Open Journal of Modern Linguistics
2013. Vol.3, No.4, 344-347
Published Online December 2013 in SciRes (http://www.scirp.org/journal/ojml) http://dx.doi.org/10.4236/ojml.2013.34044
Open Access
344
Chinese Topic Chains: An Analysis of the Topic Chains
Used in Fortress Besieged
Yuanxiu He1,2
1College of International Studies, Southwest University, Chongqing, China
2School of Foreign Languages Studies, Yangtze Normal University, Chongqing, China
Received June 26th, 2013; revised July 29th, 2013; accepted August 7th, 2013
Copyright © 2013 Yuanxiu He. This is an open access article distributed under the Creative Commons Attribu-
tion License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original
work is properly cited.
The functional typologists divide the languages in the world into two groups, one is the subject-prominent
language group and the other is the topic-prominent group. English belongs to the subject-prominent lan-
guage and Chinese to the topic-prominent language. This paper aims to discuss and summarize the types
and functions of topic chains on the basis of analyzing the topic chains used in the novel Fortress Be-
sieged written by Chinese famous writer Zhongshu Qian.
Keywords: Functional Typology; Topic Chains; Types of Topic Chains; Functions of Topic Chains
Introduction
Functional typologists Li & Thompson (1976) divided the
natural languages into two groups: one is subject-prominent
language group; the other is the topic-prominent group. They
also pointed out that Chinese language belonged to the topic-
prominent group. Just based on Li and Thompson’s research,
many Chinese scholars, such as Nansong Huang, Jiujiu Xu,
Yulong Xu etc., began to study the Chinese topics and topic
chains. They made researches about the types, the internal
structures and discourse functions of topic chains. But there is
no consensus among these researchers as far as these things are
concerned, such as the classification criteria, the discourse
functions and rationale for the choice of topic chains. In addi-
tion, these researchers resort to the qualitative research, not
based on the quantitative research. Henceforth, this paper aims
to discuss the classifications and functions of Chinese topic
chains based on the analysis of the topic chains used in the
novel Fortress Besieged written by Zhongshu Qian (1991).
Theoretical Background of the
Functional Typology
Functional typology was proposed by Li & Thompson. It is
different from word order typology proposed by Joseph H.
Greenberg in the 1960s. Through analyzing and describing the
word order about over 30 different languages, Greenberg pro-
posed six kinds of possible language types: SVO; SOV; VSO;
VOS; OSV; OVS (S stands for the subject, V for a verb, O for
an object). Functional typology is also different from the study
of typology proposed by Hawkins. Like Greenberg, Hawkins
(1994) emphasized the analysis of syntactic structure and word
order. He proposed the theory of X-Leverage for the conven-
ience of analyzing the common tendency of the word order and
of the syntactic structure. Hence, his theory is to facilitate the
processing of pure syntactic structure other than dealing with
semantic relations or semantic features. Furthermore, functional
typology is different from traditional language typology, the
so-called morphological typology. According to the relation-
ship between morphemes, morphological typology divided all
the world languages into three groups: “isolated language”,
“adhesion language” and “comprehensive language”. Unlike
Greenberg, Hawkins and also the morphologists, functional
typologists stress the semantic relations and functions among
the parts of a clause. Based on syntactic status of subject and
topic, especially the semantic functions of the segments of a
clause, functional typologists discriminated the world lan-
guages into two groups: topic-prominent language and sub-
ject-prominent language. They pointed out that Chinese lan-
guage belonged to the topic-prominent group. That is because
in each simple and complex Chinese clause, there must be a
topic not necessarily the subject, even though the topic can be
used as the subject in many cases. In other words, each Chinese
clause is developed and scrolled around a topic, see example
(1). In this example, the statement is composed of two simple
sentences. And in each simple sentence, there is a typical Topic
+ Comment structure not a typical SVO structure, in which the
comment parts (e.g. “水管漏水了”, “啥事没有”) function as a
comment for the topic “李家” and “王家” respectively, even
though these two sentential predicate phrases display the typical
SVO structure. Then what is a topic and how to define the topic
chain?
Topic and Topic Chain
Dik (1997: p. 53) defined that topic was usually placed at the
depart position of a sentence; it was the object that the com-
ment describes. In Chinese language, if the same topic is re-
peated, then the topic is most often omitted. There are two
omitted situations: backward anaphora and forward cataphora.
See example (5) and (6) respectively.

Y. X. HE
Open Access 345
Usually, Chinese simple sentence is composed of two parts,
the topic and the comment or in Halliday’s words the theme and
the rheme (2004: p. 53). When one topic-comment structure
connected with another topic-comment structure, we can get a
complicated clause. Then the topic of one topic-comment struc-
ture and the topic of another topic-comment structure can be
chained in many ways according to different needs in different
situations. These chains are called topic chains which will be
discussed in this paper. Givón (1983: p. 67) defined that topic
chain was composed of an indispensable topic, and at least one
anaphorically or cataphorically omitted topic. Fengfu Tsao
(1990: p. 45) proposed that “topic string (chain) consists of a
topic with one or a few comments.” Like Givón, Chauncey Chu
(1998) offered a similar definition about topic chain. He held
that topic chain was composed of an indispensable topic and
several anaphorically or cataphorically omitted topics. Wendan
Li (2005:67) further amended this definition as “topic chain
includes at least two sentences which are connected by an in-
dispensable NP topic and one or more anaphorically or cata-
phorically omitted NP topic.” The writer of this paper, however,
found that the above definitions are all somewhat biased. Since
Chinese is a topic-prominent language, then of course, the dis-
course of Chinese is connected by the topics. Hence, if we take
Givón’s definition, we can not explain this complex sentences
combined by independent and complete topic-comment struc-
tures, for example:
(1) 李
lǐ
家
jie
(topic1) 水
shuǐ
管
guǎn
漏
lòu
水
shuǐ
了
le
(comment1), 王
wáng
家
jie
(topic2) 却
què
啥
shá
事
shì
没
méi
有
yǒu
(comment2).
(The water pipes of Li’s house were leaking, while nothing
happened to Wang’s house.)
Here in this complex sentence, we have two sub-clauses.
And in each sub-clause, we get one topic-comment structure.
Hence, we have two different topics not just one topic. Thus
Givón’s definition failed to account for this complex clause.
And if we take the definition offered by Fengfu Tsao, then
we can not explain this discourse phenomenon. That is the dis-
course is constituted by several topics but one comment. See
example (2).
Wendan Li’s view of topic chains is also not appropriate,
because we can find topic chains not only between two sen-
tences, but also we can find topic chain in a simple sentence
and across the paragraphs.
Through above analysis, we can define the topic chain like
this: topic chain consists of two topics, of which at least one is
indispensable; and topic chain can be shown in simple sen-
tences, complex sentences, and even be shown in a cross-sec-
tion of discourse.
Types of Topic Chains
Based on the definition given above, topic chains can be put
into two groups according to the topics are indispensable or
dispensable. In the first group, all the topics of the topic chain
must be shown by themselves in each topic-comment structures,
while in the second group, all the topics except one may not be
shown, they can be omitted anaphorically or cataphorically. In
the first group, there are two subcategories. When two or more
topics are connected with one or more comment, we put this
topic chain of this topic-comment structure into the first sub-
category, we use letter A to stand for it. See example (2). When
one topic-comment structure combined with another topic-
comment structure without any topics omitted, then the topic
chain can be put into the second sub-category. We use letter B
to stand for it. See example (3) and (4). The second group in
which the topics are dispensable can also be divided into an-
other two subcategories. We use letter C to stand for the topic
chain in which some topics are anaphorically omitted. See ex-
ample (5). Then we use letter D to stand for the topic chain in
which some topics are cataphorically omitted. See example (6).
In fact, the above four topic chains are the basic topic chains
in Chinese, they can be intercrossed in real language use. We
call this type as comprehensive topic chain. We use character E
to stand for it. See example (7). Since the comprehensive topic
chain can fix the sentences flexibly, it may be most used type
compared with the other four types.
(2) 关
guān
于
yú
这
zhè
件
jiàn
事
shì
情
qíng
(topic1), 他
tā
的
de
做
zuò
法
fǎ
(topic2), 我
wǒ
不
bù
赞
zàn
同
tóng
(comment).
(About this matter and his dealing way, I don’t agree.)
(3) 许
xǔ
多
duō
女
nǚ
人
rén
(topic1) 会
huì
笑
xiào
得
dé
这
zhè
样
yàng
甜
tián
(comment1),
︳但
dàn
她
tā
们
men
的
de
笑
xiào
容
róng
(topic2) 只
zhǐ
是
shì
面
miàn
部
bù
肌
jī
肉
ròu
柔
róu
软
ruǎn
操
cāo
(comment2).
(Many women can smile just as sweetly, but their smile is
only facial muscle calisthenics.)
(4) 我
wǒ
(topic1）没
méi
有
yǒu
借
jiè
口
kǒu
(comment1), ︳我
wǒ
(topic2) 无
wú
法
fǎ
解
jiě
释
shì
(comment2)。︳我
wǒ
(topic3) 不
bù
敢
gǎn
求
qiú
你
nǐ
谅
liàng
宥
yòu
(comment3）, ︳我
wǒ
(topic4) 只
zhī
希
xī
望
wàng
你
nǐ
快
kuài
忘
wàng
记
jì
我
wǒ
这
zhè
个
ge
软
ruǎn
弱
ruò
、没
méi
有
yǒu
坦
tǎn
白
bái
勇
yǒng
气
qì
的
de
人
rén
(comment4).
(I have no excuses and no way to explain. I couldn’t ask for
your forgiveness. I only hope you will quickly forget this cow-
ard who lacks the courage to be frank.)
(5) 苏
sū
小
xiǎo
姐
jiě
(topic1) 一向
yíxiàng
瞧不起
qiáobuqǐ
这
zhè
位
wèi
寒碜
hánchěn
的
de
孙太太
sūntàitài
(comment1), ︳而且
érqiě
Ø (topic2) 最
zuì
不
bù
喜欢
xǐhuān
小孩子
xiǎoháizǐ
(comment2),︳可是
kěshì
Ø (topic3) 听
tīng
了
le
这些
zhèxiē
话
huà
(comment3), ︳
Ø (topic4) 心
xīn
上
shàng
高兴
gāoxìng
(comment4).
(Miss Su had always scorned the poor, simple-minded Mrs.
Sun and detested children, but when she heard all that, she was
quite pleased.)
(6) Ø (topic1) 坐
zuò
洋
yáng
车
chē
拉
lā
错
cuò
了
le
地
dì
方
fāng
(comment1), Ø
(topic2) 买
mǎi
东
dōng
西
xī
错
cuò
付
fù
了
le
钱
qián
(comment2) ︳两
liǎng
人
rén
(topic3)
都
dōu
没
méi
好
hǎo
运
yùn
气
qì
(comment3).
(The rickshaws took them to the wrong place; they paid the
wrong amount of money when they went shopping; neither one
had any good luck.)

Y. X. HE
Open Access 346
(7) Ø (topic1) 回
huí
到
dào
卧
wò
室
shì
(comment1), Ø (topic2) 猜
cāi
疑
yí
种
zhòng
种
zhòng
(comment2), ︳韩
hán
学
xué
愈
yù
(topic3) 一
yí
定
dìng
在
zài
暗
àn
算
suàn
自
zì
己
jǐ
(comment3), ǁØ (topic4) 就
jiù
不
bù
知
zhī
道
dào
他
tā
怎
zěn
样
yàng
暗
àn
算
suàn
(com-
ment4), Ø (topic5) 明
míng
天
tiān
非
fēi
公
gōng
开
kāi
拆
chāi
破
pò
他
tā
的
de
西
xī
洋
yáng
镜
jìng
不
bù
可
kě
(comment5).
(Feeling like a guilty thief, he crept along stealthily and re-
turned to his room filled with suspicious. Han Xue-yu was un-
doubtedly plotting against him, but how he didn’t know. The
next day, he’d just have to publicly rip the cover from Han’s
little peep show.)
(“Ø” stands for the omitted topic anaphorically or cataphori-
cally; “︱” stand for the segment of simple sentences.)
The above five types are the topic chains at the Chinese sen-
tence level. There are topic chains at the clause level and at the
paragraph level. Here, only the topic chains at the sentence
level and clause level are concerned.
Data Collection
How do these topic chains distribute and make functions in
the organization of the text? In order to make an answer to this
question, we decided to have an analysis of the topic chains
used in the novel Fortress Besieged written by Zhongshu Qian
in 1991. Then why we choose this novel? It is because the lan-
guage of a novel genre can not only best embody the character-
istics of everyday language, but also reflect the characteristics
of written language, and this book is very popular in modern
China.
How do we choose the data? Specifically, we select one
paragraph of every five pages. Then we annotated the topic
chains manually according to the classifications mentioned
above. After that, we got 365 topic chains out of 547 clauses
among which there are 26 clauses in which topic chains were
not included.
Findings and Results
Based on the analysis of the five topic chains, the actual use
of topic chains in the novel Fortress Besieged can be summed
in Table 1.
Here, we see the comprehensive type occupies 60.5 percent
of all the topic chains. Type of C occupies the second largest
proportion, it accounts for 21.4 percent followed by type B and
D respectively. Type B accounts for 9.9 percent and type D
accounts for 7.7 percent. Type A only accounts for quite lim-
Table 1.
Types of topic chains.
Types Number Frequency (%)
A 2 0.5
B 36 9.9
C 78 21.4
D 28 7.7
E 221 60.5
ited proportion, it occupies 0.5 percent.
Discussions
Why do the five basic topic chains have different distribution
of proportion? If we want to answer this question, we should
have to dig out the different functions these topic chains can
enact in the organization of the discourse. Type A has unique
discourse function. In this type of topic chain, several topics
can be connected within a short simple sentence. Thus, the
reader can get fruitful information within a few words. This
type of topic chain can make a concise text. However, the
reader or hearer needs to know more details about the topics.
That is the reason why we have other topic chains.
In type B, the topics can be emphasized by contrast or by
repetition. See example (4) again. By repeating the same topic
“我
wǒ
(I)”, the author’s inner feelings of guilt are expressed most
vividly. Nevertheless, people do not always assume this type of
topic chain to account for the events since a stressed or op-
pressed voice is taken. Maybe this is the reason why we can
find a lower proportion in Table 1.
In type C, topics are anaphorically omitted or substituted.
Hence, the reader or hearer just focuses on the activities and
events that are concerned with the main topic which is usually
put at the depart position of a clause. Like pictures, all the ac-
tivities and events scroll slowly unfolded to the reader, and give
readers a gorgeous taste of rich connotation. In contrast, in type
D, the speaker or writer do not offer the topic, he or she just
begins his or her statement by describing the activities and
events, and then cleverly mention the topic at the later part of
his or her speech. This type of arrangement of the topic can
stimulate readers’ interest, suspension and imagination by fore-
shadowing the development of the activities and events. Com-
pared with type D, type C is more easily followed by the reader
or hearer psychologically. In another word, the reader or hearer
doesn’t have to use much effort to figure out the meaning of the
clauses. On the contrary, if the speaker or writer chose the topic
chain of type D, the reader or hearer has to focus on his or her
attention and constantly imagine what the speaker or writer
really wants to express. The reader or the hearer must pay more
efforts to figure out what the speaker or the writer really intends
to say. That is the reason why people use type C more fre-
quently than type D. People do not always want to be sus-
pended or inspired psychologically.
As mentioned above, type E can be realized by intercrossing
type A, type B, type C and type D. Through this way, people
can get a more varied discourse by using type E. In addition,
this type of topic chain structure can facilitate the speaker or
writer to orderly describe the complex events. Probably this is
the reason why this type of topic chain occupies the top posi-
tion among all the other topic chains.
Conclusion
Some basic theoretic concepts, such as topic and topic chains,
are discussed in this paper. Furthermore, based on the discus-
sion of topic chains and the calculation of the topic chains used
in the novel Fortress Besieged, the types and functions of Chi-
nese topic chains have also been discussed here. Since Chinese
language is a topic-prominent language, Chinese discourse ana-
lysis should be based on the analysis of topic chains, especially

Y. X. HE
Open Access 347
on the rules and functions by which the topic chains are organ-
ized into a complete text or discourse. In this paper, only the
topic chains at the sentence level are explored, the topic chains
at the paragraph level and text level are not discussed. Probably
the distribution of topic chains has something to do with the
writing style and discourse category, but all these questions are
needed to be explored further.
REFERENCES
Chu, C. (1998). A discourse grammar of mandarin Chinese. New York:
Peterlang.
Dik, S. (1997). The theory of functional grammar. New York: Mouton
de Gruyter.
Givón, T. (1983). Topic continuity in discourse: A quantitative cross-
language study. Philadelphia, PA: John Benjamins.
Halliday, M. A. K. (2008). An introduction to functional grammar (3rd
ed.). Beijing: Foreign Language Teaching and Research Press.
Hawkins, J. A. (1994) A performance theory of order and constituency.
Cambridge: Cambridge University Press.
Li, C., & Thompson, S. (1976). Subject and topic: A new typology of
language. In Le Charles (Ed.), Subject and topic (pp. 457-489). New
York: Academic Press.
Li, W. D. (2005) Topic chains in Chinese: A discourse analysis and
application in language teaching. Muenchen: Lincom Europa Aca-
demic Publications.
Qian, Z. S. (1991) Fortress besieged. Beijing: People Literature Press.
Tsao, F. F. (1990). A functional study of topic in Chinese: The first step
towards discourse. Taipei: Student Book Co.