Two recent lines of research suggest that explicitly naming objects at study influences subsequent memory for those objects at test. Lupyan (2008) suggested that naming "impairs" memory by a representational shift of stored representations of named objects toward the prototype (labeling effect). MacLeod, Gopie, Hourihan, Neary, and Ozubko (2010) and MacLeod, Ozubko, Forrin, and Hourihan (submitted for publication) suggested that naming "enhances" memory by influencing the distinctiveness of named objects (production effect). However, these studies cannot be directly compared because they differ in several procedural details such as the format of the naming task, composition of study objects from different categories, control task, and type of lures used at test. Here we systematically manipulate those factors to better understand how using object names influences visual recognition memory. When objects belonged to unique categories, vocal naming (as used in the production effect) produced comparable memory as a non-naming task (preference rating) and both produced significantly better memory than key-press naming (as used in the labeling effect). Naming objects at study only impaired memory relative to preference rating when objects belonged to one of two categories, a condition in which names have little or no distinctiveness. Theoretically, our results pose challenges to the representational shift account that proposes special mechanisms tied to the use of object names. (Contains 7 tables and 6 figures.)