Chapter 3 Literature Review
3.1 H.-L. Lin (2003)
Lin argues that resultative, descriptive, and extent constructions are three postverbal secondary predicates in Taiwanese, which all involve the use of -kah, as illustrated by (40) (H.-L. Lin 2003, (5), (7) and (8)).
(40) a. I penn-kah khia-be-khilai. (resultative) He sick-KAH stand-not-up
‘He was so sick that he could not stand up.’
b. I penn-kah ukau giamtiong-e. (descriptive) He sick-KAH very seriously-ASP
‘He was seriously ill.’
c. Thinnting kng-kah chinchan jitsi kangkhuan. (extent) Sky bright-KAH like daytime same
‘The sky is so bright that (it is) like the daytime.’
24
She argues that the three constructions differ from each other in the scope of modification, passivization, and word order. As for the scope of modification, the resultative construction may predicate on the subject or object, as illustrated by (41) (H.-L. Lin 2003, (11) and (12)).
(41) a. I thiau-kah cin thiam. (subject-oriented resultative) He dance-KAH very tired
‘He danced himself tired.’
b. I ciong sann se-kah cin cingkhi. (object-oriented resultative) He CIONG clothes wash-KAH very clean
‘He washed his clothes clean.’
In terms of the descriptive construction, it modifies the verb (i.e., V1 in this paper) instead of the subject or object, as shown in (40b), where ukau giamtiong ‘very seriously’ refers to the degree of the person’s being ill. As for the extent construction, there are two types. Firstly, the ‘pure’ extent constructions are not predicated of the subject or object, as (40c) illustrates. Secondly, if the constructions predicate on the subject or object, the event described in the extent clause cannot be the result of the main clause, as exemplified in (42), where m cai chenn ‘not knowing waking up’ is just the extent to which the person slept.
(42) I khun-kah m cai chenn. (extent) He sleep-KAH not know wake
‘He slept so well that he did not wake up for a long time.’
In terms of the passivization, the descriptive construction does not have any
25
passive counterpart, and the extent construction only allows the object of the main verb to undergo passivization, as shown in (43)-(44).
(43) a. I ji sia-kah cin kin. (descriptive) He word write-KAH very fast
‘He writes very fast.’
b. *Ji hoo i sia-kah cin kin. (descriptive) Word HOO he write-KAH very fast
‘Words are written very fast.’
(44) a. I ka in kiann phah-kah lian chupinn long chutlai khuann. (extent) He KA his child hit-KAH even neighbor all out see
‘He hit his child so much that even the neighbors came out to have a look.’
b. In kiann hoo i phah-kah lian chupinn long chutlai khuann. (extent) His child HOO he hit-KAH even neighbor all out see
‘His child was hit so much that even the neighbors came out to have a look.’
c. *Chupinn hoo i ka in kiann phah-kah long chutlai khuann. (extent) Neighbor HOO he KA his child hit-KAH all out see
‘The neighbors came out to have a look because he hit his child so much.’
However, the resultative construction allows passivization of the subject of the resultative construction and the object of the main verb, as exemplified in (45)-(46) (H.-L. Lin 2003, (23)-(26)).
(45) a. I chio-kah ehai long lauhkhi-a. (resultative) He laugh-KAH jaw all drop-ASP
26
‘He laughed so much that his jaw dropped.’
b. Ehai hoo i chio-kah long lauhkhi-a. (resultative) Jaw HOO he laugh-KAH all drop-ASP
‘His jaw dropped because he laughed so much.’
(46) a. I ciong in kiann phah-kah hunkhi. (resultative) He CIONG his child hit-KAH faint
‘He hit his child so much that he/she fainted.’
b. In kiann hoo i phah-kah hunkhi. (resultative) His child HOO he hit-KAH faint
‘His child was hit so much that he/she fainted.’
As far as the word order is concerned, the three constructions have different word orders when the main verb is transitive. In the descriptive and extent construction, the semantic object of the main verb must precede it, whereas in the resultative construction, the object may precede or follow the main verb, as illustrated in (47)-(49) (H.-L. Lin 2003, (29)-(31)).
(47) a. I ji sia-kah cin kin. (descriptive) He word write-KAH very fast
‘He writes very fast.’
b. *I sia-kah ji cin kin. (descriptive) He write-KAH word very fast
‘He writes very fast.’
(48) a. I ka in kiann phah-kah lian chupinn long chutlai khuann. (extent)
27
He KA his child hit-KAH even neighbor all out see
‘He hit his child so much that even the neighbors came out to have a look.’
b. *I phah-kah in kiann lian chupinn long chutlai khuann. (extent) He hit-KAH his child even neighbor all out see
‘He hit his child so much that even the neighbors came out to have a look.’
(49) a. I gin-a phah-kah hunkhi. (resultative) He child hit-KAH faint
‘He hit the child so much that he/she fainted.’
b. I phah-kah gin-a hunkhi. (resultative) He hit-KAH child faint
‘He hit the child so much that he/she fainted.’
Based on the differences among the descriptive, resultative, and extent constructions above, Lin proposes three structures for each construction. Firstly, resultatives are clausal complements subcategorized for by V-kah (i.e., V1-kah in this paper), as shown in (50) (H.-L. Lin 2003, (38) and (39)).
(50) VP (resultative) NP V’
V CP
C’
C ASPP
I khau-kah chiukin-a tamkhi He cry-KAH handkerchief wet-ASP
‘He cried so much that the handkerchief was all wet.’
28
chiukin-a ‘handkerchief’ is the subject of the resultative clause because it can occur in lian…long ‘even…all’ construction, which is known as a preverbal focus position (cf.
Gu and Pan 2001), as (51) (H.-L. Lin 2003, (38b)) shows.
(51) I khau-kah lian chiukin-a long tamkhi-a. (resultative) He cry-KAH even handkerchief all wet-ASP
‘He cried so much that even the handkerchief was all wet.’
Another important issue of resultatives is the status of -kah. Lin follows T.-C.
Tang (1990)’s analysis, in which the counterpart of -kah in Mandarin, namely -de, is a suffix and it is attached to the main verb in the lexicon. That is to say, V-de is treated as a complex verb, and the analysis is supported by the fact that the whole complement following -de can be omitted, as in (6a) (repeated as (52)).
(52) Kan ni mei-de [S’
φ
]. (Mandarin) Look you beautiful-DELit. ‘Look at you! You are so beautiful!’
Based on Tang’s analysis, Lin treats -kah as a suffix attached to the main verb in the lexicon because the same type of examples can be found in Taiwanese, as illustrated by (53) (H.-L. Lin 2003, (46)).
(53) Cau-kah gua thiam-kah e. (Taiwanese) Run-KAH I tired-KAH
‘I got so tired from running.’
29
Secondly, the descriptive construction is an adverbial adjunct headed by -kah, and the AdvP is adjoined to V’, as (54) illustrates (H.-L. Lin 2003, (59)).
(54) VP1 (descriptive) NP V’
V VP2
NP V’
V’ ADVP V ADV AP I ji sia -kah cin kin
He word write-KAH very fast
‘He writes very fast.’
-kah is like an adverbial marker that changes AP into AdvP, which then modifies the main verb. Besides, -kah is a clitic-like element, which should attach to the preceding verbal element at PF. Hence, the semantic object of the main verb must precede V-kah, or the object will block the attachment of -kah, and the sentence will be ungrammatical, as shown in (47).
Thirdly, the extent construction is proposed to be an CP adjunct attached to VP, and -kah, as the head of the CP, takes a clause as its complement, as shown by (55) ( H.-L. Lin 2003, (65)).
30
(55) VP (extent) VP CP
NP V’ C’
V C ASPP
Thinnting kng -kah pro chinchan jitsi kangkhuan Sky bright-KAH like daytime same
‘The sky is so bright that (it is) like the daytime.’
As the descriptive construction, -kah in the extent construction is a clitic-like element, which attaches to the preceding verbal element at PF, so the object of the main verb has to precede V-kah, as exemplified in (48).
Lin explores the Taiwanese kah construction in detail, and proposes a diverse approach to analyze the construction. However, her analysis seems to need to be reconsidered under a closer scrutiny. Firstly, the differentiation of the scope of modification is problematic. For instance, as one reviewer points out, the resultative construction does not always predicate on the subject or object, as illustrated by (56) (H.-L. Lin 2003, (20)).
(56) I sia-kah pit long bo cui-a. (resultative) He write-KAH pen all not-have ink-ASP
‘He wrote so much that the pen ran out of ink.’
The resultative clause pit long bo cui ‘pen all not-have ink’ is not predicated on the subject i ‘he’ of the matrix clause, but on the main predicate sia-kah ‘write-KAH’.
Besides, in addition to descriptives, extent constructions also modify the main predicate penn-kah ‘sick-KAH’, as shown by the following sentences.
31
(57) a. I penn-kah ukau giamtiong-e. (descriptive) He sick-KAH very seriously-ASP
‘He was seriously ill.’
b. I penn-kah ising ma bo huatto. (extent) He sick-KAH doctor also no solution
‘He was so sick that doctor cannot help him.’
More importantly, following the classification Lin proposed, we will predict that (44a) (repeated as (58)) is ambiguous.
(58) I ka in kiann phah-kah lian chupinn long chutlai (extent / resultative) He KA his child hit-KAH even neighbor all out
khuann.
see
‘He hit his child so much that even the neighbors came out to have a look.’
The complement may refer to the extent of the event denoted by V1 phah ‘hit’ or the result of phah ‘hit’. Therefore, (58) may be an extent construction as Lin argues, or a resultative construction. Accordingly, (58) can be passivization and the semantic object of V1-kah may precede or follow it when (58) is interpreted as a resultative construction. However, the prediction is not born out, as shown by (59)-(60).
(59) *Chupinn hoo i ka in kiann phah-kah long chutlai khuann. (resultative) Neighbor HOO he KA his child hit-KAH all out see
‘The neighbors came out to have a look because he hit his child so much.’
32
(60) a. I ka in kiann phah-kah lian chupinn long chutlai khuann. (resultative) He KA his child hit-KAH even neighbor all out see
‘He hit his child so much that even the neighbors came out to have a look.’
b. * I phah-kah in kiann lian chupinn long chutlai khuann. (resultative) He hit-KAH his child even neighbor all out see
‘He hit his child so much that even the neighbors came out to have a look.’
Therefore, the classification Lin proposed should be revised. Without a correct classification, the analysis may not be absolutely right.
Secondly, as well known in the literature (C.-T. Huang 1982), the adjunct in Chinese, including Mandarin and Taiwanese, must be preverbal, as shown in (61)-(62).
(61) a. Zhangsan manman-de zou. (Mandarin) Zhangsan slowly walk
‘Zhangsan walked slowly.’
b. *Zhangsan zou manman-de (Mandarin) Zhangsan walk slowly
‘Zhangsan walked slowly.’
(62) a. Ong-e tautau-a kiann. (Taiwanese) Ong-e slowly walk.
‘Onge walked slowly.’
b. *Ong-e kiann tautau-a. (Taiwanese) Ong-e walk slowly.
33
‘Onge walked slowly.’
However, Lin argues that the descriptive and extent construction are adjuncts postverbally attached to V and VP, respectively, as illustrated in (54)-(55).
Unfortunately, the analysis is incompatible with the phrasal structure of Taiwanese.
Thirdly, the status of -kah should be reconsidered. As Lin proposed, -kah in resultatives is a suffix, which is attached to V1 in the lexicon, while -kah in the other constructions is a clitic-like head, which attaches to V1 at PF. Accordingly, her analysis predicts that the complement of -kah in resultatives can be dropped but the complement in the other constructions cannot. However, her claim is challenged by examples like (63)-(65).
(63) Descriptive constructions
a. I ciah iau, kinajit chuan cit toh, i iting ciah-kah cin pa.
He that hungry, today prepare this table he must eat-KAH very full
‘He is so hungry, and we prepare a lot of delicious food. He must eat so much that he is very full.’
b. I ciah iau, kinajit chuan cit toh, i iting ciah-kah e.
He that hungry, today prepare this table he must eat-KAH ‘He is so hungry, and we prepare a lot of delicious food. He must eat so much
that he is very full.’
(64) Resultative constructions
a. I ciah iau, kinajit chuan cit toh, i iting ciah-kah cin huannhi.
He that hungry today prepare this table he must eat-KAH very happily ‘He is so hungry, and we prepare a lot of delicious food. He must eat very
34
happily.’
b. I ciah iau, kinajit chuan cit toh, i iting ciah-kah e.
He that hungry today prepare this table he must eat-KAH
‘He is so hungry, and we prepare a lot of delicious food. He must eat very happily.’
(65) Extent constructions
a. I ciah iau, kinajit chuan cit toh, i iting ciah-kah beh He that hungry today prepare this table he must eat-KAH going-to
too.
throw up
‘He is so hungry, and we prepare a lot of delicious food. He must eat so much that he is going to throw up.’
b. I ciah iau, kinajit chuan cit toh, i iting ciah-kah e.
He that hungry today prepare this table he must eat-KAH
‘He is so hungry, and we prepare a lot of delicious food. He must eat so much that he is going to throw up.’
In fact, not only the resultative construction but also the descriptive and extent construction allow the complement following -kah to be omitted, which is contrary to Lin’s analysis.
Fourthly, the contrast of passivization and word order is actually not attributed to the classification of the three constructions. As far as the passivization (cf.
(43)-(46)) is concerned, Lin proposes that the descriptive construction does not have any passive counterpart, and the extent construction only allows the object of the main verb to undergo passivization, as illustrated by (43b) and (44c) (repeated as (66)).
35
(66) a. *Ji hoo i sia-kah cin kin. (descriptive) Word HOO he write-KAH very fast
‘Words are written very fast.’
b. *Chupinn hoo i ka in kiann phah-kah long chutlai khuann. (extent) Neighbor HOO he KA his child hit-KAH all out see
‘The neighbors came out to have a look because he hit his child so much.’
Actually, (66a) and (66b) are ungrammatical because the subject and clausal complement following the passive marker hoo do not have an adversative relationship, which is required by the passive construction in Taiwanese (cf. Cheng, L. L.-S.;
Huang, C.-T. J., Li, Y.-H. A. and Tang, C.-C. J., 1999; Kuo, P.-J., 2009). Consider the contrast below.
(67) a. Gua hoo i phah-a.
I HOO he hit-ASP ‘I got hit by him.’
b. *Gua hoo i oolo-a.
I HOO he praise-ASP ‘I got praised by him.’
The subject gua ‘I’ should be negatively affected by the clausal complement following hoo, such as i phah ‘he hit’ in (67a). Hence, (67a) is grammatical but (67b) is ungrammatical. In terms of the Taiwanese kah construction, the subject ji ‘word’ in (66a) and chupinn ‘neighbor’ in (66b) are not negatively affected by the clausal complement i sia ‘he write’ and i ka in kiann phah ‘he hit his child,’ respectively, so
36
(66a) and (66b) are ill-formed. Crucially, if the relation of the subject and predicate has the adversative reading, the Taiwanese kah construction can have the passivization counterpart, as shown in (68).
(68) a. i ka chiu-a io-kah gua ai sau kuikang-e chiuhioh.
he KA tree shake-KAH I have-to clean all-day-E leaf
‘He shook the tree so violently that I had to clean the leaves all day long.’
b. Guai [VP Opi [hoo i ka chiu-a io-kah ti ai sau kuikang-e chiuhioh].
I HOO he KA tree shake-KAH have-to clean all-day-E leaf ‘I got cleaned the leaves all day long by his shaking the tree.’
Since the relationship of the subject gua ‘I’ and the VP has the adversative meaning, it can have the passivization form no matter whether (68) is interpreted as a resultative or extent construction. Hence, whether the sentence allows passivization is related to if there is an adversative reading between the subject and the VP rather than which construction the complement belongs to. In terms of the contrast of the word order (cf.
(47)-(49)), the ungrammaticality of (47b) and (48b) is due to the syntactic structure of the Taiwanese kah construction, and I will come back to the issue in Section 4.2 of chapter 4.
Based on the empirical and theoretic challenges proposed above, the diverse analysis for the Taiwanese kah construction should be reconsidered.
3.2 C.-A. Wang (2010)
Wang (2000) explores phrasal resultatives and resultative compounds in Mandarin and Taiwanese. In this thesis, I only review the phrasal resultative part,