Acquisition of American English intonation patterns by non-native speakers: Use of real-time computer-mediated visual feedback

This dissertation reports an experimental, pre-test/post-test study of the use of a computer-based voice-interactive language learning system for teaching English as a second language.; Each of five control and eight experimental group Thai student volunteers met with the researcher in one-on-one tuition for ten minutes each week for five weeks. With one of five American English intonation patterns introduced and practiced per session, members of each group covered identical training materials. While control group members received intonation training stimuli from static, line-drawings and feedback from instructor hand gestures, experimental group members received stimuli provided by a contour display on the Visi-Pitch--with feedback provided by real-time visual display of utterance contours.; Pre- and post-treatment speech samples were collected from each participant--corresponding to each of the five training target patterns as well as a sample relating to a passage of discourse length. Samples were also gathered from seven native speakers of American English. These samples were digitized and electronically filtered with low-pass filtering to remove segmental data (above 300 Hz.) and were randomly presented to 17 native English speaking raters who were asked to assign a nativeness value of 1 through 7 for each sample (1 = clearly foreign, 7 = clearly native).; With sample length not significantly changed from pre- to post-treatment and with interrater reliability high, there were several significant findings of the study: (1) native-speaker raters were clearly able to differentiate native from non-native speech on the basis of suprasegmental information; (2) post-treatment scores for both control and experimental group members were significantly improved over pre-treatment ones; (3) improvement for each of the training targets was similar and significant; (4) improvement for the long, paragraph-length passage was nearly twice that of the short, training-target passages; (5) there was no significant difference between the improvement of the control and experimental groups; and (6) trainees in the experimental group reported a positive attitude toward the learning experience.