20世紀台灣與日本的常見命名漢字與性別意識：詞彙語意學與跨文化之研究

Abstract

人的命名在一個社會文化中扮演非常重要的角色，因為它不僅可能影響一個人的形象，也反應了社會的價值與期望，因此命名的選擇帶有特定的意識形態，而其中性別的概念在名字中尤其顯著（Alford, 1987; Li, 2002）。本研究想探討台灣與日本二十世紀的常見命名漢字與性別所呈現的意識型態以及其歷時演變，並討論兩個文化之間的相同與差異之處。台灣與日本的名字都主要以華語詞彙構成，這些詞彙都擁有多個語意，因此語意在命名中顯得十分重要（Lee, 1998; Lu Millward, 1989; Tang, 2009; Watanabe, 2005），而本研究最主要就是要討論常見命名中詞彙語意的概念。 本論文根據Van Dijk (1995, 1998)所定義的言談（discourse）以及意識形態（ideology）建立了理論模型，將命名視為一種言談，也就是其中一種會影響性別意識形態的語言使用。本研究主要分析台灣與日本名字中的詞彙語意（lexical semantics）並以名字語法風格（syntactic styles）的分析補充詞彙語意分析所忽略的部分。在詞彙語意的分析中，我們將詞彙語意分類並使用兩種分類方式：WordNet類別（WordNet categories）以及社會語言類別（sociolinguistics categories），使用這兩種類別分析能更完整地展現名字在本體論以及社會文化層面中的認知結構，以彌補只使用單一分類方式的不足。此外，語法風格分析欲了解名字中詞彙組合所帶有的性別傾向。由於先前Liu的研究（2018）發現不同性別詞彙的組合會影響命名中的性別傾向，進一步分析該組合能了解女性或男性名字中想傳達的性別傾向。 方法上，本研究選取20世紀（1920年代至2010年代）台灣與日本十大常見男女名字作為分析資料。所有命名用字都被分開處理，先透過問卷請台灣及日本母語人士判斷每個命名用字的性別，因此根據問卷結果，台灣與日本的命名用字及其詞彙語意被分為三個性別組：中性、陰性及陽性。隨後以字典篩選的每個用字所使用的語意會根據問卷的性別分類。之後，每個性別組使用WordNet以及社會語言類別分類，以統計分析不同性別組中各個類別的分佈比例，並使用卡方檢定來判斷不同性別組是否有分布上的顯著差異。語法風格的分析上，本研究觀察台灣與日本男女命名用字性別組成的歷時演變，並以統計分析來計算不同年代所使用各個性別用字比例的變化。 本研究根據分析結果整理了兩大發現：第一，無論於整體或是歷時分析中，台灣與日本常見命名漢字的性別意識形態都展現了性別刻板印象以區分性別，但兩個文化在區分性別時依賴的面向不同，因此就語意分析上台灣的傳統性別刻板印象比日本強。這代表台灣的常見命名漢字更仰賴語意區分性別，而日本除了語意以外，音韻及特殊文化背景也是常見的區分方式。第二、歷時分析上，雖然兩個文化常見命名皆展現中性化趨勢，但該趨勢依然難以改變社會中性別不平等的事實。無論男女名，台灣命名的中性化更強調陽性特質的價值；而日本男性名的中性化則不夠有力削弱性別二元化的現象，因為日本女性名非但但沒展現中性化而更強調傳統女性刻板印象。 此研究有兩個主要貢獻，首先，在方法上結合了Van Dijk的論述與意識形態的概念以及詞彙語意的分析，補足過去詞彙語意的研究主要只對語言本身而較少應用在意識形態分析上的不足，而使用分類且量化的方式可以幫助我們處理更大量且詳細的語料。第二，本論文的研究結果補足過去主要只聚焦在台灣與日本個別的研究，並沒有跨文化與歷時演變的比較。文化上，本研究發現台灣與日本常見命名漢字的性別意識形態的展現有文化差異，因為兩者在性別區分上所依賴的面相不同，以至於結果發現語意上日本命名的性別刻板印象較台灣低。歷時分析上，雖然過去研究都認為兩個文化的常見命名有中性化的現象，但幾乎沒有討論中性化的影響與其機制。透過本論文的分析與討論，希望可以讓人更了解命名對於社會的影響力以及台灣與日本的社會狀況，並能將命名作為一種社會實踐方式來創造更多元包容的社會。

Personal naming plays an essential role in society since it can not only influence personal images but also reflect the values and expectations of any society. Hence, the choices of personal names express ideologies, while gender is a particularly essential concept in the names (Alford, 1987; Li, 2002). The present research attempts to investigate the gender ideologies and diachronic development of popular Chinese characters in personal naming and gender ideologies in Taiwan and Japan in the 20th century. Taiwanese and most Japanese personal names are constructed with Chinese characters. Since these characters carry multiple lexical senses, semantics has significant importance in personal naming (Lee, 1998; Lu Millward, 1989; Tang, 2009; Watanabe, 2005). Hence, this thesis mainly discusses the lexical semantics in personal names. According to Van Dijk’s (1995, 1998) definition of discourse and ideology, this study constructs a theoretical model that sees personal naming as a discourse, i.e., a kind of language use, that may influence gender ideology. This thesis primarily conducts a lexical-semantic analysis with a supportive analysis of syntactic styles to compensate for neglected parts of the previous investigations. In the lexical-semantic analysis, two kinds of classification systems are adopted to categorize lexical senses: WordNet categories and sociolinguistics categories. Using both systems can show more explicit cognitive structures of personal names in the ontological and social aspects. Furthermore, the analysis of syntactic styles presents the gender bias carried by gendered-characters collocations in the names. In Liu’s study (2018), it has been found that different gendered characters’ compositions may affect the distinctive gender bias in the names, and the preferred gender biases in women and men’s names can be understood. Methodologically, the top ten most common names for women and men in Taiwan and Japan from the 1920s to 2010s are selected as the research data. Each character in the names is dealt with separately. First, Taiwanese and Japanese native speakers are asked to identify the gender of each character through questionnaires. Based on the questionnaire results, Taiwanese and Japanese naming characters and their senses are classified into three gender groups: neuter, feminine, and masculine. In addition, the lexical senses of each character are selected by dictionaries according to specific criteria. Next, each gender group’s lexical senses are categorized by WordNet and sociolinguistic categories. Statistical analysis is conducted to understand the distribution percentage of the categories in three gender groups in overall data and diachronic development analyses. Additionally, a chi-square test is carried out to test the significant differences among the three gender groups. In syntactic style analysis, the compositions of gendered naming characters in Taiwanese and Japanese women and men’s names are examined diachronically by calculating the percentage of each character’s gender group at different time periods and observing their development. Two main findings are presented: First, both Taiwanese and Japanese names display certain gender stereotypes to distinguish genders in overall and diachronic analyses but with different gender differentiation methods. Semantic analysis has shown that the traditional stereotypes are more robust in Taiwan than in Japan since Taiwanese names rely more on semantic meanings, while phonology and specific cultural backgrounds are also evident features in Japanese names, besides semantics. Second, in diachronic analysis, the trend of neutralization in both cultures still does not seem to change gender inequality in modern times. Taiwanese neutralization of personal names focuses on the value of masculinity more, whereas Japanese male names’ neutralization may not be strong enough to reduce gender binary since Japanese female names contrastively show the emphasis on stereotypes rather than neutralization. This research has two main contributions. First, methodologically, we integrate Van Dijk’s idea of discourse and ideology with lexical-semantic analysis. Since most of the past lexical semantics studies largely focus on the language itself and seldom examine ideology, this thesis compensates for this gap. Moreover, adopting the method of senses’ categorization provides a way to quantify a large number of lexical senses. Second, this study’s cultural and diachronic comparisons build upon previous studies, which only investigate Taiwanese and Japanese personal names independently of each other. Culturally, different expressions of gender ideologies between the two cultures’ popular naming characters are found. Since two cultures depend on distinctive aspects to distinguish gender, stereotypes are less intense in Japan than in Taiwan with semantic analysis. Diachronically, although previous studies consider both cultures to reflect the phenomenon of neutralization in common personal names, its influences have not been discussed. Through the analysis and discussion of this study, we hope that people can understand the influences of personal naming toward gender ideologies in contemporary Taiwanese and Japanese society, as well as consider personal naming as a social practice to create a society that can hold diverse traits of gender.