Diachronic Shifts in Gendered Occupational Associations: A Distributional Semantic Analysis of English Occupational Nouns, 1950–2020

Abstract

This study employs distributional semantic methods to investigate the evolution of gendered associations in English occupational nouns between 1950 and 2020. Situated at the intersection of diachronic corpus linguistics, gender studies, and computational semantics, the research examines how linguistic representations of occupations have changed alongside broader social transformations in gender roles and labor market participation. The analysis is based on diachronic word embeddings trained on a 10-billion-word corpus compiled from Google Books and the Corpus of Historical American English (COHA). Eighty occupational nouns were selected across three prestige categories: high-prestige occupations (e.g., surgeon, professor, lawyer, engineer), mid-prestige occupations (e.g., teacher, journalist, accountant, police officer), and low-prestige occupations (e.g., nurse, cleaner, secretary, waitress). The study measures cosine similarity between occupational noun vectors and gender-associated lexical vectors constructed from male and female pronouns and role nouns. These semantic associations are tracked decade by decade to identify patterns of continuity and change in occupational gender representation. The findings reveal substantial reductions in masculine bias among high-prestige professions, particularly from the 1970s onward, reflecting increasing female participation in professional and managerial fields. Mid-prestige occupations exhibit moderate shifts toward gender neutrality, while low-prestige and traditionally gendered service occupations demonstrate slower rates of semantic change and retain stronger gender associations. The results suggest that linguistic representations of occupations evolve unevenly across occupational categories, mirroring broader social and institutional patterns of gender integration. The study contributes to methodological developments in distributional semantics by demonstrating the value of diachronic word embeddings for tracking long-term ideological change. It also provides empirical evidence for the relationship between language, occupational identity, and shifting gender norms in modern English-speaking societies.