Agreement features in the nominal domain as classifiers: theoretical and experimental perspectives

Number and gender features are often assumed to differ in agreement processing, yet psycholinguistic evidence does not consistently support this distinction. This paper examines optional agreement in complex NPs (e.g., Italian Un centinaio di senatori si è dimesso / si sono dimessi, “A hundred senators has/have resigned”), where verbs may agree with either the singular head noun or the plural embedded noun. We argue that number and gender operate similarly as feature bundles, functioning like nominal classifiers. Crosslinguistic data show that optional agreement arises from nominal class valuation rather than from inherent gender/number differences. Psycholinguistic experiments further reveal that speakers' choices reflect structural prominence and processing ease, rather than morphological categorical distinctions. Our analysis unifies gender and number within a classifier-based framework, challenging traditional divisions and offering a streamlined account of agreement variability.