Open Journal of Modern Linguistics
2013. Vol.3, No.1, 79-86
Published Online March 2013 in SciRes (
Copyright © 2013 SciRes. 79
Semantic-Communicative Structure and
Word Order in Mandarin Chinese
Antoine Tremblay1, David Beck2
1Dalhousie University, Halifax, Canada
2University of Alberta, Edmonton, Canada
Received November 5th, 2012; revised December 4th, 2012; accepted December 11th, 2012
In some languages more than in others, communicative considerations—such as what a message is about,
what information is new or old, and whether this or that participant is in the Speaker’s focus of atten-
tion—constrain the structure of a sentence. The goal of the present paper is to describe how different Se-
mantic-Communicative Structures affect word order in simple mono-transitive sentences without coverbs
or adverbial phrases in Mandarin Chinese. The discussion is couched in the Meaning-Text framework,
relevant parts of which are clarified at the onset of the paper. We argue that Subject-Verb-Object (SVO)
sentences are communicatively unmarked in that they do not signal any particular communicative consid-
eration. Other word orders, however, specifically encode certain communicative considerations. This is
the case of Prolepsisi-Subjecti-Verb-Object (PiSiVO) and Object-Subject-Verb (OSV) sentences, which
are discussed here.
Keywords: Semantic-Communicative Structure; Word Order; Mandarin Chinese; Meaning-Text Theory
It has been argued by a number of linguists, among others,
Chao (1968), Hu (1995), LaPolla (1988, 1993, 1995), Li and
Thompson (1975, 1976, 1989), Li (2005), and Van Valin and
LaPolla (1997), that word order in Mandarin Chinese (hence-
forth MC) is determined to a great extent by informational/
communicative considerations (we will use the term “commu-
nicative” throughout this paper). In other words, a certain com-
municative structure will give rise to a certain order of con-
stituents within the sentence. Li and Thompson (1976), among
others, argue that the Topic-Comment communicative opposi-
tion is the determining factor affecting word order. Li and
Thompson (1976: pp. 461-465) define the notion of Topic as a
definite NP (i.e., one which the Addressee already knows and
can identify) that specifies “the domain within which the pre-
dication holds” (i.e., it is the “centre of attention”). They main-
tain that Topics always occur in sentence-initial position1.
Although the sentence-initial element may be, and often is,
definite and the centre of attention, in some instances it is nei-
ther. Consider the example in 1), where the interlinear glosses
PFV and DE stand for perfective aspect and possessive respec-
tively. Note that throughout the paper, the context in which
examples are used is given between square brackets. Each con-
text establishes a specific communicative structure and conse-
quently constrains the set of sentences that can be used within it.
Also note that the acceptability of each sentence in each context
given in this paper was checked against linguistic intuition of
nine native speakers of Mandarin Chinese. It is assumed our
informants were able to extract the communicative structure
from the contexts and provide acceptability judgments accord-
1) [A 走进客厅, 看到许多糖果包装纸洒在四周. 她问 B
以下问题: “A comes into the living room and sees many candy
wrappers lying all around. She asks B the following question”]:
A: ?
[sheiComment] [chi le wo de tang]Topic
who eat PFV I DE candy
“Who ate my candy?”
B: 张三 .
[zhang.san]Comment [chi le ni de tang]Topic
Zhangsan eat PFV you DE candy
“Zhangsan ate your candy”
In 1), the Topicexpressed by the phrase consisting of the
Verb 吃了 chi le “ate” and the definite NP 你的糖 ni de tang
“your candy”appears after the Comment, which is expressed
by 张三 Zhangsan. Smith’s (1991) test of Topichood, cited in
Mel’čuk (2001: p. 105), can be used to demonstrate that 张三
Zhangsan is not a Topic. Smith’s test can briefly be described
as follows. A sentence can be paraphrased using the expression
Speaking of X, ··· if X expresses the Topic of the sentence (or
part of it). This is exemplified by the two following English
2) a) [Where]Comment is [the rabbit?]Topic.
b) [The rabbit]Topic [is probably in Alices garden]Comment.
c) Speaking of the rabbiti, [hei]Topic [is probably hiding in
Alices garden]Comment.
1Also see Lambrecht (1994: Chapter 4), and especially Section 4.7, pages
199-205, for a discussion of the “Topic-first principle”. It is similar to La-
Polla’s (1995: 310) proposal, according to which topical elements occur in
pre-verbal position while focal elements appear in post-verbal position.
d) *Speaking of Alices gardenj, [the rabbit]Topic [is proba-
bly hiding therej]Comment.
(adapted from Mel’čuk, 2001: p. 105)
The rabbit in 2c) passes the test because it expresses the
Topic, whereas Alices garden in 2d) does not because it ex-
presses the Comment. It is assumed here that in MC, 说到
shuodao ··· is used in the same manner as English Speaking of
··· Turning back to our example in 1), it is shown in 3) that
Zhangsan is not a Topic.
3) a)*说到 张三 .
shuo.dao zhang.san ta chi le ni de tang
speaking.of zhang.san he eat PFV you DE candy
“Speaking of Zhangsan, he ate your candy”
b) 说到 张三
shuo.dao ni de tang zhang.san chi le
speaking.of you DE candy zhang.san eat PFV
“Speaking of your candy, Zhangsan ate them”
张三 Zhangsan in 3a) fails the test because it is a Comment,
while the sentence in 3b) is grammatical given that 你的糖 ni
de tang “your candy” is part of the Topic, which, we wish to
stress, occurs after sentence-initial 张三 Zhangsan.
In addition to being definite, the sentence-initial element can
be indefinite, which runs counter to Li and Thompson’s (1976)
claim stated above. Consider the sentence shown in 4), where
the interlinear gloss CL stands for classifier.
4) [一群人围成一堆在看什么东西, 张三看不到, 问其中
一个围观者发生什么事了, 这个人跟张三说: “People are
gathering around something. Zhangsan cannot see and asks one
of the bystanders what happened. The latter tells him the fol-
工人 受伤
[[yi ge]Indefinite shou.shang le]Comment
one CL worker injure PFV
“A worker was injured”
The sentence in 4) is an all-Comment sentence. Neither
个工人 yi ge “a worker” or 受伤了 shou.shang le
“be injured” pass the 说到 shuo.dao ··· test (not shown here).
In this paper, we argue that word order in MC is deter-
minedamong other thingsby a number of different commu-
nicative considerations, which are called Semantic-Communi-
cative-oppositions within the Meaning-Text Theory (Žolkovskij
& Mel’čuk, 1967; Mel’čuk, 1988, 2001). It is important to
stress the fact that Mel’čuk’s (2001) perspective on communi-
cative organization synthesizes, insofar as possible, the huge
body of literature on the subject matter. As he mentions in the
introduction to his book, he is in no way re-inventing the wheel.
Rather, Mel’čuk integrated work done by many researchers
spanning many decades to form what he argues to be a set of
eight (semantic) communicative oppositions, which, depending
on the language, affect the “translation” of meaning (semantic
context) into text (a sentence).
In this study, our goal is to describe how different Seman-
tic-Communicative Structures affect word order in simple
mono-transitive sentences without coverbs or adverbial phrases2.
We constrain our discussion to those Semantic-Communicative
Structures that affect the initial position of a sentence3. In Sec-
tion 2, we review relevant aspects of the Meaning-Text Theory.
In Section 3, we discuss simple SVO, PiSiVO, and OSV sen-
tences. Finally, we give concluding remarks in Section 4.
Meaning-Text Theory
In this section we introduce relevant aspects of the Mean-
ing-Text Theory (henceforth MTT; see Mel’cuk, 1988, 2001,
for details). In the MTT framework, every utterance has a Se-
mantic Structure (SemS), which encodes the propositional
meaning of a sentence; this is the “objective” meaning of an
utterance, which is structured as a connected oriented network
of labeled nodes. A sample SemS is shown in Figure 1 (tense
and number are not shown; single quotation marks are used to
indicate the signified, that is, the semantics, of a linguistic
In Figure 1, the numbers labeling the arcs differentiate the
arguments of a functor (e.g., a verb, a preposition, etc.). That is,
“John” is the 1st argument of the functor “meet1” and “doctor”
is its 2nd argument, whereas “meet1” is the 1st argument of the
functor “place” and “airport” is its 2nd argument. As the English
sentences in 5) show, a single SemS can give rise to many sur-
face syntactic forms.
5) a) John met the doctor at the airport [neutral prosody].
b) JOHN met the doctor at the airport [heavy stress on
c) It was John who met the doctor at the airport.
2Coverbs are defined in Po-Ching and Rimmington (2004) as verbs that are
similar to English prepositions and that generally occur in conjunction with
other verbs (e.g., dui “towards, facing”, xiang “heading, towards”, and zi
3Word order in simple intransitive sentences has been discussed to some
extent in Tremblay and Beck (2007). We relegate the analysis of di-transi-
tive sentences and SOV sentences to future resea
Figure 1.
A sample Semantic Structure (adapted from Mel’čuk, 2001: p. 5).
Copyright © 2013 SciRes.
The surface forms given in 5) depend on the Semantic-
Communicative Structure (Sem-CommS) of each utterance,
which specifies the manner in which the Speaker wishes to or-
ganize a message against the backdrop of a linguistic and/or
extra-linguistic context (including world knowledge). Though
the sentences in 5) all convey the same basic message, that is,
that a man named John met a specific doctor at a specific air-
port, the Sem-CommS provides additional meaning, which is
“superimposed” on top of the basic message. In 5a), the Speak-
er might simply be reporting on John’s activities. In 5b), the
Speaker may have vehemently wished to be the person assigned
to meet a very famous doctor at the airport; instead another co-
worker named John, who couldn’t care less about the doctor,
was given the task (John is emotionally prominent for the
Speaker). The sentence in 5c) could be uttered in response to a
statement, which the Speaker knows to be false, about Mary
meeting the doctor at the airport (John would be, in this case,
logically prominent).
In order to formally capture this “extra-layer of meaning”,
Mel’čuk formulated eight Semantic-Communicative oppositions
by integrating, as much as possible, the colossal body of litera-
ture on communicative organization. These oppositions are: i)
Thematicity (roughly the Topic-Comment dichotomy); ii) Given-
ness; iii) Focalization; iv) Perspective; v) Emphasis; vi) Pre-
supposedness; vii) Unitariness; and viii) Locutionality. An ex-
ample of a SemS with a partial Sem-CommS superimposed on
it is given in Figure 2.
In Figure 2, “John” is the Rheme of the SemS, while
“meet1”, “place”, “airport”, and “doctor” are part of the Theme.
In addition, “John” is Focalized: He is presented by the Speaker
as being logically prominent (s/he is communicating that it is
precisely John and no one else).
The Sem-CommS determines the Deep-Syntactic Structure
(DSyntS) of a sentence, which in turn determines its surface
form. The partial DSyntS corresponding of the partial Semantic
Representation (which is shown in Figure 2) is given in Figure
3. Note that small caps indicate lexemes, that is, the pairing of a
signified to a signifier.
The arrows indicate Deep-Syntactic Relations (DSyntRels)
such as the actantial DSyntRels I, II, the Attr(ibutive) DSyntRel,
and the Coordinate DSyntRel (see Mel’čuk, 1988: pp. 63-67 for
details). The dashed bi-directional arrow shows obligatory co-
reference between the two occurrences of the lexeme JOHN. In
Figure 3, i) MEETActive has a DSyntRel I relation to IT-BE; ii)
JOHN has a DSyntRel II relation to IT-BE as well as a DSyn-
tRel I relation to MEETActive; iii) DOCTOR has a DSyntRel II
relation to MEETActive; iv) AT—> AIRPORT has an Attributive
relation to MEETActive; and v) AIRPORT has a DSyntRel II
relation to AT (see Mel’cuk, 1988, for more information). In
addition, the superimposed Deep-Syntactic-Communicative
Structure indicates that vi) JOHN is the Rheme; vii) MEETactive,
DOCTOR, AT, and AIRPORT are part of the Theme, and viii)
JOHN is Focalized4.
Of the eight Sem-Comm oppositions, four are relevant to the
present discussion namely, Thematicity, Givenness, Focaliza-
tion, and Perspective; they are characterized in the following
Figure 2.
The SemS + partial Sem-CommS of the sentence It is John who
met the doctor at the airport.
Figure 3.
A partial Deep-Syntactic Representation of Figure 2 (adapted
from Mel’čuk, 2001: p. 9).
Thematicity has three values: Theme, Rheme, and Specifier,
the last of which will not be considered here. The Theme-
Rheme opposition is the most universal and relevant in that a
proposition necessarily says something (the Rheme) about
something (the Theme). The Semantic Theme of a sentence can
be defined as the part of its SemS that corresponds to what the
message is about, and the Semantic Rheme as what is stated
about the Theme by means of the sentence. By way of example,
let us consider the short English dialogue shown in 6).
6) a) [What]Rheme [did John cat c h?]Theme.
b) [John caught]Theme [a rabbit]Rheme.
The configuration John caught expresses the Theme and a
rabbit expresses the Rheme. That is, the event of John catching
something is talked about and it is said about it that what was
caught is a rabbit. Note that a sentence (i.e., a finite clause)
necessarily has a Rheme. If no Rheme is present, the initial
SemS gives rise to a nominal or infinitival phrase.
The notion of Theme is related to that of Topic, though the
relation depends on how Topic is defined. For some, Theme
and Topic are one and the same, while for others Theme is
merely a portion of what constitutes a Topic. This will be dis-
cussed in more detail in Section 4.
4The Deep-Syntactic-Communicative Structure specifies the division of the
sentence into Theme-Rheme, Given-New, etc. Part of what constitutes the
Sem-CommS is encoded, at this level, in the lexical choices made: For
instance, choosing the frame IT-BE X WHO Y to indicate that X expresses
a Focalized Rheme in English.
The Sem-Comm-opposition of Givenness is composed of
three values: Given, New, and Not-Applicable. The value Given
Copyright © 2013 SciRes. 81
can be defined as the part of the SemS of a sentence that the
Speaker presents as being in the Addressee’s current con-
sciousness or at least easily accessible for the Addressee. To
say that something is in the Addressee’s consciousness is to say
that the Addressee can foresee it coming up next in the dis-
course and can uniquely identify its referent. This state of af-
fairs can arise from i) linguistic context (e.g., preceding utter-
ances) or ii) extra-linguistic context (e.g., world/encyclopedic
knowledge shared by both the Speaker and the Addressee
and/or from the situation in which the discourse takes place).
The value New, on the other hand, characterizes words that are
not in the Addressee’s consciousness. That is, a New configura-
tion is not accessible from either linguistic or extra-linguistic
context nor can a pre-existing identity be found for its referent.
Finally, Not-Applicable means that the Given-New division
does not apply for a certain semantic configuration.
The Given-New opposition is an Addressee-oriented Sem-
Comm-opposition in that the Speaker chooses which part of the
SemS is Given and which part is New according to what infor-
mation he believes the Addressee can or cannot access. Some
English examples are given below.
7) a) [The brothers had the quarrel over the book]Given.
b) [(Some) brothers]New [had the quarrel]Given [over the
c) [The brothers]Given [had a quarrel]New [over the book]Given.
d) [(Some) brothers had a quarrel over a book]New.
(adapted from Mel’čuk, 2001: pp. 161-162)
In 7a), the whole sentence is Given, that is, the Speaker be-
lieves that the Addressee can foresee what is coming up in the
discourse and/or can assign unique identities to the referents. In
7b), however, the Speaker believes that the Addressee does not
know “the brothers’ identities”; this part of the sentence is
therefore encoded as New. In 7c), the Addressee is believed to
know nothing about the quarrel, which is coded as New. Finally
in 7d), the Speaker thinks that the identity of the brothers, the
quarrel, and the book are all unknown to the Addressee and the
whole sentence is coded as New.
Focalization has two values: Focalized and Non-Focalized.
The former is defined as the part of a proposition that the
Speaker presents as being logically prominent for him, that is,
which is in the Speaker’s focus of attention. A logically promi-
nent configuration excludes any other possibilities. A Non-
Focalized element is simply not logically prominent (it is the
unmarked value of the division). An example of Focalization in
English was given in 5c), repeated here under 8).
8) [A is reporting on Mary’s activities at a meeting. B knows
John went to the airport to meet the doctor, not Mary].
A: ··· and Mary went to the airport to meet the doctor.
B: It was John who met the doctor at the airport.
In 8), John is logically prominent for B), who is telling A)
and the people at the meeting that the person who met the doc-
tor at the airport was precisely John, not anyone else.
Before moving on to the main portion of the paper, let us de-
fine the following two important concepts. The first one is the
notion of “Prolepsis”. The Prolepsis (P) is a sentential element
that always occurs to the left of a sentence. It is syntactically
very loosely connected to it and allows a pause to separate it
from the rest of the sentence (Mel’čuk, 2001: p. 130). The second
concept pertains to the “communicative markedness” of a word
order. A communicatively unmarked word order is one that
neutralizes different Sem-CommSs, that is, it can be used in
sentences that express a number of different Sem-CommSs.5 A
word order is marked with respect to another one if it en- codes
fewer Sem-CommSs (potentially only one) than the word order
it is compared to. By way of example, let there be word orders
a) and b), and Semantic-Communicative Structures α and β
where i) the Subject in Sem-CommS α is Thematic, and ii) the
Subject in Sem-CommS β is Rhematic. Now, let us suppose
that word order a) can encode Sem-CommSs α and β whereas
word order b) can only be used to encode Sem-CommS α. Then
word order b) is said to be marked relative to word order a). By
way of example, let us consider the English SVO sentence in 9),
where the Subject John expresses the Rheme.
9) a) [Who]Rheme [met the doctor at the airport?]Theme.
b) [John]Rheme [met the doctor at the airport]Theme.
The Subject in English SVO sentences can also express (part
of) the Theme, as in 10).
10) a) [Who]Rheme [did John meet at the airport?]Theme.
b) [John met]Theme [the doctor]Rheme [at the airport]Theme.
These two examples show that the Subject in English SVO
sentences can be either Rhematic or Thematic. The Subject in
Subject-clefted sentences, however, can only be Rhematic.
Consider the short dialogue in 11), where A’s answer is unac-
11) [Two coworkers are talking about John, another co-
worker who is the boss’ favorite employee. A famous doctor
will be giving a presentation in the next few days and John
asked to meet him at the airport. The two coworkers, who are
big fans of the doctor’s, know that John couldn’t care less about
A: ··· John was sent to the airport yesterday to meet someone.
Do you know who?
B: No, who?
A: *It was [[John] Focalized who met]Theme [the doctor]Rheme [at
the airport]Theme.
B: Why didnt they send one of us two? They know were
huge fans of the doctors!
In 11), John is a Theme because he is (part of) what the
message is about; he is also Focalized because he and no one
else went to the airport to meet the doctor. A’s answer is unac-
ceptable because in English Focalized Thematic Subjects are
encoded by Pseudo-cleft sentences, as instantiated in 12), while
Focalized Rhematic Subjects are encoded by Subject-cleft sen-
tences, as shown in 13).
12) [The one whom John met at the airport]Theme,Focalized [was
the doctor]Rheme.
13) a) [Who]Rheme [did John meet at the airport?]Theme.
b) [John met]Theme [Jack]Rheme [at the airport]Theme.
c) [It was the doctor]Rheme, Focalized [whom John met at the
Given that SVO sentences can encode Thematic and Rhe-
matic Subjects whereas the (Focalized) Subject in Subject-
clefted sentences can only be Rhematic, it can be said that SVO
sentences in English are communicatively unmarked with re-
spect to Subject-clefted sentences. We are now in a position to
begin our discussion of word order in Mandarin Chinese.
Simple Mono-transitive Sentences in Mandarin
Word order in Mandarin Chinese, though flexible, is con-
5See Beck (2002: pp. 20-41) for a discussion of markedness.
Copyright © 2013 SciRes.
strained by Semantic-Communicative factors. In this section,
we describe the Semantic-Communicative Structure of simple
mono-transitive Subject-Verb-Object (SVO), Prolepsisi-Sub-
jecti-Verb-Object (PiSiVO), and Object-Subject-Verb (OSV)
sentences. We argue that SVO sentences are communicatively
unmarked, while the other two sentence types are marked. That
is, whereas SVO sentences neutralize a number of Sem-
CommSs, PiSiVO sentences encode Given Focalized Subjects,
and OSV sentences specifically encode Focalized Objects. We
will begin the discussion by demonstrating that the Sem-
Comm-opposition of Thematicity does not affect word order in
mono-transitive sentences.
First consider the sentence in 14), where 小李 xiao li ex-
presses the Rheme and 打破了花瓶 da po le huaping “broke
the flower vase” the Theme.
14) a) 打破 花瓶?
[shei]Rheme [da.po le]Theme
who hit.break PFV flower.vase
“Who broke the flower vase?”
b1) 打破 花瓶.
[xiao li]Rheme [da.po le]Theme
little li hit.break PFV flower.vase
“Little Li broke the flower vase”
b2) 花瓶 小李 打破 .
[]Theme []Rheme [da.po le]Theme
flower.vase little.Li hit.break PFV
“The flower vase little Li broke”
In 14), either the Subject (expressing the Rheme) or the Di-
rect Object (expressing part of the Theme) can appear in sen-
tence-initial position. Now consider the following sentences,
where the Semantic Subject 张三 Zhangsan, which is the
element of the Semantic Representation expressed both as the
Prolepsis (P) and the Syntactic Subject (S) at the Deep-Syntac-
tic Representation, can be a Theme 15) or a Rheme 16).
15)a) 张三 打破 什么?
[zhang.san [da.po le]Theme []Rheme
zhang.san hit.break PFV what
“What did Zhangsan break?”
b) 张三 打破
[zhang.sani tai da.po le]Theme
zhang.san he hit.break PFV
[ni de]Rheme ([PSV]ThemeORheme)
you DE flower.vase
“Zhangsan he broke your flower vase”
16) a)怎么 ?
[ le]Rheme
what happen PFV
“What happened?”
b) 张三 打破
[zhang.sanitai da.po le]Theme
zhang.san he hit.break PFV
[ni de]Rheme ([PSV]ThemeORheme)
you DEflower.vase
“Zhangsan he broke your flower vase”
The Prolepsis Zhangsan in 15) and 16) occurs in sentence-
initial position because it is Given and Focalized (see below),
not because it expresses the Theme or the Rheme. The data
presented here clearly shows that Thematicity is not sufficient
to describe word order in MC transitive sentences; other Se-
mantic-Communicative oppositions are needed to account for
the facts. In the following section, we discuss SVO sentences in
more detail.
SVO Sentences
As was shown above, the Subject in SVO sentences can ex-
press either a Theme or a Rheme. In 19) and 20), it is shown
that the Subject can also be Given or New, respectively.
17) [说到小王 “Talking about little Wang”]
编辑 这 杂志
[ta]Theme, Given bian.ji zhe ben za.zhi (STheme, Given VO)
he edit this CL magazine
“He edits this magazine”
18) [小王对老朱说: “Little Wang announces the following
to old Zhu”]:
有人 吃 了 你 盒饭.
[]Rheme, Newchi le ni de (SRheme, New VO)
someone eatPFVyou DE
“Someone ate your lunch box”
In the following section, we discuss PiSiVO sentences.
PiSiVO Sentences
As mentioned earlier, the Prolepsis (P) is a sentential element
that always occurs to the left of a sentence; it is syntactically
very loosely connected to the rest of it and allows a pause
separating it from the rest of the sentence (Mel’čuk, 2001: p.
130). When the Semantic Subject (in the Semantic Representa-
tion) is Given and Focalized (i.e., it is logically prominent), it is
encoded in the Deep-Syntactic Representation both as a Prolep-
sis (P) and a Subject (S), giving rise to PiSiVO sentences. The
examples shown in 19) and 20) demonstrate that the Semantic
Subject in PiSiVO sentences is necessarily Given.
19) [战场上一个军营受到敌人的猛烈攻击, 士兵张三开
枪误伤了刘中尉. 战斗结束后, 朱战士对王战士说: “At war,
a military camp sustained a heavy assault by the enemy. Zhang-
san, a soldier, mistakenly shot lieutenant Liu. After the assault,
soldier Zhu tells soldier Wang the following”]:
Copyright © 2013 SciRes. 83
张三 开枪误伤
[zhang.sani tai] Given, Focalized kai.qiang.wu.shang
zhang.san he shoot.wrongly.hurt
le liu zhong.wei ([PiSi]New, FocalizedSO)
PFV liu lieutenant
“Zhangsan he shot lieutenant Liu”
In 19), the Semantic Subject is Given and the sentence is ac-
ceptable (it was deemed acceptable by 77.8% (7/9) of the native
speakers of Mandarin Chinese we asked for grammaticality
judgments). The sentence in 20), however, is unacceptable be-
cause the Focalized Semantic Subject is New. Note that only
77.8% (7/9) native speakers of Mandarin Chinese deemed this
sentence unacceptable in the given context (i.e., 22.2% deemed
it acceptable).
20) [战场上一个军营受到敌人的猛烈攻击, 一个士兵开
枪误伤了刘中尉. 战斗结束后, 中士向上校汇报了这个情况:
“At war, a military camp sustained a heavy assault by the en-
emy. A soldier mistakenly shot lieutenant Liu. After the assault,
the sergeant reported this to the colonel”]:
* 士兵 开枪误伤
 [yi ge shi.bin]New, Focalized tai kai.qiang.wu.shang
one CL soldier he shoot.wrongly.hurt
中尉. ([PiSi]New, FocalizedVO)
le liu zhong.wei
PFV liu lieutenan t
“A soldier he shot lieutenant Liu”
Given that PiSiVO sentences specifically encode Focalized
Given Semantic Subjects, this sentence type is communica-
tively marked with respect to SVO sentences. We now turn to
OSV sentences.
OSV Sentences
It is commonly believed that the Object in OSV sentences
expresses the Theme of the sentence (Wei, 1989; Li et al., 1992;
Mel’čuk, 2001; Paul, 2002). This is shown in 21).
21) [秘书看到自己的花瓶碎了. 她问小刘怎么回事儿.
刘说: “The secretary sees that her flower vase is broken. She
asks little Liu what happened to it. Little Liu says the follow-
花瓶 老板 打破
[ni de]Theme [lao.ban da.po le]Rheme
you DE flower.vase boss hit.break PFV
“Your vase, the boss broke”
Although the Object in OSV sentences can be Thematic, it
can also express the Rheme. This is illustrated in 22).
22) [李四问张三吃不吃鱼鳍, 张三说不吃. 李四问他吃不
吃鱼尾, 张三说也不吃, 然后说: “Lisi asks Zhangsan if he
eats a fish’s fins. Zhangsan says he doesn’t. Lisi then asks him
if he eats the tail. Zhangsan says he doesn’t and then says”]:
[yu.tou]Rheme [wo chi]Theme (ORhemeSV)
fish.head I eat
“Fish heads I eat”
That 鱼头 yu.tou “fish heads” is a Rheme can be shown
with the help of the 说到 shuo.dao “speaking of” for The-
maticity in 23).
23) *说到 鱼头 .
shuo.dao yu.tou wo chi
speaking.of fish.head I eat
“Speaking of fish heads I eat them”
It is claimed here that Objects in OSV sentences occur in
sentence-initial position not because they are what the message
is about (i.e., Themes), but rather by virtue of being Focalized
(i.e., logically prominent for the Speaker). In other words,
的花瓶 ni de “your flower vase” in 21) occurs in
sentence-initial position because the Speaker wishes to com-
municate that it is the flower vase that the boss broke and not
something else, and in 22) 鱼头 yu.tou “fish heads” is the first
element of the sentence because the Speaker eats precisely fish
heads, not the fins or its tail.
The Object in OSV sentences cannot be New and Focalized
in all-Rhematic sentences. This is illustrated in 24).
24) [老赵回家了. 他发现小周在哭. 老赵说: “Old Zhao
comes back from work. He sees little Zhou crying and says the
老赵: 怎么 ?
old Zhao: le
what.happen PFV
“What happened?”
小周: *一 棵 政府
little Zhou:[[yi ke lao shu]New,
Focalized shi zheng.fu
one CL old tree city government
kan le]Rheme (ONew, FocalizedSV
cut PFV
“An old tree the city cut down”
Rather, in such a situation a passive bei constructions
will be used and the Semantic Object (i.e., the 2nd actant of the
functor kan “cut” in the Semantic Representation) will be
encoded at the Deep-Syntactic Representation as a Syntactic
Subject; this is shown in 25).
25) 一 棵 老树 市 政府
[[yi ke lao shu]New, Focalized bei shizheng.fu
oneCL oldtree PASSIVE citygovernment
砍 了.
kan le]Rheme
“An old tree was cut down by the city”
Copyright © 2013 SciRes.
If the Object is Given and Focalized, as in 26), then an OSV
sentence is acceptable.
26) [前院里有一棵小周特别喜欢的老树, 老赵回到家发
现小周在哭. 老赵说: “Little Zhou has an old tree in the front
yard she loves a lot. Old Zhao comes back home from work. He
sees little Zhou crying and asks”]:
老赵: 怎么 ?
old Zhao: le
what.happen PFV
“What happened?”
小周: 政府
little Zhou: [[wo de lao shu]New, Focalized shi zheng.fu
I DE old tree citygovernment
kan le]Rheme (ONew, FocalizedSV
cut PFV
“My old tree the city cut down”
It can be shown that the Object in OSV sentences is Focal-
ized but not in SVO sentences by using the construction
“even” (where lian means “even” and dou means
“all”). This construction has been said to indicate focus/em-
phasis/contrast (Zhang, 2000; Chen, 2004; Shyu, 2004; Wang,
2008) and is roughly taken to mean “in addition to X, and
against expectations, also Y”. Consider the sentences given in
27) [John, 一个加拿大人, 和李四在一家餐馆吃饭. John
问李四他吃不吃鱼. 李四说他吃, 然后接着说: “At a restau-
rant, John, a Canadian, asks Lisi if he eats fish. Lisi says that
she does and adds”]:
a) * .
wo dou chi lian [tou]Focalized (SVOFocalized)
I all eat evenhead
“I eat even the head”
b) .
lian [tou]Focalized wo dou chi (OFocalizedSV)
even head I all eat
“Even the head I eat”
The sentence in 27a) is ungrammatical because Focalized
Objects cannot appear in post-verbal position. The sentence in
27b), on the other hand, is acceptable given that the Focalized
Object occurs in pre-verbal (sentence-initial) position where
··· lian ··· dou highlights a contrast between the body of the
fish and its head: John would expect Lisi to eat the flesh, but
that Lisi also eats the head goes against his expectation (Wang,
2008, pp. 878-879).
Note that lian ··· dou “even” is not a Focalizer. As
shown in 28), where the context is the same as in 27), the Ob-
ject does not need lian ··· dou “even” to be Focalized.
[tou]Focalized wo chi (ORhemeSV)
head I eat
“the head I eat”
Given that OSV sentences specifically encode Focalized Ob-
jects and that they encode fewer Sem-CommSs than SVO sen-
tences, the OSV sentences are communicatively marked.
The goal of the present paper was to describe how different
Semantic-Communicative Structures affect word order in sim-
ple mono-transitive sentences without coverbs or adverbial
phrases in Mandarin Chinese. We first demonstrated that the
Theme-Rheme opposition is not sufficient to account for word
order in MC, that is, the sentence-initial element in mono-tran-
sitive sentences can express the Theme or the Rheme. Subse-
quently, we showed that the Given-New opposition in concert
with the Focalized-Not-Focalized dichotomy dictate whether a
sentence exhibited the SVO, the PiSiVO, or the OSV word or-
der. Whereas the Subject in SVO sentences can be a Given
Theme, a New Theme, a Given Rheme, or a New Rheme, the
Subject in OSV sentences can be Given and Focalized or New
and Focalized. In PiSiVO sentences, however, it has to be Giv-
en and Focalized.
Word order in Mandarin Chinese has been a topic of discus-
sion in modern linguistics at least since the 1950’s. Since then,
a great number of studies have attempted to understand the
underpinnings of word order in this language. Many researchers,
including ourselves, agree that it is determined to a great extent
by communicative consideration. To this date, theories of MC
word order have relied on the Topic-Comment dichotomy.
However, it remains unclear what a Topic is exactly. For Dik
(1980), Topic is equivalent to the notion of Theme used in this
paper. For Li and Thompson’s (1976), the concept of Topic
encompasses the Semantic-Communicative values of Theme,
Given (which is equal at least in part to their definite), and in
some cases Focalized. We hope to have convincingly demon-
strated that it is preferable to dissociate these values in order to
properly account for the different word orders in mono-transi-
tive sentences in Mandarin Chinese.
Beck, D. (2002). The typology of parts of speech systems: The marked-
ness of adjectives. New York and London: Routledge.
Chao, Y.-R. (1968). A grammar of spoken Chinese. Berkeley and Los
Angeles, CA: University of California Press.
Chen, C.-F. (2004). Asymmetrical behaviors of presuppositions of lian-
constructions in Chinese. Unpublished Manuscript.
Dik, S. C. (1981). Functional grammar. Dordrecht and Cinnaminson:
Hu, W. (1995). Functional perspectives and Chinese word order. Ph.D.
Dissertation, Columbus: Ohio State University.
Lambrecht, K. (1994). Information structure and sentence form: Topic,
focus, and the mental representation of discourse referents. Cam-
bridge: Cambridge University Press.
LaPolla, R. J. (1988). Topicalization and the question of lexical pas-
sives in Chinese. Third Annual Conference on Chinese Linguistics,
13-14 May 1988, Columbus: Ohio State University.
LaPolla, R. J. (1993). Arguments against “subject” and “direct object”
as viable concepts in Chinese. Bulletin of the Institute of History and
Philology, 63, 759-813.
LaPolla, R. J. (1995). Pragmatic relations and word order in Chinese. In
P. Dowing, & M. Noonan (Eds.), Word order in discourse (pp.
297-329). Amsterdam and Philadelphia: John Benjamins.
Li, C. N., & Thompson, S. A. (1975). The semantic function of word
Copyright © 2013 SciRes. 85
Copyright © 2013 SciRes.
order: A case study in Mandarin. In C. N. Li (Ed.), Word order and
word order change (pp.163-195). Austin and London: University of
Texas Press.
Li, C. N., & Thompson, S. A. (1976). Subject and topic: A new typol-
ogy of language. In C. N. Li (Ed.), Subject and topic (pp. 457-489).
New York: Academic Press.
Li, C. N., & Thompson, S. A. (1989). Mandarin Chinese: A functional
reference grammar. London and New York: University of California
Li, P., Bates, E., Liu, H., & McWhinney, B. (1992). Cues as functional
constraints on sentence processing in Chinese. In C. Hsuan-Chih, &
A. J. L. Tzeng (Eds.), Language processing in Chinese (pp. 207-
233). Amsterdam: North-Holland.
Li, W. (2005). Topic chains in Chinese: A discourse analysis and ap-
plications in language teaching. Muenchen: Lincom Europa.
Mel’čuk, I. A. (1988). Dependency syntax: Theory and practice. New
York: SUNY Press.
Mel’čuk, I. A. (2001). Communicative organization in natural lan-
guage: The semantic-communicative structure of sentences. Amster-
dam and Philadelphia: John Benjamins.
Po-Ching, Y., & Rimmington, D. (2004). Chinese: A comprehensive
grammar. New York and London: Routledge.
Shyu, S. I. (2004). (A)symetries between Mandarin Chinese liandou
and shenzhi. Journal of Chinese Linguistics, 32, 81-128.
Smith, C. (1991). Sentences in text: A valediction for sentence topic. In
C. Georgopolous, & R. Ishihara (Eds.), Interdisciplinary approaches
to language (Essays in Honour of S.-Y. Kuroda) (pp. 545-564). Dor-
drecht: Kluwer Academic Press. doi:10.1007/978-94-011-3818-5_29
Tremblay, A., & Beck, D. (2007). Word order in Mandarin Chinese
intransitive clauses. In K. Gerdes, T. Reuther, & L. Wanner (Eds.), In
Meaning-Text Theory 2007. Proceedings of the 3rd International
Conference on Meaning-Text Theory (pp. 417-426). München and
Wien: Wiener Slawistischer Almanach.
Van Valin Jr., R. D., & LaPolla, R. J. (1997). Syntax. Structure, Mean-
ing and Function. Cambridge: Cambridge University Press.
Wang, X. (2008). The semantics and pragmatics of liandou/ye, lian,
dou, and ye. In M. K. M. Chan, & H. Kang (Eds.), Proceedings of the
20th North American Conference on Chinese Linguistics (NACCL-
20). Volume 2 (pp. 875-891). Columbus: The Ohio State University.
Wei, H. M. (1989). Word order of Mandarin Chinese. Ph.D. Disser-
tation. East Lansing, MI: Michigan State University.
Zhang, N. (2000). Object shift in Mandarin Chinese. Journal of Chi-
nese Linguistics, 28, 201-246.
Žolkovskij, A., & Mel'čuk, I. A. (1967). O semanticeskom sinteze [On
semantic synthesis]. Problemy Kybernetiki, 19, 177-238.