On naming a giraffe a zebra: Picture naming errors across different object categories.

Abstract

Two experiments examined naming to deadline for pictures of objects from categories with structurally similar or dissimilar exemplars. In Experiment 1, a wider range of naming errors were made for objects from structurally similar categories, consistent with those items having a broader set of structural neighbors. Also, only visual + semantic errors were made to structurally similar objects, whereas modal errors for some structurally dissimilar objects were either purely visual or purely semantically related to targets. In Experiment 2, the names of visual + semantic primes occurred as perseverative responses to objects from both category types. The data fit a cascade model of picture naming, in which errors under deadline reflect the rate-limiting processes affecting naming. The naming of clearly drawn, high-contrast line drawings of objects is usually a relatively effortless task. However, such effortless performance masks a number of interlocking component processes, including those involved in accessing different types of stored knowledge. Techniques are needed to prize apart these processes to understand better the nature of the representations and transmission processes underlying picture naming. Consider what happens when, instead of being allowed sufficient time to recognize the object and retrieve its associated name, subjects are forced to respond rapidly, perhaps with the first name that comes to mind. Can the nature of the errors that occur add to one's understanding of the components of the picture naming system? Response deadline techniques have been used previously on various research topics, including studies of choice reaction time (Pachella, 1974), visual attention (Humphreys, 1981), visual word recognition (Antos, 1979), and shape perception (Humphreys & Quinlan, 1988). Such techniques can provide a useful tool for examining the earliest information that becomes available to response processes. In recent work (Vitkovitch & Humphreys, 1991), we have used a speeded naming technique to examine the interference effects of categorically related primes on the naming of picture targets. Errors to pictures reflected mainly the names of other items from the same category and, in particular, those of previously named related (prime) pictures; that is,