Across three experiments (Gygax et al., 2008; Gabriel et al., submitted; Sarrasin et al., in prep.) conducted in three different languages (English, French and German) the GREL group has documented the dominance of grammatical gender in the mental representation of gender during reading comprehension. More specifically, GREL has examined the joint influence of two potential biasing sources, namely the use of the masculine form, supposedly interpretable as a generic form, in gender marked languages (French and German) and gender-stereotypical information, when inferring gender from role names. The findings suggest that when grammatical gender cues are available, readers use them to construct a mental representation of gender. When not available, readers rely on stereotypical information. There are some exceptions, but these are few. Sarrasin et al. (in prep) and Gabriel et al. (submitted) examined possible sources that could attenuate the impact of the use of the masculine form by increasing the stereotypicality of the role names (Sarrasin et al., in prep) and by providing readers with additional and different grammatical cues (Gabriel et al., in prep.). Both studies showed that the bias introduced by the masculine form was rather strong and difficult to attenuate. Gygax and Gabriel (2008) even identified cases in which the bias was strengthened. At least, the latter study, although demonstrating an undesirable effect even more discriminatory to women, showed that the representation of gender was somehow flexible.In this present project, we propose to manipulate the conditions that may constrain both grammatical and stereotypical sources and result in a mixed representation of gender, which can be considered as a gender-fair representation. In a sense, we are trying to see if readers may be able to consider the masculine form as a generic one. We propose four experiments in French, based on Gygax and Gabriel (2008), in which different manipulations are implemented to force readers into a true generic interpretation of the masculine form. Compared to the work that has been conducted so far by the GREL group, this project raises more applied issues. If past research was primarily aimed at exploring the different biases that were responsible for the construction of a mental representation of gender, this project provides us with (a) clear indications of the flexibility of those biases and (b) directions that may be undertaken to modify those biases, in intervention programs, for example.