Feng 2009 Minimal Word and Its Function in Mandarin Chinese

3
Minimal word and its function in Mandarin

Chinese
Shengli Feng
All rights reserved. May not be reproduced in any form without permission from the publisher, except fair uses permitted under U.S. or applicable copyright law.
The notion of Minimal Word (MinWd) has always been a fundamental concept
in the Prosodic-Morphological systems developed since McCarthy and Prince
(1990). It is a prosodically circumscribed domain which may be selected as the
locus of morphological transformation in lieu of the whole domain (McCarthy and
Prince 1990, 1993, 1998). Theoretically, the notion of MinWd is derived from the
interaction of both Prosodic Hierarchy and Foot Binarity, as stated in the following
(taken from McCarthy and Prince 1998: 284):
1. Prosodic hierarchy
Prosodic Word PrWd
|
Foot Ft
|
Syllable σ
|
Mora μ
2. Foot binarity: Feet are binary under syllabic or moraic analysis.
The Prosodic Hierarchy impinges on every prosodic word to contain at least one
foot, while the Foot Binarity demands that every foot be bimoraic or disyllabic.
As a result, a prosodic word must contain at least two moras or syllables according
to the transitivity of the Prosodic Hierarchy. The Minimal Word is therefore a
Copyright 2009. Hong Kong University Press.
single PrWd in the system.1 As we will see below, the Minimal Word is of singular
importance in characterizing a wide range of Prosodic-Morphological phenomena
not only in languages known in previous studies, but also, as I would like to argue
in this chapter, in Mandarin Chinese.
EBSCO Publishing : eBook Collection (EBSCOhost) - printed on 4/6/2023 8:14 AM via UNIVERSITY OF CAMBRIDGE
AN: 359479 ; Edited by Janet Zhiqun Xing.; Studies of Chinese Linguistics
Account: s3859159.main.ehost
stud of Chinese Book-3.indd 47 2008/11/17 12:51:52 PM

48 Shengli Feng
Of course, we are fully aware that in Chinese morphology, there are complexities
which may not be accounted for only by the MinWd hypothesis.2 However, what I
am going to demonstrate is this: there are phenomena that can only be explained in
terms of a Minimal Word analysis. This is the very purpose of the present study.
This chapter is organized as follows. The section titled “Minimal word as
a condition for VO compound” consists of a study on the verb-object structure
and shows that among all VO forms in the language, only the ones that meet the
minimal word requirement exhibit word properties while longer forms are all
on a par with phrases. The section, “Minimal word as a condition for category
changing,” shows that a process of category changing from a [Aux V] verbal
expression to a [Aux-V]adjective compound is conditioned strictly on whether or not
the [Aux V] is a minimal word. “Minimal word as a condition for morphosyntactic
operation” demonstrates that there is a clear distinction between MinWd and non-
MinWd [A+N] forms differentiated syntactically as well. The last section is a
summary of this study.
Minimal word as a condition for VO compound

In Chinese traditional linguistics, it has long been a problem to distinguish VO
compounds from VO phrases. For example:
3 a. 關心 b. 担心
guān-xīn dān-xīn
concern heart carry heart
“concern” “worry”
a’. 我關心他。 b’. 我很担心他。
Wŏ guān-xīn tā Wŏ hĕn dān-xīn tā
I concern him I very worry him
“I am concerned about him.” “I am very much worried about him.”
a”. 你關什麼心？ b”. 他担了三年心？
Ni guān shénme xīn? Tā dān le sān nián xīn?
You concern what heart he carry ASP three years heart
“What on earth are you concerned about?” “He has been worried for three years.”
c. 跑步 d. 睡覺
păo-bù shuì-jiào
run-feet, sleep-wake
“jog” “sleep”
EBSCOhost - printed on 4/6/2023 8:14 AM via UNIVERSITY OF CAMBRIDGE. All use subject to https://www.ebsco.com/terms-of-use

Minimal word and its function in Mandarin Chinese 49
c’. 他跑了三次步。 d’. 他睡了一天的覺。

Tā păo le sān cì bù Tā shuì le yì tiān de jiào
he run ASP three time feet he sleep ASP one day POSS wake
“He went jogging three times.” “He slept for a day.”
As we can see from the above examples, a VO form can sometimes be separated
as a phrase and sometimes be used as a word. This situation has motivated Chao
(1968) to characterize the alternative forms in terms of ionization. Since then,
great efforts have been devoted to the study of a variety of conditions by which
compounds and phrases can be distinguished (see Lu 1964, Huang 1984, Dai 1992,
Zhang 1992, Duanmu 1998, and Packard 2000). However, even though various
proposals have been made and each of them may work in certain areas for certain
cases, there is no overall generalization and constraint on what is a compound and
what must be a phrase.3 This difficulty has made linguists wonder if there is indeed
a clear-cut distinction at all. Hu (1999), for example, recently claimed that since
there is no overall satisfactory conclusion after years of theoretical investigation,
linguists should consult with native speakers about what a word is. This suggestion
was carried out by Wang (1998) in a study of a total of 647 questionnaires. The
informants were asked to make judgments about the number of words in a ten-
sentence sample in which twenty-five VO forms are distributively used. The results,
as we can see from Table 3.1, are somewhat surprising (taken from Wang 1998,
only 5 forms are given here).
Table 3.1 Judgments on wordhood for VO forms4
VO forms word (%) phrase (%)

shuì jiào “have a sleep, sleep” 95.97 4.03
xĭ zăo “take a bath, bathe” 92.52 7.48
păo bù “run feet, jog” 99.34 0.66
niàn shū “read books, study” 96.84 3.16
dān xīn “carry heart, worry” 97.01 2.99
As seen in Table 3.1, the informants’ judgments are far different from linguists’
because according to linguistic analysis, the first four forms should all be
analyzed as phrases. 5 However, the native speakers’ judgments converge to
form one conclusion: disyllabic VO forms (or more specifically, VO idioms) are
overwhelmingly considered words. For traditional linguists these results create
more puzzles than solutions because what are analyzed as phrases by linguists
are treated as words by native speakers. On one hand, linguists cannot rely on a
layman’s conception of what a word is. On the other hand, linguists cannot ignore

50 Shengli Feng
native speakers’ intuition about what a word is, especially when there is a strong
agreement among speakers on the issue. It seems that the notion of “word” as used
by non-linguists may not be the same as the one used by linguists. At present, we
are left with two fundamental questions that need to be considered immediately.
First, why are native speakers more likely to consider disyllabic forms to be words,
even if some of them are linguistically tested as phrases? Second, are there any
clear-cut distinctions at all between what must be considered compounds and what
must be considered phrases?
Given the minimal word theorem and the fundamental hypothesis P >> M
(prosody determines morphology) in Prosodic Morphology, I would argue that the
study of Chinese morphology will make a great deal of progress once the PrWd is
taken into account. As we will see later, the prosodic analysis proposed here not
only gives us an entirely new insight into Chinese morphology, but also enables us
to determine a prosodic-morphological domain in which every form is legitimate to
be or become a compound within that domain and every form beyond (bigger than)
the domain is a phrase, strictly within VO and similarly within other structures as
well. In other words, there indeed exists a clear boundary demarcating what can be
a compound and what must be a phrase or at least have phrasal properties, defined
by a prosodic qualification of being a minimal word—a new discovery which could
explain quite a wide range of phenomena in Chinese morphology and syntax.
To see how prosody works in Chinese VO compounds, let us assume, first, that
within a structure of two sister nodes labeled as in (4a), if every syllable of the two
sister nodes (V and N) is footed (by Parse-all-Syllable requirement) and the left edge
of every foot aligns with the left edge of some PrWd (All-Foot-Left requirement),
then the two sister nodes will satisfy the requirements of being a MinWd, provided
that every syllable in Chinese is a morpheme.6 This is shown in (4b):
4. a. b. PrWd
|
V Ft
σ σ
V O V O
V
Second, it is well-known that the VO compounds in Chinese are all formed with a
left-headed structure, exactly like VO phrases. That is to say, the internal argument
of a verb must be located on the right side of the V in both phrases and compounds.
Given this, if prosody indeed determines morphology in Chinese, we would expect
the interaction to give rise to a Templatic Constraint for compounding. That is,

5. Templatic constraint on VO compounds

VO-Compd = PrWd
“The VO-compound is a prosodic word.”
This is expected because a prosodically circumscribed domain (MinWd) can be

taken as the locus of morphological transformation required by P >> M, and because
compounding is the most productive morphological process in the language. As a
result, if PrWd is the most harmonic prosodic unit in the language and if it affects
morphology at all, then compounding will be the very target inevitably impinged
upon by the PrWd engaged in prosodic-morphological operations. As we will see
below, this is indeed the case.
First, we have discovered that even if not all VO forms can take an object, the
ones that can are all disyllabic (i.e., PrWd). For example:
6. 負責他負責保衛工作。
fù-zé Tā fù-zé băowèi gōngzuò
carry-responsibility he carry-responsibility security affairs
“be responsible for, be in charge of” “He is in charge of security affairs.”
關心他不關心／別人。
guān-xīn Tā bù guān-xīn biérén
concern heart He not concern-heart others
“concern, care for” “He does not care about others.”
担心他担心情况会有变化。
dān-xīn Tā dān-xīn qíngkuàng huì yŏu biànhuà.
carry heart he carry-heart situation will have change
“worry” “He worries that the situation will change.”
There is no doubt that when it takes an object, the VO must be considered a

compound, because the internal verb+object structure is syntactically opaque as
far as the phrase-structure condition is concerned. Given this, all of the above VO
forms are doubtlessly compounds. We are aware of the fact that not all disyllabic
VO-compounds can take an “outer” object. However, what is important to note here
is a categorical behavior of the trisyllabic (or polysyllabic) forms. No polysyllabic
VO forms can take an “outer” object. For instance,

52 Shengli Feng
7. *他負責任保衛工作。
*Tā fù-zérèn băowèi gōngzuò
he carry-responsibility security affair
“He is in charge of security affairs.”
*我開玩笑他。
*Wŏ kāi-wánxiào tā.
I make-joke he
“I make fun of him.”
The contrast between (6) and (7) shows that only disyllabic VO forms can have an
object, while all trisyllabic forms cannot. A question rises immediately: Why can
trisyllabic forms not take an object in the way that disyllabic forms can? Whatever
the reason, there is no question that there exists a categorical distinction between
disyllabic VO forms on the one hand, and trisyllabic VO forms on the other.
Of course, one may wonder if the trisyllabic VO forms are intransitive verbs
because, if they are, they do not take an object. However, even if this is so, the
same question still remains: Why is it that only disyllabic but not trisyllabic forms
can be transitive? It appears that the same conclusion will still pertain. There
must be a distinction between disyllabic and trisyllabic VO forms in the language.
The question, then, is: Why is there a distinction and how does it come about?
Obviously, an explanation is called for.
Furthermore, examples given below show that the trisyllabic VO forms cannot
simply be considered intransitive verbs because they cannot take aspect makers like
VO verbs do.
8. a. 他負責過保衛工作。
Tā fù-zé -guo băowèi gōngzuò
he carry-responsibility ASP security affair
“He has been in charge of security affairs.”
b. *他對保衛工作負責任過。
*Tā duì băowèi gōngzuò fù zérèn -guo
he to security affair carry responsibility ASP
b’. 他對保衛工作負過責任。
Tā duì băowèi gōngzuò fù-guo zérèn.
he to security affair carry ASP responsibility

c. 他担心着你的健康。
Tā dān-xīn -zhe nĭ de jiànkāng
he carry-heart ASP you POSS health
“He is worrying about your health.”
d. *他開玩笑着說⋯⋯
*Tā kāi wánxiào -zhe shuā: …
He make joke ASP say …
“He is making jokes while talking”
d’. 他開着玩笑說⋯⋯
Tā kāi -zhe wánxiào shuō…
He make ASP joke say
“He is making jokes while talking.”
As a general rule, a VO compound, like all other verbs in the language, can naturally
co-occur with an aspect maker. Thus, in (8a) fù-zé can occur with an experience
aspect marker -guo and in (8c) dān-xīn is with a progressive aspect marker -zhe.
What is remarkable in the above examples is that all the ones that can take an aspect
marker are disyllabic while all of the forms that cannot are trisyllabic. It is clear that
the ones that can take aspect markers must be compound verbs and the ones that
cannot must not be compounds because there is no reason for a compound verb not
to be able to co-occur with an aspect marker. Given this, we see that all compounds
are disyllabic and all trisyllabic forms are not compounds. Put differently, all
trisyllabic VO forms, unlike the disyllabic ones, cannot function as a single (or a
zero-level) verb, hence cannot be considered a compound. Thus it can be concluded
that trisyllabic VO forms must all belong to the category of phrases.
The contrast between (8c) and (8d) therefore suggests a prosodic categorization
in the Chinese morphological system. That is, only disyllabic VO forms can be
compounds while all trisyllabic VO forms lack the ability to be compounds. As seen
above, the disyllabicity perfectly meets the definition of the prosodic word, thus, can be
clearly seen that, for all VO forms, only the ones that meet minimal word requirements
are qualified to be compounds (through regular word formation or lexicalization),7
while those whose size is bigger than a PrWd are not qualified to be compounds.
Finally, separability can also be used to manifest the distinction between
disyllabic and trisyllabic VO forms. That is, only disyllabic forms cannot be
separated, while all other polysyllabic forms are separable. For example (“de” is a
possessive marker in Chinese):

54 Shengli Feng
9. 得罪得罪他 *得他的罪
dé-zuì dé-zuì tā *dé tā de zuì
get-offense get-offense he get he POSS offense
“offend” “To offend him.” “To offend him.”
開玩笑 *開玩笑他開他的玩笑

kāi wánxiào *kāi wánxiào tā kāi tā de wánxiào
make joke make-joke he make he POSS joke
“joke, make fun of” “To make fun of him.” “To make fun of him.”
As shown in the separability test above, we see, again, that there is a clear
distinction between what is revealed in the disyllabic forms and what is lacking in
the trisyllabic VO forms.
The evidence above leaves no doubt that the minimal word (PrWd) is indeed
very active in the language, otherwise there would be no explanation for why
trisyllabic VO forms are systematically different from disyllabic ones. That is
to say, within the VO structure, if the verb and its object are formed by exactly
two syllables, then the VO form will construct a minimal prosodic unit. Since the
minimal prosodic unit is the most harmonic PrWd in the language, by P >> M, all
compounds that are formed by verb+object must first be a PrWd. This is captured
by the Templatic Constraint given in (5).
Under the above analysis, we now begin to understand why disyllabic VO
forms such as niàn-shū “read books, study,” shuì-jiào “have a sleep, sleep,” etc.,
in Table 3.1 are treated as “words” by native speakers. It is because they represent
the most harmonic prosodic category of PrWd, even though they are not lexicalized
or idiomatized. On the other hand, longer VO forms such as kāi wánxiào “joke,”
fù zérèn “be in charge of,” etc., have never been perceived as words by native
speakers because they do not meet the definition of a PrWd, hence they can never
be lexicalized as compounds regardless of how highly they are idiomatized.8 This
suggests that the native speaker’s intuition about “words” in Wang’s study is in fact a
prosodic notion of PrWd, which is different from the syntactic notion of word used by
traditional linguists.
This calls for a further explanation about disyllabic VO idioms. By syntax, the
disyllabic idioms are not compounds; by prosody, however, they belong to the same
category of foot, hence they satisfy the requirement of being a PrWd, even if they
are not (yet) lexicalized as a compound in the lexicon. The native speaker’s word-
judgments about the disyllabic idioms clearly suggest that in Chinese, even phrases are
distinguished prosodically. That is to say, there are apparently two types of phrases:
one consists of disyllabic idiomatized phrases which meet the MinWd requirements
perfectly and have the potential to become compounds, and so can be interpreted by

native speakers as a single unit (i.e., a “word” in an undefined usage, but a PrWd in a
strict sense). The second one consists of freely constructed polysyllabic phrases that
are beyond the Templatic Constraint, and hence can never become compounds in the
language.9 This indicates further that the native speaker’s intuition about the disyllabic
forms is not based on an innate grammar of syntax (which sometimes can best be
detected by linguists), but primarily on an innate grammar of prosody.
In fact, if păo bù “jog,” niàn shū “read books, study,” etc., in Table 3.1 are indeed
phrases, the only way to explain the native speakers’ word-intuition about these phrases
is to admit that the function of minimal word is also active in syntax (at the PF level
or before Spell-Out as suggested in Zubizarreta 1998). Nevertheless, the fact shows
clearly that PrWd also circumscribes phrases in syntax. Given this, we may suggest
that the notion of PrWd defined in terms of prosody could function at different levels of
grammar, that is, it can apply to both morphology and syntax, though the ways it works
may be different. In morphology, every VO compound must be PrWd. In syntax, every
VO phrase that meets the PrWd requirements has the potential to become a compound
depending on its semantics and pragmatic usage in the language, but importantly,
those whose shapes are beyond the control of minimality constraint will never be
compounds. This gives rise to the distinction between disyllabic idioms that are treated
as words and trisyllabic idioms that are not considered words by native speakers. Thus,
the linguistic intuition of PrWd by Chinese speakers provides strong evidence that the
minimal word constraint applies not only to morphology but also to syntax.10
Minimal word as a condition for category changing

The minimal word effect can also be observed in auxiliary+verb compounds in
Chinese. The auxiliary ke “can” can be used to form a compound with a verb,
meaning “V-able,” for example:
10. 可笑 kĕ-xiào “can-laugh, laughable”

可怜 kĕ-lián “can-sympathize, pitiable”
可靠 kĕ-kào “can-trust, trust-able, reliable”
可怕 kĕ-pà “can-terrify, terrible”
可耻 kĕ-chĭ “can-sham, shame-able, shameful”
可行 kĕ-xíng “can-do, doable”
可愛 kĕ-ài “can-love, lovable, lovely” (ALSO “cute”?)
可恨 kĕ-hèn “can-hate, detestable, hateful”
可疑 kĕ-yí “can-suspect, suspect-able”
In Chinese, the ordinary kĕ+V compounds all consist of two syllables. Of course,
there are Aux+VV trisyllabic forms used in the language, too.11 For example:

56 Shengli Feng
11. 可造就的人
kĕ-zàojiù DE rén
can-train’s person
“a person who can be trained, a trainable person”
可加工的材料
kĕ-jiāgōng DE cáiliào
“can-process’s material” “material that can be processed, process-able material”
可閱讀的書籍
kĕ-yuèdú DE shūji
“can-read’s book “books that can be read, readable books”
However, the following contrasts show that the trisyllabic kĕ+VV forms are
different from the disyllabic kĕ+V compounds. Compare:
12. a. 可怕的事 *可以怕的事

kĕ-pà de shì *kĕyĭ-pà de shì
terrible ’s thing terrible ’s thing
“a terrible thing” “a terrible thing”
b. 可疑的人 *可以疑的人
kĕyí de rén *kĕyĭ yí de rén
suspect-able ’s person suspect-able ’s person
“a suspect” “a suspect”
c. 可加工的材料可以加工的材料
kĕ jiāgōng de cáiliào kĕyĭ jiāgōng de cáiliào
can be processed ’s material can be processed ’s material
“process-able material” “process-able material”
d. 可造就的人可以造就的人
kĕ zàojiù de rén kĕyĭ zàojiù de rén
can be trained ’s person can be trained ’s person
“a trainable person” “a trainable person”
Within disyllabic compounds, the auxiliary kĕ 可cannot be substituted with the

free standing counterpart kĕyĭ in the language, as seen in (12a-b), however within
trisyllabic kĕ+VV forms, kĕ and kĕyĭ are interchangeable. The fact that only in
disyllabic forms, kĕ cannot be changed into kĕyĭ indicates that only disyllabic forms
are compounds whereas the trisyllabic ones are equivalent to phrases. The following
examples show even more clearly that the trisyllabic kĕ+VV and the disyllabic kĕ+V
are not simply different but indeed belong to two distinct syntactic categories.

13. a. 他非常可疑。
Tā fēicháng kĕ-yí
he extremely suspect-table
“He is extremely suspect-able (suspicious).”
a’. *他非常可怀疑。
* Tā fēicháng kĕ-huáiyí
he extremely suspect-able
“He is extremely suspect-able.”
b. 他非常可靠。
Tā fēicháng kĕ-kào
he extremely trust-able
“He is extremely reliable.”
b’. *他非常可依靠。
* Tā fēicháng kĕ-yīkào
“He is extremely reliable.”
c. 他非常可信。
Tā fēicháng kĕ-xìn
“He is extremely trust-able (reliable, trustworthy).”
c’. *他非常可相信。
* Tā fēicháng kĕ-xiāngxìn
“He is extremely trust-able.”
d. *這个东西非常可加工。
*Zhège dōngxi fēicháng kĕ-jiāgōng
this thing extremely process-able
“This thing is extremely precess-able.”
As we can see, only disyllabic kĕ+V forms can be used as adjectives whereas all
longer kĕ+V adjective are either not found in the language, or are ill-formed by this
type of process, even if the verbs used in these two forms are synonyms (i.e., yí =
huáiyí “doubt,” kào= yīkào “rely” xìn=xìnrèn “trust”), as seen in (13). This shows
clearly that all longer forms are incapable of undergoing a category change from [Aux
V] verbal expressions to [Aux-V] adjectives. In other words, only disyllabic [Aux-V]
forms are allowed to form adjectives while the trisyllabic ones are prohibited from
doing so. This is clear-cut evidence that trisyllabic forms are differentiated from the
disyllabic [AuxV] forms in the language and cannot be properly explained according
to traditional morphology. In fact, this phenomenon was discovered only recently by
the application of minimal word effect in the language (Feng 2000).

58 Shengli Feng
Given the minimal word theorem and the analysis above, a Templatic
Constraint is therefore expected to be formulated in order to capture the category
changing within all Aux+V forms. It can be seen in (14).
14. Templatic constraint on categorical change12

[kĕ+V] → Adjective ⁄ [kĕ+V]PrWd
“A kĕ+V form undergoes a process of category changing, if it is a prosodic word.”
Obviously, without the notion of MinWd (PrWd), the contrast between the disyllabic
kĕ-V and the trisyllabic kĕ-VV forms revealed in the language will be lost, and most
importantly, the grammar, which would make a strict distinction by the Templatic
Constraint for the morphological process to take place, will be lacking.13
Minimal word as a condition for morphosyntactic operation

In Mandarin Chinese, there are many A(djective)+N(oun) compounds such as dàmĭ
“big+rice, rice,” dàhàn “big+man, burly fellow,” xiăobiàr “little+plait, pigtail,” etc.
Traditionally, most of the A+N forms such as dà lăohŭ “big tiger,” xiăo yŭsăn “little
umbrella,” etc., have always been considered phrases, rather than compound words,
even if it has been recognized that the A+N combinations are not freely constructed
(Zhu 1980), as shown in the following contrasts (de in Chinese is a possessive marker
and a relative clause complementizer):
15. Semantic gap

白紙白的紙
bái zhĭ bái de zhĭ
“white paper” “white ’s paper, white paper, a paper that is white”
* 白手白的手
* bái shŏu bái de shŏu
“white hand” “white ’s hand, white hand, the hands that are white”
Alternative forms (but semantically not equivalent)

大米大的米
dàmĭ dà de mĭ
big rice big ’s rice
“rice” “the rice that is big”

大车大的车
dà chē dà de chē
big vehicle big ’s vehicle
“cart” “the vehicle that is big”
Recently, Shih (1986), Dai (1992), Sproat and Shih (1991, 1996a), Duanmu (1998),
Chen (2000) and others have argued that the bare A+N forms exemplified above
should all be considered compounds, rather than phrases. Among the evidence
supporting this analysis, the strongest is this: the A in all A+N forms cannot take a
modifier like hĕn “very,” for example:
16. * 很大樹很大的樹

* hĕn dà shu hĕn dà de shu
“very big trees” “very big ’s tree, very big trees”
* 很大车很大的车
* hĕn dà che hĕn dà de che
“very big vehicle” “very big ’s vehicle, very big vehicle”
* 很小雨傘很小的雨傘
* hĕn xiăo yŭsăn hĕn xiăo de yŭsăn
“very small umbrella” “very small ’s umbrella, very small umbrella”
Such syntactic behavior, therefore, forces one to conclude that the A+N forms are
not phrases because there is no reason why the A cannot be modified if the [A+N]
is a phrase, as compared with English.
17. little umbrella very little umbrella

blackboard *very blackboard
small-pox *very smallpox
In English, the A of an [A+N] form cannot be modified if the [A+N] is a compound.

In Chinese, however, the A in all of the A+N forms is not allowed to be modified.
Thus, it is reasonable to consider them as N0 modifiers (Sproat and Shih 1991:
571). However, what we found is a clear distinction between different prosodic
entities with different syntactic behaviors among all A+N forms. That is, disyllabic
AN forms (if not all) behave differently from the longer ones systematically. To see
this, let us began with Sproat and Shih’s (1991) generalization about the adjective
ordering of “SIZE” and “COLOR” in noun phrases.14 First, to observe:

60 Shengli Feng
18. SIZE COLOR NOUN

little black umbrella *black little umbrella
xiăo hēi yŭsăn *hēi xiăo yŭsăn
小黑雨傘
big white plate *white big plate

dà bái pánzi *bái dà pánzi
大白盤子
Based on mounting evidence in different languages, Sproat and Shih (1991)

generalized an Adjectival Ordering Restriction (AOR), which says essentially that
the AOR – [SIZE > COLOR] – obtains if the adjectives involved are hierarchical
direct modifiers. This is argued to be a universal constraint for the ordering of
multiple adjectival modifiers, not only in English but also in Chinese (and many
other languages; see Sproat and Shih 1991).
Given this constraint, a clear-cut distinction emerges between disyllabic AN
forms and trisyllabic AN forms. For example (taken from Feng 2000):
19. Trisyllablic AN Disyllablic AN

* COLOR > SIZE COLOR > SIZE
* 白大盤子紅小兵
* bái dà pánzi hóng xiăo-bīng
“White big plate” “red small-soldier, the red guard”
* 紅小雨傘黑大汉
* hóng xiăo yŭsăn hēi dà-hàn
“red small umbrella” “black big-man, a black bully”
* 紅小計算机黑小辮儿
* hóng xiăo jìsuànjī hēi xiăo-biàr
“red small computer” “black little-plait, a black pigtail”
* 黑大熊猫黑大雁
* hēi dà xióngmāo hēi dà-yàn
“black big panda” “black big-goose, wild goose”
* 白大蘿卜白大米
* bái dà luóbo bái dà-mĭ
“white big radish” “white big-rice, white rice”

* 紫小蕃茄綠小葱
* zĭ xiăo fānqié lǜ xiăo-cōng
“violet small tomato” “green small-scallion, green scallion”
Clearly, only (if not all) disyllabic AN forms can violate the universal constraint
of [SIZE > COLOR] whereas three (or more) syllable AAN forms must all obey
this constraint. Once again, we see that the trisyllabic AAN forms, like all other
trisyllabic forms (V+OO, Aux+VV and V+RR), inherently lack the properties
exhibited in disyllabic forms. Why is this? Traditional grammarians would answer
this question immediately by saying: because the ones that can violate the general
constraint are compounds. This is indeed correct because in English the [SIZE >
COLOR] order can also be violated if the [SIZE+N] is a compound. For example,
20. red smallpox

black bigfoot
The compound status of the disyllabic AN forms in (19) can also be seen from
the fact that the ones that can violate the ordering requirement are all inseparable,
which shows the inherent property of being a compound. For example,
21. 黑大汉 * 大的汉

hēi dà-hàn * dà de hàn
black big-man big ’s man
“black bully”
黑小辮儿 * 小的辨儿
hēi xiăo-biàr * xiăo de biàr
black little-plait little ’s plait
“a black pigtail”
綠小葱 * 小的葱
lǜ xiăo-cōng * xiăo de cōng
green small scallion small’s scallion
“green scallion”
The inseparability of the [SIZE+N] forms in the [COLOR+[SIZE+N]] environment

confirms the analysis that the [SIZE+N] forms in (19) and (21) must all be
compounds, which makes a reasonable exception to the general constraint on
ordinary phrases, as we would expect in English.

62 Shengli Feng
However, a question arises once we adopt the above analysis, because it would
imply that the ones that must obey the general constraint are phrases. This must
be so, or there is no reason why they would behave differently from the ones that
violate the [SIZE > COLOR] constraint if both AN and AAN are compounds.
In other words, if only compounds can violate the constraint (cf. red smallpox),
the ones that must obey it should not be compounds (cf. *red small umbrella).15
However, as we have seen above, there is evidence that all A+N forms in Chinese
are compounds, because none of them can be modified by a degree adverb like hĕn
“very” (cf. *very blackboard). Given this, we are left with a paradoxical situation:
according to the [SIZE > COLOR] constraint, the ones that obey it must not be
considered compounds because only compounds can disobey it; but according to
the modifier hĕn test, all AN forms must be compounds regardless of whether they
obey the [SIZE > COLOR] restrictions or not.
How can we resolve this dilemma? While other analyses are plausible (see
note 15), I would like to suggest, first, that the ones that disobey the [SIZE >
COLOR] constraint are unquestionably compounds. However, they are not
compounds in general, rather a specific kind, say, lexical compounds. By lexical
compound I mean one that is generated in the lexicon governed by (prosodic)
morphological rules.
What about the ones that must obey the [SIZE > COLOR] constraint?
Regarding the hĕn “very” test given above, they cannot simply be treated as
phrases, but they cannot be lexical compounds either. Following Feng (2001), I
would like to suggest that they are syntactic compounds, which means that they are
formed in syntax,16 and are therefore visible to the syntactic restriction of [SIZE
> COLOR]. In fact, whatever they are is not the central issue to be addressed
here. The point which I am making concerns the following fact: only (if not all)
disyllabic forms are able to take the [COLOR+SIZE] structure, while polysyllabic
forms all fail to do so. The question then is: Why are the trisyllabic ones different
from the disyllabic ones? Why must the difference be made by different numbers of
syllables? Furthermore, how can we characterize the difference? Juxtaposed with
the analyses in the previous sections, it is clear that the difference is due to whether
an AN form is a PrWd or not. Thus, it can only be explained in terms of prosody.
Similar to the minimal word effect on VO and AuxV forms discussed above, this
phenomenon (SINGULAR) can best be accounted for in the same way as minimal
word effect. That is,
22. Minimal word constraint on AN forms in Chinese

AN-Compd = PrWd
A AN (lexical) compound is a PrWd.

Here, we do not exclude the possibility that the polysyllabic AN forms are
(syntactic) compounds, but they must be different from lexical compounds.
If we adopt the classification of lexical compounds and syntactic compounds,
we have good reason to say that the requirement of [A+N]PrWd/Compound in the
[COLOR+[SIZE+N]] structure follows directly from the minimality constraint: a
lexical compound of [A+N] forms must be a PrWd. Thus, in the environment of
[COLOR+SIZE+N], the application of the Prosodic-Morphologic constraint (22)
will give rise to the following result: [COLOR [size+N]PrWd/Compound]. The present
theory also predicts that trisyllabic AAN forms cannot be lexical compounds,
their components are still visible to certain syntactic processes (i.e., the phrasal
restriction of [COLOR > SIZE]).
If the above analysis is correct, it provides additional evidence for the
argument that the minimal word in Chinese is the most harmonic prosodic word
which is extremely active in a variety of constructions in the language.
Theoretical implications and empirical consequences

We have discussed several different syntactic and morphologic constructions in
the present study: the Verb+object, the Auxiliary+Verb and the Adjective+Noun.
We have also seen that all these different syntactic forms share a common
property. That is: the trisyllabic ones are syntactically different from the
disyllabic ones systematically. To be more specific, only the disyllabic ones
exhibit lexical properties, and therefore only the disyllabic ones can be (lexical)
compounds. Thus, for VO forms, only the disyllabic ones can take an object;
for AuxV forms, only the disyllabic ones can be modified by a degree modifier
like fēicháng “extremely”; for AN forms, only the disyllabic ones can violate
the [SIZE+COLOR+N] phrasal constraint. Contrary to the case of disyllabicity,
all trisyllabic forms of each of the three constructions syntactically deviate
from the properties of being a (true) lexical item, for example: the separable
property and the inability to take a (post-verbal) object for all trisyllabic
VOO forms; the inability to undergo the process of category changing for all
trisyllabic AuxV forms; and the obligation to strictly follow the phrasal constraint
[SIZE+COLOR+N] for all trisyllabic AN forms.
It is possible, however, for one to suggest a different analysis for the trisyllabic
forms and that is the AAN forms may reasonably be treated as phrases with a
[SIZE+COLOR+N] order, or compounds using the hĕn test. However, it does
not matter how one analyzes them, the distinction between the disyllabic and the
trisyllabic structures will still pertain. The syntactic contrast cannot be explained

64 Shengli Feng
away by any analysis that overlooks the disyllabic and the trisyllabic distinctions.
The facts brought to light in this study are quite striking: the distinctions among all
three types of forms are not only syntactic, but also prosodic. In fact, their different
syntactic behaviors were not discovered and will not be fully understood until the
prosodic distinction is brought to light. In this sense, it is reasonable to say that the
study of prosody has revealed some important prosodic-syntactic phenomena that
would otherwise be a mystery in the language.
As we have seen, the prosodic distinction in all three different structures is
centered upon the basic domain of disyllabicity. Why is this so? The minimal word
theorem provides the best explanation and it is simply the legitimacy of being
a PrWd in the language. The Chinese language is, therefore, a language that is
extremely sensitive to the prosodic domain. Actually, it is the domain that permits
only disyllabic forms to have the priority to be or become compounds, and it is
also the domain that prevents trisyllabic forms from entering into the realm of
lexical compounding. As a result, a boundary in Chinese morphology can be set
between what is a compound (PrWd applies in morphology) or permitted to be a
compound (PrWd applies in syntax), and what is forbidden from being a lexical
compound, or at most a syntactic compound.
It is well known that in Prosodic Morphology, the core area of previous
investigations has focused mainly on reduplication and infixation. The present
study, however, extends the notion of PrWd into the area of compounding and its
interaction with syntax. For all three types of forms discussed above, we have seen
that the minimality constraint indeed controls the word formation of compounding
in Chinese. That is, a compound (or more specifically a lexical compound) must
first be a PrWd, even if a PrWd is not, by necessity, a compound. This conclusion
has several implications in Prosodic Morphology. First, the Templatic Constraint
not only determines morphological operations like reduplication and infixation,
but also controls the word formation of compounding in languages like Chinese.
This raises a question for future study as to why and how the Templatic Constraint
could also control compounding. Furthermore, it is clear, by now, that the minimal
word requirement functions in Chinese. Yet, when it does, it not only affects word
formation, but also syntactic structures in a way that influences the formation
of well-formed sentences. Surprisingly, it seems that the MinWd may also
circumscribe certain syntactic phrases, so that the native speaker’s intuition about
what is considered a “word” is unquestionably affected by the prosodic notion of
MinWd. The final question then is how and to what extent prosody affects syntax,
this question is extremely important for theoretical as well as empirical inquiries
in future research.

Feng 2009 Minimal Word and Its Function in Mandarin Chinese

Uploaded by

Copyright:

Available Formats

Feng 2009 Minimal Word and Its Function in Mandarin Chinese

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Feng 2009 Minimal Word and Its Function in Mandarin Chinese

Uploaded by

Copyright:

Available Formats

3

Minimal word and its function in Mandarin

2. Foot binarity: Feet are binary under syllabic or moraic analysis.

stud of Chinese Book-3.indd 47 2008/11/17 12:51:52 PM

Minimal word as a condition for VO compound

stud of Chinese Book-3.indd 48 2008/11/17 12:51:52 PM

c’. 他跑了三次步。 d’. 他 睡了一 天的 覺。

Table 3.1 Judgments on wordhood for VO forms4

VO forms word (%) phrase (%)

stud of Chinese Book-3.indd 49 2008/11/17 12:51:52 PM

stud of Chinese Book-3.indd 50 2008/11/17 12:51:52 PM

5. Templatic constraint on VO compounds

This is expected because a prosodically circumscribed domain (MinWd) can be

There is no doubt that when it takes an object, the VO must be considered a

stud of Chinese Book-3.indd 51 2008/11/17 12:51:52 PM

stud of Chinese Book-3.indd 52 2008/11/17 12:51:52 PM

stud of Chinese Book-3.indd 53 2008/11/17 12:51:52 PM

開玩笑 *開玩笑他 開他的玩笑

stud of Chinese Book-3.indd 54 2008/11/17 12:51:52 PM

Minimal word as a condition for category changing

10. 可笑 kĕ-xiào “can-laugh, laughable”

stud of Chinese Book-3.indd 55 2008/11/17 12:51:52 PM

12. a. 可怕的事 *可以怕的 事

Within disyllabic compounds, the auxiliary kĕ 可cannot be substituted with the

stud of Chinese Book-3.indd 56 2008/11/17 12:51:52 PM

stud of Chinese Book-3.indd 57 2008/11/17 12:51:53 PM

14. Templatic constraint on categorical change12

Minimal word as a condition for morphosyntactic operation

15. Semantic gap

Alternative forms (but semantically not equivalent)

stud of Chinese Book-3.indd 58 2008/11/17 12:51:53 PM

16. * 很大樹 很大的樹

17. little umbrella very little umbrella

In English, the A of an [A+N] form cannot be modified if the [A+N] is a compound.

stud of Chinese Book-3.indd 59 2008/11/17 12:51:53 PM

18. SIZE COLOR NOUN

big white plate *white big plate

Based on mounting evidence in different languages, Sproat and Shih (1991)

19. Trisyllablic AN Disyllablic AN

stud of Chinese Book-3.indd 60 2008/11/17 12:51:53 PM

20. red smallpox

21. 黑大汉 * 大的汉

The inseparability of the [SIZE+N] forms in the [COLOR+[SIZE+N]] environment

stud of Chinese Book-3.indd 61 2008/11/17 12:51:53 PM

22. Minimal word constraint on AN forms in Chinese

stud of Chinese Book-3.indd 62 2008/11/17 12:51:53 PM

Theoretical implications and empirical consequences

stud of Chinese Book-3.indd 63 2008/11/17 12:51:53 PM

stud of Chinese Book-3.indd 64 2008/11/17 12:51:53 PM

You might also like

c’. 他跑了三次步。 d’. 他睡了一天的覺。

開玩笑 *開玩笑他開他的玩笑

12. a. 可怕的事 *可以怕的事

16. * 很大樹很大的樹