3.3 Procedures
This section can be divided into three parts. Section 3.1 summarizes the major findings and the inadequacies of the pilot study. In Section 3.3.2, procedures of the formal study, revised from the pilot study, are shown. In Section 3.3.3, the scoring and statistical analysis are introduced.
3.3.1 Pilot Study
A pilot study was conducted to investigate children’s acquisition pattern of Mandarin Chinese conditionals. The subjects consisted of thirty kindergarten students in the experimental groups (4-year-olds, 5-year-olds and 6-years-olds), with ten in each group, and ten adults in the control group.
Conditional sentences, based on Wu (1996), were categorized into five types1. The subjects participated in the comprehension task in which eighteen causal scenarios were included with different types of conditionals (5 types × 3=15) and three filler sentences. They were told that two characters in the scenarios often argued who was smarter and children had to judge the acceptability of the sentence.
The findings of the pilot study are as follows. First, each age group got almost 90% of correct responses to different types of conditionals in the task, indicating that age 4 was critical in our children’s acquisition of conditionals just like the previous
1 Type 1: Ruguo..,…jiu (if…,then…); Type 2: Ruguo…dehua,…jiu (if…, then); Type 3: …,…jiu (…, then…); Type 4: …, (pause)…. ; Type 5: negation in the front, back or in both the two positions.
54
research on English conditionals. Our Chinese children, like native English speakers, could comprehend conditionals at age 4.
Second, Type 5 (negation) in the experimental groups and in the control group was scored the lowest even though the result was not significant compared to the control group. The phenomenon showed that our subjects were not familiar with the use of negatives in conditionals since such a property was related to logical thinking.
Nevertheless, our subjects’ performance of Type 4 (i.e., non-wh-word conditionals) was noteworthy since they showed a higher percentage of production. Although conditionals without conditional adverbials are not as frequently found as Type 1 in the corpus, these sentences were still acceptable to our children. The reason why Type 4 was acceptable to the subjects might be related to the pause between two clauses which could be a contextual clue for them to infer conditionals, as stated in Li and Thompson (1981).
However, there are some inadequacies of the pilot study. First, the subject pool (30 subjects) was small; therefore, the pilot study did not yield any significant differences among different age groups and conditional types. Second, the study only focused on ‘counterfactual’ conditionals while reality and hypothetical conditionals in Mandarin Chinese were neglected. Thus, scenario differences are required to a generalization regarding the acquisition of Chinese conditionals. Third, a production task was not included while it is necessary if we would like to probe into the acquisition pattern of Chinese conditionals.
3.3.2 Formal Study
In light of the inadequacies found in the pilot study, the formal study employed a comprehension task and a production task to examine the five types of Mandarin Chinese conditionals in the hypothetical and counterfactual scenarios.
55
First, a consent form was distributed to the administration office of the kindergartens and the elementary schools for school officials to give to the subjects’
parents2. The parents were informed that all the information and data collected in the study would be used for the academic research only and would be kept confidential.
Before the experiment, the subjects were told that the two tasks would not be related to the subjects’ performance in school. To eschew the potential practice effect, the subjects were notified to do the production task (i.e., the IM task) first followed by the comprehension task (i.e., the IT task).
Before the IM task, a trial section was conducted to train our subjects to follow the procedures of the task. After the experimenter assured that the subjects were familiar with the procedures of the imitation task, the experimenter introduced the two characters in the story.
In the IM task, the story was recorded and shown on a computer screen. In the story line, the tonic emphasis fell on key words about the adverbial conditionals of Mandarin Chinese conditionals. The subjects, playing the role of the younger rabbit, responded to 24 questions in the story in which conditional sentences and fillers were randomized in order. During the period, the experimenter did not provide them with any suggestions.
After the IM task, the IT task was employed on the same day or one week later.
A training section was also implemented. The subjects were told to listen to Mandarin Chinese resultative sentences and chose the meaning of the structures from two options on the screen. After they understood the answering procedures, the experimenter introduced the main characters used in the task to them. Then, conditional episodes were played and some emphases were on the specific words.
During the period, there was no sound interference or cue.
2 Please refer to Appendix D for the consent form.
56
The subjects completed the two tasks individually in their schools and the controls took the task at NTNU or other quiet places. All the subjects’ utterances were audio-recorded for subsequent analysis.
3.3.3 Scoring and Statistical Analysis
As mentioned in Section 3.2, the present study consists of two tasks: the IT task and the IM task. In the IT task, the subjects had to choose the meaning of a target utterance from two options. If their answer was correct, one point was given.
However, if they chose the wrong answer, no point was given.
In the IM task, the subjects had to repeat what the older brother said. One point was given if the subjects correctly imitated the target type of Mandarin Chinese conditional in the episode, as shown in (1):
(1) Type 2-2
Ruguo wo dai yinliao， jiu keyi he le if I bring drinks then can drink PAR ‘If I had brought the drinks, I could have drunk it.’
Nonetheless, modifiers and adverbials were not the main factors in the IM task.
Therefore, if the subjects said some similar modifiers and adverbials in their utterances such as gaoxingdi→ kuailedi ‘happily’, one point was still given. Then, 0.5 point was given if the subjects did not correctly imitate the controlled pattern required in the experimental episodes while their conditional utterances maintained the main content, as in (2):
(2) Type1-1
Ruguo wo you caise bi ，wo jiu yao hua caihong.
if I have color pen I then want draw rainbow ｀If I have a color pen, I can draw a rainbow.＇
57
However, no point was given if they produced other sentence constructions such as questions, temporal or causal sentences. Furthermore, partial imitation, irrelevant sentences or no responses were given zero point, as in (3):
(3) Type 5-1
Xingxing hui faliang.
star can shine ‘Stars can shine.’
All the subjects’ responses were coded according to the taxonomy discussed in Section 3.2, as shown below:
Table 3-5 The Coding System Used in the Present Study
If the subjects’ imitated responses were not the target type of conditionals in the episode (i.e., the utterances were given 0.5 or 0 point ), their responses were further categorized in the production analysis. In addition, our subjects’ non-conditional responses were analyzed to see whether they comprehended the conditional sentences presented in the story and what their performances suggest in our children’s conditional acquisition. To circumvent the effect of the experimenter, both the scoring of the IM task and the frequency counts of the production types were checked by one
Conditional types Types × scenarios Hypothetical: Type 1-1
Type 1
Double conditional adverbial Counterfactual: Type 1-2 Hypothetical: Type 2-1 Type 2
Clause-initial conditional adverbial Counterfactual: Type 2-2 Hypothetical: Type 3-1 Type 3
Clause-final conditional adverbial Counterfactual: Type 3-2 Hypothetical: Type 4-1 Type 4
Non-wh conditional Counterfactual: Type 4-2 Hypothetical: Type 5-1 Type 5
Wh-conditional Counterfactual: Type 5-2
58
of her graduate school classmates and the researcher.
After coding the data, all the responses were processed by SPSS (19th edition) package. The descriptive statistics procedure was first performed to yield the mean scores and standard deviation of the IT task and the IM task. The p values, which showed the significant differences among the types and features, were also examined in these two tasks by General Linear Model. In addition, one-way ANOVA was executed to examine the extent of age effects, differences between the hypothetical and counterfactual scenarios and interactions among the types in the IT task.