Chapter 3
Research Design and Results
In this chapter, we will introduce the subjects and research design of the present study. Section 3.1 describes the background of the subjects participating in this study. In section 3.2, materials and research methods are reported. In section 3.3, a pilot study, the procedures for conducting the tasks of this study and the scoring criteria are provided. Section 3.4 presents the results of the present study. Finally, a summary of this chapter will be given in section 3.5.
3.1 Subjects
The present study adopted a quantitative method to investigate Chinese-speaking children’s L1 competence. Sixty Chinese children aged from 4 to 6 participated in this study; each age group contained 20 subjects. Besides, there was one control group, which consisted of 20 adult Chinese native speakers. The background information of the subjects is shown in Table 3-1.
Table 3-1: Background of Subjects
Group Age Number of Subjects
Group 1 4 (3;8 – 4;7) 20
Group 2 5 (4;8 – 5;7) 20
Group 3 6 (5;8 – 6;7) 20
Control Group 23 20
These subjects were all Chinese native speakers. Some of them spoke
Taiwanese, while others spoke Chinese only. These subjects were selected from
Ai-Er kindergarten. The subjects’ classroom teachers were inquired before the
experiment to ensure that the subjects selected did not have sensory, emotional, and
learning problems. The subjects were preschoolers and they attended the kindergarten five days a week. The schooling at the kindergarten was from 8 a.m. to 4 p.m. The language used for communication in the kindergarten was Mandarin Chinese, and English was taught for 1 hour per week in the kindergarten. The results of the performance of each age group were compared to see whether age is a crucial factor affecting the acquisition of Chinese topics.
3.2 Materials and Methods
“Obviously one can find out about competence by studying performance, but this study must be carried out in devious and clever ways, if any serious result is to be obtained” (Chomsky 1964). Thus, a methodological design plays an important role in examining children’s syntactic development.
There are numerous methods to assess children’s syntax, such as elicited
imitation, elicited production, and intermodal preferential looking paradigms. The
present study designed tasks to examine the types of topics classified in Table 3-2.
Table 3-2: Topics Examined in the Present Study Syntactic Semantic Examples
Generic Shuiguo, xiangjiao hen pianyi.
fruit banana very cheap
‘As for fruits, bananas are very cheap.’
Based-generated
Definite Zhexie chezi, hongche zui da.
these car red car most big
‘As for these cars, the red car is the biggest.’
Generic Daxiang, bizi chang.
elephant nose long
‘Elephants, (their) noses are long.’
Possessive-like
Definite Na-ge wawa, yianjing da.
that-CL doll eye big
‘That doll, (her) eyes are big.’
Generic Katong, jiejie meitian kan.
cartoon older sister everyday watch
‘Cartoon, older sister watches (it) everyday.’
Moved
Definite Nabei shui, xiao tuzi he le.
that-CL water little rabbit drink PFV
‘That glass of water, the little rabbit drank (it).’
As Table 3-2 shows, there were six types of topics in total. In order to better
understand how the subjects comprehended and used these topics, two tasks (i.e., an
elicited imitation task (EI task) and a sentence selection task (SS task)) were
employed.
3.2.1 The Elicited Imitation Task
The EI task (see Appendix A) was a production task used to test whether the
subjects could use the six types of topics in a single sentence and in a topic chain. It
was conducted before the SS task in order to prevent the subjects from being exposed
to the correct sentences that would be used in the SS task. This task basically
followed Chien and Lust’s (1983) design of sentences with redundant topics to
investigate children’s use of topics. The experimenter said sentences with redundant
topics. Pictures were also shown to provide more contextual cues to children when
the subjects listened to the sentences. Then, the subjects were asked to repeat the
sentences according to the pictures. If they correctly produced the sentence-initial
topic and the topic chain in their repetition, it meant that they have acquired the topic
and the topic chain. Two kinds of sentences, a single topic construction and a topic
chain, were used for the six types of topics. In addition to the test sentences, four
fillers were included; thus, there were sixteen sentences in total. Table 3-3 shows the
number of test sentences used in the EI task.
Table 3-3: Number of Test Sentences for Each Syntactic Construction in the EI Task
Syntactic Semantic The Length of the Sentence Sentence Number Single Topic Construction 1
Generic
Topic Chain 2
Single Topic Construction 3 Based-generated
Definite
Topic Chain 4
Single Topic Construction 5 Generic
Topic Chain 6
Single Topic Construction 7 Possessive-like
Definite
Topic Chain 8
Single Topic Construction 9 Generic
Topic Chain 10
Single Topic Construction 11 Moved
Definite
Topic Chain 12
13 14 15 Fillers
16
An example of test sentences used in the elicited imitation task is given in Table 3-4.
Table 3-4: A Test Sample Given in the Elicited Imitation Task Single Sentence:
Shuiguo, shuiguo de xiangjiao hen pianyi.
fruit fruit DE banana very cheap
‘As for fruits, bananas are very cheap.’
Topic Chain:
Shuiguo, shuiguo de xiangjiao hen pianyi, shuiguo de pingguo hen yingyang, fruit fruit DE banana very cheap fruit DE apple very nutritional shuiguo de putao hen haochi.
fruit DE grape very delicious
‘As for fruits, bananas are very cheap, apples are very nutritional, and grapes are very delicious.’
3.2.2 The Sentence Selection Task
The sentence selection task (see Appendix B) was a comprehension task used to test whether children could comprehend the six types of topics and topic chains in Chinese. This task was a grammaticality judgment task because it assumed implicitly ‘that the participants understand the notion of a sentence being “good”/
“right” or “silly”/ “wrong”’ (Gordon 1996). The grammaticality judgment task has
been regarded as an inappropriate method for language acquisition study because it is
believed that young children would not be able to make grammaticality judgments
(McDaniel and Cairns 1996). However, in the present study the grammaticality
judgment task was modified by providing subjects with background scenarios. After
being told the scenarios, participants listened to sentences said by two puppets, Lion
and Dog. Participants were asked to choose the puppet that said the sentence in a
right order according to the scenario. If the subjects chose the right puppet, it meant that they have acquired the topic. There were totally sixteen pairs of sentences, that is, two pairs for the six types of topics and four pairs of fillers. Each type of topics contained one pair for a single topic construction and another pair for a topic chain.
Table 3-5 shows the number of test sentences for the SS task.
Table 3-5: Number of Paired Test Sentences for Each Syntactic Construction in the SS Task
Syntactic Semantic The Length of the Sentence Pair Number Single Topic Construction 1 Generic
Topic Chain 2
Single Topic Construction 3 Based-generated
Definite
Topic Chain 4
Single Topic Construction 5 Generic
Topic Chain 6
Single Topic Construction 7 Possessive-like
Definite
Topic Chain 8
Single Topic Construction 9 Generic
Topic Chain 10
Single Topic Construction 11 Moved
Definite
Topic Chain 12
13 14 15 Fillers
16
Table 3-6 is an example of the scenario and the sample sentences for the base-generated generic topic in a single sentence.
Table 3-6: A Test Sample Used in the Sentence Selection Task
You yi tian, mama qu caishichang mai shuiguo. Ta tiao le hen jio,
have one CL mother go market buy fruit she pick CRS very long zuihou ta mai le xiangjiao yinwei xiangjiao hen pianyi.
finally she buy PFV bananas because bananas very cheap
Xianzai xiao shizi gen xiao gou dou shuo tamen zhidao fasheng le sheme shi.
now little lion and little dog both say they know happen PFV what thing Danshi tamen chizhong zhiyou yi-ge ren jiang dui le. Ni tingting, shei
but they among only one-CL person say right CRS you listen who shuo dui ne?
say correct PRT
‘One day, Mother went to the market to buy some fruit. She picked the fruit for a while, and finally she bought bananas because they are very cheap. Now, the puppy and the little lion both say they know what has happened. However, only one of them gets it right. Please listen to them and tell me who is correct?’
Lion:
Shuiguo, xiangjiao hen pianyi.
Fruit banana very cheap
Dog:
*Xianjiao, shuiguo hen pianyi.
banana fruit very cheap
‘As for fruits, bananas are very cheap.’
3.3 Procedures
Before the formal study, a pilot study was employed to test the subjects’
responses to base-generated, possessive-like, and moved topics. The procedure and the result of the pilot study are reported in this section. The procedure for conducting the formal study is also introduced.
3.3.1 Pilot Study
Twelve Chinese children participated in the pilot study and they were further divided into three age groups (i.e., 4, 5, and 6 years old). There were four children in each group (two males and two females).
This study basically followed Chien and Lust’s (1983) design by using redundant topics to investigate whether children have the concept of topics. If children reduced the redundant topic in the second clause, it was assumed that they have acquired the concept of the topic. The higher proportion of the redundant topics they reduced, the more salient the topics were. A story retelling task was adopted to elicit the subjects’
use of the three topics. The subjects were told twelve stories containing the redundant sentences. In order to prevent them from imitating the sentences, each story contained 8 sentences to influence their working memory. Moreover, in the training session, they were taught to retell the story in their own sentences, and they were asked to retell the story following the picture.
It was found that possessive-like topics were acquired before base-generated
topics, followed by moved topics. The subjects reduced the possessive-like topics
most frequently. All age groups were able to reduce possessive-like topics. The
6-year-olds almost fully acquired them. The production rates of the base-generated
topics were in the middle. Not until age 5 did the subjects know how to use Chinese
topics. Even the 6-year-olds did not reduce base-generated topics often. Moved
topics were the most difficult type and acquired late. Only the 6-years-olds had the concept of moved topics. However, their production rate was still low.
In addition to the developmental sequence of the acquisition of the three types of topics, it was found that the subjects tended to reduce sentence-initial topics, instead of redundant topics in main clauses. According to Chien and Lust (1983), their subjects reduced sentence-initial topics because they were sensitive to the pragmatic constraint that topic usually cannot be mentioned more than once. Thus, the findings had showed that their subjects already have the pragmatic concept.
However, there are some problems with the pilot study. First, only one task was employed in the pilot study. The story retelling task can only be used to observe the production of topics by children. Children’s comprehension of the three types of the topics was not examined. Thus, a comprehension task is required to prove the effectiveness of the result found in the story-retelling task. Second, the subject pool of the pilot study was too small. Twelve subjects were not enough to validate the developmental sequence of Chinese topics. More children should be recruited to reach a convincing generation. Due to the problems with the pilot study, the formal study included a comprehension task and recruited more subjects to participate in the tasks.
3.3.2 Formal Study
The formal study adopted two tasks: an elicited imitation task and a sentence
selection task. The procedure and the scoring of each task are introduced below.
3.3.2.1 The Elicited Imitation Task
The formal elicited imitation task was preceded by a set of training sentences, which allowed the participants to get used to the task. First, the experimenter played a game with the subjects: “I will tell you some sentences. Can you say what I say?”
Then, two sentences were said one by one. After one sentence was said, the child was asked to repeat it. The experimenter encouraged the subjects with guided words when they forgot the sentence. The sentences ensured that children expect to be asked to repeat sentences of similar length and complexity as the experimental test sentences.
After the training sentences, the formal task was conducted. The subjects were asked to repeat the sentence, as they did for training sentences. However, no intervention was made in the formal task. The experimenter gave the subjects enough time to think and produce the sentences. When the subject repeated only part of a sentence, this was considered as a response, and the experimenter asked, “Is that all?”
For the scoring criteria, one point was given if the children produced the sentence-initial topic, as (3a) shows. If the topic was omitted or appeared in a different position, as illustrated by (3b), no point was given.
(3) a. Shuiguo, mama mai le (shuiguo de) xiangjiao.
fruit mother buy PFV (fruit DE) banana ‘Fruits, Mother bought bananas’
b. *t, Mama mai le shuiguo de xiangjiao.
mother buy PFV fruit DE banana
As for topic chains, one point was given if the children correctly produced a topic chain, as (4) shows.
(4) a. Suiguo, xiangjiao hen pianyi, putao hen haochi.
fruit banana very cheap grape very delicious
‘As for fruits, bananas are very cheap and grapes are very delicious.’
b. *xianjiao hen pianyi, putao hen haochi.
banana very cheap grape very delicious
c. *Suiguo, xiangjiao hen pianyi, suiguo de putao hen haochi.
fruit banana very cheap fruit DE grape very delicious As shown in (4a), a topic chain consists of a topic and the subsequent comment sentences. If the children did not produce the sentence-initial topic as (4b) or did not apply the Topic-NP deletion as (4c), no point was given.
3.3.2.2 The Sentence Selection Task
The formal sentence selection task was preceded by a set of training sentences.
The experimenter played a game with the subjects: “Here are two animals, Lion and Dog. They sometimes say nonsense sentences, but I don’t know who says the correct one. Now, please listen to their sentences, and tell me whose sentence sounds better to you.” The subjects listened to two pairs of sentences from the puppets. The training ensured that children knew how to select the correct sentence.
In the SS task, each participant was given sixteen background scenarios in a random order. Then two puppets provided sentences for the subjects to choose according to the background scenario.
For the scoring criteria, one point was given if the subjects selected the puppet that said sentences in right order, as in (5):
(5) Shuiguo, mama mai le xiangjiao.
fruits mother buy PFV banana
‘Fruits, Mother bought bananas.’
No point was given if the subjects chose the puppet that said sentences in wrong word order, as in (6):
(6) *Mama, shuiguo mai le xianjiao.
mother fruit buy PFV banana
3.4 Results
As mentioned in Chapter 2, Chinese topics can be classified into different types according to their linguistic properties. In this section, the results of the SS task are reported in section 3.4.1 and the EI task in section 3.4.2.
3.4.1 The Sentence Selection Task
The sentence selection task was a comprehension task. In the SS task, the subjects were given two sentences to choose, along with one picture and one stimuli question. If the subject chose the correct sentence, which meant s/he could comprehend the meaning of the topic in the given sentence, then one point would be given for the test item. However, if the subject failed to choose the correct sentence, no point would be given. Then, the subjects’ mean scores were calculated and one-way ANOVA and the Scheffe analysis were applied to further analyze the mean scores.
3.4.1.1 Semantic Topics
The subjects’ responses to generic topics and definite topics are discussed in this
section. Table 3-7 shows how the subjects comprehended generic topics in the SS
task.
Table 3-7: Subjects’ Comprehension of Generic Topics
Group 1 Group 2 Group 3 Control Group
Mean SD Mean SD Mean SD Mean SD
0.50 0.21 0.59 0.16 0.68 0.16 0.93 0.11
As can be seen in Table 3-7, the mean scores of the correct responses by Group 3 and Group 2 were higher than the mean scores obtained from Group 1. One-way ANOVA indicated that all the three groups and the control group performed significantly differently (F(3,73)=27.089, p=0.000), which meant no subjects have reached the adult’s proficiency level. In addition, Group 1 and Group 2 (F(3,73)=27.089, p=0.359) did not respond to generic topics significantly differently;
neither did Group 2 and Group 3 (F(3,73)=27.089, p=0.368). However, significant difference was found between Group 1 and Group 3 (F(3,73)=27.089, p=0.009).
With regards to definite topics, Table 3-8 presents the results of definite topics in the SS task.
Table 3-8: Subjects’ Comprehension of Definite Topics
Group 1 Group 2 Group 3 Control Group
Mean SD Mean SD Mean SD Mean SD
0.51 0.18 0.65 0.19 0.71 0.15 0.94 0.82
As Table 3-8 shows, Group 2 and Group 3 also outperformed Group 1, which was similar to the results of generic topics. One-way ANOVA indicated that all the three groups and the control group responded significantly differently (F(3,73)=26.586, p=0.000), suggesting that no subjects have attained adults’ proficiency. In addition,
the difference between Group 1 and Group 2 in response to definite topics was
significant (F(3,73)=26.586, p=0.050). However, no significant difference was
found between Group 2 and Group 3 (F(3,73)=26.586, p=0.760), implying that a
developmental gap between Group 1 and Group 2 while Group 2 and Group 3 were more or less at the same proficiency level.
To sum up, the 6-years-olds have not achieved the adults’ level of proficiency in comprehending both generic topics and definite topics. Moreover, the 4-year-olds, 5-year-olds, and 6-year-olds performed similarly in comprehending generic topics.
As for definite topics, a developmental gap was found between ages 4 and 5.
3.4.1.2 Syntactic Topics
This section discusses how our subjects comprehended different syntactic topics, including base-generated topics, possessive-like topics, and moved topics. Table 3-9 shows the subjects’ responses to base-generated topics in the SS task.
Table 3-9: Subjects’ Comprehension of Base-generated Topics
Group 1 Group 2 Group 3 Control Group
Mean SD Mean SD Mean SD Mean SD
0.54 0.22 0.60 0.26 0.68 0.21 0.98 0.08
As can be seen in Table 3-9, Group 2 and Group 3 performed better than Group 1.
One-way ANOVA showed that there was a significant difference between Group 3 and the control group (F(3,73)=17.993, p=0.001). However, no significant difference was found among the three experimental groups (Group 1 and Group 2:
F(3,73)=17.993, p=0.886; Group 2 and Group 3: F(3,73)=17.993, p=0.732, Group 1 and Group 3: F(3,73)=17.993, p=0.244). Thus, the three age groups had similar performance on base-generated topics, and none of them has acquired the adult grammar.
Table 3-10 shows how the subjects comprehended possessive-like topics.
Table 3-10: Subjects’ Comprehension of Possessive-like Topics
Group 1 Group 2 Group 3 Control Group
Mean SD Mean SD Mean SD Mean SD
0.54 0.17 0.68 0.20 0.85 0.18 0.99 0.06
As can be seen in Table 3-10, Group 3 had most correct responses to possessive-like topics and there was no significant difference between Group 3 and the control group (F(3,73)=30.248, p=0.099), suggesting Group 3 should have acquired the adult grammar of possessive-like topics. Group 1 and Group 2 did not perform significantly differently (F(3,73)=30.248, p=0.069), but a significant difference was found between Group 2 and Group 3 in responding to possessive-like topics (F(3,73)=30.248, p=0.014), showing that there was a developmental gap between ages 5 and 6 in their comprehension of possessive-like topics.
Table 3-11 reports the subjects’ comprehension of moved topics.
Table 3-11: Subjects’ Comprehension of Moved Topics
Group 1 Group 2 Group 3 Control Group
Mean SD Mean SD Mean SD Mean SD
0.44 0.27 0.59 0.25 0.55 0.19 0.89 0.21
As Table 3-11 shows, Group 2 had slightly higher mean scores among the three
experimental groups, which was different from the results of base-generated topics
and possessive-like topics discussed earlier. Though Group 2 received the highest
mean score, no significant difference was found between Group 2 and Group 3
(F(3,73)=11,285, p=0.986). Moreover, one-way ANOVA showed that the gap
between Group 2 and the control group was significant (F(3,73)=11,285, p=0.007),
indicating that no subjects of this group have indeed acquired the adult grammar of
moved topics. However, the difference in the responses by the three experimental
groups was not significant (Group 1 and Group 2: F(3,73)=11,285, p=0.247; Group 2 and Group 3: F(3,73)=11,285, p=0.986; Group 1 and Group 3: F(3,73)=11,285, p=0.473), showing that all the subjects performed quite similarly on moved topics in
the comprehension task.
To sum up, the 6-year-olds have achieved the adult’s level of proficiency in comprehending possessive-like topics, but not base-generated topics and moved topics. All the subjects had similar performance on base-generated topics and moved topics, but not possessive-like topics. A developmental gap was found between ages 5 and 6 in the subjects’ comprehension of possessive-like topics.
3.4.1.3 Topics Chains
In this section, we will discuss how our subjects comprehended topic chains.
Table 3-12 shows the subjects’ comprehension of topic chains.
Table 3-12: Subjects’ Comprehension of Topic Chains
Group 1 Group 2 Group 3 Control Group
Mean SD Mean SD Mean SD Mean SD
0.50 0.23 0.57 0.22 0.57 0.17 0.92 0.14
As can be seen in Table 3-12, Group 2 and Group 3 outperformed Group 1. The difference between Group 3 and the control group was significant (F(3,76)=19.008, p=0.000). Therefore, Group 3 has not reached the adult’s level of proficiency. In
addition, the difference among the three groups was not significant (Group 1 and Group 2: F(3,76)=19.008, p=0.675; Group 2 and Group 3: F(3,76)=19.008, p=1.000;
Group 1 and Group 3: F(3,76)=19.008, p=0.757), indicating the three groups were at a similar stage of development.
In sum, the 6-year-olds have not reached the adult’s level of proficiency in the
comprehension of topic chains. And all the three groups had similar performance on
comprehending topic chains.
3.4.2 The Elicited Imitation Task
The EI task, as mentioned earlier, was a production task. In the EI task, for each test item, the subjects were given one sentence to imitate. If the subject produced the semantic or syntactic topics when s/he imitated, one point would be given. However, if the subject omitted the topics in his/her production, no point would be given. As for topic chains, if the subject correctly produced the topic chains ‘topic + zero topic,’ one point would be given. The production of the subjects who suffered from the imitation effect was excluded for data analysis in the scoring
1. Then, the subjects’ mean score of correct production were calculated and analyzed by one-way ANONVA and the post hoc Scheffe analysis.
3.4.2.1 Semantic Topics
The subjects’ performance on generic topics and syntactic topic is presented in this section. The results of their production of generic topics are shown in Table 3-13.
Table 3-13: Subjects’ Production of Generic Topics
Group 1 Group 2 Group 3 Control Group
Mean SD Mean SD Mean SD Mean SD
0.42 0.23 0.59 0.21 0.69 0.15 0.81 0.19
Generally speaking, Table 3-13 shows that Group 2 and Group 3 outperformed Group 1. The mean scores of the correct production of Group 2 and Group 3 were higher than that of Group 1. One-way ANOVA showed that Group 3 and the adult controls
1 There were three subjects of Group 3 imitated the experimenter’s sentences. They production was the same as the input provided in the elicited production task.