we shoved along as fast as the mules could go, but the
last twenty li or so were done in pitch darkness and bitter cold. The
gates of the suburb were closed when we arrived, but we got them
open with some difficulty, after considerably exaggerating the
importance of the high official posts we both occupied. We had done
nearly fifty miles since morning, so were glad to get an excellent
Chinese concoction of eggs, and tumble into our blankets as quickly
as might be.](https://image.slidesharecdn.com/25132899-250520074500-7d793756/85/Plurality-And-Classifiers-Across-Languages-In-China-Dan-Xu-Editor-70-320.jpg)
Plurality And Classifiers Across Languages In China Dan Xu Editor
- 1. Plurality And Classifiers Across Languages In China Dan Xu Editor download https://ebookbell.com/product/plurality-and-classifiers-across- languages-in-china-dan-xu-editor-50265798 Explore and download more ebooks at ebookbell.com
- 2. Here are some recommended products that we believe you will be interested in. You can click the link to download. Plurality And Perspective In Psychoanalysis Adam Rosencarole https://ebookbell.com/product/plurality-and-perspective-in- psychoanalysis-adam-rosencarole-5537270 Plurality And The Poetics Of Self 1st Ed Bruce Bond https://ebookbell.com/product/plurality-and-the-poetics-of-self-1st- ed-bruce-bond-10487922 Astrologies Plurality And Diversity In The History Of Astrology Nicholas Campion https://ebookbell.com/product/astrologies-plurality-and-diversity-in- the-history-of-astrology-nicholas-campion-46710856 Religious Plurality And Pragmatist Theology Openness And Resistance Henriksen https://ebookbell.com/product/religious-plurality-and-pragmatist- theology-openness-and-resistance-henriksen-47486042
- 3. Verbal Plurality And Distributivity Patricia Cabredo Hofherr Editor Brenda Laca Editor https://ebookbell.com/product/verbal-plurality-and-distributivity- patricia-cabredo-hofherr-editor-brenda-laca-editor-50367568 Religious Plurality And The Public Space Joint Christian Muslim Theological Reflections Simone Sinn https://ebookbell.com/product/religious-plurality-and-the-public- space-joint-christian-muslim-theological-reflections-simone- sinn-5449332 Verbal Plurality And Distributivity Hardcover Patricia Cabredo Hofherr Brenda Laca https://ebookbell.com/product/verbal-plurality-and-distributivity- hardcover-patricia-cabredo-hofherr-brenda-laca-10005432 Medieval Damascus Plurality And Diversity In An Arabic Library The Ashrafiya Library Catalogue Konrad Hirschler https://ebookbell.com/product/medieval-damascus-plurality-and- diversity-in-an-arabic-library-the-ashrafiya-library-catalogue-konrad- hirschler-51969736 Postcrash Economics Plurality And Heterodox Ideas In Teaching And Research 1st Edition Omar Feraboli https://ebookbell.com/product/postcrash-economics-plurality-and- heterodox-ideas-in-teaching-and-research-1st-edition-omar- feraboli-6841514
- 5. Dan Xu (Editor) Plurality and Classifiers across Languages in China
- 6. Trends in Linguistics Studies and Monographs Editor Volker Gast Founding Editor Werner Winter Editorial Board Walter Bisang Hans Henrich Hock Heiko Narrog Matthias Schlesewsky Niina Ning Zhang Editors responsible for this volume Niina Ning Zhang Walter Bisang Volume 255
- 7. Plurality and Classifiers across Languages in China Edited by Dan Xu
- 8. ISBN 978-3-11-029382-1 e-ISBN 978-3-11-029398-2 ISSN 1861-4302 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.dnb.de. 6 2012 Walter de Gruyter GmbH, Berlin/Boston Typesetting: RoyalStandard, Hong Kong Printing and binding: Hubert & Co. GmbH & Co. KG, Göttingen ♾ Printed on acid-free paper Printed in Germany www.degruyter.com
- 9. Table of contents Preface vii Dan Xu Introduction: Plurality and Classifiers across languages of China 1 I Correlations between different types of quantification Walter Bisang 1 Numeral classifiers with plural marking. A challenge to Greenberg 23 Dan Xu 2 Reduplication in languages: A case study of languages of China 43 II Numeral classifiers and their diachronic development Shengli Feng 3 The Syntax and prosody of classifiers in Classical Chinese 67 Lin Jang-Ling and Alain Peyraube 4 Individuating classifiers in Early Southern Min (14th–19th centuries) 101 III The expression of plurality Christoph Harbsmeier 5 Plurality and the subclassification of Nouns in Classical Chinese 121 Barbara Meisterernst 6 Number in Chinese: a diachronic study of zhū 諸 From Han to Wei Jin Nanbeichao Chinese 143 Marie-Claude Paris 7 Bu-tong ‘different’ and nominal plurality in Mandarin Chinese 183 Jingqi Fu 8 Plurality in the pronominal paradigms of Bai dialects 203
- 10. Hongyong Liu and Yang Gu 9 Frequentative aspect and pluractionality in Nuosu Yi 223 Thomas Hun-tak Lee 10 Quantificational structures in three-year-old Chinese-speaking children 243 Author index 283 Subject index 287 vi Table of contents
- 11. Preface Quantification is one of the most basic notions for human beings, and conse- quently it is one of the most commonly marked notions in languages. It is indi- cated with different devices, phonological, morphological, syntactic, and others. In this book, quantification will be observed and studied from three different perspectives: synchronic, diachronic and typological. Several topics are studied in the book, with a focus on plurality. The book will take advantage of recent results from research on quantification in languages of China to deal with more concrete and more targeted subjects, such as relationships between plurality and classifiers, plurality and reduplication, plurality expressions in non-Han languages, and plurality concepts in child language acquisition. This book originated from a project awarded by the French ANR, Quantifi- cation and Plurality ANR-06-BLAN0259, then by the IUF (University Institute of France). Without their substantial financial support, this book could not have been completed. The editor expresses her sincere gratitude to these institutions. The authors are from France, Germany, Norway, China, Hongkong and the United States. Thanks to collective research and to the contribution of each scholar, the book progressively took shape and became what it is today. Completion of the project is actually marked by the publication of two books: one was published in China, Liàng yǔ fùshù de yánjiū 量与复数的研究 ([Quantification and plurality] 2010), and the second (the present book) is being published in Europe. In the research, as well as in the editing work, my colleagues have helped me in one way or another and have given me pertinent remarks and useful advice. The editor’s thanks also go to Murielle Fabre who did the formatting work for this volume, which is time-consuming labor. Many thanks also to Marie-Claude Paris and Jingqi Fu for helping me with checking work. Undoubtedly, all faults are mine. Dan Xu Paris September 2012
- 13. Dan Xu Introduction Plurality and Classifiers across languages of China1 1 Background Plurality has drawn the attention of scholars (Sanches [1971] 1973; Greenberg 1972, 1988; Lü 1984; Dryer 1989, 1992; Bisang 1993, Heide 1995, Ortmann 2000, Comrie 2008, Xu 2010, among others) for several decades. Papers on plurality have most often discussed it in terms of numeral classifiers and the absence of obligatory plural marking. But numeral classifiers and plural marking are not the only strategies for grammatical quantification; reduplication is a third device for quantification. In fact, plural marking, numeral classifiers and reduplication constitute the main means of quantification marking in the domain of grammar. There are many more lexical devices for expressing quantification, but this book is interested in grammatical systems. The present introduction will first con- centrate on the typological correlation between the three different strategies for quantification, as well as some general issues. In the second section, it will address individual devices for quantification in Chinese, concentrating on plurality and reduplication, which have been studied less frequently than classi- fiers. The third part will go beyond Chinese and will present a short typology of how quantification is expressed in the other languages of China. A description of how the book is organized will conclude this introduction. Languages in which all three strategies of quantification are marked obliga- torily are rare, while languages lacking all of them are even less frequent (see Chapter 2 in this volume). It is important to distinguish two notions here, co- occurrence and co-existence. I refer to the co-existence of three devices (plural markers, reduplication and classifiers) instead of their co-occurrence, which is 1 This book is one of the outcomes related to the ANR projects Quantification et Pluralité ANR- 06-BLAN0259 and Space and Quantification (IUF). We express our gratitude for their financial support. Our thanks also go to the anonymous reviewer for his/her constructive comments on the Introduction and to all participants who have provided their contribution to the present volume. My debts to Walter Bisang are numerous. He has given me not only effective support and help but also very useful suggestions for this Introduction. Finally thanks to Craig Baker for his help with proofreading.
- 14. impossible in most languages. They are correlated in expressing quantity: if one is present and obligatory, the other two are often absent or optional. This also assumes that in human languages, quantification has to be explicitly or implicitly indicated by different means. If a language does not possess obligatory plural marking, it tends to exploit reduplication as a grammatical device. In this sense, plural marking and reduplication are comparable in syntax. More concretely, one or the other is frequently attested in the nominal structures of human languages. Reduplication often co-exists with classifiers in Mainland Southeast Asian languages. This is one of the characteristics of this linguistic area (Enfield 2005). In fact, two patterns are found in languages: obligatory plural marking on the one hand vs. reduplication-classifier marking on the other. Reduplication is rarely obligatory (even though it is fully productive in many languages, cf. Rubino 2005) and adds specific semantics to plurality. For that reason, it can easily co- exist with the numeral-classifier systems of Chinese and Mainland Southeast Asian languages. Even though the mutual exclusion of numeral classifiers and obligatory plural marking has often been discussed in typological work since Sanches ([1971] 1973) and Greenberg (1972), the theory is not as water-tight as is often claimed. Even Greenberg (1972) does not treat this correlation as having no exceptions. Some publications, especially a monograph on Japanese by P. Downing (1996), have pointed out that the correlations between plurality and classifiers must be more widely investigated and better described, since counterexamples exist. Based on Japanese data, Downing (1996: 28) indicates that “the language [Japanese] possesses not only numeral classifiers, but several plural markers which may appear with nouns which are accompanied by classifiers as well.” Bisang (in this volume) also shows potential counterexamples with data on Northern Kam and Weining Ahmao. Reduplication and the occurrence of numeral classifiers are compared much less in the literature. In contemporary Chinese, plural marking and reduplica- tion are not obligatory, while classifiers are when using numerals. One device cannot co-occur with the others as is illustrated by the following examples: (1) 人们， 人人， 五个人 rén men rén rén wǔ ge rén person-PL, person-person, five-CL-person ‘people, everybody/people, five people’ (2) *人人们， *五个人们 rén rén men wǔ ge rén men person-person-PL, five-CL-person-PL 2 Dan Xu
- 15. In fact, Chinese seems to have become a nominal classifier language not as early as has been believed, and the order “numeral + classifier + noun” won as a dominant order only after the 10th century AD (Wu 2006). This means that Old Chinese was not a classifier language for a long time. It is thus interesting to investigate how plurality marking developed during the periods when classifiers were not obligatory (see Feng in this volume). Plural markers in contemporary Chinese are obviously based on word plurality (one of the categories of plurality denoted by Dryer [1989]), while Old and Middle Chinese used other means (Xu 2010). In Chinese, nouns and verbs can be reduplicated to express plurality, pluractionality, distributivity, etc. Not all nouns can be reduplicated, and not all reduplicated nouns indicate plurality. Also, productive reduplication is rela- tively new in Chinese, and very few examples can be found in Old Chinese and Middle Chinese. In modern Mandarin Chinese, nouns cannot be freely redupli- cated. Not as many nouns can be reduplicated (such as 人人 rénrén ‘everyone’, 天天 tiāntiān ‘everyday’) as adjectives and verbs. Classifiers, like nouns, can be repeated indicating distributivity: (3) 人人， 年年 rénrén, niánnián ‘everyone, every year’ (4) 个个, 张张 gègè, zhāngzhāng ‘every one, every piece (of paper)’ (5) 家家户户 jiājiāhùhù ‘each family’ Example (3) is often cited to show noun reduplication in Chinese. In fact, this is far less productive than adjective or verb reduplication. Temporal words such as 年 nián ‘year’ can be reduplicated. In (4), classifiers are reduplicated to express distributivity. In general, classifiers can be reduplicated when the con- text is clear. Example (5) shows that some reduplicated words have been com- pletely lexicalized. These examples are abundant and we shall pass over them here. One thing is clear: distribution/distributivity is in fact another way to express plurality; and without a concept of collectivity in which plurality is implied, the expression “everyone” is not tenable. Quantification and many other grammatical phenomena are better described and studied in Indo-European languages because comparative methods have Introduction: Plurality and Classifiers across languages of China 3
- 16. been established by European scholars since the 19th century. The investigation of the languages of China seems to lag far behind. The linguistic situation in China is also more complex, since more languages from more families are involved. The languages of China consist of the Han languages (Mandarin and at least nine other Sinitic languages), and the Han are surrounded by different ethnic groups and languages. It is thus important and necessary to study the Han languages as well as the non-Han languages of China. Since European lan- guages tend to have obligatory plural marking and lack numeral classifiers, and since they use reduplication less productively and less frequently (cf. Rubino 2005) than many languages of China, a better understanding of the quantifica- tion strategies in the languages of China will certainly enrich our comprehension of human language and thought. In addition, the study of Chinese and the other languages spoken in China will have its impact on the study of other language families. The languages of China deserve more investigation, especially the non- Han languages, because many of them remain unknown or poorly studied. Plurality is “an area that has been rather neglected, especially in typological work” (Dryer 1989: 888); and “A clearer understanding of the diachronic origin (or origins) of plural words is necessary in order to better understand their synchronic properties.” (Dryer 1989). This book tries to remedy the general lack of information on grammatical quantification in the languages of China by pro- viding papers by scholars who have been working in these fields for a long time and have made important contributions to the study of the languages in China. A book edited by Li Jinfang and Hu Suhua (2005) observes non-Han languages of China and studies classifiers. In a recent book Liàng yǔ fùshù de yánjiū 量与 复数的研究 ([Quantification and plurality], Dan Xu [ed.], 2010) which has been published as a companion to the present volume, seventeen authors (two papers are written by co-authors) give their contributions to these problems. The authors of these two books do not overlap, and the first one was written in Chinese while the present one is in English. These two books aim to review Chinese and some non-Han languages in depth to better understand the correla- tion between plurality and classifiers. 2 Devices for marking quantity in Chinese2 Chinese is one of the languages that does not require plural marking in noun phrases (see Feng Shengli and Harbsmeier in this volume for Old Chinese). Plu- rality is expressed in the clause by numerous adverbs, or indicated in nominal 2 Studies on the origin and development of the plurality suffix -men in Chinese can be found in the sister book Quantification and Plurality (2010) mentioned earlier. 4 Dan Xu
- 17. phrases by numerals, adjectives or noun modifiers (Paris in this volume). Plurality in Chinese is not obligatorily expressed by a morpheme or several morphemes as in Indo-European languages, but by abundant adjunct words. Thus, Chinese seems to follow Greenberg’s (1972) universal even though the complementary distribution of obligatory plural markers and numeral classifiers has been chal- lenged by more recent data (see Bisang, this volume). Chinese differs from morphologically rich languages, but might have used morphology for expressing quantification in its early history. Eventually, plural words (lexical items) became prominent. Though contemporary Chinese does not have obligatory marking of every NP for plurality, it possesses some adverbs to indicate plurality (see Lee in this volume), such as 都 dōu ‘all’, 皆 jiē ‘all’, 全 quán ‘all’, etc. The list was much longer in Middle Chinese: 都 dōu, 皆 jiē, 全 quán, 並 bìng, 盡 jìn, 具/俱 jù, 悉 xī, 咸 xián, 縂 zǒng, 擧 jǔ, etc. (see also Meisterernst in this volume). It seems that the Chinese language prefers marking plurality in the VP to marking it in the NP. In previous studies on plural marking in Chinese, scholars have focused on the plural marker men which is only applicable to noun phrases with the feature [+human] and it is not obligatory especially in the object position. In Early Old Chinese (11th–5th centuries BC), some words like 眾 zhòng, 群 qún, 諸 zhū, marked plural meaning in the prenominal position. Only 余 yú was a post- nominal plural word. These prenominal plural words became lexicalized later (see also Meisterernst in this volume) and some postnominal plural words like 曹 cáo, 輩 bèi, 等 děng took over the function of plural marking. Due to its post- nominal position, yú survived and continued to express plural meaning. There is no major debate among scholars about the status of these plural words which are not full-fledged plural markers. The real plural marker men evolved during the Song Dynasty (960–1279 AD). From diachronic data, Dan Xu (2009a) suggests that the emergence of postnominal plural marking is correlated with the develop- ment of classifiers which moved from the postnominal position to the prenominal position, while plural marking did the reverse (see also Jang-Ling Lin and Alain Peyraube in this volume). In this book, authors will observe the relationship between plural marking and classifiers from different angles to better under- stand the problem. The investigation will not be limited to Chinese, but will extend to other languages spoken in China in order to work with different per- spectives on this topic. Reduplication is a frequent phenomenon in Asian and African languages (for example, Malay in Indonesia, Lepcha in India, Lomongo in the Congo, Amharic in Ethiopia, Malagasy in Madagascar, Yoruba in Western Africa, etc.). It can indicate a variety of concepts such as plurality, distributivity, iteration, and intensity (cf. Rubino 2005 for a more exhaustive list). One of its meanings Introduction: Plurality and Classifiers across languages of China 5
- 18. in Chinese is that of the diminutive. In this case, it has to be considered as a morphological device by which the iconic function of repeating phonological material has a symbolic meaning (see detailed analysis of Dan Xu in this volume). Some reduplicated nouns express an intimate meaning3; again these nouns are not very prevalent and are often found in adult-child dialogue. When adjectives are reduplicated, the degree of the adjective’s quality is, in general, intensified. But the intensification interpretation is not always present in contemporary Chinese: for example, color perception can be subjective and variable, therefore adjectives describing color are prone to subjective interpreta- tion. More precisely, reduplicated adjectives, in the beginning, were iconically motivated to emphasize degree. Some of them can emphasize the subjective per- ception of the author (see Montaut 2009); thus the diminutive interpretation is possible. The ABB type of reduplication is very well-developed in contemporary Chinese. Cheng Xiangqing (1992: 95) points out that the ABB form began to become more prevalent in the Tang Dynasty (618–907 AD). This phenomenon can also be seen in Modern Hindi. Singh (2005: 268) reveals that in some cases, a reduplicated adjective can express an emphatic distribution if the noun is plural, while the same word can signify an “approximation” in other cases. Compare his two examples: (6) hariiharii green green (leaves) ‘very green (leaves)’ (7) hariiharii saarii green green sari ‘a greenish sari’ These examples suggest that the subjective interpretation of colors is frequent in languages, and the subjectivity plays a key role in favoring the evolution of the plus-degree towards the minus-degree. However diachronic examples of Chinese ABB confirm that at the beginning of its development, reduplication marked the plus degree in an iconic way. When this device is full-fledged, it can become a morphological device, losing its initial meaning. In fact, reduplicated verbs marking diminutive meaning in Chinese illustrate this change well. More con- cretely speaking, in the first stage, they generally indicate emphasis, reinforce- ment and the plus-degree, while in the second stage, they almost always express diminutive meaning and the minus-degree. Here are some examples: 3 For example: 勺勺 sháoshao ‘spoon’, 兜兜 dōudou ‘pocket’, 包包 bāobao ‘bag’. 6 Dan Xu
- 19. (8) 看看， 想想， 说说， 走走 kànkan, xiǎngxiang, shuōshuo, zǒuzou ‘to take a look, to think a bit, to say something, to take a walk’ In the above examples from contemporary Chinese, the reduplicated verbs mean “to go to do something a little bit, to try”. The grammatical meaning “unaccomplished, atelic, imperfective, short duration action” is in fact an inter- pretation of the new status of the reduplicated part: it indicates the future. Reduplicated verbs are not compatible with the past tense, adverbs implying the past, or elements indicating accomplishment. An aspect marker [+accomplish] le has to be inserted between the two parts if one wishes to express the past, a telic action. Again, I think that some contemporary non-iconic meanings have to pro- gressively lose their initial iconic signification. Now we will illustrate this point of view with some diachronic examples: (9) 看看似相识 (孟浩然 Mèng Hàorán, 689–740 AD) kànkàn sì xiāngshí see see like be acquainted ‘Seeing each other, it seems that we are acquainted.’ (10) 望望不见君 (李白 Lǐ Bái, 701–762 AD) wàngwàng bú jiàn jūn look look NEG see you ‘I am looking and looking but I cannot see you’ In these sentences, the reduplicated verbs intensify the continuance of the action in repeating the same verb. The iconic meaning is clear. According to Cheng Xiangqing (1992) and Liu Xiaonong’s separate descriptions (1992) and to Wang Ying’s (1996: 234) studies, reduplicated verbs from before or during the Tang never indicated diminutive meaning. All of them express high frequency, augmented quantity, continuance, and so on. It is interesting to note that in modern Yangzhou (located in Jiangsu Province) dialect, reduplicated verbs have kept the iconic meaning indicating high frequency of the action (Zhu Jingsong cited by Wang Ying 1996: 240), in contrast with Mandarin Chinese. How did these iconically motivated reduplicated verbs lose their initial mean- ing and become non-iconic, i.e. morphological devices, in Mandarin Chinese? Again diachronic data and contemporary dialects provide us with comple- mentary evidence. I believe that the rise and the merger of the patterns V1 (V2) kàn ‘see’ (try to do something) and Vyī ‘one’ V (do something once) caused Introduction: Plurality and Classifiers across languages of China 7
- 20. reduplicated verbs to lose their iconic meaning. Cai Jinghao (1990) and Wu Fuxiang (1995) provide a diachronic study of V1 (V2) kàn. Wu indicates that V yī ‘one’ V (do something once) is found in the Song and may come from “V + Numeral + classifier” in which the verbal classifier was copied from the same verb. Observe some examples: (11) 可去拜一拜，作别一声。 (碾玉观音 Niǎn yù Guānyīn 12th–13th centuries) kě qù bài yí bài zuòbié yì shēng can go worship one once, bid-farewell one sound ‘You can worship them and bid them farewell.’ (12) 掀起帘子看一看 (碾玉观音 Niǎn yù Guānyīn) xiānqǐ liánzi kàn yí kàn raise curtain look-one-look ‘When he raised the curtain and took a look inside . . .’ In these cases, the verbal classifiers already express a short duration action meaning “once”. In contemporary Chinese the reduplicated verbs VV signifying a diminutive meaning can reincorporate yī ‘one’ between them. All examples cited in (8) can be rewritten as V yī ‘one’ V instead of VV. This suggests that some contemporary instances of VV indicating diminutive meaning came from the V yī ‘one’ V pattern but not directly from iconically motivated ones like (9) and (10). These instances of VV are just the shortened form of “V one V”. Another construction also favors the diminutive interpretation of VV: V1 (V2) kàn ‘look’ in which the verb kàn ‘look’ was gradually grammaticalized as a particle indicating a future meaning, since verbs which do not denote per- ception can be combined with kàn, originally meaning ‘look’. Here are some examples cited by Wu (2006): (13) 试尝看 (白居易 Bái Jūyì, 772–846 AD) shì cháng kàn try taste see ‘Try it’ (14) 收拾看 (祖堂集 Zǔtángjí, 10th century) shōushi kàn put-in-order see ‘Try to put it in order’ 8 Dan Xu
- 21. (15) 你猜猜看 (西游记 Xīyóu Jì, 16th century) nǐ cāi cai kàn you guess guess see ‘Try to guess it’ (16) 等我尝尝看 (西游记 Xīyóu Jì) děng wǒ chángchang kàn wait me taste taste see ‘Let me taste it’ The above sentences also indicate an atelic action expressing “try to do”. Even today VV kàn exists in Mandarin Chinese, retaining the same meaning. In the construction VV kàn, the verb kàn ‘see’ has undergone a process of gramma- ticalization, evolving from a full verb to a grammatical element, and finally became completely optional. Example (15) is already entirely similar to today’s Mandarin Chinese in which the particle kàn can be omitted. The rise and development of these two constructions, which originally indi- cated a one-time, attempted action, facilitated replacing iconically motivated verb reduplication. In other words, these constructions’ shortened form VV coincided with the reduplicated verbs indicating high degree and intensifica- tion. The latter did not survive in Mandarin Chinese and gave way to newly arisen constructions which took the VV form but modified the initial meaning. This hypothesis can be indirectly confirmed by contemporary dialect data (Cao Zhiyun, ed. 2008). Except for Mandarin in the Northeast (Heilongjiang, Jilin, Liaoning, Hebei, Shanxi and Shandong) and dialects in the South (concentrated in Guangzhou and Hainan), most dialects lack the VV form and use another con- struction to express a short time action or a diminutive meaning: V 一下 yíxià ‘once’. This construction coexists with VV in Beijing Mandarin indicating a diminutive meaning or a short duration action. These facts suggest that contem- porary VV in the North with minus-degree meaning comes from other construc- tions, but does not imply that reduplicated V was not iconic at the initial stage. Diachronic data in Chinese shows us that some contemporary instances of redu- plication may have originally come from other constructions, and sometimes a modern reduplicated form signifying diminution is the residue of an ancient non-reduplicated form. 3 Plural typology in languages of China Three types of plurality are attested in China’s non-Han languages (Dan Xu in this volume): Introduction: Plurality and Classifiers across languages of China 9
- 22. (a) Tibeto-Burman type; (b) Hmong-Mien (Miao-Yao), Tai and Austro-Asian type; (c) Altaic type. Some languages such as Mandarin share some characteristics with the first and second types. In other words, they actually belong to a mixed type. Their plural marking is applicable to nouns (not obligatory) and pronouns (obligatory)4. (a) Tibeto-Burman type In this group, all languages, except the Bai language (see Jingqi Fu in this volume), have an OV word order, and classifiers. Differences in plural marking are seen in this group. For example, in the Tibetan, Yi (see Hongyong Liu and Yang Gu in this volume), and Jingpo subgroups, plural marking exists and is sensitive, in most cases, to human or animate features of the noun, such as in Baima, Lahu, Naxi, Rouruo and Dulong. The Qiang group, belonging to the same family as these languages, behaves in a different manner: almost half of the plural markers use the same form for animate or non-animate noun phrases. (b) Hmong-Mien (Miao-Yao), Tai and Austro-Asian type The Hmong-Mien (Miao-Yao), Tai and Austro-Asian languages form the second group. All languages in this group have a VO word order and classifiers, but the plural marker is not widely used. In most languages of China, plural markers are found after an NP, while in some languages in this group such as Mulao, Dai (Dehong), Bunu and Jiongnai, plural markers can be put before NPs when NPs refer to parental terms (Sun, Huang and Hu 2007). It is clear that this kind of plural marker is not widespread, but it still merits attention. Geographically, these languages are sprinkled around the south of China and do not present a coherent region like Tibeto-Burman or Altaic languages. This also partly ex- plains why they often present Chinese language features despite being geneti- cally distant or unrelated. For the same reason, some of these languages show typologically rare plural-marking systems which are expressed morphologically on the classifier (cf. Bisang in this volume on Kam and Weining Ahmao). (c) Altaic type In this type, plural markers are rather well developed and classifiers are absent (except some languages such as Salar, Korean, and Manchu). In these languages 4 The plural marking for noun and pronoun is not always the same. For example in the Shixing language, the marking is distinct (Dan Xu 2009b). 10 Dan Xu
- 23. the use of the plural marker does not depend on the semantic feature [+animate] and all nouns can be marked by plural suffixes. However, the plural marker is not always obligatory as in European languages. The Salar, Korean, and Manchu languages employ classifiers due to influence from the Chinese language. Geo- graphically they have been surrounded by Chinese languages for centuries and language contact has been very intense. The evidence for this contact-induced change is that classifiers are not yet obligatory in Salar, Korean, or Manchu, and the word order of numeral and noun is not fixed. For these reasons, we suppose that classifiers do not present a significant feature for this type of language, even though some of them have begun to develop classifiers due to long contact with Chinese languages. It is clear that the Sinitic languages have VO order like type (b) and use classifiers like (a) and (b). It is important to notice that in groups (a) and (b), almost all languages (42/44) use reduplication devices and all of them employ classifiers. This fact confirms our hypothesis that reduplication and classifiers form one pattern with regard to plural marking. Additionally, in languages using classifiers, plural markers are often distinct in that they are sensitive to the feature [animate]. 4 Organization of the book In this volume, quantification will be observed and studied from different per- spectives. The book studies several topics focusing on plurality. Based on the present results of research on quantification, different chapters deal with con- crete and targeted subjects such as relations between plurality and classifiers, plurality and reduplication, expression of plurality in non-Han languages, and plurality concepts in child language acquisition. The papers are divided into three sections. The first one takes a typological perspective and consists of two papers in which correlations between different types of quantification are discussed. Bisang deals with the correlation between the presence of numeral classifiers and the non-obligatoriness of plural mark- ing as claimed by Greenberg (1972); Xu describes reduplication in its interaction with plural marking and numeral classifiers. The second section deals with numeral classifiers and their diachronic development in Mandarin Chinese (Feng) and Early Southern Min (Lin and Peyraube). The expression of plurality is the topic of the third section. It starts with the expression of plurality in the history of Chinese, first in its semantic form in classical Chinese (Harbsmeier), then in the form of the marker 諸 zhū from Han to Wei Jin Nanbeichao Chinese (Meisterernst). Both of these papers are looking at the domain of nouns. This Introduction: Plurality and Classifiers across languages of China 11
- 24. also applies to the next two papers on the use of 不同 bùtóng ‘different’ (Paris) and the expression of plurality in the pronominal system of Bai (Fu). The paper by Liu and Gu is dedicated to the expression of plurality in the domain of the verb. Finally, Lee’s paper looks at the acquisition of quantification by three- year-old children. 4.1 Correlations between different types of quantification Chapter 1, “Numeral classifiers with plural marking. A challenge to Greenberg” by Walter Bisang, starts out with data on Northern Kam and Weining Ahmao, in which numeral classifiers obligatorily mark number. These findings are prob- lematic to Greenberg’s (1974: 25) universal claim that “[n]umeral classifier lan- guages generally do not have compulsory expression of nominal plurality, but at most facultative expression.” Even though plural marking is not fully obliga- tory in these languages, the fact that there are processes of grammaticalization which support a classifier system that carries obligatory information on number shows the weakness of Greenberg’s claim. After a discussion of the diachronic language-internal processes that produced the classifier systems of Northern Kam and Weining Ahmao, the paper will offer some general observations con- cerning the status of implicational universals. It supports Newmeyer’s (2005) view that typological generalizations must be seen as “holistic” in the sense that there is only an indirect correlation between implicational universals and rules, constraints and the properties of constructions in individual languages. Chapter 2, “Reduplication in languages: A case study of languages of China” by Dan Xu, deals with reduplication from a typological perspective. Reduplica- tion is broadly attested in human languages, especially in the Southern Hemi- sphere. In languages where reduplication and classifiers are found extensively, plural marking is not well-developed, and is sensitive to the semantic feature [+human]. The more plural marking is developed, such as in Indo-European languages and Altaic languages, the less this semantic feature is required. The more a language possesses developed plural-marking morphology, the less it needs reduplication and classifiers. Reduplication is iconically motivated. Posi- tive degree constitutes its core meaning. Reduplication with negative meaning characterizes the grammaticalization process from icon to symbol. 4.2 Numeral classifiers and their diachronic development Chapter 3, “The Syntax and Prosody of Classifiers in Classical Chinese,” is written by Shengli Feng, who scrutinizes what mechanism might have triggered the rise of 12 Dan Xu
- 25. classifiers in Chinese. Based on Borer’s theory of nominal structure (2005: 95), the author notes a parallelism between light verbs and light nouns, i.e. classifiers. Plural marking and classifiers are in fact two sides of the same coin. Given the theory presented here, the author argues that pre-Archaic Chinese, though possessing neither plurality nor classifiers as traditionally observed, actually had some lexical markers that served to individualize nouns, which may confirm the hypothesis that there is no language which has neither a plural morphology nor a classifier system. It is further argued that the emergence of classifiers in Late Archaic Chinese, though syntactically licensed, was prosodically motivated, and in turn, constitutes a sub-case of a typological change from the synthetic characteristics of Pre-Archaic Chinese (before 1000 BC) to the analytical charac- teristics of Post-Archaic Chinese, around the time of the Eastern Han (25–220 AD). Chapter 4, “Individuating Classifiers in Early Southern Min (14th–19th cen- turies)” by Lin Jang-Ling and Alain Peyraube, analyzes the evolution of the nature and function of the individuating classifiers in Southern Min from the 14th c. to the 19th c. The paper is a comparative analysis of different versions of Lao Qida (15th–18th c.), which represents a Chinese Northern dialect and several versions of Li Jing Ji (16th–19th c.), an opera in Southern Min dialect. The authors conclude that the classifiers used in the Southern Min documents are more diversified than those seen in the Northern Chinese of the same period. Furthermore, the occurrences of classifiers found in the former are more numer- ous than those attested in the latter. The hypothesis of cyclic change of classi- fiers put forward by Peyraube (1998) obviously seems to be less evident in the southern dialects than in the northern ones. 4.3 The expression of plurality Chapter 5, “Plurality and the subclassification of nouns in Classical Chinese” by Christoph Harbsmeier, provides a survey of the role of the semantic categories “plural” and “singular” in Classical Chinese, an uninflected language which does not generally use explicit markers for the plural. Harbsmeier gives a meticulous analysis of commonly used words in classical Chinese revealing their singular/ plural meaning, which has often been ignored by readers and even by researchers. The author’s work is inspired by the Danish linguist Otto Jespersen, who, in his historical grammar of English, elaborated a subclassification of English nouns that can provide inspiration for a proper treatment of the subclassification of nouns in Chinese historical grammar. Chapter 6, “Number in Chinese: a diachronic study of the case of 諸 zhū from Han to Wei Jin Nanbeichao Chinese” by Barbara Meisterernst, discusses Introduction: Plurality and Classifiers across languages of China 13
- 26. the syntactic and semantic constraints of the plural word zhū. The author indi- cates that semantically the function of zhū is to refer to a well-defined plural set of items, and that syntactically, zhū has to be analyzed as an adjective rather than a determiner since it can combine with determiners such as demonstratives and personal pronouns. Using the Animacy Hierarchy framework, Meisterernst shows that personal pronouns, which are situated at the top of the Animacy Hierarchy, are not compatible with zhū. Zhū underwent an important extension from [+animate] [+human] to [–animate]. According to the data from the Wei Jin Nanbeichao (220–589 AD) periods, and especially from the early 5th century Buddhist text Gaoseng Faxian zhuan, the author concludes that pluralization obviously was not the only function of zhū, at least during that period. Chapter 7, “Bu-tong ‘different’ and nominal plurality in Mandarin Chinese” by Marie-Claude Paris, studies a pair of two Mandarin (apparent) synonyms — 不同 bùtóng ‘different’ and 不一样 bù yíyàng ‘different’. Although their mean- ing seems at first sight to be equivalent, bùtóng and bù yíyàng actually behave very differently, both at the syntactic and at the semantic levels. While they can be substituted for each other in some cases, in other cases they cannot; hence their difference is established. Bù yíyàng can only be used as a scalar relative predicate or as a predicate modifier, whereas bùtóng can function both as an absolute predicate and as a noun determiner. This paper concludes that (i) plurality and number constitute two different markings, whose distribution is not complementary in Mandarin and that (ii) bùtóng can, in some contexts, be labeled a ‘redundant’ plurality marker. Chapter 8, “Plurality in the pronominal paradigms of Bai dialects” by Jingqi Fu, investigates plurality in Bai, a Sino-Tibetan language whose linguistic affilia- tion is debatable. Examining how plurality and case are expressed in the prono- minal paradigms of Bai dialects, she proposes to derive the plural pronouns, possessive and objective pronouns by a complex form of pronoun base followed by an ending: plural morpheme, possessive and locative markers respectively, contrary to the analysis where the inflected plural is independently derived via vowel inflection. In doing so, she accounts for the different forms of plural pronouns by the differences in the plural endings across dialects. She also inves- tigates how plural formation interacts with case (objective, subjective and geni- tive). This account is consistent with the general characteristic of Bai of being a non-inflected language and puts Bai on a par with other Tibeto-Burman languages which developed the objective case later. Chapter 9, “Frequentative Aspect and Pluractionality in Nuosu Yi” by Hong- yong Liu and Yang Gu, explores plurality’s relationship with frequentative aspect. Nuosu Yi belongs to the Tibeto-Burman language group and possesses three different means to encode a frequentative event: by using a preverbal tem- 14 Dan Xu
- 27. poral adverb, a postverbal frequentative aspect marker (like the West Green- landic suffix “-tar”), or by both of them co-occurring in the same sentence. The authors propose a unified way to explain these three different means. Following Van Geenhoven (2004, 2005), the authors argue that the Nuosu Yi frequenta- tive aspect marker is a pluractional operator, but the preverbal adverb is not. Because of semantic overlap between the preverbal adverb and the frequenta- tive aspect marker, the authors study in detail the semantic and syntactic con- straints on the non/co-occurrence of these two markers in Nuosu Yi. Chapter 10, “Quantificational structures in three-year-old Chinese-speaking children” by Thomas Hun-tak Lee, investigates the realizations of existential quantification, universal quantification, negation and modals in children’s lan- guage. The experiments are performed on Mandarin-speaking and Cantonese- speaking children from one and a half to three and a half years of age. Cognitive structures relative to these quantifications are explored. The standard quantifica- tional structures have been much expanded to describe restricted quantification, binding of sets and predicates, questions and various kinds of modality. Based on rich data on child language, the author makes some remarks on how the logical operators are realized in Chinese, providing evidence for these operators. Lee’s data show that existential quantification occurs earlier than universal quantification. This book is expected to have an impact on the study of linguistic typology, language contact, and patterns of the evolution (grammaticalization, lexicaliza- tion) of languages and linguistic areas. More and more linguists are realizing that forming conclusions on patterns in general linguistics without a better understanding of Chinese and languages of China is risky and dubious. This volume may also provide useful material for cognitive sciences and anthropol- ogy. Typological studies require a solid description of Chinese and the languages of China. Investigation of quantification in diachronic Chinese and non-Han lan- guages will offer general linguists a schema complementary to other languages of the world. References Andler, Daniel. 2004. Reprint. Introduction aux sciences cognitives [Introduction to cognitive sciences]. Paris: Editions Gallimard, 1992. Behr, Wolfgang. 2006. “Homosomatic juxtaposition” and the problem of “syssemantic” (huì yì ) characters. In Ecriture chinoise. Données, usages et représentations [Chinese writing, data, use and representations], Françoise Bottéro and Redouane Djamouri (eds.), 75–114. Paris: EHESS/CRLAO. Bisang, Walter. 1993. Classifiers, Quantifiers and Class Nouns in Hmong. Studies in Languages 17: 1–51. Introduction: Plurality and Classifiers across languages of China 15
- 28. Borer, Hagit. 2005. In Name Only. Oxford: Oxford University Press. Bybee, Joan, Perkins Revere and Pagliuca William. 1994. The Evolution of Grammar. Chicago/ London: University of Chicago Press. Cai, Jinghao. 1990. Chongtan yuzhuci ‘kan’ de qiyuan [Again on the origin of the particle kan], Zhongguo Yuwen 1: 75–76. Cao, Zhiyun (eds.). 2008. Hanyu fangyan ditu ji [Chinese dialects Atlas]. Beijing: Commercial Press. Carlson, Robert. 1991. Grammaticalization of Postpositions and Word Order in Senufo Languages. In Approaches to Grammaticalization, Elizabeth Closs Traugott and Bernd Heine (eds.), vol. II: 201–223. Amsterdam/Philadelphia: John Benjamins Publishing Company. Chao, Yuen Ren. 1968. A Grammar of Spoken Chinese. Berkeley/Los Angeles: University of California Press. Chen, Chusheng. 2004. Jinwen changyong zidian [Dictionary of inscriptions on bronze in common use]. Xi’an: Shanxi renmin chubanshe. Cheng, Xiangqing. 1992. Bianwen fuyinci yanjiu [dissyllabic words in the Bianwen]. In Sui Tang Wudai hanyu yanjiu [Studies of Chinese in the Sui, Tang and Wudai], Cheng Xiangqing (ed.), 1–132. Jinan: Shandong Jiaoyu chubanshe. Clark, Herbert H. 1973. Space, Time, Semantics and the Child. In Cognitive Development and the Acquisition of Language, Timothy E. Moore (ed.), 27–63. New York/London: Academic Press. Comrie, Bernard. 2008. The Areal Typology of Chinese: Between North and Southeast Asia. In Chinese Linguistics in Leipzig CLE 2, Redouane Djamouri, Barbara Meisterernst and Rint Sybesma (eds.), 1–21. Paris: EHESS/CRLAO. Corbett, Greville. 2000. Number. Cambridge: Cambridge University Press. Croft, William. 1990. Typology and Universals. Cambridge: Cambridge University Press. Dai, Qingxia. 1990. Yimianyu de jiegou zhuci [The grammatical markers in Tibeto-Burman languages]. In Zangmian yuzu yuyan yanjiu [Research on Tibeto-Burman languages], 83–97. Beijing: Minzu chubanshe. Downing, Pamela. 1996. Numeral Classifier Systems: The case of Japanese. Amsterdam/ Philadelphia: John Benjamins Publishing Company. Dryer, Matthew. 1989. Plural words. Linguistics 27: 865–895. Dryer, Matthew. 1992. The Greenbergian word order correlations. Language 68 (1): 81–138. Dryer, Matthew. 2003. Word order in Sino-Tibetan languages from a typological and geographical perspective. In The Sino-Tibetan languages, Graham Thurgood and Randy J. LaPolla (eds), 43–55. London/New York: Routledge. Enfield, Nick J. 2005. Areal Linguistics and Mainland Southeast Asia. Annual Review of Anthropology 34: 181–206. Feng, Chuntian. 1991. Jindai hanyu yufa wenti yanjiu [Grammatical studies in Modern Chinese]. Jinan: Shandong jiaoyu chubanshe. Greenberg, Joseph. 1972. Numeral classifiers and substantival number: problems in the genesis of a linguistic type. Language Universals 9: 1–39. Greenberg, Joseph. 1974. Numeral classifiers and substantival number: problems in the genesis of a linguistic type. In Proceedings of the 11th International Congress of Linguistics, Bologna – Florence, Aug–Sept 1972, Bologna 1974: 17–37. (Reprinted in Greenberg 1990: 16–93). 16 Dan Xu
- 29. Greenberg, Joseph. 1988. The first person inclusive dual as an ambiguous category. Studies in Language 12: 1–18. Greenberg, Joseph. 1990. On language. In Selected writings of Joseph H. Greenberg, Denning, K. and S. Kemmer (eds.). Stanford: Stanford University Press. Guo, Pan. 2004. Hanyu sheshu wenti yanjiu [Studies on numbers in Chinese]. Beijing: Zhonghua shuju. Guo, Wengao. 1995. ‘Yu’ qianzhi yu shuci de yongfa. [‘Yu’ used before numbers]. Zhongguo Yuwen 5: 367–368. Hagège, Claude. 2010. Adposition. Oxford: Oxford University Press. Haiman, John. 1980. The Iconicity of grammar: Isomorphism and motivation. Language 56 (3): 515–540. Haiman, John (ed.). 1985. Iconicity in Syntax. Amsterdam/Philadelphia: John Benjamins Publishing Company. Heide, Wegener. 1995. The German plural and its acquisition in the light of markedness theory. In The Development of Morphology Systematicity. A Cross-Linguistic Perspective, Hanna Pishwa and Karl Maroldt (eds), 247–261. Tübingen: Gunter Narr. Heine, Bernd and Tania Kuteva. 2002. World lexicon of grammaticalization. Cambridge: University of Cambridge. Heine, Bernd, Ulrike Claudi and Friederike Hünnemeyer. 1991. From Cognition to Grammar- Evidence from African Languages. In Approaches to grammaticalization, Elizabeth Closs Traugott and Bernd Heine (eds.), Volume 1: 149–187. Amsterdam/Philadelphia: John Benjamins Publishing Company. Hopper, Paul J. and Sandra A. Thompson. 1980. Transitivity in grammar and discourse. Language 56: 251–299. Iljic, Robert. 2001. The problem of the suffix-men in Chinese grammar. Journal of Chinese Linguistics 29 (1): 11–68. Jiang, Lansheng. 1995. Shuo “mo” yu “men” tongyuan [“Mo” and “men” have the same origin] Zhongguo Yuwen 3: 180–190. Li, Jinfang and Suhua, Hu. 2005. Han Zang yuxi liangci yanjiu [Studies on classifiers in Sino- Tibetan]. Beijing: Zhongyang Minzu daxue chubanshe. Li, Yanhui and Yuzhi, Shi. 2000. Hanyu liangci xitong de jianli yu fushu biaoji “men” de fazhan [The establishment of the classifier system and the development of the plural marker “men”]. Dangdai yuyanxue 2: 27–36. Liu, Xiaonong. 1992. Tang Wudai ci dieyin chutan [Preliminary study of reduplicated sounds in Tang and Wudai]. In Sui Tang Wudai hanyu yanjiu [Studies of Chinese in Sui, Tang and Wudai], Cheng Xiangqing (ed.), 440–471. Jinan: Shandong Jiaoyu chubanshe. Lü, Shuxiang and Jiang Lansheng. 1985. Jindai hanyu zhidaici [Demonstratives and pronouns in Modern Chinese]. Shanghai: Xuelin chubanshe. Lü, Shuxiang. 1984. Hanyu yufa lunwenji [Collected papers on Chinese grammar]. Beijing: Commercial Press. Ma, Guoquan. 1979. Liang Zhou tongqi mingwen shuci liangci chutan [A preliminary study on numerals and classifiers in bronze inscriptions during the Eastern and Western Zhou]. Gu wenzi yanjiu 1: 126–136. Mei, Tsu-lin. 1986. Guanyu jindai hanyu zhidaici [About demonstratives and pronouns in Modern Chinese]. Zhongguo Yuwen 6: 401–412. Montaut, Annie. 2009. Réduplication et constructions en écho en hindi/ourdou [Reduplication and echo constructions in Hindi/Urdu]. Faits de Langue/Les Cahiers 1: 9–44. Introduction: Plurality and Classifiers across languages of China 17
- 30. Newmeyer, Frederick J. 2005. Possible and probable languages: A generative perspective on linguistic typology. Oxford: Oxford University Press. Ortmann, Albert. 2000. Where plural refuses to agree: Feature unification and morphological economy. Acta Linguistica Hungarica, Volume 47: 249–288. Peyraube, Alain. 1998. On the History of Classifiers in Archaic and Medieval Chinese. In Studia Linguistica Serica, Benjamin K. Tsou (ed.), 39–68. Hong Kong: City University of Hong Kong. Peyraube, Alain and Wiebusch Thekla. 1993. Le rôle des classificateurs nominaux en chinois et leur évolution historique: un cas de changement cyclique [The role of nominal classifiers in Chinese and their historical development: a case of cyclical change]. Faits de langues 2: 47–57. Rubino, Carl. 2005. Order of subject, object and verb. In The world atlas of language structures, Martin Haspelmath, Matthew S. Dryer, David Gil and Bernard Comrie (eds.). Oxford: Oxford University Press. Sagart, Laurent. 1999. The Roots of Old Chinese. Amsterdam/Philadelphia: John Benjamins Publishing Company. Sanches, Mary. 1973. Numeral classifiers and plural marking : an implicational universal. Language Universals 11: 1–22. Singh, Rajendra. 2005. Reduplication in Modern Hindi and the theory of reduplication. In Grammar, Comparative and general–Reduplication, Bernard Hurch (ed.), 263–281. Berlin/ New York: Mouton de Gruyter. Sun, Hongkai. 2001. Lun Zangmian yuzu zhong de qiangyuzhi yuyan [On the Qiangic branch of the Tibeto-Burman languages]. Languages and linguistics 2: 1, 157–181. Sun, Hongkai, Huang Xing and Hu Zengyi (eds). 2007. Zhongguo de yuyan [Languages in China]. Beijing: Commercial Press. Svorou, Soteria. 1993. The Grammar of Space. Amsterdam/Philadelphia: John Benjamins Publishing Company. Tsou, Benjamin. 1976. The structure of nominal classifier systems. In Austroasiatic Studies, Philip N. Jenner, Laurence C. Thompson, and Stanley Starosta (eds.), Part II: 1215–1247. Honolulu: University of Hawaii Press. Talmy, Leonard. 2000a. How Language Structures Space. Toward a Cognitive Semantics, Volume 1: 177–254. Cambridge/Massachusetts: Massachusetts Institute of Technology Press. Talmy, Leonard. 2000b. A Typology of Event Intergration. Toward a Cognitive Semantics, Volume 2: 213–288. Cambridge/Massachusetts: Massachusetts Institute of Technology Press. Van Geenhoven, Veerle. 2004. For-adverbials, frequentative aspect, and pluractionality. Natural Language Semantics 12: 135–190. Van Geenhoven, Veerle. 2005. Atelicity, pluractionality and adverbial quantification. In Perspectives on Aspect, Henk Verkuyl, Henriëtte de Swart and Angeliek van Hout (eds.), 107–124. Netherlands: Springer. Wang, Ying. 1996. Tangshi zhong de dongci chongdie [Reduplicated verbs in Tang poetry], Zhongguo Yuwen 3: 233–234, 240. Wang, Huayun. (preprint). Fushu ciwei “men” de laiyuan [Origin of the plural suffix -men]. Wu, Fuxiang. 1995. Changshitai zhuci “kan” de lishi kaocha [Diachronic examination of the tentative aspect particle “kan”]. Yuyan yanjiu 2: 161–166. Wu, Fuxiang. 2006. Wei Jin Nanbeichao shiqi hanyu mingliangci fanchou de yufahua chengdu [Degree of grammaticalization of nominal classifiers during the Wei Jin Nanbeichao periods]. In Linguistic Studies in Chinese and Neighboring Languages: Festschrift in Honor of Professor Pang-Hsin Ting on His 70th Birthday. Language and Linguistics Monograph Series Number W-6: 553–571. 18 Dan Xu
- 31. Xing, Fuyi. 1960. Lun “men” he “zhuwei” zhilei bingyong [On the coexistence of “men” he “zhuwei”]. Zhongguo Yuwen 6: 289, 292. Xu, Dan. 1994. Guanyu hanyu li dongci + X + didianci de juxing [On the V + X + locative pattern in Chinese]. Zhongguo Yuwen 3: 180–185. Xu, Dan. 2006. Typological change in Chinese syntax. Oxford: Oxford University Press. Xu, Dan. 2009a. Cong Zhanguo zonghengjia shu kan Xi Han chuqi fushu gainian de biaoda [Plurality in Chinese during the early Han as seen in the Zhanguo zonghengjia shu], Lishi yuyanxue yanjiu, Volume 2. Beijing: Commercial Press. Xu, Dan. 2009b. Xiayou Shixingyu de mouxie tedian [Some syntactic features of Down-river Shixing]. Minzu Yuwen, 1: 25–42. Xu, Dan. 2010. Cong yuyan leixing kan hanyu fushu xingshi de fazhan [The development of plurality from a typological perspective]. In: Dan Xu (ed.), Liang yu fushu de yanjiu [Quantification and plurality], 90–112. Beijing: Shangwu yinshuguan. Xu, Dan (ed.). 2008. Space in Languages of China: Cross-linguistic, synchronic and diachronic perspectives. Dordrecht: Springer. Xu, Dan (ed.). 2010. Liang yu fushu de yanjiu [Quantification and plurality]. Beijing: Shangwu yinshuguan. Xu, Dan, and Fu Jingqi. 2012. Classifiers and some typological considerations. To appear in Breaking down the barriers: Interdisciplinary studies in Chinese linguistics and beyond. (Language and Linguistics Monograph Series) Taipei: Institute of Linguistics, Academia Sinica. Xu, Zhongshu (ed.). 1989. Jiaguwen zidian [Dictionary of bone inscriptions]. Chengdu: Sichuan cishu chubanshe. Zhang, Meilan. 2001. Jindai Hanyu houzhui xingrongci [Suffixed adjectives in Modern Chinese]. Guiyang: Guizhou jiaoyu chubanshe. Zhao, Jie. 1989. Xiandai Manyu yanjiu [Study on Contemporary Mandchou], Beijing: Minzu chubanshe. Introduction: Plurality and Classifiers across languages of China 19
- 33. I Correlations between different types of quantification
- 35. Walter Bisang 1 Numeral classifiers with plural marking. A challenge to Greenberg Abstract: Classifiers of Northern Kam languages show a clear distinction between singular and plural. Classifiers in Weining Ahmao do the same but they not only inflect for number (singular vs. plural) but also for definiteness and size (Gerner and Bisang 2008). The cooccurence of systematic number marking and numeral classifiers casts some doubt on Greenberg’s (1974) universal on the incompatibil- ity of numeral classifiers with obligatory plural marking. This paper will show how language internal processes of grammaticalization can produce structures which seem to be typologically rare or even impossible. Since these processes are characterized by the interaction of various levels of grammar (at least phonetics/ phonology, morphology, syntax, semantics), the paper also shows that a typology that singles out individual levels of grammar is doomed to failure. Keywords: numeral classifiers, numeral classifiers in Niger-Congo languages, obligatory plural marking, mass/count distinction, (in)definiteness, classifier inflection, typological universals 1 Introduction: The compatibility of numeral classifiers with number marking on the noun 1.1 The general assumption Numeral classifiers are an important areal feature of East and mainland South- east Asian languages. In most languages belonging to this area, nouns can only be counted if they are combined with a numeral classifier. Thus, a Chinese noun like xìn ‘letter’ must occur with its classifier fēng (1a) in the context of counting and with some other operations of quantification, otherwise (1b) the construc- tion is ungrammatical. The same applies to the Thai noun còtmǎay ‘letter’ with its classifier chabàp in (2): (1) a. sān fēng xìn b. *sān xìn three CL letter three letter ‘three letters’ I.M.: ‘three letters’
- 36. (2) Thai a. còtmǎay sǎam chabàp b. *còtmǎay sǎam letter CL three letter three ‘three letters’ I.M.: ‘three letters’ The classical explanation for examples like (1) and (2) and for the existence of numeral classifiers in general was given by Greenberg (1974, also cf. Sanches and Slobin 1973), who claims a universal implicational relation between numeral classifiers and the lack of obligatory plural marking: Numeral classifier languages generally do not have compulsory expression of nominal plurality, but at most facultative expression (Greenberg 1974: 25). 1.2 Numeral classifiers are compatible with number marking on the noun Even though Greenberg’s (1974) universal claim had and still has an important impact on typological research in numeral classifier systems, it does not apply universally. Greenberg (1974) himself is aware of this when he adds the adverb ‘generally’ to his above implicational universal and thus signals that there may be exceptions. And indeed, there are exceptions. Aikhenvald (2000: 100) points out in her compendium of classification systems that there are even quite a few languages all over the world in which numeral classifiers and obligatory number marking coexist. Aikhenvald (2000: 100) mentions South Dravidian languages, Nivkh (Paleosiberian isolate), some Algonquian languages, some South American languages (Tucano, North Arawak) and various languages “which combine numeral classifiers with noun class systems.” Since it is not possible to illustrate each language in a short paper like this, only the last type of languages will be briefly discussed. This type is represented in some Bantoid languages (Niger- Congo: Benue-Congo) with their noun class system as one of their genetic characteristics. Since many noun classes of these languages consist of pairs of obligatory prefixes in which one stands for singular and the other for plural, number is firmly integrated into the system. If a language of this type with its inbuilt singular/plural distinction additionally has numeral classifiers the impli- cational universal claimed by Greenberg (1974) is automatically violated. One of the languages in which this is the case is Ejagham, a language spoken in the Cross River Basin in Nigeria and Cameroon (Watters 1981). This language has nineteen noun classes. Twelve of them are double class, i.e., they consist of a singular marker and a plural marker. In the case of noun class 3/6 for long thin objects and for natural objects, the singular prefix (class 3) is N- and the plural prefix (class 6) is a- as in ɴ̀-čɔ̂g ‘pestle’ vs. à-čɔ̂g ‘pestles’ (Watters 1981: 294). 24 Walter Bisang
- 37. In addition to these noun classes, Ejagham has five numeral classifiers which are derived from nouns and have their own noun-class prefixes. One of them is the classifier ɴ̀-sûm (singular)/à-sûm (plural) of noun class 3/6 for “any fruit or root which is long” (Watters 1981: 310). As can be seen from (3) on the overall structure of the noun phrase, the numeral classifier takes the initial position of the noun phrase: (3) CL GEN-LINK N NUM The genitive linker (GEN-LINK) between the classifier and the noun is con- trolled by the noun class of the classifier (CL). The numeral, which is situated at the end of the construction1, also agrees with the class of the classifier. In the following example, the noun -tə̀bɛ̀ ‘okra pod’, which belongs to noun class 9/14 (singular: ɴ̀-tə̀bɛ̀; plural: ɔ̀-tə̀bɛ Watters 1981: 300) is presented in com- bination with the numerals -d ‘one’ and -bá1ɛ́ ‘two’. As can be seen, the head noun and the classifier are both marked for number: (4) The classifier Ǹ-sûm (sg.)/à-sûm in Ejagham a. ɴ̀-sûm ì ɴ̀-tə̀bɛ́ yə́-d NCL3:SG-CL:fruit GEN-LINK NCL9:SG-okra.pod NCL3:SG-one ‘one okra pod’ b. à-sûm ´ ɔ́-tə̀bɛ́ á-bá1ɛ́ NCL6:PL-CL:fruit GEN-LINK NCL14:PL-okra.pod NCL6:PL-two ‘two okra pods’ (Watters 1981: 310) The numeral classifier system of Ejagham is not pervasive. As Watters (1981: 313) points out, “most nouns do not use a classifier when being enumerated, but for the various types of nouns listed above, e.g. seeds, grains, kernels, nuts, long fruits and roots, round fruits and roots, clusters of fruits, plants, trees and vegetables a classifier is generally required”. 1 The structure of the classifier construction in Ejagham additionally contradicts Greenberg’s (1974: 31) claim that the classifier and the numeral must form a continous constituent. As can be seen from (3) and (4), word order in the classifier construction of Ejagham is [CL-N-NUM]. While classifiers form a coherent unit with the numeral in East and mainland Southeast Asian languages, they establish a more coherent relation with the noun to be counted in Ejagham and other Niger-Congo languages. In Ejagham, this can be seen from the genitive-linker, which agrees with the classifier and creates a head-modifier relation between the classifier and the classified noun. The coherence between the numeral and the classifier is maintained by the agreement relation as it is controlled by the class-membership of the classifier. Numeral classifiers with plural marking. A challenge to Greenberg 25
- 38. There are other Niger-Congo languages in which the numeral-classifier sys- tem is pervasive, i.e., each count noun needs to have a classifier in the context of enumeration. One language of this type is Kana, another Cross River language spoken in the South of Nigeria, which has lost the Bantu noun class system (Ikoro 1994, 1996):2 (5) Kana a. bàɛ̀ bēè ŋwíí two CL:fruit child ‘two children’ b. tāà té ítòbtòb three CL:tree motorcycle ‘three motorcycles’ (Ikoro 1996: 91–92) Even if the use of classifiers is limited to certain semantic domains in Ejagham, the combination of obligatory number marking and numeral classifiers remains a challenge to Greenberg’s (1974) claim. This challenge may be mitigated to a certain extent if it turns out that fully pervasive systems of numeral classifiers are limited to those Niger-Congo languages in which the obligatory opposition of singular vs. plural reflected in the noun-class system has vanished (see Section 4). In spite of a considerable amount of counter-evidence against Greenberg’s implicational claim from various languages and areas across the globe, there hasn’t been much counter-evidence from the area of East and mainland South- east Asia. More recent research on Northern Kam (Gerner 2006) and on Weining Ahmao (Gerner and Bisang 2008, 2009, 2010) shows that there is evidence against this universal claim even in the isolating languages of that area. In fact, the few instances of morphology that developed in these two languages are concerned with the morphophonological form of the classifier and – what is even more remarkable – with the emergence of singular/plural-distinctions that are marked on the classifier itself. As will be shown in the next section, classifier morphology is the result of fusion with other words or markers which lose their syntactic independence. Such processes are not limited to the development of classifier inflection, they are also attested in other domains of grammar. One of 2 As can be seen from (5), Kana does not violate the word-order constraints claimed by Greenberg (1974: 31) (cf. endnote 1). Its word order is [NUM-CL-N]. In spite of this, the fusion between CL and N is higher than between NUM and CL. There is a genitive linker in the form of a floating low tone which takes the final syllable of the head of the construction, i.e., the classifier (CL) followed by the noun (N) as its dependent. Thus, the form bēè ‘CL:fruit’ in (5a) is a combination of [bēē plus low tone]. In the case of té ‘CL:tree’ in (5b), the low tone does not show any effects. 26 Walter Bisang
- 39. them is discussed by Fu (in this volume) on the integration of plural markers into pronominal paradigms of Bai. It is needless to say that the integration of plural marking into the morphophonology of classifiers or pronominal systems considerably weakens Greenberg’s universal claim. The paper is organized as follows: Section 2 presents the classifier systems of Northern Kam (see Subsection 2.1) and Weining Ahmao (see Subsection 2.2). Section 3 describes the factors that operate against Greenberg’s (1974) universal claim. It will turn out that none of these factors are associated with cognitive motivations that are generally mentioned in linguistic typology. A short con- clusion in section 4 will take up the discussion of the validity of typological generalizations where there are various factors that potentially undermine them. 2 The data from Northern Kam and Weining Ahmao 2.1 Number-sensitive numeral classifiers in Northern Kam Northern Kam belongs to the Kam-Tai branch of the Kadai language family.3 Like most of the other East and mainland Southeast Asian languages, Northern Kam is an isolating language. There is only one exception, and that exception shows up in the numeral-classifier system. Since this is nicely described by Gerner (2006), I will only summarise his findings in the present subsection. Numeral classifiers in Northern Kam show a regular morphophonological contrast between singular and plural forms. The initial consonant of the singular form is almost always a glide ([w], [j]) or a voiced fricative ([ʑ], [ɤ]), while the plural form of classifiers starts with a voiceless stop or a nasal. This distinction is made only in Northern Kam. The numeral classifier system of Southern Kam does not have such a distinction. Since the single classifier form of Southern 3 There are two different classifications of this language family. The older one is from Edmondson and Solnit (1988) and distinguishes three subfamilies, i.e., Kam-Tai, Hlai (= Li) and Geyang. In this classification, which is also called the “Kadai hypothesis”, Kam-Tai is further divided into Kam-Sui (to which Kam belongs) and Tai. A more recent classification by Chamberlain (1997), which is sometimes dubbed “Tai-Kadai hypothesis”, splits Kam-Tai into two independent subbranches and thus distinguishes four subfamilies, i.e., Tai, Kam-Sui, Hlai and Kadai. This classification is known under the heading of “Zhuang-Dong” in the Chinese context. Since Gerner’s (2006) comparative analysis of core classifiers in 22 Kam-Tai languages provides supportive evidence for the Kadai hypothesis of Edmondson and Solnit (1988), I adopt this classification. My above description of the genetic situation is from Gerner (2006: 239). Numeral classifiers with plural marking. A challenge to Greenberg 27
- 40. Kam systematically corresponds to the plural form of Northern Kam, the singu- lar form of Northern Kam must be a special, historically derived form which only developed in that language (Gerner 2006: 244–247). The derivational rules are described in detail in Gerner (2006). For that reason, only a few examples will be presented in this paper: (6) Singular vs. plural form of Northern Kam (Gerner 2006: 244–245)4 Derivation Class meaning CL:SG CL:PL a. [w] <– [p] human wəu45 pəu45 b. [w] <– [m] dual body parts waŋ55 maŋ55 c. [w] <– [m] clothes wəi31 məi31 d. [j] <– [ɕ] entities with handle jaŋ45 ɕa ŋ45 e. [j] <– [t ˛] 1-dim entities jiu22 t ˛iu22 f. [ʑ] <– [n] 3-dim entities ʑən45 nən45 g. [ɣ] <– [ʔ] small diverse range ɣa55 ʔa55 As is to be expected from the above description, the singular form of the classifier is used with the numeral i45 ‘one’ in (7a). For numbers greater than one, the plural form is employed in (7b): (7) Northern Kam (Gerner 2006: 243–244; for the form of the classifier, cf. [5e]) a. i45 jiu22 na45 one CL:SG river ‘one river’ b. ham11 t ˛iu22 na45 three CL:PL river ‘three rivers’ Numeral classifiers also occur in [classifier + noun] constructions. In this construction, they mark indefiniteness. The singular form stands for singular indefinite (8a) and (9a), while the plural form expresses plural indefinite in (8b) and (9b): 4 The tones are represented by number combinations. The numbers represent relative pitch on a scale from 1 (lowest) to 5 (highest). The first number represents the beginning of the tonal contour, the second number stands for its end. 28 Walter Bisang
- 41. (8) Northern Kam (Gerner 2006: 249; for the form of the classifier, cf. [6c]) a. wəi31 tu33 b. məi31 tu33 CL:SG garment CL:PL garment ‘a garment’ ‘garments’ (9) Northern Kam (Gerner 2006: 249; for the form of the classifier, cf. [6g]) a. ɣa55 hoŋ22 b. ʔa55 hoŋ22 CL:SG loom CL:PL loom ‘a loom’ ‘looms’ In other grammatical contexts, only the plural form of the classifier is used. In the case of the demonstrative construction, the classifier marks singular, even though the plural form is used: (10) Northern Kam (Gerner 2006: 251; for the form of the classifier, cf. [6b]) maŋ55 ta45 ai33 CL:PL eye DEM:this ‘this eye’ The situation as it has been described so far is remarkable in at least two respects. First, it is typologically rare that the plural form provides the basic form from which the singular is derived. Of course, singulative forms are attested in a considerable number of languages but this is usually limited to a subset of nouns. In Northern Kam, the derivation of the singular from the plural is pervasive. Second, the form which explicitly marks plural in the [numeral + classifier]-construction (7b) and in the [classifier + noun]-construction ([8b] and [9b]) marks singular in the demonstrative construction (10). Both of these prop- erties can be accounted for if one looks at the historical development of the number distinction in Northern Kam classifiers. Based on Shi (1997), Gerner (2006: 244) convincingly argues that the singular/ plural-distinction of numeral classifiers developed “when the ancestor classifiers collocated with the numeral i45 ‘one’”. The new classifier forms that emerged from the collocation with that numeral were reanalysed as singular forms in contrast to the original form which collocated with all other numerals and did not undergo any morphophonological change. Since this unchanged form occurred with all numerals higher than one, it was only natural to associate this form with plural Numeral classifiers with plural marking. A challenge to Greenberg 29
- 42. meaning. Thus, the typologically remarkable situation that the singular is derived from the plural synchronically is due to a morphophonological process of change. Once the distinction of singular vs. plural was established with numerals, it moved into the [classifier + noun]-construction with its indefinite meaning. This scenario is quite plausible if one takes into account that the con- struction [one + classifier + noun] is used for expressing indefiniteness in a vast number of languages (Dryer 2005). As soon as the [singular classifier + noun]- construction was associated with indefiniteness as a short form of the [one + classifier + noun]-construction, the ancient form of the classifier was reanalysed as a plural indefinite marker in the [plural classifier+ noun]-construction. As can be seen from example (10), the number distinction is not fully gram- maticalized through all constructions which take a numeral classifier. In these constructions, the plural classifier is used even though it does not necessarily mean plural. If one takes the plural form as the original form of the classifier it does not come as a surprise that this form occurs in all grammatical contexts in which the singular/plural-distinction remains irrelevant. As I will argue in the remainder of this subsection, the fact that it actually marks singular in the demonstrative construction can be accounted for by a look at Thai as another Kam-Tai language. In Thai, numeral classifiers are optional in the demonstrative construction. If they occur with demonstratives there is a very strong tendency to interpret them as singulative markers (Hundius and Kölver 1983, Becker 2005). Thus, (11a) with no classifier can be singular or plural, while (11b) with a classifier is most likely to be singular: (11) Thai a. rót níi b. rót khan níi car this car CL this ‘this car/these cars’ ‘this car’ If one assumes that the situation in Northern Kam was similar to that in Thai before the emergence of number-sensitive classifiers, the occurence of a classifier in the demonstrative construction also triggered a singular interpreta- tion in Northern Kam. Since the singular/plural-distinction did not go beyond the numeral construction and the [classifier + noun]-construction, the original form is still the only option for the demonstrative construction. The seemingly paradoxical case that the form used with plural function in some constructions triggers a singular interpretation in the demonstrative construction can thus be resolved by the argument that there is no other form available for the demon- strative construction and that that form triggers singular interpretation as it did before the singular/plural-distinction took shape. 30 Walter Bisang
- 43. 2.2 Classifier inflection in Weining Ahmao Weining Ahmao is a Miao language spoken in the Weining County of Western Guizhou. Its classifier system is described since Wang Fushi (1957 [1972])5 and Wang Deguang (1986, 1987) and it seems to be unparalleled in the world’s lan- guages (for more data and the analysis of the inflectional paradigm, see Gerner and Bisang 2008, 2010). Each of its classifiers is inflected for the categories of number (singular vs. plural), reference (definite vs. indefinite) and size (augmen- tative, medial, small). The forms concerning size also express social deixis in dialogues (not in narratives). Roughly speaking, male speakers use the augmen- tative form of the classifier, female speakers the medial form and children the diminutive form (for more details, see Gerner and Bisang 2008, 2010). The following table shows how the form of Ahmao classifiers varies within the twelve-slot paradigm determined by these three categories: Table 1: The inflectional paradigm of classifiers in Weining Ahmao (Gerner and Bisang 2008: 721; Gerner and Bisang 2010: 79) Gender/Age Register Size Singular Plural Definite Indefinite Definite Indefinite Male Augmentative CVT C*VT ti55 a11 CVT’ di31 a11 C*VT’ Female Medial Cai55 C*ai213 tiai55 a11 CVT’ diai213 a11 C*VT’ Children Diminutive Ca53 C*a35 tia55 a11 CVT’ dia55 a11 C*VT’ The definite singular augmentative form with its structure of CVT is the basic form of the classifier. It consists of a consonant (C), a vowel nucleus (V) and a lexically determined tone (T)6. Each of these elements can be subject to morphophonological change in the other slots. The symbol C* stands for changes in the properties of the consonant in some classifiers. Depending on the individual classifier, the properties involved are voicing and aspiration to distinguish indefinite from definite classifiers. The vowel quality [ai]7 is charac- teristic of medial size, while [a] marks diminutive size. Medial-definite forms are typically expressed with [55] tone, medial-indefinite forms exhibit a [213] tone. Diminutive-definite forms display the tones [53] (singular) or [55] (plural) and 5 The description of Wang Fushi (1957) slightly differs from the one presented in Gerner and Bisang (2008). Wang Fushi (1957) does not mention the distinction between definite and indefinite augmentative/male forms in the singular (cf. Tables 1 and 2). 6 On the notation of tones by the combination of numbers, cf. fn. 2. 7 Wang Fushi (1957) has [ae] instead of [ai]. Numeral classifiers with plural marking. A challenge to Greenberg 31
- 44. finally diminutive-indefinite forms are associated with the tones [35] (singular) and [55] (plural) (Gerner and Bisang 2008: 722). The plural forms are characterized by the marker ti55, which undergoes similar changes as the singular classifier forms described in table 1. This marker is followed by a11 plus the classifier in its form of C*VT’, in which T’ indicates that the tone takes a special form at least with some classifiers. There are some 40 to 50 numeral classifiers in Weining Ahmao. To illustrate how the pattern described above works with a concrete classifier, table 2 shows the inflectional forms of the classifier lu55, one of the most common classifiers which is used with nouns denoting inanimates. Table 2: The inflectional paradigm of the classifier lu55 in Weining Ahmao (Gerner and Bisang 2008: 722) Gender/Age Register Size Singular Plural Definite Indefinite Definite Indefinite Male Augmentative lu55 lu33 ti55 a11 lu55 di31 a11 lu55 Female Medial lai55 lai213 tiai55 a11 lu55 diai213 a11 lu55 Children Diminutive la53 la35 tia55 a11 lu55 dia55 a11 lu55 The following example shows how the different classifier forms are used in a text. The text in (12) is a Weining Ahmao version of the well-know fable of the fox and the crow. In the first line, the classifier tu44 for animals occurs in its indefinite (medial singular) form, since the noun a33 dy33 ‘fox’ is introduced into the text8. Later, it is taken up in the corresponding definite form tae33. The noun nG‘ae35 ‘meat’, which stands for the cheese in the western version of the fable, is already known from previous context and is thus marked by the definite (ts‘ae33: singular, medial) form of the classifier. The crow is also marked by the definite (ta33: singular, diminutive) form of the classifier because it has been introduced earlier. Since classifiers also occur in possessive constructions, we find the classifier lu55 in its definite (singular, diminutive) form la55 on the last line.9 Possessive constructions have the form [possessor CLi possesseei], in 8 In narratives, the selection of size depends on the real size of the objects described. In (12), the fox is marked by the medial-size form, not by the augmentative form. In principle, the selection of the size form depends on the view of the speaker. If she wants to point out the big size of a given nominal concept she can chose the augmentative form. Since foxes are not extremely big, the use of the medial form corresponds to the default. Once the size of a given element of a narrative is determined, this feature contributes to reference tracking. 9 Wang Fushi (1957) reports a different tone for this classifier form. As can be seen from table 2, the medial definite singular form of lu55 is la53 in Gerner and Bisang (2008) instead. 32 Walter Bisang
- 45. which the classifier reflects the properties of the possessee. The definiteness of the possessee a33 ‘ndʑ‘au35 ‘mouth’ can be inferred from its inalienable relation to the possessor. (12) Weining Ahmao t‘au33 i55 m’a35 i55 dae35 a33 dy33 d‘œy31 d‘a35. time that there.is one CL:INDEF fox exit come tae33 a33 dy33 ɲi55 la11 ae55 tʂ‘ae55 daɯ11, i55 vie33 ɲ‘i13 CL:DEF fox this also very hungry PF but he nʄ‘ie55 hi33 tau33 qɯ55 qa55 ʂi33 n’au35. ɲ‘i13 b‘o31 ts‘ae33 look.for not get food anything eat he see CL:DEF nG‘AE35 ku11 ɲo55 v’ae31 ta33 li55 a55 la55 a33 ‘ndʑ‘au35 i55, . . . meat REL at place CL:DEF crow CL:DEF mouth that ‘At that time a fox came out. He too became very hungry, but he had been unable to find anything to eat. When he saw the piece of meat in the crow’s mouth, . . .’ (Wang Fushi 1957: 106–107; 1972: 161–162) The text in (12) is a narrative text. In dialogues as the one in (13) below, the size value of classifiers is additionally associated with social deixis. Thus, male speakers normally use the augmentative forms, female speakers the medial form and child speakers the diminutive form (cf. above). While this is roughly the standard rule, the system is more flexible and can trigger a lot of politeness effects in specific situations. Example (13) presents a dialogue between two male speakers. The first speaker A uses the augmentative form of the classifier lu55 as one would expect it of a male speaker. In contrast, the second male speaker B, reacts with the corresponding female form of the classifier (lai55). The reason for this selection is straightforward if one takes into account that speaker B reacts to a compliment. Due to politeness, he opts for being modest and thus selects the female classifier which is associated with lower social status if used by a male speaker: (13) Weining Ahmao (Gerner and Bisang 2008: 728) A: gɦi31 lu55 ŋgɦa55 ɲi55 zau44 ku11. 2:PL CL:AUG:SG:DEF house DEM:PROX good very Male Speaker A: ‘Your house is so nice.’ B: qha55 tsau55 ku55 lai55 ŋgɦa55 ɲi55. NEG:IMP praise 1.SG CL:MED:SG:DEF house DEM:PROX Male speaker B: ‘Don’t praise my house.’ Numeral classifiers with plural marking. A challenge to Greenberg 33
- 46. Apart from its inflectional classifiers, Weining Ahmao is basically an isolat- ing language as most other East and mainland Southeast Asian languages. This begs the question of how the inflectional classifier system emerged. Gerner and Bisang (2009: 19–29; 2010: 86–90) suggest a two-stage scenario. The first stage induced the development of the three size distinction. It is charachterized by a rebracketing process followed by further processes of erosion. The second stage, which was responsible for the distinctions of definite vs. indefinite, started out from the combination of the numeral i55 ‘one’ with a classifier. The first stage is based on derivational markers which are used to form aug- mentatives and diminutives. In the case of Weining Ahmao, these markers are a55 ɲie53 (augmentatives) and ŋa11 (diminutives). The augmentative marker is related to the noun ɲie53 ‘mother’, while the diminutive marker is still present in the noun ŋa11 ʑau11 ‘child’. Both markers can take the initial position of nominal compounds. If these markers are combined with nouns for animals, their initial meaning is still visible – the augmentative marker produces a noun referring to the female gender of animals, while the diminutive marker derives a noun denoting the juvenile members in adult-young animal pairs (14a). If com- bined with inanimate nouns the two markers express a vague idea of largeness (physically or metaphorically) and smallness, respectively (cf. example [14b]): (14) Weining Ahmao (Gerner and Bisang 2010: 87) a. mpa44 ‘pig, hog’ → a55 ɲie53 mpa44 ‘sow [augmentative, female]’ → ŋa11 mpa44 ‘piglet [diminutive, child]’ b. tɕa44 ‘wind’ → a55 ɲie53 tɕa44 ‘storm’ [augmentative]’ → ŋa11 tɕa44 ‘breze of wind [diminutive]’ The development of different classifier forms for size started out with a pro- cess of reanalysis in which the augmentative and diminutive markers no longer formed a unit with the noun that followed them but were rebracketed into a new unit consisting of the classifier plus the augmentative/diminutive marker: (15) NUM CL [a55 ɲie53/ŋa11–NOUN] → NUM [CL–a55 ɲie53/ŋa11] NOUN After this change of word boundaries, the sequences [CL–a55 ɲie53] and [CL– ŋa11] were further reduced through various stages (loss of the nasals ɲ and ŋ, etc.) into [ai] and [a], respectively. The association of the initial augmentative 34 Walter Bisang
- 47. form in [ai] as a marker of medial size has to do with the use of these markers in the social setting between men, women and children (for details, cf. Gerner and Bisang 2009: 13–18). In this context, males used the original classifiers, while the forms in [ai] and [a] basically remained for women and children, respec- tively. Through this process, the size values of the various forms were redistrib- uted. The original form was associated with augmentative and male speakers, while the forms in [ai] and [a] were downgraded: [ai] was newly associated with medial size and female speakers and [a] became the marker for diminutive size and child speakers. The initial meanings of ‘mother’ and ‘child’ might have contributed additionally to this process of redistribtion. The second stage is based on the function of numeral classifiers as markers of definiteness in the bare-classifier construction [CL + N]. This is illustrated by ts’ae33 nG‘ae35 [CL meat] ‘the meat’ and ta33 li55 a55 [CL crow] ‘the crow’ in example (12) from Weining Ahmao. Since Weining Ahmao is accessible only in its contemporary form with its fully-fledged classifier-inflection system, it is not possible to illustrate this stage without the simultaneous integration of size. The vast majority of Miao languages do not express size on the classifier but they use bare-classifier constructions for marking definiteness. One of these languages is Hmong, which will be used to illustrate classifiers as definiteness markers in bare-classifier constructions without the interference of size. In (16), the two nouns niam ‘wife’ and txiv ‘husband’ are introduced in the first sentence and are then taken up with their classifier tus as definite NPs in the next sentence: (16) Hmong Thaum ub muaj ob tug niam txiv. Tus txiv tuag lawm. Long.ago there.are two CL wife husband CL husband die PFV Tus niam quaj quaj, nrhiav nrhiav tsis tau tus txiv. CL wife cry cry look.for look.for NEG get CL husband ‘Long ago there were a wife and a husband. The husband died. The wife was crying a lot but no mattter how hard she tried she was not able to find her husband.’ (Mottin 1980: 200) Indefinite nouns are marked by the numeral ib ‘one’ as in the Kam language described in subsection 2.1. (17) Hmong Ua ciav nws pom ib lub nkauj npuas. suddenly he see one CL pigpen ‘Suddenly, he discovered a pigpen.’ (Mottin 1980) Numeral classifiers with plural marking. A challenge to Greenberg 35
- 48. Parallel to the situation in Northern Kam, the classifier changed its form under the influence of the preceding numeral one, which is i55 in Weining Ahmao, written ib10 in Hmong, and can be reconstructed with a closed syllable as *iet which became *iʔ55 at a later stage. As is argued by Gerner and Bisang (2009, 2010), it was the interaction of the glottal stop with the classifier in sequences of the type [*iʔ55 CL] which triggered processes such as voicing, aspi- ration or tonal change in the phonological structure of the classifier.11 The oppo- sition that was formerly expressed by [CL + N]definite vs. [one + CL + N]indefinite is now reflected by different forms of the classifier itself and can be represented as [CLdef + N] vs. [CLindef + N]. Since this opposition was developed in the context of the singular form of [one + CL + N]indefinite, the formal distinction was limited to the singular. For expressing the same distinctions (plus the three size distinc- tions) in the plural, the language followed another strategy. The opposition between singular and plural is based on some sort of a plural classifier which must have undergone the same changes as the normal classifiers described in table 1. The details of this development need more research. From the look at another Miao language like Hmong, one can clearly say that there are plural classifiers. In Hmong, there is even a distinction between the plural classifiers cov and tej, which roughly mark definiteness and indefiniteness, respectively. In (18), the definite plural classifier cov in line 2 refers to the seven wives (poj niam) mentioned before in line 1. The plural clas- sifier tej in (19) refers to an unspecified plural number of widows (poj ntsuag): (18) Hmong Ces ob tug mus txog chaw uas Yawm Pus then two CL go to country/place REL Yao Pu muaj xya tus poj niam. Ces cov poj niam quaj dheev tias: . . . have seven CL wife then PL:CL wife cry very.much QUOT ‘Then, the two of them arrived at the country in which Yap Pu had seven wives. The wives cried and said: . . .’ (Mottin 1980) 10 There are no syllables with consonantal codas in Hmong (the nasal ŋ is the only exception). The Latin-based writing system of Hmong thus uses the post-nuclear position for encoding tones by various consonant letters. The consonant -b, for instance, represents the high-level tone [55]. 11 This process is known as “change in glottal articulation” (Lehmann 1992: 193; also cf. Gerner and Bisang 2009, 2010). 36 Walter Bisang
- 49. (19) Hmong Kuv niam tias: kom kuv tuaj nrhiav tej poj ntsuag I mother QUOT IMP 2.SG come seek CL:PL widow tsev nyob xwb. house live only ‘My mother told me: you only frequent houses of widows.’ [During your journey, you shall only ask for hospitality in houses of widows.] (Mottin 1980) 3 Discussion: Factors that operate against Greenberg’s (1974) universal claim Greenberg’s (1974) implicational universal discussed in Subsection 1.1 excludes the cooccurrence of obligatory plural marking and numeral classifiers. From such a strict perspective, the classifier systems of Northern Kam and Weining Ahmao provide no counter-evidence against Greenberg (1974), since the expres- sion of number is still not compulsory in these two languages, i.e., the bare noun alone with no classifier is still sufficient in certain contexts and if it is clear that plurality is meant. This conclusion is certainly correct if one looks at Green- berg’s claim from a purely synchronic angle that disregards the historical devel- opment of grammatical systems. However, a solid universal claim should not only be observable at the synchronic level but also at the diachronic level. If combinations of properties that are excluded by a universal claim can be tied together in historical processes of language change this casts serious doubt on the solidity of that universal. In the case of Chinese, the findings from diachronic research basically sup- port the claim that there is a correlation between transnumerality and numeral classifiers (cf. Meisterernst, this volume; Harbsmeier, this volume). The marker zhū does not exclusively express plural, it also quantifies over unmarked plural nouns and it expresses referentiality (Meisterernst, this volume). Even the exis- tence of unmarked nouns with plural meaning (Harbsmeier, this volume) is no straightforward argument against the overall transnumeral character of nouns as long as plural nouns are a clearly defined subset of lexicalized plurals of a limited size. Personal pronouns seem to be the only instance in which obligatory plural marking is of some importance (cf. Meisterenst, this volume; also cf. Fu, in this volume, on number in the pronominal system of Bai). But this only affects the highest level of the animacy hierarchy as discussed by Corbett (2000) and thus concerns only a small number of nominals. Numeral classifiers with plural marking. A challenge to Greenberg 37
- 50. The examples of Northern Kam and Weining Ahmao represent a much stronger challenge to Greenberg’s universal claim. If it were a rigid universal it should operate against the fusion of number marking with a numeral classifier system. In addition, this development produced a situation on the synchronic level in which the use of a classifier automatically forces the speaker to select a value for number (singular vs. plural). A system in which the use of a classifier entails number marking is additional evidence against the strength of Greenberg’s claim. From what has been said so far, one can distinguish two types of lan- guages which are problematic for Geenberg’s universal claim: (i) Languages with numeral classifiers and obligatory plural marking (ii) Languages in which the use of a classifier obligatorily calls for number marking Both types are attested in the world’s languages. Ejagham as described in section 1 represents a straightforward case of type (i), while Northern Kam and Weining Ahmao stand for type (ii). The processes that favour the emergence of such structures are not uniform. In spite of this, all of them are related in one way or another to the specific grammatical constellations that existed in the languages concerned immediately before the change took place. In the case of Ejagham, it is important to know that its numeral-classifier system is certainly much younger than the noun-class system with its obligatory distinction of singular vs. plural. The noun-class system is genetically well-established in Bantoid languages and belongs to the reconstructed inventory of Proto-Bantoid. If a language with such a noun-class system develops numeral classifiers it does not necessarily lose the property of obligatory number marking and, if that is the case, contradicts Greenberg’s (1974) universal claim. Of course, such a combination of properties should not emerge at all if Greenberg’s claim would be an absolute universal with no exceptions. In the case of Northern Kam and Weining Ahmao, the most important moti- vation for the emergence of inflectional classifiers is of a purely phonetic and phonological character and has nothing to do with cognitive motivations as they are generally discussed in linguistic typology. The whole development was instigated by the numeral ‘one’ (i45 in Northern Kam, i55 in Weining Ahmao) and its phonetic influence on the consonant of the numeral classifier, sometimes also on its vowel nucleus and its tonality in Weining Ahmao (see Section 2). Interestingly enough, the functional oppositions associated with the changes on the classifier differ in the two languages. In Northern Kam, the opposition is interpreted in terms of number (singular vs. plural), while it is employed for marking number (singular) and definiteness vs. indefiniteness in Weining Ahmao. The reason for that difference lies again in the specific grammatical constellations 38 Walter Bisang
- 51. immediately before the new systems developed, or, to be more precise, in the grammatical functions of the classifiers before their form was influenced by the numeral ‘one’. In Weining Ahmao, the classifier developed its different morphophonological realizations at a time when the classifier was used for marking definiteness. In Northern Kam, the expression of definiteness was certainly not a central function of the classifier. As a consequence, there was little chance for classifiers of Northern Kam to be associated with definiteness, while this was a real option for classifiers in Weining Ahmao. In Weining Ahmao, a second process of reanalysis followed by morphophonological erosion led to the three-way distinction between augmentative/male, medial/female and diminutive/child. Since the distinctions of definite vs. indefinite and singular vs. plural equally operate on all three elements of the size distinction, it is reasonable to assume that size distinction took place first. 4 Conclusion Functional explanations of the incompatibility of numeral classifiers and obliga- tory plural marking are based on the assumption that nouns in numeral-classifier languages are transnumeral, i.e., they are not specified for number in the lexicon (Greenberg 1974, Hundius and Kölver 1983, Bisang 1999, 2002). Due to their transnumerality, nouns cannot occur in immediate combination with numerals, they have to be individualized by the numeral classifier which provides them with the necessary conceptual boundaries by which they can be perceived as countable entities. From the perspective of formal semantics, Chierchia (1998a, b) argues that all nouns in Chinese are mass nouns in the sense that they do not make a distinction between sets of atoms and sets of sums of atoms (also cf. Li and Bisang 2012). The denotation of a noun like Chinese māo ‘cat’ includes both the set of atoms of cats and the set of sums of atoms of cats. As a consequence, mass nouns are inherently plural and do not need additional plural marking. If they have to be counted they have to be atomized, i.e., they have to be singled out as individual atoms by a numeral classifier. However, the analyses and motivations of the type discussed so far are not uncontroversial. Cheng and Sybesma (1999), for instance, argue that there is a mass/count-distinction in the lexical properties of Chinese nouns. The incompatibility of numeral classifi- ers and number marking is not only reflected in semantic approaches, it is also integrated into syntactic theories. One of them is Borer’s (2005) nominal struc- ture. In her theory, plural morphology and numeral classifiers cannot cooccur because they both are dividers and thus express the same function (cf. Feng, this volume). Numeral classifiers with plural marking. A challenge to Greenberg 39
- 52. No matter how appropriate the motivations for the presence of numeral classifiers in a language are in general, the present paper shows that there are various factors that do not depend on transnumerality and the existence of a mass/count-distinction. Thus, it seems rather implausible to argue that all nouns in a Bantoid language such as Ejagham are mass nouns if that language makes a systematic distinction between singular and plural in its noun-class system. In spite of this, the language is about to develop a numeral classifier system. The data from Kana with its fully developed classifier system and its loss of the old noun-class system (cf. example [5]) may support the hypothesis that there is a correlation between transnumerality and the existence of numeral classifiers because the singular/plural distinction associated with the noun- class system has vanished in the course of time. The question is whether the disappearance of the noun-class system is directly linked to the development of a fully-fledged numeral-classifier system. An alternative explanation for this loss comes from phonetics and phonology and is related to the erosion of phonological substance, a process which may well be completely unrelated to transnumerality. A lot of additional research will be needed to understand the emergence of the numeral-classifier systems of Niger-Congo languages. In East and mainland Southeast Asian languages such as Northern Kam and Weining Ahmao, the singular/plural-distinction developed even if it is plausible to argue that they were transnumeral at least before they integrated number dis- tinctions into their respective classifier systems. From a typological perspective which emphasizes the relevance of whether there is a count/mass-distinction in a language, the shift from a simple classifier system based on transnumeral nouns to a system with number-distinctions which somehow presuppose a count/mass-distinction looks rather great. However, the phonetic process that was ultimately responsible for this shift does not depend at all on that func- tional motivation. This leads to the ultimate question of the status of typological universals and their cognitive motivations. The findings of this paper seem to support Dryer’s (1997) view that exceptionless universals are extremely rare and that typologists should better spend their time on statistical universals. They also argue rather for what Newmeyer (2005) calls holistic functionalism than for atomistic functionalism. Atomistic functionalism assumes that “[t]here is direct linkage between properties of particular grammars and functional motivations for those properties” (Newmeyer 2005: 174). In contrast, holistic functionalism sees a far more indirect relation between grammars and motivations. As Newmeyer states, “grammars as wholes reflect the ‘interests’ of language users, but there is no question of parceling out rules, constraints, constructions, and so on of individual grammars and assigning to them a functional motivation” 40 Walter Bisang
- 53. (Newmeyer 2005: 178). The examples shown in this paper illustrate how individ- ual structures of languages depend on a multitude of individual factors which make universally valid generalizations extremely difficult. Thus, Greenberg’s (1974) universal on numeral classifiers is certainly not exceptionless and its link- age to functional motivations can only be an indirect one in the sense of holistic functionalism. If one takes into account the considerable amount of evidence against it and the fact that it does not exclude processes of grammaticalizaton which combine classifiers with plural marking, it is safe to assume that Green- berg’s universal claim is a rather week universal. References Aikhenvald, Alexandra Y. 2000. A Typology of Noun Categorization Devices. Oxford: Oxford University Press. Becker, Neele. 2005. Numeralklassifikatoren im Thai. Ph.D. diss., Department of English and Linguistics, University of Mainz (Germany). Bisang, Walter. 1993. Classifiers, quantifiers and class nouns in Hmong. Studies in Language 17: 1–51. Bisang, Walter. 1999. Classifiers in East and Southeast Asian languages: counting and beyond. In Numeral Types and Changes Worldwide, Jadranka Gvozdanovic (ed.), 113–185. Berlin/ New York: Mouton de Gruyter. Bisang, Walter. 2002. Classification and the evolution of grammatical structures: a universal perspective. Sprachtypologie und Universalienforschung 55: 289–308. Borer, Hagit. 2005. In Name Only. Oxford: Oxford University Press. Chamberlain, James R. 1997. Tai-Kadai anthropods: A preliminary biolinguistic investigation. In Comparative Kadai, the Tai Branch, Jerold Edmondson, Jerold and David Solnit (eds.), 291–326. Dallas: Summer Institute of Linguistics and the University of Texas at Arlington. Cheng, Lai-Shen Lisa and Rint Sybesma. 1999. Bare and not-so-bare nouns and the structure of NP. Linguistic Inquiry 30: 509–542. Chierchia, Gennaro. 1998a. Reference to kinds across languages. Natural Language Semantics 6: 339–405. Chierchia, Gennaro. 1998b. Plurality of mass nouns and the notion of ‘semantic Parameter’. In Events and Grammar, Susan Rothstein (ed.), 53–103. Dordrecht: Kluwer Academic Publishers. Corbett, Greville G. 2000. Number. Cambridge: Cambridge University Press. Dryer, Matthew S. 1997. Why statistical universals are better than absolute universals. Chicago Linguistic Society 33: 123–145. Dryer, Matthew S. 2005. Indefinite articles. In The World Atlas of Language Structures, Martin Haspelmath, Matthew S. Dryer, David Gil and Bernard Comrie (eds.). Oxford: Oxford University Press. Edmondson, Jerold and David Solnit. 1988. Introduction. In Comparative Kadai: Linguistic Studies Beyond Tai, Jerold Edmondson and David Solnit (eds.), 1–26. Dallas: Summer Institute of Linguistics and the University of Texas at Arlington. Numeral classifiers with plural marking. A challenge to Greenberg 41
- 54. Feng, Shengli. 2012. The syntax and prosody of classifiers in classical Chinese. This volume. Fu, Jingqi. 2012. Plurality in the pronominal paradigms of Bai dialects. This volume. Gerner, Matthias. 2006. Noun classes in Kam and Chinese Kam-Tai languages: their morpho- syntax, semantics and history. Journal of Chinese Linguistics 34: 237–305. Gerner, Matthias, and Walter Bisang. 2008. Inflectional speaker-role classifiers in Weining Ahmao. Journal of Pragmatics 40: 719–731. Gerner, Matthias, and Walter Bisang. 2009. Inflectional classifiers in Weining Ahmao: mirror of the history of a people. Folia Linguistica Historica 30: 1–36. Gerner, Matthias, and Walter Bisang. 2010. Social-deixis classifiers in Weining Ahmao. In Rara and Rarissima. Documenting the Fringes of Linguistic Diversity, Jan Wohlgemuth and Michael Cysouw (eds.), 75–94. Berlin/New York: Mouton de Gruyter. Greenberg, Joseph H. 1974. Numeral classifiers and substantival number: Problems in the genesis of a linguistic type. In Proceedings of the 11th International Congress of Linguistics, Bologna – Florence, Aug–Sept 1972. Bologna. 17–37. Reprinted in Greenberg, 1990: 16–93. Greenberg, Joseph H. 1990. On Language. Selected Writings of Joseph H. Greenberg, ed. by Denning, K. and S. Kemmer. Stanford: Stanford University Press. Harbsmeier, Christoph. 2012. Plurality and the subclassification of nouns in classical Chinese. This volume. Hundius, Harald, and Ulrike Kölver. 1983. Syntax and semantics of numeral classifiers in Thai. Studies in Language 7: 165–214. Ikoro, Suanu M. 1994. Numeral classifiers in Kana. Journal of African Languages and Linguistics 15: 7–28. Ikoro, Suanu M. 1996. The Kana language. Leiden: Research School CNWS. Lehmann, Winfred P. 1992. Historical Linguistics. An Introduction. London: Routledge. Li, Xuping, and Walter Bisang. 2012. Classifiers in Sinitic languages: From individuation to definiteness marking. Lingua 122: 335–355. Meisterernst, Barbara. 2012. Number in Chinese: a diachronic study of the case of zhu from Han to Wei Jin Nanbeichao Chinese. This volume. Mottin, Jean. 1980. Contes et légendes Hmong Blanc. Bangkok: Don Bosco Press. Newmeyer, Frederick J. 2005. Possible and Probable Languages. A Generative Perspective on Linguistic Typology. Oxford: Oxford University Press. Sanches, Mary, and Linda Slobin. 1973. Numeral classifiers and plural marking: an implica- tional universal. Working Papers in Language Universals 11: 1–22. Shi, Lin. 1997. 侗语汉语语法比较研究 Dongyu Hanyu yufa bijiao yanjiu [Dong-Han Com- parative Grammar Research]. Beijing: Central University of Nationalities Press. Wang, Deguang. 1986. 威宁苗语话语材料 Weining Miaoyu huayu cailiao [Language material in the Weining dialect of the Miao language]. Minzu Yuwen 民族语文 3: 69–80. Wang, Deguang. 1987. 贵州威宁苗语量词拾遗 Guizhou Weining Miaoyu liangci shiyi [Find- ings on numeral classifiers in the Weining dialect of the Miao language in Guizhou]. Minzu Yuwen 民族语文 5: 36–38. Wang, Fushi. 1957. 贵州威宁苗语量词 Guizhou Weining Miaoyu liangci [The classifier in the Weining dialect of the Miao language in Guizhou]. Yuyan Yanjiu 语言研究 1957: 75–121. Translated 1972 as: The classifier in the Weining dialect of the Miao language in Kwei- chou. In Miao and Yao Linguistic Studies: Selected Articles in Chinese, Herbert Purnell (ed.), 111–185. Ithaca, NY: Cornell University. Watters, John Roberts. 1981. A phonology and morphology of Ejagham – with notes on dialect variation. Ph.D Diss., University of California, Los Angeles. 42 Walter Bisang
- 55. Random documents with unrelated content Scribd suggests to you:
- 56. BUDDHIST TEMPLE. OUTSIDE KUEI HUA CHENG, CHINA. Next day was uneventful, but we ascended gradually to a little village called Cha-Ha-Pa-La, 100 li, and got quite into the mountains again, very different country from what we had passed through since leaving Pao T'eo. The soil here appeared very rich, and every available patch of land was cultivated. Every day we used to pass through numerous villages, and in one we were lucky enough to come in for a theatrical entertainment. No village is without its theatre, and performances are frequent in the season; but as it was now the dull time we only saw one. The theatre is always situated close to, generally opposite, the city temple, a great convenience to the Taoist priests, who are also the theatrical managers. The play did not appear exciting, and neither of us understood a word of it, which was perhaps just as well, so we left after a very few minutes. The theatres being in the open, no entrance money is charged. The audience come and go as they please, but are expected to drop a
- 57. small contribution into a collection box which is continually being handed round. This system has the further advantage of enabling the actors to converse freely with friends below when not otherwise engaged. From Cha-Ha-Pa-La we made an early start, as we heard that the night's halting-place was to be a fair-sized town, and we wanted to arrive early so as to renew our supplies. We were pretty high up, and as it was now very near the end of November, the mornings were uncommonly cold. Rijnhart, Malcolm, Lassoo, and Esau all started walking, while Shahzad Mir and I followed with the carts. Not far from the village we crossed a low pass into a lovely wide valley, and a little further on the party ahead managed to take a wrong turn. I went on with the carts, and got to our midday halting-place about ten o'clock, and there I had to sit and wait till the others should come up. I managed to make out, by drawings in the dust and other devices, that there was another road to Pekin, which the others must have taken, and by which they would get to our destination in seventeen days. Not thinking they were likely to be quite so long in finding out their mistake, I resolved to have my breakfast and await their arrival. After a couple of hours' waiting I was beginning to get a little bit anxious, and was considering what steps I had better take if, by any chance, they did not arrive before nightfall. However, at about one o'clock they hove in sight, so I immediately had the mules harnessed, while Shahzad Mir warmed up some tea. In a few minutes Rijnhart and Malcolm were at the inn, but the two men were some way behind, and while waiting for them to come up, I heard how they had managed to miss us. It appeared that, when they missed the road, all the Chinese they had met had misled them by saying that they were right, and it was not till late in the day that they got into the proper road and arrived at the halting- place, after walking fully thirty miles.
- 58. "We were soon ready for a fresh start, but the carters, anxious for any excuse to back out of their agreement as long as the fault could be laid at our door, were very reluctant, but threats to cut off half their hire prevailed, and we were quickly on the move. The march was a long one, over rolling grassy plains, out of Chinese into Mongolian country. As the sun set a cold wind rose, and for some time we sat shivering on the carts. The carts were travelling very fast, as the going was excellent. We thought we should never get to our destination, but all the time we were buoyed up with the thought that, at all events, we should get to a big place where there was sure to be a good inn. At last they told us that Ho-Lo-Si-T'ai, our goal, was close by, and in a few minutes we saw, not the town with castellated walls that we had imagined, but two miserable wayside inns with large yards full of sheep and camels, as unpromising a spectacle as one could wish to see. On inquiry, things proved even worse than they had at first appeared, for both inns were crammed full, and we could not find a spot to sleep in. At last we heard that there was a room in a small house close by, so round we went, only to find it tenanted by a man, two women, a sheep, and some children, a fair supply for a place about twelve feet square, but into it we squashed, until we had revived our circulation, and then, after a considerable amount of squabbling, we arranged that we would sleep in the carts, while the carters occupied the already overcrowded room, for the simple reason that the carters flatly refused to take us any further unless they slept inside the house and we outside. Perhaps, after all, the cold blast was preferable to dirt and a close atmosphere. The following day our road lay over more grassy plains, in which were herds of antelope—called by the Chinese the "yellow sheep"— some sand-grouse, and a few great bustards, none of which we shot. There were several Mongol encampments, with their neat little
- 59. circular dwelling-places and sheep-folds, much more civilized than those we had seen in the far west. That evening we left the Mongol country, possibly for ever, and got into the cultivated district again. The change was very sudden, and showed us admirably how the Chinese are slowly, but surely, encroaching upon the country of their pastoral neighbours. Cha-Ha- La-Po, our home for the night, was an insignificant little place, and chiefly to be remembered for the fact that we there managed to get rid of some six hundred very inferior cash, brought from Shih-Tsui- Tsi. At noon the next day we had to change our axles, the roads from here being much narrower than they had hitherto been; and while doing so we had a row with the carters. A crowd promptly assembled, but Rijnhart soon managed to get them on our side, and we gained our point. We were now well within our time, and the carters knew that they could fulfil their contract without difficulty, so they loitered over the axle changing and made a very short march to a brand-new inn, where we put up in an excellent, clean, and airy room. There was still plenty of daylight left, of which we took advantage to have a real good wash in warm water.
- 60. MONGOL ENCAMPMENT. As if to make up for this early halt, we were awakened soon after midnight, and started about 1 a.m.; but this energy, we had every reason to believe, was owing to a desire on the part of our jehus to pass through the next village in the dark, owing to monetary or some other form of trouble they were in there. At first it was very dark, and as our road was very rough we went stumbling along till the moon came out from behind some heavy clouds. Rising rapidly, we were soon on a small pass, where some pious individual, in hopes of future reward, had built a neat little temple. The descent on the far side was steep, and the rocky gorge through which our road lay was very picturesque in the brilliant moonlight. About daylight we got down to the bed of a stream, which we crossed and recrossed several times till we arrived at the small town
- 61. of Hsing-Ping-Ho, the gate of which forms part of the Great Wall of China. We had already seen the Great Wall, near Chong Wei, where, however, it is little more than a turf embankment, but as we approached the capital we saw how it gradually improves, both in its original construction and its later preservation; but even now we were very much disappointed, and were unable to realize how it ever managed to gain its world-wide reputation. At Hsing-Ping-Ho the wall is of earth, faced with brick, generally not more than eight to ten feet high, and quite narrow, very different from the imposing structure one had always pictured it to be, and which we were yet destined to see. But although we could not yet look upon the wall with the respect one had hoped to feel for it, neither of us could help admiring the dogged perseverance with which it has been carried over mountain tops and down valleys, the most forbidding natural obstacles being treated as nought in comparison with the orders of the emperor. After breakfast our road ran parallel to the stream we had crossed in the morning, which had now grown to a considerable size, and was called the Wo-Ku-Shan-Ho. Every few miles we passed a large walled city, evidently very old, and now almost entirely deserted; they must all have seen better times and been places of importance—probably about the time that the Great Wall was being built. I photographed the walls of one of these cities, and the temple outside tenanted by some wooden figures. The name of the city was Si-Yang Ho.
- 62. TEMPLE OUTSIDE SI-YANG HO. Showing Stages from Pao T'eo to Pekin. 14th November. —Pao T'eo to Tour-tsi 40 " November. " Sa-la-ki 50 15th November. " Mi-tour-chon 30 " November. " Ta-ri-tsi 80 16th November. " Peh-sie-ki 50 " November. " Kuei Hua Cheng 70 17th November. " Shih-Rong-Wa 90 18th November. " Cha-Ha-Pa-La 100 19th November. " Mongol Camp (long) 50 " November. " Ho-Lo-Si-T'ai 70 20th November. " Mongol Camp (short) 50 " November. " Cha-Ha-La-Po 80 21st November. " Cha-ka-ri 60
- 63. " November. " Teo-tao-keo 35 22nd November. " Hsing-Ping-Ho 75 " November. " Uen-tsi-tao-ri 50 23rd November. " Io-kia-t'ong 65 " November. " Nant-sing-ho 75 24th November. " Shuen-hua-fu 40 " November. " Siao-Si-Fu 35 25th November. " Ki-mio-si 35 " November. " Sa-cheng 45 " November. " Huai-lai-hsien 60 26th November. " Cha-tao 50 " November. " Nan-Kou 45 27th November. " Sha-ho 50 " November. " Pekin 45 From Pekin to Tien Tsin is 80 miles.
- 64. CHAPTER XXXI. HOW TO MANAGE INN-KEEPERS AND CARTERS—SHUEN-HUA-FU—"SPIRIT'S PAPER"—SHAHZAD MIR LOST AND FOUND—ESAU'S PRESTIGE. About dusk we got to a very small village, with a most charming little inn. We got two very comfortable little rooms, with excellent furniture, consisting of tables, chairs, and looking-glasses. This must have been quite one of the longest day's journey we did, but li are very erratic things to go by. As far as we could make out one hour's journey was always counted as ten li, quite irrespective of whether we went fast or slow; thus ten li in one place would be nearly five miles, and in another very little over three. Delighted as we had been on arrival with our little inn, we were still more so on our departure, for we got off without a word of dispute at the price to be paid for our accommodation. At this time our rule was to pay one hundred cash for each room occupied, and about forty cash each for water and for the use of the fire; when nearer the capital we increased this as the accommodation improved, just as in the rough places further west we had given less. Having a fixed rule like this, which was liberal without being excessive, saved us an infinity of trouble, as the innkeepers saw at
- 65. once that Rijnhart knew what he was about, and was not likely to be cheated or bluffed into making exorbitant payments. But there was always one thing they could not get over, and that was that Rijnhart always handed them whatever money we took out of the bag and gave him. They saw that we put ourselves in his hands, and could not make out why he exacted no squeezes from us or them; never was such honesty heard of! BRIDAL CHAIR. This day, the 23rd November, was evidently a very lucky one: every one was out visiting friends, dressed in their best, the big ceremonial hat was seen on all sides, and several marriages were in progress. As if to make up for our peaceful morning, we had a most truculent fellow to deal with at breakfast; our cash were all too small and our fees were quite insufficient; in fact, he found a casus belli in
- 66. our every act, and crowned his impertinences by selling us some two dozen real bad eggs, at eight cash apiece. Sometimes a traveller is apt to find himself in a very awkward position, if his carters and an offensive innkeeper happen to be in league. It may happen that the carters owe the innkeeper money, and the latter threatens to get them punished unless they help him in extorting money from the unfortunate foreigner; the innkeeper then demands a perfectly unreasonable sum from his victim, which the latter refuses; upon this the carters come up and say that they cannot possibly leave till the money is paid, and the helpless traveller finds himself between the devil and the deep sea. Luckily for us, our agreement effectually stopped any nonsense of this kind, for if the carters said they could not go on, we at once replied, "All right, please yourselves—but if you do not arrive by the 27th, you lose half your hire." Every day, now, villages became more numerous, and the country was very thickly cultivated. We stopped a long time at our midday halt at the city of Shuen-Hua-Fu,[15] as the carters had to change some silver and transact some business before proceeding. After leaving, we passed a very large number of gravestones and memorial stones, with old inscriptions of the rounded characters; the hillsides on the left of the road were studded with them for several miles, and in many places the coffins were sticking out of the ground, where the earth had been washed away from over them. On the road we met a convoy of mules carrying boxes containing tins of oil from Batoum, and another with silver; the latter had an escort armed with spears decorated with red tassels, each mule, too, was decorated with a red flag, the whole making quite an imposing show. Towards evening the road got very bad, and in one place we all had to get out and steady the carts as they passed down a steep,
- 67. rocky path. Luckily we got down without mishap, but one of the carts was within an ace of slipping down a decline, where there was nothing to stop it for a considerable distance, and where both it and the mules must inevitably have been smashed. As we went down, a string of carts was coming up, two or three teams being hitched on to one cart and then returning for the next. The shouts of the drivers, cracks of whips, and creaking of wheels, made a terrible din, while the motley teams of ponies, mules, donkeys, and bullocks, all straining every nerve and then only managing to progress some ten or fifteen yards without a rest, made a picture. I should very much like to have been able to perpetuate it, but the light had already failed, and I was unable to do so. It was dark when we got down to Siao-Si-Fu and into our inn, but we had really made quite a short day's journey. Next morning we were up long before it was light, and made a start with lanterns. It was just as well we did so, for our landlord, who was evidently accustomed to foreigners, and looked upon them as fair prey, refused to open the gates of the inn yard unless we paid him far more than was his due; but he reckoned without his host, for as there was no crowd about, and no fear of exciting a disturbance, we were able to threaten, and telling him that if he did not undo the lock, we would blow it open with a revolver, he soon saw that the game was up, and let us go in peace. The village of Siao-Si-Fu lies just at the entrance of a narrow gorge, through which runs a rapid stream; the road winds along the hillside, generally at a good height above the water, but it has been carefully made, and the going is excellent. We met a large number of camels coming through, mostly carrying tea into the interior; and in one place one of them had managed to slip some distance down the hillside, but did not appear to have hurt himself much—he had been wise in choosing a gradual slope, where he could not fall far. The defile lasted nearly ten miles, but we missed most of its
- 68. beauties, owing to our early start and the darkness, which was also responsible for our party getting separated, Shahzad Mir having somehow managed to lose the rest of us. This did not cause us much anxiety at the time, as we thought that he had merely lagged behind, and would soon be up with us. A GORGE WITH GREAT WALL IN DISTANCE. When, however, we had made our halt, and breakfast was being prepared, and there were still no signs of him, we began to get anxious, and in making inquiries, we heard that he had passed our halting-place, and been met by carters coming from the Pekin direction. We still consoled ourselves with the thought that he would soon find out his mistake and retrace his steps, but when it was time to yoke up and make a fresh start, nothing had been seen of him.
- 69. We were now very anxious about him, so divided ourselves into two parties, the carts and servants going one road, while we went another, which passed through several small towns and ran parallel to the road taken by the carts. Both parties made inquiries as they went along, and for some thirty li we still heard of him; he seemed to have walked straight ahead, never stopping except to buy some food. His clothes and gait, as well as his dark face, made him very conspicuous, so that the various guards on the town gates always noticed him; but after going about ten miles we lost all trace of our faithful Duffadar, and all we could do was to give notice of his disappearance in different places, and ask the people to show him which way we had gone. The most unlucky part of the business was that he could not speak a solitary word of Chinese, and might get himself into trouble, but fortunately we were in northern, and not central or southern China, so we felt sure that sooner or later he would turn up.[16] In some of the towns we passed through we saw quantities of "spirit's paper" lying about the streets, or hanging up in the shops for sale. This spirit's paper is simply round pieces, cut to represent cash, each sheet of paper representing any number, or gilt to represent silver or gold. These are bought by the people and scattered about the houses, or put in holes in walls and tree trunks, where they are found by restless, wandering spirits, who, poor creatures, are easily deceived into thinking them offerings of great value, and consequently refrain from injuring the pious, but economical offerer. They also burn large quantities of this paper money, and effigies of carts, horses, and other signs of worldly wealth, all of which go to improve the social status of their ancestors in the nether world.[17] The Chinaman, though decidedly superstitious and pious, according to his lights, rarely allows his regard for his ancestors to overcome that for his own pocket.
- 70. This was to be another very long day, so we halted for an hour, about 4 o'clock in the afternoon, and fed the animals, we ourselves buying some splendid grapes for forty cash a pound. These are grown south of the town of Shuen-Hua-Fu; also quantities of millet (huang-mi[18]), which is ground up and eaten like rice by the poor classes. While wandering round, telling people of Shahzad Mir's disappearance and looking at the various shops, we saw hung up in a cage on the city gate the head of a man who had been found guilty of theft, and next to this gruesome relic were the shoes of several officials, whose ability or zeal in discharge of their duties had merited this special distinction; thus are the best and the worst of men's deeds kept fresh in the minds of the Chinese people. We were just about to make our third start for the day when two Russian merchants entered our inn. We were much struck by the promptitude with which they extracted their revolvers from the sedan chairs in which they were travelling, and placed them in handy and conspicuous positions on top of their luggage. Curiously enough the only language in which we could converse with them was Chinese, which one man spoke fluently; and it was a great relief when we learnt that we need have no fear of the Gulf of Pe Chili being frozen over, for we were still in plenty of time. We still had sixty li to do before halting for the night at Huai-Lai- Hsien, so we shoved along as fast as the mules could go, but the last twenty li or so were done in pitch darkness and bitter cold. The gates of the suburb were closed when we arrived, but we got them open with some difficulty, after considerably exaggerating the importance of the high official posts we both occupied. We had done nearly fifty miles since morning, so were glad to get an excellent Chinese concoction of eggs, and tumble into our blankets as quickly as might be.
- 71. Next morning we had to wait till the main gates were open, as our road lay through the city. The innkeeper coolly demanded 1,160 cash for the use of his rooms, but after a very few moments was more than satisfied with 360, in addition to which we of course gave a douceur to the man who had waited on us, as is the custom. A biting cold wind was blowing from the west, and as it was also freezing hard, it was quite the most unpleasant day of the whole journey, but it was not till we got in the open country, outside the walls, that we really felt the full force of the gale. The road was crowded with travellers going to and from Pekin, and one could not but envy the wealthy individuals who, sitting snugly in their neatly- made sedan chairs, could despise both wind and cold. Usually we used to wonder how even a Chinaman could sit all day in one of these conveyances, from which every breath of fresh air was carefully excluded, but to-day I think we would gladly have got into the stuffiest of them had we got the chance. But if it were bad for us, what must it have been for the people travelling against the wind? Time after time we congratulated ourselves on the wind being at our backs. Several unfortunate Russians whom we met looked as though the wind was nearly cutting them in two, while Mongols and Chinese sat with their backs towards the heads of their camels or mules as they rode along. One poor man, whose whole attention was fixed on keeping himself warm, was very nearly run over by our carts in spite of all our warning shouts. Somehow, watching the misery of others always makes one forget one's own, and we were able to laugh at the ludicrous though pitiable faces of those we met, and in this way the fifty li to Cha-Tao passed less unpleasantly than it could otherwise have done. Just before reaching this village we passed through another of the many ramifications of the Great Wall, behind which we cowered until the carts came up, wondering all the time what had happened
- 72. to the unfortunate Shahzad Mir. Where could he have passed the night? What was he doing now? How would he manage for food? These and many other questions we asked one another, but decided that the only way in which we could help him was to get to Pekin quickly, and give notice at the yamen there of his disappearance. From the wall to the inn was a very short way, and soon after the carts had come up we found ourselves in a crowded inn, where every one was busy; all the travellers were shivering like ourselves, and all clamouring for hot tea and other luxuries. At length we managed to get something, and had just got well into a substantial meal, when, to our surprise and delight, Shahzad Mir strode into the yard. While some food was being got ready for him he told us his tale. After missing us in the dark, on the morning of the 25th, he had quite made up his mind that we must be ahead, so on and on he went, eventually leaving the right road to his left. By midday he realized that he was lost, but this did not seem to have bothered him in the least. He knew that we were making for Pekin or Peh Chine, as the people call it, and guessed that by repeating that word in an inquiring tone of voice, he would get people to show him the road. He also knew that Pekin lay pretty well due east, so made up his mind to hold his course in that direction until he got to the sea, if necessary, after which he hoped to fall in with some European who would help him. At nightfall he thought he ought to try and get some food and lodging. He was in a fair-sized town, so walked into the biggest shop he could see, and made the owner understand his wants. Food they gave him willingly, but it was some time before he could get them to let him stay the night, his dark-coloured face apparently filling them with awe; but at last he attained his object, and got not only a good night's rest, but a substantial meal before leaving in the morning, for all of which his host absolutely refused payment.
- 73. THE GREAT WALL AT CHA-TAO. Somehow or other the dark complexions of our retainers always struck awe into the minds of the Chinamen, and time after time Rijnhart had to answer questions as to whether these men were "really wild or not." Esau especially, with his flowing Ladakhi locks and earrings, was an object of fear and wonder, being taken for either a woman or a cannibal, greatly to his own disgust, though luckily he could not understand the remarks aroused by his strange appearance, unless they were explained by Rijnhart afterwards. After breakfast the wind dropped a bit, and the cold was much less than it had been, for which we had reason to be thankful, as about two li beyond Cha-Tao we passed through the Great Wall again, and were able to see it at its best. It is here made of rubble, and splendidly faced with stone; towers are placed at regular intervals, and the top is broad enough for any troops to pass one
- 74. another without the slightest difficulty or confusion. Just by the gate a flight of steps leads up to the wall, from which we looked down the valley towards Cha-Tao, and I also managed to take a photograph showing the wall as it runs away over the hill tops to the north. One could now for the first time realize the magnitude of the original design, and had the whole northern frontier of China been divided off from Mongolia by a wall, such as this is near Pekin, and some 1,600 miles in length, there would have been nothing in the world to compare with it. Just behind the wall we found an old cannon, the inscription on which was copied by Rijnhart, as he believed that it proved the gun to be some three thousand years old, but unluckily we had no books at hand to corroborate this idea, or to tell us how old it really was. The name of the Emperor in whose reign it was made was very clear, and a book of dates would show at once between what years the gun was made. Meanwhile our carts had got some way ahead of us, and we had to hurry after them. Ascending steadily we reached the summit of the Nan Kou Pass, from which a rapid descent into a warmer climate commenced. The road was at this point by far the best we saw in China, and showed signs of immense labour and no inconsiderable amount of engineering skill having been expended on it. Our road lay down a narrow valley, high on each side of which rose rocky mountains, while along the northern skyline ran the Great Wall.
- 75. CHAPTER XXXII. A TRUCULENT INNKEEPER—A SEDAN CHAIR—CHINESE WOMEN AND THEIR FEET —PEKIN—DEPARTURE OF RIJNHART—CARTERS EARN A BEATING. As we descended we passed several small shrines and numerous inscriptions cut in the face of the rock, many of which were in ancient Tibetan characters, but more remarkable than these was a large figure of Sakya Muni, cut on a rock, which stands in a very conspicuous position high above the road. Rijnhart and Malcolm climbed by a steep staircase to a small shrine, some eighty feet above the road; but although it was evidently very old, they found nothing more remarkable than the names of several Russians and a few Englishmen who had been there in previous years, among others being Grey, Walker, and Allen, 1868 and 1879. We did not add our own to the list.
- 76. BUDDHIST ARCHWAY BY NAN KOU. Still descending, we passed under several arches, which looked as though they had at some time been gates of different fortified camps, and then came to a magnificent Buddhist arch covered with carvings and representations of Buddha; this must be a great age, and is one of the most interesting relics in Northern China. A short way further on we came to the village of Nan Kou (southern valley), where visitors to the Great Wall almost always stop the night after leaving Pekin. Knowing this, we were quite prepared to pay more than usual for our accommodation, but the demands with which we were met far exceeded our expectations. In the first place, we could get no fire on which to cook our own food, but must take what the inn provided; then our servants could not cook their food; and lastly, we must pay for everything over-night, and not wait till the morning, as had been our invariable custom. Eventually these little difficulties were settled, but not until our landlord had worked himself up into a towering passion, such as only a Chinaman can, and had stamped up and down the room like a mad man.
- 77. Welcome to our website – the perfect destination for book lovers and knowledge seekers. We believe that every book holds a new world, offering opportunities for learning, discovery, and personal growth. That’s why we are dedicated to bringing you a diverse collection of books, ranging from classic literature and specialized publications to self-development guides and children's books. More than just a book-buying platform, we strive to be a bridge connecting you with timeless cultural and intellectual values. With an elegant, user-friendly interface and a smart search system, you can quickly find the books that best suit your interests. Additionally, our special promotions and home delivery services help you save time and fully enjoy the joy of reading. Join us on a journey of knowledge exploration, passion nurturing, and personal growth every day! ebookbell.com