Neural Evidence That Gender-Inclusive Language Attenuates Automatic Gender Predictions

Abstract

Languages vary in whether and how they encode gender. As societies pursue greater inclusiveness, gender-inclusive language (GIL) has emerged as a debated innovation, introducing pronouns and morphological strategies to neutralize grammatical gender (e.g., todxs, for todos/todas, "all"). Yet whether GIL actually reduces gender expectations remains empirically underexplored. Evidence at the neural level is especially scarce. To address this gap, we asked Italian speakers to categorize masculine (lui, "he") and feminine (lei, "she") pronouns following nouns with canonical (masculine -o, feminine -a) or neutralized endings (-*, -{schwa}). Noun gender-stereotypical association was independently manipulated (male-oriented, e.g., chirurgo/a/{schwa}/*, "surgeon"; female-oriented, e.g., maestra/o/{schwa}/*, "teacher"; neutral, e.g., vicina/o/{schwa}/*, "neighbor"). Recording response times and event-related potentials (ERPs) to target pronouns allowed us to assess their automatic integration and processing costs. Gender-incongruent pronouns after canonical forms elicited longer response times and larger N400 and P300 amplitudes, reflecting increased processing difficulty and reanalysis. Pronouns following GIL forms showed an intermediate processing profile. Response times fell between consistent and inconsistent conditions, N400 amplitude was larger than in the consistent condition, and P300 amplitude was reduced compared to the inconsistent condition. Together these findings indicate that both pronouns remained viable interpretations and could ultimately be integrated. Stereotype-based knowledge shaped early processing without preventing integration, predominantly for masculine pronouns. Overall, our findings provide the first neural evidence that, despite processing costs, GIL forms can foster gender-neutral representations, with relevance to ongoing societal debates and broader cross-linguistic implications.