近年來，以語料庫為基礎的英文研究對其語言描述及教學帶來大量嶄新的啟發。英語流行歌曲及歌詞經常應用在各級英文課堂上，但使用有系統之語料庫研究尚不多見；少數前人研究使用的歌詞素材也僅及於2015年前發行之歌曲。本研究分析較新之英文流行樂曲歌詞的詞彙面向，以提供英語教師教學建議。我們自美國告示牌流行音樂榜單中挑選出479首英文歌曲的歌詞(取樣分布於2007~2017年, 約二十二萬字)，根據內容分成八類；並使用電腦軟體協助分析單字難易度及找出常見三字片語，輔以人工檢視。其結果根據字表研究顯示當學習者熟知常見英文3,000字時，對本語料庫歌曲單字能達95.43%理解程度；若識字數提升至8,000字時，則能了解近97.93%內容(比Tegge, 2017發現稍難)。我們同時找出常見三字片語(例如I don’t, I can’t, 和 you don’t)；也含常用的口語化非正式用字(wanna, gonna)及書寫手法(‘cause, *in’)。歌曲語料庫中所含的女性及男性名詞，和前人研究比較結果發現兩者常用性別名詞的數量排序近似，但我們資料較無先前的研究多樣。綜合以上，本文納入較新英文流行樂歌詞，經語料分析479首歌詞後之整體語言特徵展現當代英文流行歌詞特有的口語化文本特質。

Recent corpus research has provided much evidence guiding the description of the English language and pedagogy. English song lyrics as a corpus have received limited attention among researchers (including older songs released only before 2015) in spite of the observation that they are a common source of pedagogical materials for language teaching. The current study aimed to conduct corpus analyses on song lyrics in order to offer pedagogical implications to English teachers. We adopted a corpus-based approach by compiling 479 song lyrics taken from Billboard as the collection (with sampling years 2007 to 2017, about 220,000 words), and classified them into eight types. Lexical coverage, phrasal units, informal contractions, writing conventions, and gendered nouns were also analyzed using a combination of automatic processing and manual examination. We found that the knowledge of 3,000 English word families based on vocabulary research of wordlists could reach 95.43% lexical coverage for comprehension in the song corpus, and with 8,000 words, the comprehension coverage would rise to 97.93%, slightly more difficult than Tegge (2017). More frequent three-word phrases were located in our song corpus (I don’t, I can’t, or you don’t). Further, informal contractions (wanna, gonna) and writing conventions (‘cause, *in’) of top frequencies were also identified. Concerning lexical gendered nouns, our findings of frequency orders were similar to prior research, but the variety was not as many. Overall, the linguistic make-up of these 479 song lyrics shows the unique spoken corpus features of updated English songs.