行政院國家科學委員會專題研究計畫 成果報告
兒童談話中的連貫性
計畫類別： 個別型計畫 計畫編號： NSC91-2411-H-004-023- 執行期間： 91 年 08 月 01 日至 92 年 07 月 31 日 執行單位： 國立政治大學語言學研究所 計畫主持人： 黃瓊之 報告類型： 精簡報告 處理方式： 本計畫可公開查詢中 華 民 國 92 年 11 月 1 日
兒童談話中的連貫性
Coherence in Children’s Conversation
中文摘要
本研究探討以漢語為母語的兒童如何在言談中建構連貫性,也就是如何在談 話內容上表現關連性的能力。語料包括六個兩歲到四歲的幼兒與其母親或父親的 自然對話。語料的分析著重在幼兒如何聯繫他們的話語和之前父母的話語。分析 中發現幼兒的話語多鄰接父母的話語。年齡較大的幼童的鄰接語較能連續之前父 母話語的主題並加上新的訊息。另外，年齡較小的幼童話語與父母的話語多為語 境上的關連，而年齡較大的幼童則亦能表現語言上的關連。 關鍵字： 連貫性，關連性，語言習得Abstract
This study investigated Mandarin-speaking children’s abilities to demonstrate
coherence in conversation, i.e. the ability to reflect relevance relationships in the
structure of conversation content. The data consisted of natural interactions between six children (2 to 4 years old) and their parents. The conversation between each child-parent dyad was analyzed to examine how the children related their utterances to the preceding adult utterances to achieve coherent conversational exchanges. It was found that the children tended to produce more adjacent utterances than nonadjacent utterances to prior adult speech. Older children’s adjacent utterances were more contingent; that is, they tended to share the same topic with the preceding utterance and add new information. In addition, while younger children’s contingent utterances demonstrated contextual relevance, older children’s contingent utterances may also show linguistic contingency.
2
Introduction
The development of children’s communicative competence has received much attention in recent years. To communicate effectively, young children must know how to participate in conversation with others. One of the crucial conversational abilities that children must acquire is to demonstrate coherence in conversation. That is, in the dynamic course of conversation, children’s contribution must reflect relevance relationships in the structure of conversation content. Such ability requires important cognitive, linguistic, and conversational skills which children must acquire.
In child language acquisition research, coherence has often been studied in terms of contingency. For example, Foster (1982) studied early coherence in child language and found that English-speaking children begin with non-verbal interaction within which they can show contingency relations in adjacency pairs. In Foster (1990), the author further distinguished two kinds of relevance relationships. The first kind is
horizontal relevance, in which consecutive utterances are relevant to each other. The
second kind is vertical relevance, in which each utterance is relevant to the overall topic being developed. It was found that children as young as 2;6 begin to demonstrate both types of relevancy relationships in conversation, constructing hierarchical structure of coherence.
Another type of early coherence, as shown in Ochs et al. (1979), is achieved when a single piece of information is conveyed across separate utterances, sometimes across different speakers’ turns. In the sequence, each utterance provides one component of the proposition, such as the argument, predicate, and modifier. Thus, the different utterances are jointly encoding the single proposition. In Ochs (1983), it was further demonstrated that children also use the strategy of repetition to produce relevant contributions. They may repeat all or part of the preceding utterance as an attempt to respond appropriately to particular types of utterances.
Bloom, et al. (1991) provided a detailed analysis of the utterance-to-utterance
relevancy relationships in children 1;9-3;0. The study showed that initially contingency relations are basically contextual rather than linguistic. Contingency, in general, and linguistic contingency, in particular, are the major developments in the children’s discourse. Bernstein (1981), on the other hand, looked at the expression and organization of propositional information, and demonstrated various aspects involved in the development of contingent propositional relations between utterances.
In addition to the studies of contingency, some studies have focused on children’s use of cohesive devices in constructing coherence. The emergence of intersentential links has been reported to emerge at two to three years (Peterson and Dodsworth, 1991). In addition, it has also been reported that children by four years old have acquired various surface markers of cohesion such as ellipsis, proforms and conjunctions to mark relations across utterances and speakers (McTear, 1985).
In Mandarin child language acquisition research, only a few studies have touched upon coherence in children’s conversation. Huang (2003) studied coherence in Mandarin child-mother conversation within a temporal reference framework. Chang (1998) investigated the development of autonomy in Mandarin-speaking children’s narratives; Hickman & Liang (1990), on the other hand, focused on how Mandarin-speaking children used clause structures to organize narrative structures. However, to the best of the researcher’s knowledge, no studies have been done to systematically investigate Mandarin-speaking children’s conversation within a coherence framework and from a developmental perspective. Therefore, little is presently known about Mandarin-speaking children’s developmental course of coherence in conversational interaction.
The purpose of this study was to explore how Mandarin-speaking children, aged two to four, accomplish and acquire coherence in child-parent conversational interaction. This study focused particular on how children construct utterance-to-utterance relevance relationships. Quantitative and qualitative analyses were conducted to examine how children of different ages relate their utterances to preceding adult utterances to achieve coherent conversational exchanges.
Methods
The subjects of this study included two two-year-olds, two three-year-olds, and two four-year-olds. The children were visited in their homes. Natural child-parent conversations were audio- and video- taped to capture both the linguistic data and the contextual information. Each child-parent dyad was recorded for six hours. The data analyzed for the study included two hours of recording from each dyad.
The analysis framework followed Bloom et al. (1991). Child utterances were first classified into the categories of Adjacent utterances and Non-adjacent utterances.
4
pause after a previous adult utterance.
ii. Non-adjacent utterances: child utterances that occurred without a previous adult utterance or with a definite pause after a previous adult utterance
Adjacent utterances were further classified as Repetitive, Contingent, or
Non-contingent.
i. Repetitive utterances: child utterances that shared the same topic with the preceding adult utterance but did not add information; that is, all or part of the preceding utterance was repeated with no change.
ii. Contingent utterances: child utterances that both shared the same topic with the preceding utterance and added information to the preceding utterance.
iii. Non-contingent utterances: child utterances that did not share the same topic as the preceding adult utterances
Contingent utterances were further classified as Contextually contingent or
Linguistically contingent.
i. Contextually contingent: child utterances that were contingent to the preceding utterance based on situational context only.
ii. Linguistically contingent: child utterances that were contingent to the preceding utterance based on formal intraclausal or interclausal relations as well as possible situational context.
Results and Discussion
The analysis showed that except for one two–year-old, who produced almost equal percentages of adjacent and non-adjacent speech, the other children all produced more adjacent speech than nonadjacent speech. The children appeared to have learned some basic turn-taking rules. They knew that conversation was
reciprocal, and they were able to supply utterances after adults had said something.
Further analysis of the children’s adjacent utterances revealed that the children’s adjacent utterances were quantitatively and qualitatively different. Younger children produced higher percentages of repetitive utterances and non-contingent utterances than older children. On the contrary, older children supplied higher percentages of
contingent utterances than younger children. Thus, the ability to talk contingently developed with age; older children were more capable of relating their utterances to preceding adult utterances by providing relevant new information.
The children’s repetitive utterances were further examined to investigate their communicative functions in constructing coherent conversation. It was found that the children’s repetitive utterances functioned differently in the speech of the different age groups. Older children’s repetitive utterances were more communicatively appropriate and they also served a larger variety of communicative functions. Two-year-olds’ repetitive utterances often functioned to imitate preceding adult utterances; these repetitions sometimes were not appropriate in the interactions, revealing the children’s failing to understand the preceding adult utterances. Three-year-olds also imitated adult utterances, but they also used repetitions to show acknowledgement or agreement. As for four-year-olds, a more interactive function was found. That is, four-year-olds may repeat part or all of the preceding information and use it as the topic of their subsequent utterances.
The children’s contingent utterances can be contextually or linguistically related to the preceding adult utterances. Further analysis showed that contextual contingency decreased with age while linguistic contingency increased with age. Older children were more capable of expanding the verb relations from the prior adult utterances; they may expand the relation with replacements, additions, or other modifications. In addition, it was also found that interclausal relations were a late development, which rarely occurred in two-year-olds’ speech. Some occurrences were found in
three-year-olds’ and four-year-olds’ speech.
The findings of the study were generally consistent with the results in Bloom et
al. (1991). Our study further revealed some qualitative characteristics of the children’s
utterances in constructing coherence conversation. Further analysis is needed to analyze coherence in child language from a more global coherence perspective.
References
Berman, R. A. & Slobin, D. I. (1994). Relating Events in Narrative: A Crosslinguistic
Developmental Study. Hillsdale, NJ: Lawrence Erlbaum Associates.
Bernstein, L. (1981). Language as a product of dialogue. Discourse Processes, 4, 117-47.
6
interaction between information processing and linguistic knowledge. In Bloom, L., Language Development From Two to Three. Cambridge: Cambridge University Press.
Chang, C-J. (1998). The development of autonomy in preschool Mandarin Chinese-speaking children play narratives. Narrative inquiry, 8, 77-111.
Foster, S. (1982). Discourse topic and children’s emerging ability to handle it.
Proceedings of the Berkeley Linguistic Society.
Foster, S. (1990). The Communicative Competence of Young Children. New York: Longman.
Hickman, M. & Liang, J. (1990). Clause-structure variation in Chinese narrative discourse: a developmental analysis. Linguistics, 28, 1, 167-200.
Huang, C-C. (2003). Talking about past events in conversation: An analysis of Mandarin mother-child and adult-adult discourse. Taiwan Journal of Linguistics, 1(1), 123-157
McTear, M. (1985). Children’s Conversation. New York: Basil Blackwell.
Ochs, E. O. (1983). Making it last: repetition in children’s discourse. In Ochs, E., & Schieffelin, B. Acquiring Conversational Competence. Boston: Routledge & Kegan Paul.
Ochs, E. O., Schieffelin, B., & Platt, M. (1979). Propositions across utterances and speakers. In Ochs, E., & Schieffelin, B. (eds.) Developmental Pragmatics. New York: Academic Press.
Peterson, C. & Dodsworth, P. (1991). A longitudinal analysis of young children’s cohesion and noun specification in narratives. Journal of Child Language, 18, 397-415.
