| As important features of oral discourse, oral-interactional language functions(shortened as OLF) play a substantial role in reflecting oral ability (See Weir,1993;O’Sullivan et al.,2002, etc.). However, their actual values and the methods and meansby which to assess them have hardly been touched upon.This dissertation attempts to provide a feasible solution to the above concern byestablishing an OLF model. The model is designed on the theoretical basis ofpragmatic competence description by Bachman&Palmer (1996) and Bygate (1987)’soral skill theory, and on the empirical basis of O’Sullivan et al.(2002)’s practicalmodel of oral language functions. After7phases of experimental application,modification and evaluation, the finalized version comes into being. The OLF modelis composed of three main functions: informational functions, interactional functions,and managing interaction. Informational functions refer to those language functionsthat convey information during oral interactions. It comprises9item functions,namely, providing personal/objective information, expressing personal opinions,justifying opinions, providing examples for opinions, comparing/contrasting,speculating, summarizing, expressing personal suggestions, and expressing personalpreferences. Interactional functions refer to those language functions that makeconversations interactive during oral interactions. It includes12item functions,namely, agreeing, disagreeing, asking for opinions/information, insisting on personalopinions, modifying personal opinions, supporting, challenging opinions, persuading,asking clarification, providing clarification, checking understanding, and providingunderstanding. Managing interaction refers to those language functions that manageand control the progress and direction of conversations during oral interactions. Itcovers6item functions, namely, initiating conversation, greeting, transitioning,developing, changing topic, and terminating conversation.Based on the OLF model, this dissertation selected240pieces (around150,000words) of the transcriptions of the role-playing task of TEM4oral test samples from the corpus SWECCL after balancing the factors of oral proficiency, gender, and roleof the test takers in three consecutive years. With proper encoding, the corpus sampleswent through two phases of trial tagging, the results of which have shown that theinter-rater reliability and intra-rater reliability reach up to.982and.998respectively,indicating that the tagging system is reliable. With the aid of software SPSS17.0andAnt Conc3.2.1, the author has done descriptive, t test and correlation analysisseparately on the tagged samples. Coupled with sample discourse analysis, this paperdiscusses the factor influence of oral proficiency, gender, role, and testing year onOLF elicitations, and expands on the most pertinent reasons accordingly. This paperpoints out that the establishment of the OLF model enriches modes of languageassessment that are not micro-linguistic, and provides an empirically-based theory forassessing English learners’ oral communicative ability, as well as makes it possible toassess such an ability in real time. Also, with a proper rating criterion in place, thethree main functions, i.e. informational functions, interactional functions, andmanaging interaction are capable of reflecting oral richness, interactiveness, anddirectness respectively; the joint elicitation of informational functions andinteractional functions mirrors the communicative effectiveness of oral interaction.The research results have found that a significant variation exists in elicitationfrequencies of different OLF item functions, among which the most frequently usedincludes expressing personal opinions, justifying opinions, asking foropinions/information, transitioning, etc. In respect of factor influence, the differencesof oral proficiency, gender, role and testing time do not exert any significant impacton OLF elicitations. Yet, t test shows that difference does exist on some OLF itemfunctions, especially for the factor of oral proficiency. It has been found thathigher-level groups use significantly more interactional functions than lower-levelgroups.Also, an OLF-based teacher questionnaire is designed. With the aid of severalexperts, the questionnaire receives two revisions before being sent out. The surveyinvolves180teachers from more than100colleges and universities across China. Themain part of this questionnaire including29questions, relates to the importance of OLF. The first2questions pertain to the value of OLF as a whole while the rest27ones correspond to the27item functions in the OLF model. The choices for eachquestion adopt the6-point Likert scale. This questionnaire is aimed at finding out theuse values of OLF from teachers’ perspective. SPSS analysis shows that thequestionnaire’s Cronbach’s Alpha reaches.934, indicating a rather high reliability, anddesirable correlation coefficients indicate satisfactory validity. Research findingsdisplay that all of the27OLF item functions register relatively high values. Yet,compared with the mean scores of informational functions and managing interaction,that of interactional functions is relatively low, indicating that interactional functionsare less often used in real communication from teachers’ standpoints. Questionnaireanalysis indicates that the values and significance of OLF are clearly confirmed by theteacher and different OLF item functions exhibit different value weightings.Finally, the author designs a system of conversion formulae and attempts tocorrelate students’ OLF elicitation patterns with teachers’ OLF evaluations by puttingthem on the same rating scale. The design of the formulae is based primarily on thedifference of actual OLF elicitation frequencies. The research results discover that thecorrelation coefficients between students’ and teachers’ OLF value assessmentamounts to.306which is not significant. This finding indicates that some degree ofdifference exists between the two sides’ assessment on OLF values, which isintensively manifested in8item functions: providing personal/objective information,challenging opinions, persuading, insisting on opinions, providing examples foropinions, summarizing, providing understanding, and developing. The difference ismainly caused by test task requirements, test time restraints, candidates’ oralexpression habits, teachers’ over-evaluation, etc. At the end of this dissertation, theauthor highlights that introducing the OLF model into classroom help teachers knowstudents’ oral communicative and critical thinking habits better, thus contributing tothe enhancement of students’ oral communicative ability. Besides, the relativeweighting of each item OLF can be penned into teaching syllabus, enriching itsdescriptions on oral communicative functions. |
PDF Full Text Request