Fahn (1993)

Fahn investigated the acquisition of the ba construction and proposed that the ba construction is subject to five constraints. First, the Progressive constraint states that the progressive marker cannot occur in the construction. Second, the Verb Selection constraint says that only the ‘Accomplishment’ type of verb can occur in this construction. Third, the Modifier constraint states that appropriate modifiers are needed for the other three types of verbs, i.e., states, activities, and achievements, to occur in this construction. Fourth, the Compound Verb constraint says that only resultative verb compounds can occur in the ba construction. Finally, the Definiteness constraint says that the ba-NP must be definite or generic. Furthermore, to account for the characteristics of the ba-NP, that is, it can be a non-argument, the position it occupies is not always a theta position and it does not occur in a sentence-initial position, Fahn argued that the ba-NP should be a second topic.

To understand how children acquire these properties, three tasks, a puppet selection task, a question-answer task and an imitation task, were executed in this study. The puppet selection task was used to examine children’s understanding of the five constraints. Two puppets were used with one uttering a grammatical sentence and the other an ungrammatical one. The subjects were asked to choose the puppet which produced the grammatical sentence. In the question-answer task, the fact that the ba-NP can be the antecedent of a reflexive was inspected. Eight sentences, four ba sentences and four non-ba counterparts, were used in the task.

The participants had to answer which person, mentioned in the sentence, is the antecedent of the reflexive. The imitation task investigated children’s treatment of regular topics or the ba-NPs in a chain, that is, to see how children formed a topic chain. Did they prefer to use the skill of maintenance, deletion or pronominalization?

Five ba- and five non-ba sentences were used in the test. The subjects repeated what

they heard after the experimenter.

The participants were 100 Mandarin-speaking children aged from 2;6 to 7;5, and they were further divided into ten groups at half-year intervals, with ten children in each group.

The results of the puppet selection task showed that age seemed to play a central role in the experiment. Different age groups differed as to how well they mastered the five constraints. Also, the results indicated that age five was a crucial demarcation point for the Progressive, Verb Selection, and Modifier constraints whereas age six was a significant point for the development of the Compound Verb and the Definiteness constraints. In addition, the Progressive, Verb Selection, and Modifier constraints seemed to form a dimension while the Compound Verb and Definiteness constraints form the other dimension. The constraints on the first dimension were acquired earlier than those on the second dimension. However, whether there was an order among the three constraints in the first group or between the two constraints in the second group was not sure.

As regards the question answer task, the results showed that younger children (below the age of four) tended to select the non-subject as antecedent while the older ones (above the age of five) preferred the subject. As for the children between the ages of four and five, the situation was not so straightforward. This might designate a transition between the two preferences.

Concerning the imitation task, the results indicated that the children preferred the topics or the ba-NPs to form a topic chain and this tendency increased as they got older. Also, the deletion of the topics and the ba-NPs decreased with ages;

pronominalization increased in the ba-NP chain while it decreased in the regular topic chain.

To account for that certain constraints were acquired earlier than others, the theory

of conservativism (Pinker 1989) and its implication of rapid learning were applied in the study. Based on the theory of conservativism, Fahn argued that the Definiteness constraint would not emerge until the distinction between definite and indefinite articles were acquired. Once the contrast was acquired, it would apply to the ba-construction according to the theory of rapid learning. Likewise, the Progressive constraint would not emerge until the progressive marker zai was acquired. Given this assumption, the relatively late acquisition of the Definiteness constraint was attributed to the relative difficulty of definite and indefinite articles compared to the progressive marker. In addition, evidence was given to show that the children acquire verb meanings, especially ‘Accomplishment’ verbs, relatively early (Erbaugh 1982). Also, the early acquisition of the Modifier constraint was attributed to the early emergency of verbs and aspect markers, and the frequent occurrence of modifiers in salient position. All these explained why the Verb selection and Modifier constraints were acquired before the Definiteness constraint. Cummulative Complexity (Brown 1973) and Developmental Law (O’Grady 1987) were also applied to account for the later occurrence of the Compound Verb constraint than the Progressive, Verb Selection and Modifier constraint. According to the theories, the late acquisition of the Compound Verb constraint was attributed to a more complex form and meaning of compound verbs.

In sum, this study gives us a comprehensive analysis of how the ba construction is formed in the development of children’s grammatical system. Nevertheless, the number of the subjects in each age group was small. In addition, this study examined the acquisition of the Mandarin ba construction but not the Taiwanese ka construction. Based on the previous studies, the ka construction is an even more complicated structure than the ba construction. To get a better understanding of the developmental order of the Taiwanese ka construction, more study is needed.

2.4.3 Summary

Two studies of the acquisition of the Mandarin ba construction have been reviewed in this section. Cheung’s (1992) findings showed that the ba construction should be treated as a verb-like construction in children’s grammar and that the Object Affectedness Linking Rule alone could not explain children’s acquisition of the construction. Fahn’s (1993) study showed that the Progressive, Verb Selection, and Modifier constraints emerged before the Compound Verb and Definiteness constraints.