Abstract
Previous eye-tracking research has shown that, during spoken-word recognition, gender marking on preceding articles restricts the competitor set to gender-matching nouns: Upon hearing “Cliquez sur le[masc] bouton” (‘Click on the button’), French listeners did not take the picture of a gender-mismatching ‘bottle’ (bouteille[fem]) into consideration, despite onset similarity between bouton and bouteille (Dahan et al., 2000) In the interpretation of the gender effect, two issues need to be distinguished: When does gender information infl uence noun recognition? Do gender cues pre-activate gendermatching nouns, or does the effect set in when the noun onset is heard? At what level of processing does gender have an effect?
What types of representations are involved: shallow co-occurence frequencies or deeper morphosyntactic gender categories? In another experiment, Dahan et al. (2000) found that,
by itself, a gender marked article does not seem to prime all gender-matching nouns: Hearing “la[fem] louche” (‘the ladle’) did not increase fixations to a non-onsetoverlapping ‘sock’, chaussette[fem]. This goes against a pre-activation account.
