Testing Gender Markedness of Nouns with Self a Paced Reading Study
- Ethan Wilcox, Department of Linguistics, Harvard University, Cambridge, Massachusetts, United States
AbstractSome English nouns occur in gender-marked pairs, which fall into two classes: In the Superordinate class, the unmarked (masculine) form is available to refer to female referents ("Allison Janey is a good actor"), whereas in the specific class it is not (*"Diana is a good prince"). Two theories account for this alternation: The Featural Theory proposes that the unmarked are unspecified for gender features. The second, Frequency Theory proposes relative frequency of the marked vs. unmarked forms are responsible (Haspelmath, 2006). This work provides evidence against the frequency theory by employing a self-paced reading study that tests relative processing times of anaphoric pronouns referring to gendered nouns. If noun pairs are split along Specific/Superordinate class lines, a processing slowdown is found for processing processing pronoun gender mismatches, except for nouns like 'actor', as expected. However, when the noun pairs are split by relative frequency the effect disappears.