Procedure

The subjects were tested in the production task before the comprehension task.

During both production and comprehension tasks, regardless of how many test items have done, a 3-minute break is given every 20 minutes to avoid possible noisy effects brought by tiredness.

In the production task, each child participant and the experimenter sit on one side of the table looking at the computer screen. The helper, to whom the child needed to describe the picture, sit on the other side of the table and wore a pair of ear covers so that s/he was unable to see the computer screen and hear the test items. In front of the helper, there were four pictures for the selection: one correct picture that matched the expected answer and three incorrect pictures. Of the three incorrect pictures, one picture had an action that was different from the one mentioned in the matrix clause of the expected answer. Another incorrect picture had an action that was different

43

from the one mentioned in the CC of the expected answer. The other incorrect picture includes a character that was different from the one mentioned in the CC of the expected answer. These pictures were designed for making sure that the subjects would generate complete complex sentences.

After the elicited question of each test item was given, the child had to answer the question by describing the picture shown in the computer screen in order to assist the helper to choose the identical picture with the one shown in the computer screen.

The child would tell the helper whether s/he had selected the correct one. The experiment began with practice trials to help the child become familiar with the content they were going to participate in, and then the main session that contained experimental trials and filler trials began. The whole experiment was recorded and transcribed for the analysis.

In the comprehension task, the subjects had to decide whether the picture presented in the computer screen matched with the statement broadcast by the computer. The child subjects were told that if the statement matched with the picture, then they should judge the picture as ‘true.’ On the other hand, if they thought that the statement did not match with the picture, then they should judge the picture as ‘false.’

When children thought the statement was wrong, they would have to provide an explanation to justify their answers. The scenarios of both tasks are given in Appendix V.

For the sake of assuring the formal study of a feasible research design, a pilot study was conducted. The relevant reports of the pilot study are presented in the following section.

2.3.3.1 Pilot Study

9 child subjects, evenly divided into three age groups (3;10, 5;8 and 7;5),

44

participated in the production and comprehension tasks. When collecting the data, we discovered that the attention span of the child group aged 3;10 was shorter than the period they needed for completing the production task. Though a break was then provided, it was difficult to get some of them back to the task. One child of this group (age: 3;11) did not seem to understand what she had to do in the comprehension task and responded to almost all test items with ‘true.’ On the contrary, the other two children of the group (age: 3;8 and 3;10) could correctly judge 55% of the CC complex sentences. In the production task, the collected data showed that children of this age were able to answer to questions given by the trials although the response might only consist of one clause rather than a complex sentence. These observations made us decide not to recruit children under 3 years old since they might have even greater difficulties participating in the tasks. As for the oldest children, that is, the 7-year-olds, they were able to correctly comprehend 83% of the test items and accurately produce 69% of the test items. Although it seemed that 7-year-olds have not yet completely acquired complex sentences with CCs, we would temporarily focus on the CC development until this age and leave the acquisition of older children for further studies because the current results were able to reveal a tendency in response to the research questions. Additionally, we also found that in the comprehension task some of the younger children had difficulties using the terms,

‘true’ and ‘false,’ yet they could easily judge if the utterance they heard was ‘the same as’ or ‘different from’ the given picture. Consequently, in the formal study, if a subject was not able to correctly judge at least half of the practice trials by saying

‘true’ or ‘false,’ they would be asked to decide again whether the pictures and the utterances in the practice trials were ‘the same’ or ‘different.’ Regarding the materials, all the vocabulary used in the study has already existed in children’s mental lexicon,

45

yet a few actions depicted in the pictures needed to be modified because they misled the subjects to judge a sentence with other messages brought by the actions. The revised pictures were then used in the formal study.

2.3.3.2 Formal Study

Since the procedures were implemented and proved to be unproblematic in the pilot study, the formal study adopted the same procedures for investigating the CC acquisition. For the adult group, the procedure was similar to that in the child group.

Nevertheless, in the production task, there was no helper in the experiment with adults.

The adult subjects were merely asked to describe the picture corresponding to the elicited question to the experimenter.

2.3.4 Data Analysis