On the basis of the theoretical background taken up earlier, this study aims to address the following questions:
1. Are there any property effects on L1 Hakka passives?
2. Do Hakka children perform similarly on each property’s sub types?
3. Do Hakka children perform similarly on different tasks on Hakka passives?
4. Are there any particular errors that Hakka children make in acquiring passives?
5. Are there any age effects on L1 Hakka passives?
1.4 Significance of the Thesis
Passives acquisition has drawn lots of researchers’ attention for years (Baldie 1976, Chang 1986, De Villiers and de Villiers 1978, Harwood 1959, Kuo 1995, Marchman 1991, Tse et al. 1991, Tseng 1997, Sudhalter Vicki and Braine 1985, etc.). However, there is room for further explorations. Moreover, L1 acquisition of Hakka passives is still an undeveloped frontier. The concerned properties of the study, therefore, will range from those mainly on the focus, i.e., verb types, animacy, reversibility, and truncation (Kuo 1995, Tseng 1997), to syntactic complexity and other related issues.
Therefore, the present study aims to reveal Hakka children’s language acquisition of passives in detail, and find out their difficulties in language development. It is hoped that the findings of the present study will help to reflect the acquisition of the Hakka passive construction and arouse Hakka native speakers’ awareness of language preservation and passing-on.
1.5 Organization of the Study
The study is organized as follows. Chapter Two discusses main properties of Hakka passives and reviews relevant theoretical and empirical research. Chapter Three is dedicated to the research design of this study, and Chapter Four reports and discusses the experimental findings. Finally, Chapter Five offers conclusions as well as suggestions for further research.
Chapter Two
Linguistic Properties and Previous Studies of Hakka Passives
In this chapter we aim to analyze linguistic properties of Hakka passives and to review related studies and empirical research. Linguistic properties of Hakka passives will be discussed in Section 2.1. In Section 2.2, three theoretical studies of Hakka passives will be reviewed (Chiang 2006, Hsiao 1996, Lai 2001). Two empirical studies on the acquisition of Mandarin passives (Kuo 1995, Tseng 1997) will be discussed in Section 2.3. Finally, a brief summary of this chapter will be given in Section 2.4.
2.1 Linguistic Properties of Hakka Passives
Before we discuss about linguistic properties of the Hakka passive, it is necessary to introduce its basic form. The Hakka passive is marked by the morpheme bun and thus is also called the bun construction. Its pattern can be seen in (3).
(3) a. NP1 bun NP2 VP.
b. Sinsan bun eniateu hak do. (Luo 1988:300) teacher BUN we frighten DO
‘The teacher was frightened by us.’
In (3a), NP1 is a structural subject and usually serves as a patient; while NP2, the
“post-bun NP,” is a structural object and acts as an agent of the sentence. Descriptively, in (3b), the subject sinsan ‘teacher’ is acted upon by a behavior of frightening, which is performed by the object eniateu ‘we.’ The morpheme bun has signified the passive voice in the sentence.
After understanding the basic pattern, now let us turn to the important properties of
Hakka passives. First, the constraint of no truncation in Hakka passives is discussed in Section 2.1.1. Furthermore, the syntactic complexity and the resumptive pronouns are examined in Sections 2.1.2 and 2.1.3, respectively. Then, adversity of the Hakka passives is introduced in Section 2.1.4., followed by an issue of animacy/inanimacy stated in Section 2.1.5. Finally, the degree of transitivity is explored in Section 2.1.6.
2.1.1 No truncation
In a Hakka passive, the object can not be truncated (cf. Chiang 2006, Hsiao 1996, Lai 1989, 2001, Luo 1988, Shi 2001). In other words, only when an overt post-bun NP occurs can a passive sentence be grammatical, as illustrated below:
(4) a. Didi bun baba ma.
brother BUN father blame
‘Brother was blamed by Father.’
b. Didi bun ngin ma.
brother BUN person(s) blame
‘Brother was hit by someone(s).’
c. *Didi bun ma.
brother BUN blame ‘Brother was blamed.’
Example (4) depicts a contrast between non-truncated and truncated passive sentences. In (4a) and (4b), both agentive objects are overt NPs, which qualifies their grammaticality. Their difference lies in that the object in (4a) is a specified NP: baba
‘father’, while in (4b) is unspecified: ngin ‘person (s),’ which is a person unknown in the discourse. By contrast, (4c) is ungrammatical due to its lack of an explicit object. In
general, example (4) shows that whether an object is specified or not, it needs to be obligatorily presented in the post-bun position in Hakka.
2.1.2 Syntactic complexity
Now let us consider an issue of syntactic complexity in Hakka passives: the short-distance passive vs. the long-distance passive.
A short-distance passive is a simple, mono-clausal passive (Feng 1990, Ting 1998).
Due to its simplicity, it is regarded as an unmarked form of the Hakka passive, as illustrated in (5) and (6).
(5) a. Fai ngin bun A-Bo zok ziu le.
bad man BUN A-Bo catch away ASP ‘The bad man was caught away by A-Bo.’
b. [IP1Fai ngin1 bun [IP2 A-Bo zok ziu t1 le]].
(6) a. Tong bun cengin siit tet le.
candy BUN kid eat TET ASP
‘The candy was eaten by a kid.’
b. [IP1Tong1 bun [IP2 cengin siit tet t1 le]].
In (5a), the animate subject fai ngin ‘the bad person’ is a patient, who undergoes a catching-away by the agent A-Bo. The deep structure of (5b) shows that there is only one IP, i.e., IP2, in the post-bun clause. In (6a), the inanimate subject tong ‘candy’ is a theme, which is eaten by a kid. (6b) shows that there is only one IP, i.e., IP2, in deriving a short-distance passive like (6a). As a short-distance passive, its syntactic structure is simple and the meaning is straightforward.
On the other hand, a long-distance passive, also called a non-local passive (cf. Ting
1998), has posed certain problems for its syntactic nature1 and thus needs more efforts to understand its meaning, as shown in (7) and (8).
(7) a. Fai ngin bun A-Bo hem gincha zok ziu le.
bad person BUN A-Bo ask police catch away ASP ‘The bad person was caught away by A-Bo’s asking for the police’s efforts.’
b. [IP1 Fai ngin bun [IP2 A-bo ham gincha [IP3 zok-ziu le]]].
(8) a. Tong bun amei hem cengin siit tet le.
candy BUN mother ask kid eat TET ASP
‘The candy was eaten by Mother’s asking for the kid’s efforts.’
b. [IP1 Tong bun2 [IP2 amei hem cengin [IP3 siit tet le]]].
In (7a), the bad person was arrested by the police, but it was actually A-Bo who asked the police to catch the bad person away. The real agent of the entire event is A-Bo, while the police should be the agent of a sub-event of arresting the bad person. Example (7b) shows that in the deep structure, where at least two IPs, i.e., IP2 and IP3, are involved in the post-bun construction. The same applies to (8), where the real agent of the entire event is amei ‘mother,’ and cengin ‘kid’ as the agent of a sub-event of eating the candy.
Two IPs can be found in the bun clause as well. From the examples above, we know that the long-distance passive is more complex in its syntactic representation and semantic
1Accordingly, a long-distance passive exhibits “unbounded” dependency. This is regarded as counterevidence against A-movement and is characteristic of A’-movement (cf. Feng 1990, Huang et al.
2008, Li 1991, Ting 1998). The gap-to-be position associated with the subject of the bun construction is filled with the complement of an object control verb, which is the fact entirely unexpected under an A-movement approach. Under the argument, the underlying form for example (7) is as follows:
(i) [IP1 Fai-ngin1 [VP bun [CP1 Op1 [IP A-Bo ham gin-cha2 [CP2 t”1 [IP2 PRO2 zok-ziu le t’1]]]
In this underlying structure, an A’-movement of a null operator is involved.
Though the treatment for the long-distance passives of Hakka is reasonable to some extent, it still needs more confirmations. In the present study, this is not the major concern. Therefore, we put it aside for further research.
2 Some suggest that the morpheme bun should be located inside the IP2 bracket (cf. Feng 1990).
interpretation.
2.1.3 Resumptive pronouns
The third property we are concerned with is the distribution of resumptive pronouns in the so-called non-gapped passive (cf. Feng 1990, Ting 1998, Huang et al. 2008). Like the long-distance passive which we have discussed in Section 2.1.2, a non-gapped passive is also uncertain in its syntactic derivation3. Moreover, it also contains an overt object which occupies the internal argument position of the “passivized” control verb.
But it differs from the long-distance passive in that this overt object co-indexes with the structural subject (cf. Feng 1990, Li 1991, Ting 1998).
To be precise, the word order of the non-gapped passive can be considered as “NP1+
bun + NP2+ V+ NP3…,” which seems identical to the long-distance passive. However,
unlike different referents which NPs refer to in the long-distance passive, the internal argument of the passivized verb in the non-gapped passive, i.e., NP3, in fact is co-indexed with NP1. Since NP3 is usually a pronoun, it is called a resumptive use (cf.Huang et al. 2008), as shown in (9) and (10)4.
(9) Meimeii bun baba fak gii so tinai.
sister BUN Father punish she clean floor ‘Sister was punished by Father to clean the floor.’
(10) Gieulei bun gincha hem gii hi zok fai ngin.
dog BUN police ask it go catch bad person ‘The dog was asked by the police to catch the bad person.’
3 Like the long-distance passive, the syntactic nature of the non-gapped passive is not our focus in the present study. For further information, reader may refer to Feng (1990), Li (1991), Ting (1998), and Huang et al. (2008).
4 To our knowledge, no studies explicated the constraint of subject animacy in the non-gapped passive.
However, an inanimate subject will make the sentence awkward. Therefore, both examples here are illustrated with animate subjects.
In (9), the pronoun gi ‘she’ refers to meimei ‘sister,’ who was punished by baba
‘Father.’ Due to the same referent of meimei and gi ‘she,’ gi in this structure is a resumptive pronoun. The clause beginning from baba to the sentence-final position, on the other hand, describes the passive event in an active voice. The same pattern can be seen in (10), where the pronoun is also a resumptive pronoun, referring to the non-human animate subject gieule ‘dog.’ The involvement of the resumptive pronoun in the Hakka passive may be more or less complicated for children to comprehend and produce.
2.1.4 Adversity
The next property to be discussed is a semantic constraint on the use of Hakka passives: adversity (cf. Hsiao 1996), which implies inflictive or unfortunate situations (Chao 1968, Li and Thompson 1981). A passive sentence without any negative meaning would be considered unnatural, as can be seen below:
(11) a. Didi bun baba ma.
brother BUN father blame
‘Brother was blamed by Father.’
?b. Didi bun baba kuatsong.
brother BUN father praise
‘Brother was praised by Father.’
A verb ma ‘to blame’ in (11a) expresses the meaning of adversity, which meets its semantic requirement for the passive. By contrast, (11b) sounds awkward unless the speaker refers to something ironic5 (cf. Hsiao 1996:78-79).
5 To meet the semantic requirement of adversity, when there is a seemingly positive message conveyed by verbs such as complimenting, it is usually regarded as sarcasm or humiliation in the passive construction.
The constraint of adversity is met even if the meanings of verbs are neutral, such as verbs of perception or cognition. Such verbs in Hakka include
koindo ‘to see,’ tangdo ‘to
hear’ and so forth, as illustrated in (12)-(13).(12) Ngioule bun cegin koi do le.
kitty BUN kids see DO ASP
‘A kitty was seen by a kid.’
(13) Gi gon gai fa bun pet ngin tang do le.
he speak CSC words BUN other people hear DO ASP
‘What he said had been heard by other people.’
In (12), the passive construction conveys unwillingness or reluctance of the kitty seen by the kid. In (13), the implication of adversity is also transmitted through the passive construction where the subject may not expect the over-hearing by other people. Both sentences illustrate the semantic constraint of adversity even the verbs are neutral in meaning.
On the whole, Hakka passives imply adversity, even if the verbs carry non-adversative meanings. Therefore, the unmarked usage of Hakka passives should be set under certain unpleasant or with some negative propositions.
This is exemplified in the following.
(i) Ngai bun ngin hien tso pantson. (Hsiao 1996:78-79) I BUN people elect be the-leader-of-the-class
‘I was elected as the leader of the class.’
(ii) Gi bun ngin kuatsong.
he BUN people praise ‘He was praised by people.’
In (i), the bun construction may imply the unwillingness of ngai ‘I’ to be elected as a class leader, and in (ii), the speaker may suggest that gi ‘he’ does not deserve the praise, or that the praising is an irony.
2.1.5 Animacy/Inanimacy
Animacy of the structural subject (the patient role) and object (the agent role) is an issue considered in the present study. Animacy has been widely discussed in the literature in that it exerts certain influences on the acceptance for passives (Bock 1986, Dewart 1979, Harris 1978, Kuo 1995, Lempert 1984, 1989, 1990, Tseng 1997). Generally speaking, it is believed that the prototypical subject should be conceptualized as a category, which always comprises cognitively salient properties or components such as animate, cause, and dynamism (cf. Bates and MacWhinney 1982, Hinton 1981, Lempert 1989). Accordingly, an animate subject in passives should be more qualified than an inanimate subject. On the other hand, when we look into the thematic role, the prototype of an agent is also believed to be animate (Givon 1984). Under this proposition, it is suggested that a passive with an animate agent should be the unmarked form (cf. Kuo 1995). In the present study, we will discuss the relation between the passive and subject/object animacy.
Four possible sequences of combinations are found in Hakka passives according to animacy of the subject and object: animate-animate, animate-inanimate, inanimate-animate, and inanimate- inanimate.
(14) Didi bun meimei da. (animate-aniamte) brother BUN sister hit
‘Brother was hit by Sister.’
(15) Mama bun coido got do le. (animate-inanimate) mother BUN knife cut DO ASP
‘Mother was cut by a knife.’
(16) Tong bun didi siit tet le. (inanimate-animate) candy BUN brother eat TET ASP
‘The candy was eaten by Brother.’
(17) Bozhi bun tai fung chui zeu le. (inanimate-inanimate) newspaper BUN strong wind blow away ASP
‘Newspaper was blown away by a strong wind.’
In (14), the subject didi ‘brother’ is an animate NP that is hit by another animate NP—meimei ‘sister’. In (15), the subject mama ‘mother’ is an animate NP that functions as an unwilling patient undergoing a cutting by an inanimate NP: knife. Examples (16) and (17), on the other hand, both consist of inanimate subjects is passives. In (16), an inanimate subject tong ‘candy’ is eaten by an animate NP didi ‘brother.’ In (17), bozhi
‘newspaper’ is an inanimate subject, which is a theme and is blown away by the inanimate natural force: wind. These four types will all be examined in the present study.
2.1.6 Degrees of Transitivity
Verbs of the Hakka dialect encompass two main kinds
6: transitive verbs and intransitive verbs (Luo 1985). A feature unique to transitive verbs is that they can be used both in active and passive voices (cf. Chomsky 1965, Luo 1985). Intransitive verbs, on the other hand, can only be used in active sentences7, as shown in (18).
6 Auxiliary verbs are excluded here due to their unweightedness in Hakka passives.
7 Some languages have intransitive verbs in the passive construction, e.g., Djeoromitxi or Yabuti, the Macro-Ge language (cf. Nascimento 1998). Some dialects of Chinese in the Mainland China also exhibit intransitive passives, e.g., the Jieyang Dialect of Chaozhou (cf. Matthews et al. 2005). However, this is not the case in Hakka passives.
(18) a. Yit gai cengin ki lan.
one CL kid stand LAN ‘A kid stands.’
b. *Bun yit gai cengin ki lan.
Based on different distribution, four main uses of transitive verbs can be listed in the following, where (19) to (21) are adopted from Luo (1985:153-154):
(19) a. S+ Vt+ O (Active) b. Lochu to gai ngau butoi.
rat TO there bite bag ‘A rat was biting a bag there.’
(20) a. S+ bun+ O+ Vt (Passive) b. Nggunchung bun keikung siit tet le.
centipede BUN rooster eat TET ASP
‘A centipede was eaten by a rooster.’
(21) a. S+ Vt+ IO+ DO or
S+ Vt+ DO+ P+ IO (Ditransitive Active voice) b. Gi bun gia laie it gap tien.
he give his son one CL field.
‘He gave his son a field.’
c. Gi bun it gap tien bun gia laie.
he give one CL field to his son ‘He gave a field to his son.’
(22) a. IO+ bun+ S+ V+ DO or
DO+ bun+ S+ V+ IO (Ditransitive Passive voice) b. It gap tien bun gi sung gia laie.
one CL field BUN he send his son ‘A field was sent to his son by him.’
c. Gia laie bun gi sung it gap tien.
his son BUN he send one CL field ‘His son was sent a field by him.’
From the above sentences, it is evidenced that a qualified verb type is necessary in the passive construction. In fact, instead of focusing on syntactic properties of transitive verbs, Hopper and Thomson (1980) propose parameters of transitivity according to
“different facet of effectiveness or intensity with which the action is transferred from one participant to another.” Maratsos et al. (1985) further claims that verbs fall into “a continuum of transitivity” in that the more transitive a verb is, the more qualified it is used in a passive construction (cf. Kuo 1995, Maratsos and Chalkley 1980, Maratsos et al.
1985, Pinker et al. 1987, Sudhalter and Braine 1985, Tseng 1997). Therefore, the present study will examine the relationship between verbal transitivity and the passive construction.
In this study, transitive verbs in Hakka passives are divided into three types8 according to degree of transitivity from high to low: resultative verb, action verb, and experiential verb, as can be seen in (23)-(25).
8 The classification is adopted from Tseng (1997).
(23) a. Resultative verbs: compound verbs that consist of a main verb and a resultative state, e.g., zok ziu ‘to catch away’, siit tet ‘to eat to the end’
b. Fai ngin bun A-Bo zok ziu le.
bad man BUN A-Bo catch away ASP ‘The bad man was caught away by A-Bo.’
(24) a. Action verbs: verbs which involve kinesics, e.g., ma ‘to blame’, dui ‘to chase’
b. Didi bun baba ma.
brother BUN Dad blame ‘Brother was blamed by Dad.’
(25) a. Experiential verbs: verbs which include both perceptual verbs and cognition verbs, e.g., nou ‘to hate’, koin ‘to see’
b. Cetle bun gincha nou.
thief BUN police hate ‘The thief was hated by the police.’
Accordingly, a resultative verb should be the most qualified category as a passive verb.
This is mainly due to its disposal nature in the passive construction (cf. Luo 1985, Tseng 1997). On the other hand, the most difficult category of verb type, i.e., an experiential verb, is attributed to its “low degree” in transitivity (cf. Maratsos et al. 1985). An action verb is supposed to be in-between.
To sum up, six properties of the bun construction are discussed. First, the Hakka passive does not exhibit truncation. Second, the contrast between the long-distance passive and the short-distance passive involves syntactic complexity. Third, a resumptive pronoun is focused in the non-gapped passive in Hakka. Fourth, generally speaking, Hakka passives denote adversity. Fifth, an animate subject is supposed to be prototypical in Hakka passives. Finally, a resultative verb is found to be the most transitive, followed
by an action verb and an experiential verb.
2.2 Linguistic Studies of Hakka Passives
Three studies of Hakka passives will be reviewed in this section. The first study provides a general picture of the descriptive forms and meanings of the bun construction (Hsiao 1996). The other two studies are more concerned about semantic development of
bun in Hakka passives (Chiang 2006, Lai 2001).
2.2.1 Hsiao (1996)
Bun9 in passives has been descriptively discussed in Hsiao (1996). The basic structure of the Hakka passive and its implied meanings are two main parts of the discourse. Concerning the basic form, bun acts as an agent marker in the structure of
“NP1 bun NP2 VP”; it marks the second NP as the one that triggers the affectedness of NP1, as shown in (26):
(26) a. Gi dalan bile. (Hsiao 1996:71) he break glass
‘He broke the glass.’
b. Bile bun gi dalan le.
Glass BUN he break ASP ‘The glass was broken by him.’
The morpheme le, which is a perfective aspect marker and also an expression of
“state-changing,” is usually needed in passives. However, the attachment of the aspect
9 In Hsiao’s (1996) thesis, the transcription system is simplified from Luo’s (1984), so bun is transcribed
9 In Hsiao’s (1996) thesis, the transcription system is simplified from Luo’s (1984), so bun is transcribed