Pilot Study

In the pilot study, only five subtypes of English it-clefts, i.e., the types of subject NP, object NP, PP, ADV P, and ADJ P, were designed in the tasks, which differs from the formal study that classifies English it-clefts into fourteen subtypes. The subjects were fifteen students in total at one of the top male high schools in Taipei, and all of them were second graders in the gifted class of the social studies¹. The experiment was conducted in a self-study session, during which the students had 50 minutes to complete the tasks.

Due to the fact that most of the previous empirical studies put emphasis on either the perception ability (Carrell 1977) or the production ability (Callies & Keller 2008) of the subjects, the pilot study combined both of them in order to have an integrated exploration on the acquisition of the structure. A Grammaticality Judgment Task (GJ task) was designed as the perception task, and a Discourse Completion task was the production task. Among the fifteen students, seven of them did the GJ task, where they had to judge whether or not the sentences were correct. As for the test items, there were totally 24 questions involved². Each of the five subtypes of clefts consisted of four questions, along with four fillers (the complete instruction, test items, and fillers of the task are in Appendix C). On the other hand, the other eight students did the Discourse Completion Task, in which three turn takings of conversation were designed for each question. Regarding the test items of the DCT, there were four items for each subtype of cleft construction, as well as four fillers randomly laid among the test questions. In total, there were 24 questions in this task. Besides, two of the four questions of the same type showed contrast in the sentences, the other two not

1 All the students in the gifted class of the social studies were selected by a screening test designed by the school. There is only one gifted class of the social studies in one grade, with around 30 students in the class.

2 Most of the test items are modified from the examples sentences in Collins (1991).

(the complete instruction, test items, and fillers of the task are in Appendix D).

As for the scoring of the Grammaticality Judgment Task, one point was given when the answer was correct, and zero point for wrong answers. As for the Discourse Completion Task, there were four parts to focus on for each test sentence, as the segments between the slashes shown in (1).

(1) It is / cleaved XP / that/ cleft clause

Each segment was given 0.5 point, so there were totally 2 points for one test sentence.

A scoring classification of the possible answers based on the criteria shown in (1) is listed in Table 3-4 as follows.

Table 3-4 A Classification of Scoring

Answer Score Note

Ex. It is my father that does the laundry.

2

expected answer Ex. It is my father who does the laundry.

1. Ex. It is my father does the laundry. 1.5 The complementizer that is missing.

Ex. It is my father.

1

Two target segments are missing.

Ex. My father does the laundry.

Ex. My father. 0.5 Only one segment is correct.

Ex. The laundry is not done by my mom. 0 No target segment is found.

The results of the pilot study are shown herein (see Appendix E for the mean scores of each question). For the Grammaticality Judgment Task, a difficulty hierarchy of comprehending it-cleft was found as in (2). The type at the very left was the easiest for the subjects, and the one at the right was more difficult than the left, and so on.

(2) NP-Subject > PP > ADV P > NP-Object > ADJ P

As shown above, while the NP-Subject type was the easiest one to comprehend, the ADJ P type was the most difficult. As for the Discourse Completion task, another difficulty hierarchy was observed, as in (3).

(3) ADV P > NP-Object > NP-Subject > PP > ADJ P

The easiest type for the subjects to produce was ADV P, which was different from the comprehension task. However, the most difficult type was the same as the GJ task, i.e.

ADJ P. The reason why the hardest one in both tasks was ADJ P might be that the adjective phrase is not allowed to be cleaved in Chinese, or the frequency rate in the corpuses is low (i.e., 0.1%, Collins 1991).

There were some inadequacies in the pilot study. Firstly, our subjects performed much worse on certain questions of the Grammaticality Judgment Task, and thus a detailed investigation of the questions is required to see whether the modification of the questions is necessary. Secondly, the scores of the Discourse Completion Task were generally low for each type. This might be due to the fact that the tone expressed by the English it-cleft was not strong enough in the turn takings of the conversations to make the subjects use the construction, or, the English it-cleft itself was such a marked structure that only a few data could be elicited. Thirdly, not every type of English it-cleft had test items with contrast information, for instance, the incorrect test items in both the GJ task and the Discourse Completion task. Thus, the test items with contrast information should be designed as many as those without contrast information, so as to make a systematic investigation of the significance of contrast information in Chinese learners’ acquisition of English it-clefts. Fourth, the subject pool in the pilot study was not big enough to show a generalization. Therefore, more subjects must be recruited in the formal study instead. Lastly, we did not look into the subjects’ English proficiency in the pilot study, thus how L2 proficiency influences

the subjects’ performance could not be discussed.