2.3 First Language Acquisition of Modal Verbs
2.3.4 Moscati et al. (2017)
Some researchers (Well 1979, Palmer 1986, Noveck 2001, Chen & Ma 2017) have discussed about the acquisition of different meanings of modal verbs for a long time.
For example, Palmer (1986) states that 3 years old children have been already able to use the dynamic, deontic and epistemic meanings. However, Noveck (2001) argues that children are still confused about the epistemic modals when they are 5 years old. Thus, Moscati et al. (2017) investigated children’s utterances of the epistemic modal verbs might and must. In their study, a total of 14 child subjects aged from 4;6 to 5;6 were
30
recruited together with 16 adults as the control group.
The experiment (i.e., Mysterious Box) consisted of four steps. In Step 1, the participants were asked to have a look at the animals in green boxes. In Step 2, the green boxes were closed so that the participants could not see the animals. In Step 3, the participants were asked to guess the animal in an orange box by producing an utterance
with a modal might or may. Finally, in Step 4, all the boxes were opened, and the participants could check whether their answers matched the test items or not. Figure 2-3 shows the Mysterious Box used in Moscati et al.’s study:
Figure 2-3 The Mysterious Box (Moscati et al. 2017:1031)
During the whole process, the participants were wearing a visual world paradigm, which recorded their eye ball movements10 . After they finished the experiments, Moscati et al. (2017) analyzed the participants’ data with three different focuses for each sentence. Firstly, they analyzed the beginning of the sentence and the end of the modal (i.e., A monkey). Second, the modal and the main verb (i.e., must/might be) were
10 All the pictures have been divided into 4 interest areas (IAs): 1. An orange box, 2. A green box with mentioned animals, 3. A green box with unmentioned animals, and 4. A plate with a fruit.
31
examined. Third, the rest of the sentence like the preposition or object (i.e., in the orange box) was scrutinized.
Overall, Moscati et al. (2017) found that the participants could distinguish and manipulate the modal verbs might and must. The percentages for the two modal verbs were 86.9 % for might and 97.6 % for must. However, some of the participants were
confused with the game. For example, they did not know which modal was suitable to use. In the eye tracking experiment, Moscati et al. (2017) found that both groups of the participant (i.e., children and adults) had different results. For instance, the adults were more careful about producing the preposition and the object (i.e., in the orange box).
On the contrary, some younger children could produce sentences with modal verbs, but they seemed unsure about the meanings of the modals might and must.
To sum up, Moscati et al. (2017) concluded that the participants had ability to produce sentences with epistemic modals, since most of them performed as well as the adults. Moreover, they claimed that the 5- year-old children already could master the epistemic modal verbs.
Nonetheless, there are still some limitations of Moscati et. al.’s (2017) study. First, they did not explain the age effect on the acquisition of modals clearly. Moreover, different researchers might classify a modal verb differently. For example, the modal must is said to have both the meanings of deontic and epistemic in previous studies
32
(Lyon1977, Palmer1986). Moscati et al. (2017) mainly used it as an epistemic meaning in their research. Therefore, a study on children’s acquisition of the deontic meaning can be further conducted.
2.3.5 Summary of the Previous Empirical Studies
The major findings and limitations of the empirical studies (Modyanova et al.
2010, Gaidagri 2014, Wang 2014, and Moscati et al. 2017) reviewed in Section 2.4:
Table 2-1 Major Findings and the Limitations of the Previous Empirical Studies
Major Findings Limitations
Dynamic hui (3 yrs old), Deontic hui (5 yrs old), Epistemic hui (8 yrs old)
33
Wang (2014) and Gaidagri (2014) were more concerned about the production of modal verbs. They both conducted a qualitative longitudinal study to investigate children’s
utterances of modal verbs. By contrast, Modyanova et al. (2010) and Moscati et al.
(2017) concentrated on the study of children’s interpretations of modal verbs. It was found in these four studies that the interpretation and production of older children were better than those of younger ones. Moreover, Modyanova et al.’s (2010) and Wang’s (2014) findings indicated that children’s acquisition order was dynamic> deontic>
epistemic.
However, these four studies had different findings about the age effect. For example, Modyanova et al. (2010) and Wang (2014) found that the children acquired the dynamic, deontic and epistemic meanings when they were 3-, 5- and 8-year-old. On the contrary, Gaidagri (2014) and Moscati et al. (2017) concluded that their children could acquire the epistemic modal at the early ages (i.e., 2-year-old and 5-year-old, respectively).
Nevertheless, there are some limitations of these four studies. Firstly, the number of their participants was small (Modyanova et al. 2010, Wang 2014, Gaidagri 2014).
34
Moreover, some of them (Modyanova et al. 2010, Moscati et al. 2017) did not concern the different meanings of modal verbs in their studies. Lastly, many of them (cf.
Modyanova et al. 2010, Wang 2014, Gaidagri 2014) only conducted one task in their experiments.
2.4. Summary of Chapter Two
In Chapter Two, we have introduced the theoretical and empirical studies of modal verbs in English and in Chinese. The Chinese modal hui can have five basic meanings:
the dynamic hui, the deontic hui, the epistemic hui, the generic hui and the futurity hui.
Moreover, the development of dynamic > deontic > epistemic is the most convincing one according to previous studies (i.e., Guo 1995, Wu 2009, Wang 2014).
With regard to the empirical studies, we have found that children acquired the dynamic meaning before the deontic and epistemic ones. They not only could acquire the dynamic hui with ease, but also mastered it in the early ages. However, most of the children were found to get confused or have difficulties in acquiring the deontic and epistemic modals. In addition, it was found that children acquired the deontic and epistemic meanings at the ages of 5 and 6; however, some of the researchers found that children acquired these meanings after 8. In the following chapter, the research design will be introduced to investigate these issues.
35
Chapter Three Research Design
In this chapter, the research design of the present study is discussed. Section 3.1 introduces our subjects, methods and materials, and procedures of the experimental study. The background of the subjects and research procedures of the corpus study are
presented in Section 3.2. Finally, a summary of this chapter is provided in Section 3.3.
3.1 Study I: Interpretation Task (Truth Value Judgement)
In language acquisition, a truth value judgement (TVJ) task has been commonly used to test whether children can interpret the meanings of a target sentence (cf. Crain
& McKee 1985). In such a task, researchers may provide toys, pictures, or videos to ask subjects questions, and the subjects need to respond with ‘True’ or ‘False.’ In this way, data can be quickly and easily analyzed (Eisele and Lust 1996). Therefore, Hirst
& Weil (1982), Noveck (2001), and Hsu (2011) used a TVJ task to investigate children’s modal acquisition and found that their subjects acquired modal verbs successfully by the age of 6- to 8-year-old. Hence, the present study designed a TVJ task to examine children’s interpretation of the Chinese modal hui.
36
3.1.1 Subjects
In their studies of the Chinese modal hui (Guo 1995 and Wu 2014), some researchers recruited children aged 3-, 5- and 8-year old as their subjects and argued that children acquired the meanings of the dynamic hui at the age of 3, the deontic hui at the age of 5, and the epistemic hui at the age of 8. Nevertheless, they could not claim
that 4-, 6- and 7-year-olds might have acquired these three meanings of hui since they skipped these age groups. In the study of English modals, Noveck (2001) and Moscati et al. (2017) invited children aged 4- and 5-year-old as their experimental subjects and found that some of their subjects neither comprehended test items nor answered questions correctly.
Hence, we recruited the subjects aged from 3 to 6 years old to see if they could interpret the Chinese modal hui11. The number of the subjects in the experimental study (ES) was sixty-four Taiwanese children in kindergartens in Tainan, and sixteen adults were also recruited as the control group, as can be seen below:
11 Seven- and eight-year-old subjects will not be included in the present study since the results of the pilot study showed that these subjects had already acquired all the meanings of hui by the age of 6.
37
Table 3-1 Background Information of the Subjects in the Experimental Study
Group Age Range Number
ES-1 3-year-olds 3;0~3;11 16
ES-2 4-year-olds 4;0~4;11 16
ES-3 5-year-olds 5;0~5;11 16
ES-4 6-year-olds 6;0~6;11 16
ES-Controls 19-year-olds 19;0 or above 16
As Table 3-1 shows, there were five age groups of the experimental subjects, and each group consisted of 16 subjects. All the subjects had no physical problems, such as deaf, mute, blind, etc.
3.1.2 Methods and Materials
In the literature of language acquisition, there are two main approaches to collect data: quantitative and qualitative (Larsen-Freeman & Long 1991). A qualitative approach can be used to collect subjects’ utterances, sounds or speeches during a long period of time (Janet Holland, Rachel Thomson and Sheila Henderson 2016). However, this approach usually takes researchers a lot of time and effort (Hiroshi 1976). On the contrary, a quantitative approach not only provides us with an opportunity to collect data in a short period of time, but also helps us analyze data with ease (Hopkins 2000).
38
Hence, the present study followed Moscati et al. (2017) to conduct a quantitative study with an experiment consisting of truth value judgement questions to investigate children’s modal acquisition. To avoid ambiguity and misunderstanding, conversations in stories with pictures were provided for each test item.
In the literature, some researchers like Liu (2009) and Hsu (2011) have conducted
syntactic and pragmatic studies of the Chinese modal verbs via a quantitative approach;
however, they did not test Chinese children’s interpretation of the modal hui. Regarding the meanings of hui, we examined the five meanings of it, as discussed in Chapter Two:
the dynamic hui, the deontic hui, the epistemic hui, the futurity hui and the generic hui.
Four test items were designed for each modal meaning, yielding a total of 20 test items, as shown below.
39
Table 3-2 The Test Design of the Chinese Modal Hui in the Experimental Study Type Modal
40
As can be seen in Table 3-2, the 20 test items embedded in two stories, each of which had an introduction, a trial question, and pictures for the 10 test items along with True/False questions. Table 3-3 illustrates a sample question in the TVJ task:
Table 3-3 A Sample Question of the TVJ Task Scene 1
The subject saw:
The subject heard:
Xiao Shi chidao le.
Baby Lion is late.
Laoshi: Bu keyi chidao, zhidao ma?' Teacher: "Don’t be late, OK?"
Xiao Shi: Hao, wo hui tinghua de.' Baby Lion: "OK, I'll be obedient."
Scene 2
The subject saw:
The subject heard:
Qingwen Xiao Shi daying laoshi guaiguai tinghui, dui-bu-dui?
‘Baby Lion promises the teacher that he will be obedient, doesn’t he?’
For a complete version of the TVJ task in the experimental study, please refer to Appendix A. It is expected that with the findings of our experiment we can see which
41
meaning of hui is easier for children to interpret, and whether they acquire these meanings at different ages.
3.1.3 Procedure 3.1.3.1 Pilot Study
A pilot study was conducted on December 26, 2017 to ensure the quality of the
research design and test items. There were 10 True/False questions designed in the TVJ task (See Appendix B) for the five meanings of hui. The subjects were 15 Chinese students from Ming Dao Elementary School in Taipei (5 students each for the age group:
8-, 10- and 12-year-old), and 5 adults as the control group.
The results of the TVJ task showed that different age groups interpreted the Chinese modal hui differently. Of the five meanings of hui, there was no clear pattern shared by the three age groups regarding futurity hui and generic hui. As for the other three meanings of hui, the 8- and 12-year-old subjects scored higher on the dynamic hui and the deontic hui than on the epistemic hui, i.e., dynamic > deontic > epistemic
(from high to low). Their responses were close to those of the adult group. This order supports the arguments of Lyons (1977) and Palmer (1986) that the acquisition order of modal meanings is dynamic > deontic > epistemic. Nevertheless, the order for the 10-year-old subjects was deontic > dynamic > epistemic, which was not found in the previous literature. There are some limitations of the pilot study. For example, the
42
subject pool was small, and the test items were not enough. As a result, the findings might not be reliable; therefore, it is necessary to conduct a further study to investigate children’s acquisition of the modal hui.
3.1.3.2 Formal Study
In order to improve the reliability and validity of the present research, more
subjects were recruited, and more test items were included to investigate the acquisition order and the meanings of hui among different age groups.
All the child subjects in the formal study were selected from kindergartens in Tainan, and the adult participants were recruited from National Taiwan Normal University. Since our subjects were kindergarteners, it is rather important to provide a consent form (see Appendix C) and receive approvals from their parents. This form reported the research motivation, expectation and details of the present study. After the parents and subjects signed the consent form, we kept the subjects’ personal information confidential and started to conduct the experiment (i.e., the TVJ task) individually.
As mentioned in Section 3.1.2, there were two stories in the TVJ task, and each story had an introduction and a trial question, together with 10 pictures and 10 T/F questions. For example, the subjects were shown an introduction as in (1) before Story I was told:
43
(1) Yi tian you yi tian, Ahe manmandi changda le. Tutu gege xiwang Ahe xuehui zhaogu ziji, suoyi Tutu gege gei Ahe henduo renwu. Women yiqilai bang Ahe wancheng renwu, hao ma?
‘Day after day, Ahe has gradually grown up. Brother Tutu hopes that Ahe can take care of himself. So, he gives Ahe many tasks. Let’s help Ahe to complete the tasks, OK?’
After the introduction, we started our trial question to ensure our subjects understood what they were supposed to do during the task. Table 3-4 presents the trial question of
Story 1:
Table 3-4 The Trial Question of Story 1 Scene 1
The subject saw:
The subject heard:
Ta hui tiaosheng.
‘He can jump rope.’
Scene 2
The subject heard:
Nai chang tu, dui-bu-dui?
‘Is that picture correct?’
44
After the subjects finished the trial question, we began with Story I, which consisted of scenarios and conversations, and they judged the test items by answering True or False.
All their answers were recorded by the experimenter. This task took each subject approximately 30 minutes to complete.
3.1.3.3 Scoring and Data Analysis
As we discussed in Section 3.1.3.2, there were 20 True/ False questions in the TVJ task. If the subjects answered a question correctly, one point was given. However, if their answer was wrong, they received zero point. The scores were keyed into an Excel file and later ran by an R-Studio software, using the ANOVA to analyze the data. By this function, we obtained the analytic statistics and values of each group. It is to see if children’s interpretation of hui varies among different age groups, and to explore children’s acquisition order of hui within each age group (Tang 1979, Guo 1994, Huang 1999, Chang 2001).
3.2 Study II: Corpus Study
One of the convenient platforms for Taiwanese researchers (i.e., Huang et al. 2000, Lee 2015) to collect data is the Academia Sinica Balanced Corpus of Modern Chinese.
However, there are some restrictions on this corpus. First, factors like gender, age or education backgrounds are not shown. Second, there are no Chinese children’s token
45
or utterance. Therefore, the present study did not consider the Sinica Corpus because it cannot provide any utterance of the Chinese modal hui produced by Chinese children.
Consequently, the present study used the Chinese Corpora – CHILDES, which is a multilingual corpus established by Carnegie Mellon University and the University of Pennsylvania with the grant from the National Science Foundation in 1999. It contains
a wild range of linguistic fields, such as language acquisition, discourse, sociolinguistics, and so on. The Chinese Corpora –CHILDES is one of the important corpora, which contains children’s utterances produced by 728 child subjects collected by one Taiwanese, seven Chinese, and three Cantonese researchers. The child subjects are 158 Cantonese, 566 Chinese and 4 Taiwanese children. Hence, the present study obtained the utterances from The Chinese Corpora –CHILDES to investigate children’s production of the Chinese modal hui.
3.2.1 Subjects
As we mentioned, the present study conducted an experimental study with an interpretation task and a corpus study. To our knowledge, few researchers have conducted a language acquisition study of hui with a corpus approach, although Chen and Ma (2017) have employed a study of the language acquisition of the Chinese modal verbs like hui ‘will,’ keyi ‘can,’ xiang ‘want,’ etc. However, they investigated 1- and 2-year-old children and found few tokens of hui produced by the subjects. Apparently,
46
their findings were not sufficient to support their claim that hui should have three meanings. Therefore, the present study collected the utterances of hui from 3-, 4-, 5-, and 6-years-old subjects in the Chinese Corpora –CHILDES, as Table 3-5.
Table 3-5 Background Information of the Subjects in the Corpus Study
Group Age Range Number
CS-1 3-year-olds 3;0~3;9 16
CS-2 4-year-olds 4;1~4;8 16
CS-3 5-year-olds 5;7 ~5;10 16
CS-4 6-year-olds 6;2~ 6;7 16
As Table 3-5 shows, each age group consisted of sixteen children. (i.e., children in the group CS-1 were all 3-year-old children). They were kindergarteners interviewed individually in the library at school in Hsinchu, and Chinese is their native language.
3.2.2 Procedure
In the formal study, we collected the utterances from the 3-, 4- , 5-, 6- year-old children and the adults in the collection of Chang 1 and Chang 2 in the Chinese Corpora --- CHILDS. In this collection, we randomly selected 80 subjects, and each age group included 16 subjects. All the utterances of the subjects were checked to see if they contain the modal hui. If yes, we underlined the whole utterance, as can be seen below:
47
(2) EXP: Senlin hui fasheng shenme shiqing?
forest HUI happen what thing
‘What will happen in the forest?’
EXP: Senlin hui zenyang?
forest HUI how
‘What will the forest become?’
CHI: Hui you yige. Laohu.
HUI have a tiger
‘There will be a… tiger.’ (Chang, 1998)
The two raters who helped with the coding of the pilot study were invited to evaluate the meanings of hui according to the present classifications, as illustrated in Table 3-2. They mainly focused on the utterances of hui produced by the children in the dialogues selected by the researcher and decided on the type of its meaning. If the two raters disagreed on the coding, a third rater was asked to take another look at the utterances. After the coding was completed, we counted the frequency of each meaning of hui, and further keyed in all the types of hui to an Excel file. We later ran an analytic software R-studio, using Chi-square to calculate children’s utterances of hui within each age group and each meaning of hui to see if children acquire the meanings of the modal hui at different ages. We also saw if the acquisition order found in the corpus supports the previous literature (Guo 1995, Noveck 2001, Moscati et al. 2017).