Abstract:
The role of classroom interaction in second language acquisition (SLA) has been the subject
of extensive research in recent years. The purpose of this study was to investigate the claimed superiority of communication tasks involving required information exchange (split information tasks) over tasks involving optional information exchange (shared information tasks) on the basis of how much negotiation of meaning learners produce when performing each type of
task. The study also sought to analyze qualitative aspects of negotiation and to assess the
theoretical claims made for negotiation in the light of the analysis. Subjects for the study included eight adult students from an English proficiency course who were assigned to two groups each containing four subjects. Over a period of six days the groups performed four communication usks of which two were split information tasks and two were shared information tasks. Full transcriptions of the task performances provided data for the study. Results confirmed that significantly more negotiation and repetition occurred in split information tasks. There was a small movement towards more even distribution of negotiation among interlocutors in split information tasks although the consistency of the differential contributions of specific interlocutors was noticeable across both types of task. The qualitative analysis distinguished six main types of negotiating questions in the data,
some of which were shown to be more effective than others in generating comprehensible
modifications to input or in extending the language output of the subjects. ln addition,
negotiating questions dealt with five broad dimensions of meaning: the form of the message,
grammatical and lexical meaning, content, opinions, and procedures. Of these five dimensions, only the first and second sometimes involved new or unfamiliar linguistic features in the input, thus fulfilling a requirement of the interaction hypothesis suggested by Ellis (1991).
Significant post-test gains in the subjecrs' knowledge of vocabulary embedded in the tasks
suggested that the negotiation of lexical meaning results in measurable learning of new words.
Overall however, negotiation dealt more with non-target language features of output than with
unfamiliar input and it was this which provided the more promising interactional route to
language development. An investigation of other features of interaction revealed no significant difference in the amount of talk produced in split and shared information tasks. Talk was more evenly distributed among interlocutors in the split information tasks although inequalities persisted,
with panicular interlocutors dominating interaction across all tasks. In the shared information tasks, turns and utterances were significantly longer, and conjunctions were used more frequently. Prepositions on the other hand were used more frequently in the split information
tasks. These results suggest that the greater need to express links between propositions in the
shared tasks results in discourse of grearer synracric complexiry. While the study supported the claim that split information tasks produced more negotiation than shared information tasks, a qualitative analysis of the negotiation, and of other aspects of interaction, suggested that more negotiation does not necessarily provide superior
conditions for language development,
