From words to gender: Quantitative analysis of body part descriptions within literature in Portuguese

• Published in: Information processing & management 2024-05, Vol.61 (3), p.103647, Article 103647
• Main Authors: Silva, Mariana O., Melo-Gomes, Luiza de, Moro, Mirella M.
• Format: Article
• Language: English
• Subjects: Data mining; Gender representation; Literature in Portuguese; Natural language processing
• ISSN: 0306-4573; 1873-5371
• DOI: 10.1016/j.ipm.2024.103647

This article presents a quantitative analysis of gender representation within literature in Portuguese, focusing on the descriptions of male and female body parts. We investigate a corpus of 34 literary works from our 80,000-sized dataset. By leveraging Natural Language Processing techniques, we analyze over 50 body part descriptions of 315 unique characters identified through predetermined lists from Wikipedia and Todo Estudo. To assess gender, we consider two different gender detection approaches that achieve F1 scores above 90%. Overall, our analyses quantify the frequency, specificity, and objectification of body part descriptions and provide empirical evidence of gender portrayal in literature written in Portuguese. The findings reveal specific differences in the frequency and choice of adjectives used for male and female body parts, shedding light on prevalent gender stereotypes in literary works. This research advances the discourse on gender representation, employing quantitative methods to expand our understanding of gender dynamics within a distinct literary dataset. It may further serve as a resource for gender studies, literature analysis, and computational linguistics. •Quantitative analysis of male and female body part descriptions in Portuguese literature.•Natural Language Processing techniques to analyze a diverse corpus of Portuguese literary works.•Identification of potential biases within portrayals of male and female characters.•Streamlined character identification process through predetermined lists.•Contribution to the literature on gender representation and challenging stereotypes.