4.4 Other Patterns Elicited
The last question is to examine patterns that other than the target sentences Hakka children made in the PP Task. Analysis of the children’s production would facilitate our understanding of the characteristics and the process of children’s language development (e.g. Bever 1970, etc.). Six major patterns were found in response to our target Hakka
bun sentences, including 1) no elicitation, where the subjects remained silent; 2) key
words, where only one related noun or verb was produced; 3) active expressions, where
the active patterns were used; 4) passive expressions, where the children were able to utter passive sentences, but with some omissions or substitutions; 5) code-switching, where Mandarin Chinese was applied; and 6) ungrammatical sentence, where the utterances disagreed with Hakka grammar. Table 4-5 exemplifies sentences under each category and compares the error frequency of the experimental group with that of the control group.Table 4-5: Other Responses Elicited in the PP Task
Omission of post-bun NP+main verb
Fai ngin bun ginchak zok ziu le. (G3S2) bad person BUN police catch away ASP
‘The bad person was caught away by the police.’
11 (1.42%) 0
Omission of resumptive pronouns
Didi bun baba fak so tinai.(G3S14) Brother BUN Father punish clean floor
‘Brother was punished by Father to clean the floor.’ A-Bou BEI bad person catch away ASP
‘A-Bou was caught away by the bad person.’
6 (0.77%) 0
As shown in Table 4-5, there were totally 775 unexpected responses produced by the experimental group, and 8 items by the control group. These 8 items were the omission of resumptive pronouns (75%) and the substitution of key words (25%). The responses of the experimental groups, on the other hand, displayed the six patterns. Among them, no
elicitation (48.26%) took up the largest proportion of all patterns, followed by key words
(26.58%) as the second, and then active expressions (15.35%), ungrammatical responses (6.32%), passive expressions (2.71%), and finally code-switching (0.77%). With regard to the categories of key words, it was found that nouns (17.68%) were produced more than twice as many as verbs (8.9%). With respect to active sentences, the pattern“Topic-(S)-VP” (9.55%) was found more than “S-VP” (5.80%). As for passive
expressions, 11 items came from the deletion of the post-bun NP together with the main
verb (1.42%), 6 (0.77%) were the resumptive pronoun omissions, which were quantitatively equal to the adults’ adoption (n=6), and 4 substitutions of key words (0.52%).From the above results, we noted that the 8 items found in the control group all derived from the non-gapped passives. Among them, three-forth of the responses were the omission of resumptive pronouns and one-fourth were using a verb phrase to reply, i.e., chufak ‘to punish’ or sevon ‘to wash dishes,’ to substitute the complex construction of “verb1+ resumptive pronoun+ verb2+ noun.” This phenomenon again suggested an infrequent use of the non-gapped construction in Hakka passives. When dealing with the same construction, our children adopted similar strategies. However, more ungrammatical sentences were made. For instance, both object and resumptive pronoun were truncated, forming an awkward utterance, e.g., *Didi bun fak so tinai, or the verb and the resumptive pronoun were deleted together, e.g., *A-moi bun mama sevon.
Another noticeable point lied in the distinct weight among nouns and verbs produced by the subjects where the former was twice as many as the latter. Literature has
indicated that nouns are conceptually simpler than verbs (Gentner 1982), accounting for the findings that children’s first words are primarily nouns rather than other categories (Genter 1978, Macnamara 1972, Nelson 1973). An abundant of evidence is drawn to make this ground solid: acquisition of English vocabulary (Greenfield and Smith 1976, Huttenlocher 1974), and cross-linguistic studies, i.e., Mandarin Chinese, Japanese, and German (cf. Gentner 1982). Our study lent a support to the subjects’ preference for nouns when they were unable to make full sentences.
Third, the production of an active pattern “Topic-(S)-VP,” which occupied almost one-tenth of answers given by the experimental group, was also an interesting issue. The pattern accords to what is called “the topic construction13” in Mandarin Chinese, i.e., topic +comment, which is considered “the grammatical meaning of subject and predicate in a Chinese sentence” (Chao 1968:69). Topics always occur in sentence-initial position to establish a “special, temporal, or personal frame/domain” for the following comment (Chafe 1976). When the grammatical subject (S) was saved in the comment, for instance, our subjects tended to produce A-Bo fai ngin zok ziu le ‘A-Bo, the bad person caught away’ instead of A-Bo bun fai ngin zok ziu le ‘A-Bo was caught away by the bad person.’
The topic is A-Bo as a personal frame/domain, while the comment is fai ngin zok ziu le, where the children were describing an event happening to the leading role. When the grammatical subject was left off from the comment, on the other hand, the sentence such as mama hak do ‘Mother, being frightened’ was produced instead of mama bun ngin hak
do ‘Mother was frightened (by someone).’ The topic is mama, which is also a personal
frame/domain, while the comment is hak do, where the children accentuated more on the description of mama’s state. Their treatment of these kinds showed the development of the topic construction, which can be seen in the Hakka dialect of Mandarin Chinese.
13 Three major types of topic-comment constructions have emerged from the heated debate about the relationship between topic and subject in Mandarin Chinese (cf. Shi 2000). Since this is not the main focus of the study, we will not go into detailed discussion here.
The other active expression was through the use of a pattern “S-VP,” which was also a strategy frequently used by the subjects. Two types were found in this pattern: normal actives and reverse actives. When “normal actives” were applied, our children conveyed correct messages but misallocated the positions of “topic” and “focus” functioning in passives (cf. Chafe 1976, Li & Thomspon 1981, Siewrierska 1984, Tseng 1997, Van Oosten 1984), where the former, the so-called “old information,” occurs in sentence-initial place, while the latter, referred as “new information,” often presents in sentence-final position. For instance, the children uttered baba cim do didi le ‘Father found Brother’ instead of the passive didi bun baba cim do le ‘Brother was found out by Father.’ In the latter, the topic is didi, which is in line with the topic of the question didi
iongban? ‘What happened to Brother?’ In their utterances, we found that they treated baba as a topic, which should be the focus in the target sentence. The collision of “topic”
and “focus” showed that the children had not fully acquired the pragmatical aspect of passives.
By contrast, when “reverse actives” were produced, the subjects exchanged the positions of “agent” and “patient” roles in their active expressions. For instance, baba
siang mama ‘Dad misses Mom’ was used instead of baba bun mama siang ‘Dad was
missed by Mom’, and celai nou cemoi ‘The boy hates the girl’ instead of celai bun cemoinou ‘The boy was hated by the girl.’ In the former sentence, the actor/agent should be mama in the target sentence rather than baba in the children’s responses, while in the
latter, the actor/agent should be cemoi other than celai. The tendency of these sorts indicated that the children interpreted test sentences as cause-and-effect questions. And thus, having reciprocal concepts in minds, they uttered active sentences with the reverse semantic roles inside. Intriguingly, verbs in these constructions are experiential verbs, i.e.,siang ‘to miss’ and nou ‘to hate,’ which are the least competent verb type used in passives.
Similar findings were found in Tseng’s (1997) study of Chinese passive acquisition,
where her young subjects were liable to reverse the semantic roles in passives with experiential verbs. Our children’s responses reinforced the assumption that experiential verbs are less prototypical as verbs in passive constructions.
Also noteworthy was a common phenomenon found in the children’s responses:
simplification, which was evident from the omission of the post-bun NP together with the main verb in the long-distance passive, becoming a short one. For instance, the children uttered fai ngin bun gincha zok ziu le ‘The bad person was caught away by the police’
instead of fai ngin bun A-Bo hem gincha zok ziu le ‘The bad person was caught away by A-Bo’s asking for the police’s help.’ Facing with this kind of construction, the children tended to shorten the target sentence by deleting the main verb and the real agent but preserving the sub-event verb, sub-event patient and agent. This indicated that syntactic complexity deterred our young children from performing well. At the same time, it was probably under the impulsion of pragmatic needs that the children paid more attention to the effect of an event rather than the cause.
Concerning the category of code-switching, it was found that two kids kept adopting Mandarin Chinese to answer certain questions, in whole sentences or the replacement of
bun by Mandarin bei. This phenomenon might be attributed to their infrequent use of
Hakka at home, resulting in their unfamiliarity with the passive construction. However, their use of Chinese passives showed that they had been already acquired the concept of passivization and they could comprehend the questions though they were unable to express fluently in Hakka.Finally, our children were likely to truncate the post-bun NP, contributing to parts of ungrammatical responses in the last category, e.g., *A-Min bun tet do (‘A-Min was kicked’). The result showed that they needed more time to acquire the notion of no truncation, which is a typical property of Hakka passives. Also, when the structural object (the agent) is unspecific in semantics or pragmatics, it would be more challenging for our
children to add an expletive pronoun in the post-bun position, e.g., A-Min bun ngin tet do
‘A-Min was kicked by someone.’ The finding confirmed what we suggested in Section 4.2 that the filling-up with the functional morpheme, i.e., the expletive, is difficult for young speakers.
The distribution of error patterns provided a primary account of the subjects’
development of Hakka passives. However, our children of different ages displayed different preferences and tendencies, signifying developmental characteristics specific to a particular age group. Table 4-6 illustrates the results of the four experimental groups.
Table 4-6: Error Responses of the Experimental Groups in the PP Task Group
Type
Group 1 Group 2 Group 3 Group 4 No elicitation 179 (70.20%) 135 (54.44%) 51 (24.29%) 9 (14.52%) Key Words 76 (29.80%) 109 (43.95%) 14 (6.67%) 7 (11.29%)
Active expressions 0 0 98 (46.67%) 21 (33.87%)
Passive expressions 0 0 7 (3.33%) 14 (22.58%)
Code-switching 0 2 (0.81%) 3 (1.43%) 1 (1.61%)
Ungrammatical response 0 2 (0.81%) 37 (17.62%) 10 (16.13%)
Total 255 248 210 62
As can be seen in Table 4-6, numbers of patterns other than the target sentences elicited in the production task were on the decrease with the growing of ages (G1: 255; G2: 248;
G3: 210; G4: 62). Of these patterns, no elicitation was found frequently highest in Group 1 (179) and then Group 2 (135), and dropped dramatically as in Group 3 (51) and Group 4 (9). Key words were produced most frequently in Group 2 (109), and Group 1 (76) as the second high, and then plunged in Group 3 (14) and Group 4 (7). Active expressions (G3: 98, G4: 21) and passive expressions (G3: 7, G4: 14) were only found in Groups 3 and 4; the former was with a decreasing trend, while the latter was the other way around.
Code-switching and ungrammatical responses were found in all groups except Group 1,
and their peaks fell in Group 3.
Looking into the production of each group, a rough line was drawn at the age of six when younger kids, i.e., Group 1 (age: 4) and Group 2 (age: 5), were nearly unable to answer the questions, either remaining silent or uttering fragments. Therefore, no
elicitation and key words took up the majority of the results. Only small proportions were
found in code-switching where Mandarin Chinese was used in a whole sentence and truncation of the post-bun NP was produced as ungrammatical responses. Furthermore, we noted that there appeared a complementary distribution between no elicitation and keywords in the younger groups. The number of key words was increasing in Group 2 with
the decrease of no elicitation, demonstrating the developmental traces.After entering the age of six, the children did significantly better. Most of them were capable of using complete sentences to answer the questions, articulated by the diving figures under no elicitation and key words, though more varieties of sentence patterns were produced. Amid them, we found active expressions the most frequently made by the older children, particularly in Group 3 (age 6). Active sentences have been pointed out easier and thus earlier for children to acquire than passives (cf. Marchman 1991, etc.).
Also, actives are more typical to be produced by contrast (cf. De Villiers and de Villiers 1978, Vicki and Braine 1985). The relative higher frequency in active expressions in the present study reconfirmed this ground. On the other hand, Group 3 was also at the starting line of making passive expressions, though the number was in the minority. The fact that Group 3 was at the summit of the production of actives and simultaneously at the beginning of passive expressions demonstrated a transitional phase, where they were still adapting themselves to the usage of passive constructions.
In addition, the 6-year-olds were at the age when the most ungrammatical responses and code-switching were produced. This is probably because the age of six was on the threshold of elementary school year, when Mandarin Chinese became the major language
to communicate in school. The high exposure to Mandarin Chinese might influence our children’s speaking of Hakka. Therefore, they made more truncated passives, which are ungrammatical in Hakka but in line with the allowance of object omission in Mandarin Chinese. And also consequently, the children who attempted to use Mandarin Chinese were relatively more than other age groups.
As for Group 4, numbers of other patterns elicited diminished to the minimum, except passive expressions. This might be attributed to the decrease of numbers in passives such as no elicitation, key words, active expressions, code-switching, and error
responses, showing a developmental progress found in Group 4, who performed almost
as well the adult. On the other hand, children at age 7 mounted to produce more passives, including target and non-target passives. Their abilities of Hakka passives were near full maturation.Three main stages to fully acquire Hakka passives can be defined accordingly: 1) Stage 1: no elicitation and key words, which were largely contributed by Groups 1 and 2;
2) Stage 2: active expressions; which were abundantly conducted by Group 3; and 3) Stage 3: passive expressions, where Group 4 was the major producers, as shown in Figure 4-4 as follows.
Stage 1 (4; 0~5; 0): No elicitation/ Key words
Stage 2 (6; 0): Active expressions
Stage 3 (7; 0): Passive expressions
Figure 4-4: Developmental Trends Found in the PP Task