No abstract

The present laboratory experiment was inspired by semantic changes reported in etymological studies: Many words that at one time had a positive or neutral connotation became negative when later applied to females. In contrast to this negativity bias, male-associated words underwent a positivity bias. Miller and Swift (1976) referred to the gender-based biases as gender-related semantic polarization. We developed the concept further by distinguishing between gender-stereotypic and gender-non-stereotypic polarizations, which involved, respectively, gender-stereotypic (e.g., dependent-independent) and gender-nonstereotypic (e.g., honest-deceitful) affective connotations. In the experiment, participants learned four-letter invented words in association with male or female pictures; the connotative meanings that had been acquired during the learning process were later measured and related to respondents' gender schema and gender attitude. The results showed that the acquired connotations of words associated with male pictures were more stereotypically masculine than were those associated with female pictures. This gender-stereotypic polarization was independent of respondents' gender schema and attitude. Gender-non-stereotypic polarization, on the other hand, was a function of respondents' gender attitude: On gender-non-stereotypic connotations, respondents who had unfavorable attitudes toward women downgraded, whereas pro-women respondents upgraded, female-associated words relative to male-associated words. The interaction effect showed that semantic change can occur in favor of women, but not always in favor of men.