Nonverbal Cues Of English Learner's Anxiety In Oral Presentation

Oral presentation is a common pedagogical activity but the most anxiety-provoking task for language learners in domestic English class.The emotion state of the presenter can be evaluated subjectively and objectively.The subjective evaluation refers to an overall perception of emotion,which can be further differentiated by self-or other-perspectives.The objective evaluation traces down to specific multi-modal indicators,such as acoustic and kinesic cues.The Brunswik's Lens Model proposes that nonverbal cues may bridge self-with other-perception of the same subject of observation,i.e.they serve as ingredients for the formation of both two perception,but different cues may contribute differently to perception from self-to other-perspectives.Therefore,the current study is to verify whether there is a consistency on magnitudes and nonverbal cues that presenters and listeners use,to perceive the English oral presentation anxiety.Seventy freshmen of non-English majors were recruited in this experiment,who took an elective course named ‘English Cultural Transmission' at University of Electronic Science and Technology of China during 2017-2018-2 semester.In the last lesson of oral test,each presenter was required to give an English presentation with audio and visual documents recorded.State Anxiety Inventory(SAI)and Likert 5-point scale have been applied respectively to obtain the self-and other-perceived anxiety.The acoustic parameters of audio data were automatically extracted by computer program,and the kinesic features of visual data were coded by independent raters with robust inter-rater reliability.The results showed a positive correlation between the two subjective perspectives on anxiety perception of presenters but a discrepancy in terms of specific nonverbal parameters.The other-perceived anxiety mainly originates from both audio and visual modalities: it is positively correlated with body tension,standard deviation of the first formant,and negatively correlated with head turnings,while self-perceived anxiety is positively correlated with two parameters of the audio modality,i.e.mean fundamental frequency and standard deviation of the first formant.Therefore,the acoustic cues indicating vocal fold vibration frequency and variations in physical features of resonance cavity are the sources of speaker's self-perceived anxiety,while other-perceived anxiety is formed by the information integration of two kinesic cues revealing body tension and head movements,and one acoustic cue indicating variations in physical features of resonance cavity of the speaker.The current research assesses the anxiety-related recognition model of speech and kinesics in a setting of English class presentation and thus provides possibility for an automatic computation on learners' negative emotion in learning a foreign language.Our findings may facilitate the construction and algorithm optimization of the intelligent system in smart classrooms,which are characterized by real-time tracking and instant feedback.