There have been contradictory reports of sex differences in language processing. A novel approach is adopted here which explores the experiential basis of such differences. Two studies examine the auditory processing of grammatical gender in Bulgarian in a gender decision (gender monitoring) task and a cued shadowing (word repetition) task. Reaction times in both experiments reveal significant two-way interactions between the grammatical gender of words (masculine vs. feminine) and the sex of the voice (male vs. female). The sex of participants in the gender decision task also interacted with word gender in terms of decision accuracy. Women were relatively more accurate on their “own reference” word gender (feminine) and less accurate on masculine gender words. A two-way interaction between word gender and participant sex on response latencies in the cued shadowing task supports the view that these effects are not strategic but have a highly automatic nature instead. Findings are interpreted in terms of individual differences in the experience of grammatical gender in such gender-marking languages.