台灣迷因文化的幽默與冒犯：解碼社群媒體吸睛關鍵

Abstract

隨著迷因成為網路傳播的重要形式，許多研究陸續揭示幽默與其熱門程度之間的可能關聯。但針對台灣迷因語境，幽默與冒犯如何交織影響其傳播力，迄今仍缺乏實證。特別是冒犯是否有助於或阻礙迷因傳播，目前文獻仍未有共識。本研究旨在 (1) 驗證幽默與冒犯兩種情緒與迷因熱門度（按讚數）的關係；(2) 從社會互動 (Hyland, 2005b)、文本分析、多模態符號學層面挖掘台灣熱門迷因中兩情感的語言學意涵。研究先自Instagram爬取迷因貼文，整理出平衡熱門度和定量冒犯比例的台灣繁中迷因語料庫。幽默和冒犯的評分則來自大眾的問卷填答，作者從語料庫中隨機抽取兩千多筆迷因製成線上問卷，控制每份問卷的熱門、冒犯迷因數量及個別填答人次，最終取得2,179筆有效迷因樣本。分析（一）使用問卷收集來的迷因探討幽默、冒犯和熱門度的關聯。分析（二）進一步篩選出圖文分離且文本類型單一（分為「片語」、「子句」、「句子」、「段落」、「對話」五類）的熱門迷因，共626張，細究迷因的幽默和冒犯語言學特徵。分析（一）結果顯示，僅幽默和熱門度呈現正相關，冒犯則無論是一次還是二次項皆無顯著關聯，兩者呼應了前人研究成果。分析（二）發現，在台灣迷因文本中，某些互動訊號傾向與具有幽默感或冒犯性的熱門迷因同時出現。若迷因創作者希望提升觀眾的共鳴效果，宜優先考量「自我揭露」與「訴諸共同知識」等互動性線索。反之，若傾向避免引發非預期的冒犯，則建議審慎檢視「對話」體裁的迷因，尤其是含有「內梗」的內容呈現。此研究貢獻在於（1）首度以純台灣繁中迷因檢驗幽默、冒犯與熱門度之關係；（2）建立涵蓋情緒評分與語言標記的台灣迷因語料庫，為後續多模態語言學與情感傳播研究提供基礎；（3）為迷因創作者與平台管理者提出互動策略與冒犯風險控管之實務建議。In recent years, numerous studies have examined meme humor in tandem with popularity; however, in the Taiwanese context, empirical evidence on how humor and offensiveness influence meme popularity is still limited, and the literature is polarized divide on whether offensiveness boosts or dampens popularity. This study therefore aims to (1) verify the relationships between the two emotions, humor and offensiveness, and meme popularity (measured by “likes”); and (2) uncover the linguistic meanings of these emotions in popular Taiwanese memes from the perspectives of social interaction (Hyland, 2005b), textual analysis, and multimodal semiotics. We first scraped meme posts from Instagram to compile a Taiwan Traditional Chinese memes corpus, balanced for popularity to capture humor variation and including a modest proportion of offensive memes. Ratings of humor and offensiveness were obtained through public questionnaires: more than 2,000 memes were randomly selected from the corpus and distributed across online surveys, with the numbers of popular and offensive memes and the final respondent counts controlled for each survey; 2,179 valid samples were ultimately collected. Analysis 1 used the questionnaire data to examine the associations among humor, offensiveness, and popularity. Analysis 2 further filtered 626 popular memes whose images and texts were separated and whose texts belonged to one of the following text types: phrase(s), clause(s), sentence, paragraph, and conversation. We then investigated the linguistic features of humor and offensiveness in detail. Results of Analysis 1 indicate that only humor positively correlates with popularity, whereas offensiveness, whether modeled linearly or quadratically, shows no significant association, echoing earlier findings. Analysis 2 suggests that certain interactional signals tend to cooccur with humorous and offensive popular memes within Taiwanese texts. Meme creators seeking greater audience resonance could consider prioritizing interactional cues such as self-disclosure and appeals to shared knowledge; conversely, if avoiding unintended offense is a priority, it may be prudent to review conversational genres, especially those embedded with insider references. The contributions of this study are threefold: (1) it is the first to test the relationships among humor, offensiveness, and popularity in Taiwanese Traditional-Chinese memes; (2) it establishes a Taiwanese meme corpus containing emotion ratings and linguistic annotations, providing a foundation for future research in multimodal linguistics and affective communication; and (3) it offers practical recommendations on interactional strategies and offense-risk management for meme creators and platform administrators.