“…Formally an antecedent is generic wherever the gender of the antecedent is unknown. However, extensive research has shown that antecedents that have a strong gender stereotype might be formally generic but they are not treated as gender-neutral during online language processing (e.g., Esaulova, Reali, & von Stockhausen, 2013;Garnham, Oakhill, & Reynolds, 2002;Kreiner et al, 2008;Oakhill, Garnham, & Reynolds, 2005;Reynolds, Garnham, & Oakhill, 2006;Sturt, 2003). At an intuitive level it follows that there might be a scale of "generic-ness" based on the realworld likelihood, or probability that an antecedent has a specific gender, with low-expectancy antecedents (a cyclist) being treated as highly generic and high-expectancy antecedents (a mechanic) being treated as somewhat generic.…”