華語誦讀時停頓偏誤分析及線上口說學習平台應用
No Thumbnail Available
Date
2018
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
本論文有7位日籍受試者，身分為在台灣就讀研究所的碩士生與碩士畢業生，主要研究項目為誦讀實際錄音(分前後測)、心理停頓、線上平台結果(分前後測)、訪談以及問卷，在前後測中間隔一個星期，此期間受試者使用線上平台自學。本論文問卷共三種，分別為受試者基本資料調查與誦讀意識調查、誦讀注意之項目調查（於前測時進行）、誦讀意識調查（於後測時進行）。後測結束之後，本研究分析使用線上平台前後測之進步幅度，練習時間、前後測進步幅度與練習次數之間的關係，分析兩個項目之相關係數分析，並分析受試者與華語教師的心理停頓與實際停頓的關係，以及由語音分析軟體來分析實際停頓表現(分析項目包括：時長、次數、不流利停頓)，並對比其與華語教師之停頓，另外也分別和受試者與5位華語教師進行了訪談。
分析結果發現，華語教師之心理停頓共23個，受試者心理停頓平均約25個。華語教師的實際停頓次數為23次，受試者前測時停頓次數平均約47次，後測時受試者停頓次數減少至26次。華語教師的實際停頓時長平均為0.54秒，受試者前測之實際停頓時長平均為0.68秒，受試者後測實際停頓時長平均為0.67秒。研究者透過訪談也發現受試者對線上平台都有正面回饋(在滿分是5分評比中的平均得分為4分)，受試者的練習時間(最長465分鐘；最短25分鐘)與前後測分數進步幅度(r=.479)以及練習次數(最多13次；最少4次)與前後測分數進步幅度(r=.638)，都具有中度相關。
透過線上平台自學，受試者不流利停頓情形改善了很多(前測時平均19個；後測時平均1.3個)，因此證明重複練習可以讓學習者提升華語語感，並減少不流利停頓，而在自學時，每次練習時間不需太久，但是次數應該越多越好。
This thesis has 7 Japanese subjects (JS) who are the students or graduates of masters program in Taiwan. The research materials consist of the recording reading-aloud tasks, the resulting scores of an online speaking training platform, the interview, and the questionnaire. This questionnaire is divided into three parts : (1) Questionnaire 1 was about the basic information on the subjects and what they noticed of reading aloud. (2) Questionnaire 2, which was conducted during The Pre-test is about which item the subjects paied attention to when they were doing the reading-aloud task. (3) Questionnaire 3, which was conducted during The Post-test was about what they noticed of reading aloud. Between Pre-test and Post-test, the subjects practiced the task by themselves through the online speaking training platform for one week. The researcher conducted not only the correlation analysis of the progress of the score, the total-study-time, total study-frequency, but also analyzed the relationships between the Internal-pauses and the Actual-pauses. The Actual-pauses of JS and Native Mandarin Chinese Teacher (NMCT) were also analyzed by the speech analysis software called Praat. The researcher interviewed all of JS and 5 NMCT. The results of the analysis found that the Internal-pauses of NMCT were 23 but the Internal-pauses of JS were 25 on average. The Actual-pauses of NMCT was 23 times; the Actual-pauses of JS on the Pre-test is approximately 47 times on average; the Actual-pauses of JS on the Post-test was about 26 times on average. The actual pause duration of NMCT was 0.54 seconds on average. The actual pause duration of JS on Pre-test is 0.68 seconds on average and the actual pause duration of JS on Post-test is 0.67 seconds on average. With regard to the online speaking training platform, the result of the interviews to the JS had positive feedback (4 based on scale of 0 to 5). The result of the correlation analysis on the study Time of JS (maximum : 465 minutes ; minimum : 25 minutes) and the progress of the score was r=.479, the result of correlation analysis on the amount of times they studied (maximum : 13 times; minimum : 4 times) was r=.638. Through the online speaking training platform, the Non-fluent-pauses of JS improved greatly (on Pre-test it was 19 on average; on Post-test it was 1.3 on average). In conclusion the amount of leaners studied by themselves was more impactful upon the duration of each study period.
This thesis has 7 Japanese subjects (JS) who are the students or graduates of masters program in Taiwan. The research materials consist of the recording reading-aloud tasks, the resulting scores of an online speaking training platform, the interview, and the questionnaire. This questionnaire is divided into three parts : (1) Questionnaire 1 was about the basic information on the subjects and what they noticed of reading aloud. (2) Questionnaire 2, which was conducted during The Pre-test is about which item the subjects paied attention to when they were doing the reading-aloud task. (3) Questionnaire 3, which was conducted during The Post-test was about what they noticed of reading aloud. Between Pre-test and Post-test, the subjects practiced the task by themselves through the online speaking training platform for one week. The researcher conducted not only the correlation analysis of the progress of the score, the total-study-time, total study-frequency, but also analyzed the relationships between the Internal-pauses and the Actual-pauses. The Actual-pauses of JS and Native Mandarin Chinese Teacher (NMCT) were also analyzed by the speech analysis software called Praat. The researcher interviewed all of JS and 5 NMCT. The results of the analysis found that the Internal-pauses of NMCT were 23 but the Internal-pauses of JS were 25 on average. The Actual-pauses of NMCT was 23 times; the Actual-pauses of JS on the Pre-test is approximately 47 times on average; the Actual-pauses of JS on the Post-test was about 26 times on average. The actual pause duration of NMCT was 0.54 seconds on average. The actual pause duration of JS on Pre-test is 0.68 seconds on average and the actual pause duration of JS on Post-test is 0.67 seconds on average. With regard to the online speaking training platform, the result of the interviews to the JS had positive feedback (4 based on scale of 0 to 5). The result of the correlation analysis on the study Time of JS (maximum : 465 minutes ; minimum : 25 minutes) and the progress of the score was r=.479, the result of correlation analysis on the amount of times they studied (maximum : 13 times; minimum : 4 times) was r=.638. Through the online speaking training platform, the Non-fluent-pauses of JS improved greatly (on Pre-test it was 19 on average; on Post-test it was 1.3 on average). In conclusion the amount of leaners studied by themselves was more impactful upon the duration of each study period.
Description
Keywords
誦讀, 華語停頓, 線上平台應用, 日籍學習者, Oral reading, Pause in CSL, Online platform, Japanese learner