한국어 신문 읽기에서 성별추론

Gender Inference in Korean Newspaper Reading

This paper investigated how gender inference is generated in Korean newspaper reading, in which there are no specific gender-markers except for gender stereotyped personal names and occupations. Three experiments were conducted to see how these two factors actually work in gender inference. In experiment 1, to check stereotyped gender information for personal names of native Koreans, I collected 20 personal names from different languages (written in Korean) from daily newspapers and asked participants about the gender they guessed intuitively. While most of the Korean names were distinctively divided into female and male, other foreign names written in Korean showed unstable division. In experiment 2, participants were instructed to read a newspaper article where a gender stereotyped occupation, 'politician(the spoke person of Grand Nation Party)' was included and to recall the gender of the person. The result showed that 68% of the participants recalled the politician as a male even though she was referred with a female name which was assessed as a female name at 98% in experiment 1. It reflected the strong effect of a stereotyped occupation in gender inference. In experiment 3, three foreign names were respectively combined with different occupations and were asked to check the gender. The result indicated that there was no difference between with and without a name in judging gender. It gave me the evidence that personal names are not reliable gender-marker when they comes with a strongly stereotyped occupation which is common in newspaper article, and in the case, gender inference is likely based on the background knowledge about the stereotyped occupations stored in a reader's mental model.