Acknowledgements

The authors would like to thank Joyce Haisma for lending her voice to the present experiment. We are grateful to Michael K. Tanenhaus and his colleagues for helpful discussions of this work and useful advice concerning data analysis. NEUTER IS NOT COMMON IN DUTCH 3 Native speakers of languages with transparent gender systems can use gender cues to anticipate upcoming words. To examine whether this also holds true for a non-transparent two-way gender system, i.e. Dutch, eye movements were monitored as participants followed spoken instructions to click on one of four displayed items on a screen (e.g., Klik op deCOM rode appelCOM, ‘Click on theCOM red appleCOM’). The items contained the target, a colour-and/or gender-matching competitor, and two unrelated distractors. A mixed-effects regression analysis revealed that the presence of a colour-matching and/or gender-matching competitor significantly slowed the process of finding the target. The gender effect, however, was only observed for common nouns, reflecting the fact that neuter gender-marking cannot disambiguate as all Dutch nouns become neuter when used as diminutives. The gender effec