本研究以台灣中級日語學習者（76名）為調查對象，實施尊敬語與謙讓語的替換測驗。藉此，探討他們於尊敬語與謙讓語學習上的問題點。具體探討的問題為誤用的傾向及是否能正確的區分使用尊敬語與謙讓語。其結果可整理如下列四點。 (1)從學習者的回答中觀察到了498個誤用例子。從該些498個例子，歸納出了33種誤用，可將其分成6個種類。而當中，直接和尊敬語與謙讓語的區分使用有關的有2個種類，17種誤用，一共為420個例子。 (2)謙讓語的正答率比尊敬語來得低。\r (3)不管是尊敬語，或是謙讓語，學習者有優先使用「特別型態之敬語語彙」的傾向。 (4)學習者似乎無法正確區分使用「謙譲語A」（需要有動作接受者才可使用），及「謙譲語B」（不需動作接受者也可使用）。\r 從本研究的結果，特別是種類繁多的學習者的誤用例子，可再次確認敬語的學習是非常不容易的。尤其對學習者而言謙讓語的學習似乎較為困難，今後有必要加強該方面的指導。 This study examines the problems in learning honorific expressions (Sonkeigo and Kenjogo) among Taiwanese intermediate Japanese learners. It is based on the substitute-test (20 questions) about Sonkeigo and Kenjogo for 76 learners, to research the tendency of errors and the usage about Sonkeigo and Kenjogo. As a result, the following four points were clarified: (1) 498 pieces of error were observed from the learners' answers. These 498 pieces of error examples can be sorted into 33 patterns, and classified into 6 categories. The errors directly related to the usage of Sonkeigo and Kenjogo were shown in 2 categories, 17 patterns, containing 420 examples. (2) The correct answer rate of Kenjogo was lower than that of Sonkeigo. (3) Both Sonkeigo and Kenjogo, the learners have the tendency of using the special honorific vocabulary (e.g. IRASSYAIMASU) as priority. (4) The usage of “Kenjogo A” and “Kenjogo B” could not be done properly. From this study, it especially shows learners' error patterns extended to many patterns. It can be reconfirmed that it is not easy to learn the honorific expressions. Though, according to the result of this study, it is unable to deny that there is limited aspect, hopefully it will be helpful on the honorific education in the future.