CHAPTER 2 GRAMMATICAL FUNCTIONS IN AN LFG
3.3 S UBCATEGORIZATION OF G RAMMATICAL F UNCTIONS
3.3.4.4 You3 'exist, have'
1. a. Wo3 you3 hen3 duo1 shu1.
I have very many book
‘I have many books.’
b. Zai4 zhuo1-shang4 wo3 you3 hen3 duo1 shu1.
at desk top I have very many book
‘On the desk I have many books.’
2. a. Zhuo1-shang4 you3 hen3 duo1 shu1.
desk- top exist very many books
‘On the desk there are many books.’
b. Zai4 wu1zi-li3 zhuo1-shang4 you3 hen3 duo1 shu1.
at house inside desk- top exist very many books
‘In the house on the desk there are many books.’
3. a. You3 hen3 duo1 shu1 zai4 zhuo1-shang4.
Exist very many book at desk- top
‘There are many books on the desk.’
b. Zai4 zhuo1-shang4 you3 hen3 duo1 shu1.
at desk- top exist very many book ‘On the desk there are many books.’
c. Zai4 wu1zi-li3 zai4 zhuo1-shang4 you3 hen3 duo1 shu1 at house inside at desk- top exist very many book ‘In the house on the desk there are many books.’
In this section we will present a discussion on the verb you3 'have, exist' in Mandarin Chinese. We will argue that it belongs to this subcategory because it subcategorizes <SUBJ , OBJ>; however, its idiosyncracies have to be accounted for uniquely in its lexical entry. We will examine some previous accounts of this verb and then present evidence for our analysis. Generally, three usages of this verb can be recognized: 1) possessive, 2) existential and 3) (subjectless) locational. To facilitate discussion we first showed the above examples. Sentences in 1 above are of the first usage, possessive; sentences in 2 show the existential usage of you3 where its subject has to be a place word, and in 3 you3 may be preceded by a locational prepositional phrase and yet here it has no overt noun phrase subject.
First we will discuss you3 in these three verbal usages. We will find a common semantic interpretation of existence to explain all three usages and also discuss the implications on the possessive construction of [NP de NP].
Then we will examine another usage of you3 where it seems to be part of an adjectival element, such as you3 yi4si 'interesting'. Finally, we will give an explicit formulation of the lexical entry accounting for all the above usages of you3 in our vLFG formalism.
There were some transformational accounts of the verb you3 during the heyday of Transformational Grammar. For example, Li (1972) links verbs denoting 'be, exist, and have' (shi4, zai4, and you3) transformationally by posing a common deep structure. We can reject this analysis for several reasons. First of all, transformations are not allowed in LFG framework.
Second, this analysis entails the grouping of shi4, zai4, and you3 into one single lexical entry and therefore the implication that the choice of shi4, zai4, or you3 is not lexical but purely phonological. This analysis contradicts the fact that these three verbs do manifest different syntactic behaviors and semantic content.
While we distinguish the preverbal locational NP, zhuo1-shang4, in 2a and the preverbal locational PP, zai4 zhuo1-shang4, in 3b, Huang (1966), along with others, such as Li and Thompson (1981:509-516), on the other hand, treats the two types of sentences as grammatically equivalent; in other words,
VERBSUBCATEGORIZATIONINMANDARINCHINESE 111
the preverbal PP in 3b is also taken to be the subject. This analysis, as Starosta (1985) protests, seems to be based on nothing more than the same English translations for these two types of sentences. Clearly, in 3a there is no preverbal subject since you3 occupies the sentence-initial position.
Furthermore, note the fact that in Mandarin Chinese non-subcategorized locational prepositional phrases usually occur preverbally, exemplified in 1b, 2b and 3c. In order to capture this generalization, sentences 3b-c, just like 3a, would have to be considered as without overt subjects, and thus all preverbal locational PP's in 3b-c function adjunctively just like the preverbal PP's in 1b and 2b. In 2, however, as both Starosta (1985) and Ding et al (1979:78-79) have pointed out, the preverbal unmarked place words should be regarded as subjects, just like the NP subjects of 1a-b. Yet, both Chao (1968:530) and Li and Thompson (1981:511) treat them as topics, which is a very ambiguous position for they do not distinguish subjects from topics syntactically.
Starosta (1985), in his localistic Lexicase analysis, makes exactly the same three distinctions; namely, he recognizes the possessive and the non-possessive you3, and he further divides the non-possessive you3 into two subtypes: existential (you31) and locational (you32). Thus, this Lexicase analysis necessarily implies that in the lexicon there are three separate lexical entries: you3, you31, and you32. Since we do recognize that there are three different usages of the verb you, it would be convenient, and it is also perfectly acceptable within the LFG framework, to simply pose three different you3's. Nonetheless, we believe doing so fails to capture some important generalizations of the verb you3 as a single lexical unit.
As Li and Thompson (1981:516) have correctly pointed out, many languages of the world, including Mandarin Chinese, express both possession and existence with the same verb. To pose two (or three) different verbs for the possessive and existential usages is to render this phenomenon accidental and trivial, at least for Chinese. The fact that in many languages the same verb expresses possession and existence is, we believe, largely because the relation of possession between two entities is in fact a kind of relation of existence as well. That A possesses B is in effect equivalent to that B exists within the domain of A. To clarify this statement, first we have to point out that possession is NOT equivalent to ownership; rather, ownership is only a specific kind of, or an instance of, possession. That A owns B can be interpreted as B exists within the domain of A's ownership. In other words,
existence subsumes possession, which in turn subsumes ownership. Therefore, while it is possible to interpret sentences 1a-b in terms of ownership (i.e., "I own many books."), it is not always appropriate to do so. We will examine the following sentences of similar usage:
4. a. Wo3 you3 hen3 duo1 qin1qi1.
I have very many relative
‘I have many relatives.’
b. Hua1 you3 hen3 duo1 zhong3lei4.
flower has very many variety
‘Flowers have many varieties.’
c. Tai2wan1 you3 hao3 qi4hou4.
Taiwan have good climate ‘Taiwan has good climate.’
d. Qi1yue4 you3 san1shi2yi1 tian1.
July have thirty-one day ‘July has thirty-one days.’
All these sentences are better interpreted in terms of existence, namely that
"A you3 B" is viewed as "B exists in the domain of A." This is precisely the kind of interpretation most suitable for sentences in 2, where the subjects are always place words that semantically function as the universe or the domain in which the object exists. Once we have shown that sentences of 1 and 2 are both syntactically and semantically equivalent in structure, we have eliminated the plausibility of posing two different you3's for 1 and 2. We thus also recognize that you3 requires two arguments: SUBJ and OBJ. Note that Cheng (1979) lists some you3 sentences like 4b and 4d above to demonstrate the "approximative" use of you3 in addition to the three uses we have listed above. However, it is clear that semantically our redefined concept of possessive and existential use of you3 also covers this "approximative" use, and syntactically sentences in 4 as well as in 1 and 2 are all equivalent in terms of their requirement of subject and object.
VERBSUBCATEGORIZATIONINMANDARINCHINESE 113
Next, we shall examine sentences in 3. We have already pointed out that sentences 2a and 3a are not syntactically equivalent because 2a has an overt subject while 3a does not. However, semantically, sentences of 3 are also best interpreted in terms of existence, namely that "you3 B" is viewed as "B exists, within an unspecified domain." The preverbal locational prepositional phrase is not a subcategorized element; rather it is an adjunctive element that denotes the general location of the existence. Compare the following sentences:
5. a. Zai4 wu1zi-li3 wo3 you3 hen3 duo1 shu1.
at house inside I have very many book
‘In the house I have many books.’
b. Zai4 wu1zi-li3 zhuo1-shang4 you3 hen3 duo1 shu1.
at house inside desk- top exist very many books
‘In the house on the desk there are many books.’
c. Zai4 wu1zi-li3 you3 hen3 duo3 shu1.
at house inside exist very many books
‘In the house there are many books.’
The preverbal prepositional phrases in all three sentences are of the same adjunctive function. Now, to claim that the you3 in 3a-c and 5c is the same verb as in 2 and 3 is in effect to also claim that all of these sentences have the same syntactic structure, that is they all have a SUBJ and an OBJ.
you3:
[ FS [ FORM 'you3'
SUBJ OPT
]
FI-V-4
]
Following the preliminary analysis of you3 of Her (1989a), we are therefore treating you3 as a single verb which subcategorizes two grammatical functions, SUBJ and OBJ, with a special default value on its SUBJ in the functional structure. Note that OPT is a placeholder and always
succeeds in unification. Thus, in this case, if the SUBJ of the you3 sentence is overt, then whatever the existing value of its SUBJ is, it will overwrite the default value OPT; only when its SUBJ is not overt and therefore with no feature SUBJ, will OPT remain to be the value of SUBJ.
By setting such a default value on you3's SUBJ we can account for the fact that it is possible for you3 not to have an overt subject, as seen in sentences of 3. In other words, the functional structures of sentences in 3 will not be considered incomplete, in LFG terminology. Note also that, although when used in existential sentences such as the ones in 2 the SUBJ must be a place word, we do not need such a constraint because in the possessive usage in 1 there is no such selectional restriction on the SUBJ. Finally, it may be worth mentioning that in all the you3 sentences thus far, negation is always marked by mei2, and in Mandarin Chinese you3 is the only verb that can be negated with mei2 regardless of what temporal aspect the verb denotes; since this phenomenon is not purely due to phonological factors (e.g., *mei2 you3shan4 'not friendly'), it certainly serves as another indication that you3 is a single lexical unit. This treatment of you3 is most economical and therefore presumably captures most generalizations of this verb.
Now we will discuss the implications of this unified treatment of you3 on the possessive construction of [NP1 de NP2]. While it is most common to interpret such a construction as "NP1 possesses NP2," we contend that, like the interpretation of [A you3 B], it is more appropriate to interpret the de possessive as "NP2 which exists in the domain of NP1." While it is absurd to interpret hua4xue2 de zhi1shi4 'the knowledge of chemistry' as "the knowledge possessed by chemistry," it is entirely appropriate to interpret it as
"the knowledge which exists in the domain of chemistry." Thus, if we take all the sentences of [NP1 you3 NP2] in 4, each of them has a semantically equivalent counterpart of [NP1 de NP2] construction, as illustrated below.
4' a. Wo3 de qin1qi1.
I DE relative
‘Relatives of mine.’
b. Hua1 de zhong3lei4.
flower DE variety
‘Varieties of flowers.’
VERBSUBCATEGORIZATIONINMANDARINCHINESE 115
c. Tai2wan1 de qi4hou4.
Taiwan DE climate ‘The climate of Taiwan.’
d. Qi1yue4 de san1shi2yi1 tian1.
July DE thirty-one day ‘The thirty-one days of July.’
Next we shall proceed to examine another usage of the verb you3, one in which you3 seems to be part of a stative idiomatic expression. Sentences listed in 6 are some examples of this kind of [you3-NP] expressions. Although Cheng (1979) has hinted that this type of you3 constructions should be considered idiomatic, we are not aware of any previous detailed account of this usage.
6. a. Ta1 hen3 you3 yi4si.
he very have meaning
‘He is very interesting.’
b. Ta1 hen3 you3 qian2.
he very have money
‘He is very rich.’
c. Ta1 hen3 you3 jiao4yang3.
he very have education
‘He is very educated.’
d. Ta1 hen3 you3 huo2li4.
he very have energy
‘He is very energetic.’
The first noticeable peculiarity of sentences in 6 is that it is semantically odd for you3 which denotes the concept of 'existence' to be modified by hen3 'very' and therefore gradable. To look at this from a different perspective, however, this of course can serve as a strong indication to us that in 6 you3
has little to do with the concept of existence. More evidence comes from the fact that only when followed by certain abstract nouns can hen3 and you3 co-occur, as exemplified in 7, and that you3 cannot be modified by hen3 when followed by a concrete noun, as we demonstrate in 8.
7. a.*Ta1 hen3 you3 xue2shuo1.
he very have theory.
‘He is full of theories.’
b.*Ta1 hen3 you3 jin1qian2.
he very have money
‘He is very rich.’
c.*Ta1 hen3 you3 jiao4yu4.
he very have education
‘He is very educated.’
d.*Ta1 hen3 you3 jian4kang1.
he very have health
‘He is very healthy.’
8. a.*Ta1 hen3 you3 shu1.
he very have book
‘He has many books.’
b.*Ta1 hen3 you3 xue2sheng1.
he very have student
‘He has many students.’
c.*Ta1 hen3 you3 mao1.
he very have cats
‘He has many cats.’
VERBSUBCATEGORIZATIONINMANDARINCHINESE 117
d.*Ta1 hen3 you3 gong1si1 he very have company
‘He has many companies.’
It seems that no syntactic or semantic generalizations can be made as to what kind of abstract NP objects can be used for the you3 phrase to be gradable, since the ones in 7 are very similar to the ones in 6 in their semantic properties. Based upon this fact, clearly we can reject the analysis where sentences in 6 are to be treated just like other you3 sentences such as the ones in 1, 2 and 3. Sentences in 6 cannot be analyzed syntactically as regular verb phrase of [you3 NP]. We will examine the two other possible alternative analyses: one, to treat them as lexicalized elements generated through a lexical process; or two, to regard them as non-lexical, idiomatic expressions.
If [you3-NP] expressions in 6 are complex lexical items with incorporated objects, then the internal syntactic structure of the [you3-NP] expression is of no relevance to the syntactic structure of the sentence, and that they behave like basic, non-phrasal verb. This is how Bresnan (1982:51, 57) treats some fixed expressions, called complex verbs, in English, such as "go over," "look on," "make fun of," "catch sight of" and "look askance at." Some Chinese fixed expressions also clearly belong to this class, such as ba3feng1 'to be on the lookout' whose literal meaning of "to hold the wind" no longer exists.
Each expression of complex verbs, for example the [you3-NP] expressions in 6, is listed as a separate entry in the lexicon. However, examples in 9 refute this lexical incorporation hypothesis.
9. a. Ta1 hen3 you3 hua4hua4 de tian1cai2.
he very have paint de talent
‘He is very talented in painting.’
b. Ta1 hen3 you3 xie3 bao4gao4 de ji4qiao3.
he very have write report de skill
‘He is very skillful in writing reports.’
c. Ta1 hen3 you3 xue2shu4 de jiao4yang3.
he very have academic de education
‘He is very educated academically.’
d. Ta1 hen3 you3 qing1chun1 de huo2li4.
he very have youth de energy
‘He is full of youthful energy.’
Since syntactically adjacency is assumed to be a necessary condition for incorporation or any other morpholexical process such as compounding (Bresnan 1982:51), we would have to stipulate that in 9a-d the entire [you3-NP] expression is a single lexical unit. The fact that in 9a-d all the NP's following you3 are syntactically complex, and in fact may be potentially infinitely so, indicates that it is impossible in this case for the entire [you3-NP]
to be analyzed as a single lexical item, for that will make our lexicon infinitely large and thus unformulatable. Furthermore, the fact that all these stative [you3-NP] expressions, when negated, still have to be marked with mei2, but never with bu4 like other state verbs in Chinese, also indicates that they are not lexicalized state verbs. This is true with all the stative [you3-NP]
expressions, including the ones whose NP can never be separated from you3 such as you3-qian2 'rich' and you3-yi4si 'interesting'. Therefore, we are left with our final option: to analyze the you3 phrases in 6 and 9 as idiomatic expressions whose internal syntactic structure is relevant to that of the sentence. Again, this is how Bresnan (1982:45-50) treats another type of
"idiom chunks" in English, idiomatic expressions such as "keep tabs on,"
"pull one's leg" and "kick the bucket."
None of the facts we have discussed thus far contradicts this analysis. Now, compare sentences in 6 and 7. We see that although the nouns following you3 in 6 are very similar to those in 7 syntactically and semantically, the you3 phrases in 7 cannot be modified by hen3. For instance, hen3 you3 qian2 'very rich' is good, while *hen3 you3 jin1qian2 is not, and while hen3 you3 jiao4yang3 'very educated' is good, *hen3 you3 jiao4yu4 is not. This observation clearly points to an idiomatic solution. Further facts indicate the same solution.
10. a. Ta1 hen3 you3 qian2.
he very have money.
‘He is very rich.’
VERBSUBCATEGORIZATIONINMANDARINCHINESE 119
b.*Ta1 hen3 you3 mai3 shu1 de qian2.
he very have buy book de money.
‘He has lots of money for buying books.’
11. a. Ta1 hen3 you3 kan4fa3.
he very have opinion
‘He is very insightful.’
b.*Ta1 hen3 you3 bu4tong2 de kan4fa3.
he very have different de opinion
‘He is full of different opinions.’
Although, as we have shown with examples in 9, the NP following you3 may be syntactically complex, whether this is allowed seems to be arbitrarily restricted, depending on each you3 expression of this sort. Thus, within the specification of the idiom you3 qian2 'rich', it has to be specified that qian2 cannot take any modifier, while the idiom you3 ji4qiao3 'skillful' has no such restriction. Only an idiomatic analysis is compatible with this finding. A similar fact can be observed in English idioms. For "kick the bucket" to have its idiomatic reading no modifier is allowed on the noun "bucket." For the same reason "he is pulling my right leg" has no idiomatic reading. Yet idiomatic expressions like "take a look" does allow certain modifiers on the noun, e.g., "take a long look" and "take a hard look." Similarly all sentences in 9 are good.
Another piece of evidence in support of the idiomatic analysis comes from topicalized and preposed positions.
12. a. Wo3 you3 qian2.
I have money
‘Money, I have.’
OR ‘I am rich.’
b. Qian2, wo3 you3.
money I have
‘Money, I have.’
c.*Qian2, wo3 hen3 you3.
c.*Qian2, wo3 hen3 you3.