當某一個人或某一部機器告訴我們一件以前我們所不知道的事情時，我們就得到了消息（Information）。嚴格地說，任何一項傳達的動作（Communicative acts），只有在它能減少我們對某一件事物的無知或不確知狀態（Ignorance or uncertainty）時，才會帶給我們消息。假定我們投擲一個兩面都是人頭的銅板，在它未著地之前，我們已確知它的答案了，因此縱然觀察到這個動作的結果（Out come），也無任何消息可言；假定我們所投擲的銅板，一面是蘭花，另一面是梅花，則它在未著地之前，我們無法預知向上的一面到底是蘭花，或者是梅花，我們就有某一程度的不確知狀態（Some degree of uncertainty）。當我們觀察其結果，原有的不確知狀態就消失了或減少了。也就是說，我們得到了消息。 從這個觀點，消息之所以發生，乃在於我們預先存有某一程度的不確知狀態。而所謂消息量，就是指不確知狀態所減少的量。因此，就量而言，這兩個概念－消息與不確知狀態－實在是指同一件事情。 Sha nnon(1948) 定義不確知狀態的量（Amount of uncertainty）為： U=-πPi 1og2 Pi， 其單位為bit。 U是不確知狀態的測量（Measure of uncertainty）。 Pi是屬於 i 組之訊號將會發生的可能率。 藉此公式，我們可以把消息加以數量化（Quantification）。 在心理學上，刺激與反應的歷程，比之傳達理論（Communication Theory）中輸入與輸出（Input and output）的歷程是頗為吻合的。我們可以把它們看成各有一個消息空間（Information spce），也就是說各有一組可供選擇的事物，互斥並各有其發生的可能率。（A set of alternatives,mutually exclusive events,with a probability assigned to each of them） 假如消息的概念對人類的行為是有意義的，則在任何抉擇的情境（Choice situation）中，可供選擇之事物的數量（Number of alternatives）對於行為必定有很重要的影響。例如我們想知道被試對某一刺激的反應有多快；他認清一刺激的準確性有多高；及在刺激消失了幾分鐘後，他還能正確回憶的可能率有多大（How quickly a subject will respound to a stimulus, how accurately he will identify a stimulus, how likely he is to remember a stimulus correctly a few minutes after its disappearance）。這些問題不但刺激的性質所影響，同時也要看所可能出現的刺激（或反應）有多少，以及各刺激（或反應）出現的可能率有多少。換句話說，上述的各種測量都與被試的不確知狀態有關。（G．A．Miller,1956;Broadent, 1958） 另一方面，假如我們希望不同程度的不確知狀態或消息量對人類行為有意義，則首先我們必須能表明出整個消息空間，亦即指出所有可供選擇之事物的數量。不確知狀態的量若為無窮，或無法表明（infinite or unspecified），則為無意義。（Cane and Horn, 1951） 從以上的論述，我們知道不確知狀態的概念意含抉擇或辨認（Choice or discrimination）。因此在任何涉及抉擇或辨認的行為情境中，最重要的莫過於對不確知狀態的測量。無疑的，大部份的學習都包含抉擇或辨認的歷程，故只要能指出有多少可供選擇之事物，則反應前不確知狀態的程度必然是學習的一個重要變數。 在一般語言學習（Verbal Learning）的實驗中，我們通常呈現給被試一列表的項目（a list of items），讓他重覆看過幾次，然後要他依項目次序預測下一個將出現的項目，如串系學習（Serial Learning）；或者是要他對出現的刺激字做正確的反應，如配對聯結學習（paired-associates Learning）。被試所可能做的預測（Anticipation）或反應，必定取樣自某一消息空間（Sampling from an information space）。這個消息空間乃由所學習列表中的可能反應所形成。因此，從被試的觀點，其每次做預測或做反應之前，必有某一程度的不確知狀態，此與消息空間有直接的關係。故消息量無疑是語言學習中的一項重要變數。 Riley在1952年利用配對聯結的方法研究在不同情境下的學習。他以八個不同的無意義音節（Nonsense syllables）為刺激字，令被試以預測的方式學習另外八個與原來刺激字各自相對的反應字。他把被試分成三組。在第一組中，如果被試所預測的反應字是不正確的，則一個正確的反應字就顯現出來；在第二組中則顯示兩個可能的反應字；第三組則顯示四個可能的反應字。然後比較三組被試學到一次完全正確反應（One perfect response）所需的平均嘗試次數（Mean trials to-criterion）。結果發現第一組成績比第二組好，而第二組又比第三組好。亦即反應不確知狀態（Response uncertainty）的程度增加，則學習的速率就減慢。 Brogden 和Schmidt在1954年也做過類似的實驗，但所用的材料為各種不同的語言迷宮（Verbal maze）。在實驗中，他們改變迷宮每一選擇點的選擇數（Number of choices at each choice point），結果發現在兩個至十二個選擇數中，學會迷宮所需的時間，和迷宮中每一單位之選擇數的多寡，幾乎是成一直線的函數關係。如果以全部的錯誤次數做為測量之標準（Criterion measure），則所得的結果大致相似。 再考慮另一個問題，我們知道在學習一列表的無意義音節或單字中，列表長度（Length of list）一向被認為是學習的重要因素。但是Adelson,Mucker和Williams（1955）卻發現在語言學習中，假如列表的長度固定，則反應不確知狀態就成為一重要之變數。在他們的實驗中，列表的長度都保持十五個項目，但是構成每一列表所用的字母數量卻各不相同。其第一個實驗，情況較為簡單，一共使用四種列表，分別用二、四、六或十五個字母所編製成。在編製列表時，各字母出現的可能率都相等（Chosen at random with equal a priori probability）。最後一個列表因此就包括了所有的十五個字母。實驗的結果指出字母數量增加，則學會該列表所需的次數就隨之增加，除了第四個列表外，幾乎是成一條直線的函數關係。 讓我們來看看第四個列表的情況。由於列表的長度固定為十五個項目，且字母的選取都會經加以平衡（Balanced），因此第四個列表就包括了所有的十五個字母。列表的項目繼續出現，則愈到後來，被試不確知狀態的程度就愈來愈小，這我們可以用統計上自由度（Degree of freedom）的概念來加以了解。故整個列表雖由十五個字母所編成，其平均的不確知狀態的程度（Average uncertainty）卻遠少於log215。例如在出現七個字母後，只剩下八個可能性，此時不確知狀態的量為3bits，而不是3.907bits。 本實驗為了避免上述的困難，使用三個單位的項目（Three units items）來編製列表。改變可供選擇之事物的數量，如五或十，並且項目之形成乃從所有可供選擇之事物中每次取三個，每一個選擇事物（Alternatives）的出現可能率都相等（Equal probabilities of occurrance）。 本實驗之目的，在於研究消息量對語言學習之影響。本實驗並借用劉氏（Liu, in press）的研究方法，在檢驗學習程度時，令被試從所學過項目中的某一片斷（Fragment）去回憶整個項目，希望研究不同大小（0,1,2）的片斷線索（Fragment cue）對回憶之影響，藉以了解學習的歷程。 此外，我們知道在一般語言學習的實驗中，列表之編製通常以英文字母（Alphabet）為單位。但我們也知道字母對被試而言，都是早已建立習慣的單位。假如我們使用被試未曾建立習慣的單位，如無意的圖形（Nonsense designs）□□□等，則對學習之影響，或與字母不同。換句話說，消息量對語言學習之影響，可能與消息本身之性質（意指被試習慣與否）有相互作用（Interaction）的關係。 底下因此而引出三個實驗。

成語的使用有助高層次的語文認知學習和應用，無論在閱讀與寫作各方面都是重要的基礎訓練，然而目前台灣大部份的成語教學、學習平台或軟體多數提供靜態的資料查詢或教學資源，缺少以學生練習為主的環境。 　　本研究之目的為發展一個以本體論為基礎的中文成語練習系統。首先從文獻中分析成語多元訊息內容以建構成語知識本體、歸納學生運用成語造句時常見的病句偏誤原因並設計電腦可行之診斷機制，再以系統發展法建構一個以本體論為基礎的中文成語練習系統，系統可提供線上成語教材、可以自動產製是非、選擇、配合題型之測驗題目，並可以用情境式造句–在既定情境下實際利用成、詞語組合成一完整句子–的方式來練習成語，而系統可以判斷答題結果或造句是否合理，再針對成語使用盲點給予即時回饋，突破現有成語線上練習僅能從預先設定好的題庫中出題的限制，並提供情境式運用成語造句的環境。 　　系統效能及滿意度評估方面，本研究以準實驗設計進行，結果發現利用本研究之系統輔助教學的學生學習成效顯著高於使用傳統教學的學生，尤其對中、低能力程度學生具有顯著學習成效。在實際成語造句方面，以本研究之系統輔助教學之學生成語語法、語義使用正確性也較傳統教學學生顯著為高。學生也覺得本研究之系統是易用及有用的，並感到滿意。另外，在系統的可行性方面，受訪教師皆認為本研究之系統可以應用於學生自學成語。 / Using idioms would benefit a higher level of language perception, learning and application. It is also a kind of important but basic training in both reading and writing. However, presently, most platforms or software of Chinese idiom teaching and learning only offer static data inquiry or teaching resources, lack the environment of relying mainly on the fact that students practice. The purpose of this research is to develop the “ontology-based Chinese idiom practice system”. At first, we probed into the meta-data of Chinese idioms from any kind of studies and references to building and forming the ontology of the Chinese idiom, and summed up the wrong sentence with Chinese idioms to analyze the reasons of missing to design the feasible diagnosis mechanism of the computer. And then, we used above-mentioned ideas to develop the ontology-based Chinese idiom practice system. The system can offer the on-line teaching material of Chinese idioms, can automatically produce true-false, multiple-choice, and matching questions to examine, can offer a platform that users can actually practice sentence-making with Chinese idioms under a provided situation utilizing situational words and idioms. The system can also judge question sheet result or diagnose user-making sentences to be reasonable and can give feedback aiming at the blind spot immediately. This research can break through the restriction that the questions of existing on-line testing of Chinese idioms are always selected from the setting up question bank and can offer the situational environment that uses idioms to make sentences. Using quasi-experimental design, one class was assigned to experimental group using the system of this research assisted learning and the other class, the control group, using traditional teaching method. Related results demonstrated the performance of the experimental group was higher than the control group, especially it made better effect to the students of middle and low abilities in the experimental group. In making actual sentences with Chinese idiom, the number of correctness of Chinese idiom grammar and meaning using of the experimental group was higher than the control group. The students of the experimental group thought the system of this research is apt to use and useful, and feel satisfied. In addition, the teachers using the system actually thought the system of this research is practicable when the students study Chinese idiom by themselves.

本研究旨在了解職前華語文教師學習外語的歷程、語言學習信念與教學信念，屬質性個案研究。對象為華語文教學研究所四位台籍研究生，聚焦於外語學習歷程、外語學習信念、外語教學信念三大主軸，以深度訪談作為資料收集方法。訪談內容及範圍主要參考了Horwitz（1985）語言學習信念研究的BALLI 模式以及Johnson（1992）與張碧玲（2000）設計的語言教師教學信念問卷。研究發現包括：（一）不同時期的外語學習經歷對個案職前教師的學習及教學信念起了「學徒式的觀察」、信念的轉變及「濾鏡」等作用。（二）透過對個案的學習及教學信念探討，可以看到信念具有多層次的特質。而其中較整體面向的信念與較個別層次的信念，呈現了相互支持、矛盾等關係。（三）教學信念隨著背後所隱藏的動機不同，不能以因果關係作解釋。（四）個案教師基於不同的學習經歷、學習信念及教學信念，在教學上建構了與社會文化情境相關，同時也具有個人意義的教學理論。最後建議華語文師資教育應提供機會讓職前教師了解自我語言學習及教學信念，以及更多教學實務的機會，以讓他們在教學行動中反思，也在反思中進行教學。華語文教師本身亦可以利用自我的外語學習經驗作為反思的工具，時時保持成長的動機與內省的能力。此外，華語文教師也應對教學現場及環境文化等因素有所知覺，這樣才能促進有效的教學。 / This qualitative case study aims to understand the TCSL ( Teaching Chinese as a Second language ) pre-service teachers’ foreign language learning experience, language learning beliefs, and teaching beliefs. The participants are four Taiwanese postgraduate students from the department of TCSL in a university in northern Taiwan. The study focus on three aspects: foreign language learning experience, language learning beliefs, and teaching beliefs. In-depth interviews are conducted as the data collection method, informed by Horwitz’s (1985) BALLI model as well as Johnson (1992) and Zhang (2000). The findings include that (1) the four pre-service TCSL teachers’ foreign language learning experiences, serving as a filter, afford “apprenticeship of observation” which leads to changes in beliefs, and (2)beliefs are multilayered, involving both supportive and contradictory relationships. (3) Teaching beliefs are not necessarily consistent with each other because of different intentions lying behind. It is difficult to regard them as having a simple cause-and-effect relation. (4) Based on different learning experiences, learning and teaching beliefs, the four pre-service teachers construct their own personal theories toward teaching, which are not only socially relevant but also personally meaningful. Finally, the study suggests that language teacher education needs to provide sufficient opportunities for pre-service teachers to understand their own learning and teaching beliefs. It is also necessary to provide opportunities to engage them in authentic teaching contexts so that they could reflect on their teaching beliefs in real-world pratice. TCSL teachers need to be encouraged to use their own learning experiences as a reflective tool to sustain their passions for teaching.

本研究針對我國學生學習日語發音進行相似度指標發展之初探，貢獻為針對目前日語發音提供一個相似度的指標可以和老師語音進行比較分析，找出分析日語發音相似度之模式。 研究從聲音數位化的角度切入，有別於過去研究使用語音辨識的方式來進行，聲音數位化後為數值的方式，因此使用指標來計算相似的程度。研究提出一套對應的聲音相似度指標，以電腦分析輔助日語學習者的發音練習。 指標建立過程由聲音取樣、正規化、端點偵測，到實際的運算，使用所蒐集的聲音資料來測試指標的穩定度與有效性，研究結果說明在以日語為母語者間的指標都很靠近，而不同日語腔調間會有一定的指標差異，對於一定日語程度的對象而言，指標落點很靠近，惟本研究此次蒐集到的聲音資料，其應用指標運算結果的分佈太過集中，如果能有更多樣化的聲音資料來測試指標應能有較漂亮的分佈圖形。 / This research includes developing a similarity index applies to the evaluation of Taiwan students learning Japanese pronunciation. The contribution of this research is that it provides a similarity index to the Japanese pronunciation comparing to the teacher’s pronunciation, finding the model of how to analysis the similarity of Japanese pronunciation. This research uses the digital audio processing to begin with, which is different from the other research that uses the speech recognition to evaluate the pronunciation. The audio will turn into numerical format after digitalize, so this research uses an index to calculate the similarity. By using this similarity index, the computer can become an assistant role that helps to analysis while learning Japanese pronunciation. The developing of index starts from audio sampling, audio normalizing, and end-point detection to the calculation of similarity index. This research collects audio data to test the stability and the validity of the similarity index. The result indicates that the similarity index of native Japanese speakers is very close；and the similarity index contains certain difference between different accents. For those Taiwan students who qualify with Japanese, their similarity index is close. Nevertheless, the result of the similarity index is too centralized, it would be better if there are more audio data to test the similarity index.

多選題測驗試題已證明能有效地評估學生的學習成效，然而，以人為方式建立題庫是一件耗時費力的工作。藉由電腦高速運算的能力，電腦輔助產生試題系統能有效率地建置大規模的題庫，同時減少人為的干預而得以保持試題的隱密性。受惠於網路上充裕的文字資源，本研究發展一套克漏詞試題出題系統，利用既有的語料自動產生涵蓋各種不同主題的克漏詞試題。藉由分析歷屆大學入學考試的資料，系統可產生類似難度的模擬試題，並且得到出題人員在遴選測驗標的方面的規律性。在產生試題的過程中導入詞義辨析的演算法，利用詞典與selectional preference模型的輔助，分析句子中特定詞彙的語義，以擷取包含測驗編撰者所要測驗的詞義的句子，並以collocation為基礎的方法篩選誘答選項。實驗結果顯示系統可在每產生1.6道試題中，得到1道可用的試題。我們嘗試產生不同類型的試題，並將這套系統融入網路線上英文測驗的環境中，依學生的作答情形分析試題的鑑別度。 / Multiple-choice tests have proved to be an efficient tool for measuring students’ achievement. Manually constructing tests items, however, is a time- consuming and labor-intensive task. Harnessing the computing power of computers, computer-assisted item generation offers the possibility of creating large amount of items, thereby alleviating the problem of keeping the items secure. With the abundant text resource on the Web, this study develops a system capable of generating cloze items that cover a wide range of topics based on existing corpra. By analyzing training data from the College Entrance Examinations in Taiwan, we identify special regularities of the test items, and our system can generate items of similar style based on results of the analysis. We propose a word sense disambiguation-based method for locating sentences in which designated words carry specific senses, and apply collocation-based methods for selecting distractors. Experimental results indicate that our system was able to produce a usable item for every 1.6 items it returned. We try to create different types of items and integrate the reported item generator in a Web-based system for learning English. The outcome of on-line examinations is analyzed in order to estimate the item discrimination of the test items generated by our system.

本質性個案研究目的在探討以國際教育網路社群IEARN進行專題導向式跨國互動EFL學習的英語教學設計。該學習方式與教學模式特性諸多符合九年一貫課程改革理念。然而，在國內文獻及研究顯示，極少英語教師進行專題式網路教學設計。本研究個案是國中兩名英語教師與其班導學生為期兩年半的跨國文化專題學習。從個案訪談與課室觀察中發現，資訊科技融入英語教學，教師具備高度科技能力並非必要考量，而是教學策略、班級經營、師生合作為首務。 依據個案進行專題式導向學習過程中所面臨的挑戰與完成的歷程，研究者對教師、學校主管、教育當局與社會提出建言，以求改善與進步。這項教學革新並非只是順應課程改革，更是為了發展學生全面的學習與實踐終身學習為目標。 / The purpose of this research is to understand the challenge of project-based telecommunication activities through the IEARN for EFL learning. The learning methods and teaching strategies of project-based learning (PBL) conform to the rationale of Grade 1-9 Curriculum reform. However, literature showed that very few EFL teachers became involved in PBL with information technology for instruction. This research, conducted as a case study in the qualitative methodology, describes how two junior high school English teachers and their homeroom students implemented PBL through international network and how they developed cross-culture awareness in half and two years. It was found that successful EFL learning infused with information technology relied much more on pedagogy, class management and teacher-student collaboration than on teacher's computer skills. Based on the challenges and implement revealed in the process of PBL through telecommunication activities, the researcher gave suggestions to teachers, school administrators, educational authority and community for further improvement. The instructional innovation not only met educational reform standards but also developed EFL students' holistic learning in a lifelong fashion.

本質性個案研究旨在從三位歐美外籍留學生的視角了解他們的華語文學習信念，探討他們在台灣留學期間所處的社會文化環境如何改變他們的語言學習信念，以及台灣作為華語文環境所能提供的機會與挑戰。本研究主要透過深入訪談（In-depth Interview），邀請三位來台留學半年至一年的歐美華語文學生敘述自身語言學習歷程，包括他們的個人華語文學習背景、來台緣由與期待、在台華語文學習歷程及信念。歷時半年四次的定期訪談均被錄音，經謄錄為逐字稿、重現（reconstruct）個別故事樣貌後，再依Lieblich等人（1998）提出的敘事分析方法，以整體——內容（Holistic-content）和類別——內容（Categorical-content）之邏輯整理成個別個案，並作為後續跨個案分析的資料依據。結果發現，研究參與者從華語為外語（Chinese as Foreign Language）學習環境，到台灣華語文為第二語言（Chinese as Second Language）學習環境的不同學習階段裡，他們的華語文目標、期待和能力不斷地與社會環境互動，而互動所產生的結果足以鞏固或動搖其舊信念。來台前，研究參與者過往的語言學習經驗和成長環境對於外語學習的觀點，無形中都型塑了他們的華語文學習信念。來台後，他們經歷固有的信念和實際與母語者交際情況之間的落差，衝擊了他們原有的信念，而且覺得自身的華語文使用者身份及所付出的努力難以受當地母語者社交圈之認同。此外，研究參與者的留學經歷也反映出台灣作為華語文學習場域的一些情況，既當他們的華語文能力變強後，才有能力善用台灣華語文環境所提供的學習機會，當他們的華語文能力仍偏弱時，他們一般較難在非正式場域與母語者互動，或從中獲得學習機會，反而需依賴課堂環境為主的學習機會。最後，由於信念會隨著學習歷程波動及變化，面向多元，具動態特質，因此本研究建議未來可選擇某一項議題做更深入的研究，若採取類似BALLI研究信念橫斷面的工具時，還可加入情景及情感面向，探討信念歷時的直斷面。在教學運用上，華語文中心能著手處理外籍華語文學生的行前準備資料，加強對留學生有關語言學習進程的輔導，在師資培訓上也應該注意如何處理初出國留學生在語言學習適應上的調整，以減少留學生因學習期待落差而受到的衝擊。另外也建議常與學習者接觸的當地母語者接受跨文化交際的教育，雙方才能創造合宜的華語文學習環境，促進外籍留學生赴台學習華語文之發展與成效。 / This qualitative case study aims to look into Chinese language learning beliefs from the perspective of three Chinese learners from Europe and America and to explore how the Taiwanese social-cultural environment changes their language learning beliefs, as well as the affordance and challenge Taiwan offers to these international Chinese learners. Three students from Europe and America who have been learning Chinese in Taiwan from six to twelve months respectively are invited to the study and receive four in-depth interviews within six month. Their personal Chinese language learning experiences over the time are narrated and recorded. The data is then analyzed using Holistic-Content and Categorical-Content method proposed by Lieblich et al., (1998). The results shows that, at different learning stages, the interaction of the participants’ Chinese language goals, expectation and proficiency levels with social environment either reaffirms or shakes their beliefs. Before studying in Taiwan, the past language learning experiences as well as the prevailing view of their native community towards foreign language basically constitutes the beliefs of the participants. At the early stage of studying in Taiwan, the gap between their expectations and the actual communication experiences with local native speakers also has critical impacts on their beliefs. The participants would also find that their efforts to make themselves live as Chinese speakers are constantly challenged by the local native community. In this case, Taiwan and its social community as the field for learning Chinese actually provide more opportunities to the learners with higher Chinese proficiency. When proficiency is low, such as at the beginning stages of Chinese proficiency or upon arrival, the participants have great difficulty getting benefits from the target language community; their learning mainly happens in the language learning classroom. It is clear that language learning beliefs is dynamic, multifaceted, and fluctuate from time to time. The study suggests that, specific topic under the multilayered nature of belief should be focused in the future research. Social environments and emotions of students should also be included in BALLI. Finally, to promote the development and effectiveness of foreign students studying in Taiwan, stronger pre-departure briefing, orientation and counseling services are needed to facilitate language development. Teachers also need training on how to reduce the impact of crashed expectations of new foreign students. In addition, local people (i.e., native speakers) who often come into contact with the learners are recommended to develop intercultural communication skills in order to provide the learners with the best possible informal communication experience.

本研究旨在了解日本學生的語言學習信念變化，並找出哪些因素在台灣和日本的TESOL環境中影響學生的學習信念的形成和變化。不同於傳統基於問卷的調查，只針對某些固定的學生的學習信念特徵的研究，最近的研究更注重學習信念的動態變化過程。社會文化理論還強調，社會文化環境在學生的思考過程中具有中介的效果。本研究旨在了解日本學生的學習信念是如何受他們的學習經驗影響，而這些經驗在不同的環境如何影響他們的學習信念。本研究使用深度訪談收集資料，採用半結構式訪談每個月針對學生的學習經歷和信念進行一學期五次的資料蒐集。通過Liblich(1998)的方法整理出訪談資料的整體與分類的內容，本研究希望找出在台灣和日本學習語言的意識形態的不同，這些差異和參與者的語言學習信念之關係，以及在台灣的日本學生面臨何種語言學習機會和挑戰。 這項研究的結果是，三位參與者經歷了不同程度的語言學習信念轉變，同時形成新的學習信念。來台灣之前，參與者對於自己的英文能力的自信心高低不一，但他們都相信能夠透過海外留學加強英文能力。在台灣，由台灣以及西方學生所組成的團體使參與者能夠練習英文並表達自己的意見，無須擔心被批評。ETP課程給他們一個機會可以使用英文來學習並溝通。來到台灣一年後，Yoshi對英文能力變得較有自信，且相信策略運用比天生能力來的重要。Taka變得較不在意自己的發音且把注意力放在文化學習上。Ko了解自己天生語言學習能力的極限並開始發展自己的學習策略。在研究建議方面，當未來的研究採用BALLI(Belief About Language Learning Inventory)(Horwitz,1999)分析時也應將親密團體和社會文化等因素加入。在英語教育方面，台灣的大學英語培訓課程應該包括更多的跨文化差異教學，所以當國際學生參加一般ETP(English taught program)時他們將面臨較少的文化衝擊。由於參與者提到與台灣和國際學生的非正式交流對她們的英語學習大有幫助，在ETP課程中教授應考量國際學生的英語水平，並把他們分配在具有相應英語水平的小組中，使他們能夠更快克服學習困難。期望這項研究的結果提供語言教學方法的調整建議，以滿足國際學生，特別是日本學生的需求。 / This study was aimed to understand the language learning belief change of Japanese students, and to find out what are the factors in the TESOL environments of Taiwan and Japan that lead to these changes. Unlike the traditional questionnaire-based research, which studied student’s learning belief based on certain critical characteristics; recent studies focused more on the changing process of language learning belief. Socio-cultural theories also stressed that the socio-cultural environment has meditational effects on the student’s thinking process (Furberg, 2009). This study tried to understand how Japanese student’s English learning belief was affected by their learning experiences and how these experiences in the different environments affected their learning belief reciprocally. This study adopted in-depth interviews to collect data, using semi-structured interviews per month for six months to discover how Japanese students’ learning experiences and belief changed in the process. By analyzing the interview data using the holistic-content method proposed by Liblich et al (1998), this study focused on how different language learning ideologies in Taiwan and Japan affected the participants’ learning belief, what opportunities and challenges the Japanese students faced in Taiwan in terms of language learning. Based on the result of this study, the three participants experienced different degrees of language learning belief changes and formed new learning beliefs in the process. Before they came to Taiwan, the participants had different levels of self-confidence and all believed that they could improve their English proficiency by studying abroad. In Taiwan, the supportive community formed by Taiwanese and Western students allowed the participants to practice English and express their ideas without being afraid of getting judged. The ETP (English Taught Program) offered them an opportunity to use English to both study and communicate on a daily basis. After one year of staying in Taiwan, Yoshi became more confident in his English proficiency and believed more in strategy use than innate language learning ability. Taka was less anxious about his Japanese accent and put more emphasis on learning cultural factors. Ko realized the limitation of his innate language learning ability and began to develop his own language learning strategies. As for theoretical implication, factors like community and socio-cultural influence should be added to BALLI’s (Belief About Language Learning Inventory)(Horwitz,1999) model in analyzing language learning belief in future studies. In terms of pedagogical implication, English training programs in Taiwan’s university should provide more cross-cultural information so that when international students take English taught classes they will encounter less cultural shock. As the participants mentioned that the informal interaction with Taiwanese and Western students assisted their English learning, professors in the ETP classes should consider international students’ English proficiency and put them within groups of appropriate English levels so they can learn with less difficulty.