It has been proposed that language is more strongly lateralized in males than in females. Recent imaging studies, however, have yielded a variety of seemingly contradictory observations. Here, we categorize these observations into three groups: (1) studies that employed sub-lexical tasks applicable to nonwords, which found sex-differences in the anterior language areas; (2) studies that employed tasks applicable to real individual words, which reported lateralized activation in both sexes (and thus no sex-differences); and (3) studies that employed passive listening to stories with a global language structure, which found clear sex-differences in the posterior language areas. We suggest that these differences in observations are explained, at least in part, by the amount of time demanded relative to the interhemispheric conduction delay.