Perception

Perception

Abstract

The experiment was designed to investigate the models proposed by Humphreys, Riddoch and Quinlan (1988) in respect of visual object processing. The experiment was based on the premis that participants would take longer to name visually presented objects whose characteristics were structurally similar compared to structurally distinct. We did not find evidence to support the cascade or sequential models for visually presented object naming.

Introduction

This investigation aims to repeat work carried out by Humphreys, Riddoch and Quinlan (1988) on the distinctions made in visual recognition and naming of real objects. These researchers suggest that real objects can be classified via visual characteristics. They propose a distinction in categories with members whose characteristics are structurally similar (SS) in one group and those whose characteristics are structurally distinct (SD) in another group. It is suggested that the way these types of objects are recognised implies differences in the recognition process itself. And therefore we can provide evidence for the theory that object naming involves a more complex cascade model than the initial theory they proposed, that of non-interactive sequential processing.

Models of object naming suggested by Humphreys et al. began with the no interactive sequential model. This model proposed that information from an object when first presented for naming would consist of firstly its structural characteristics, then semantics of the object and finally its name. Different studies have shown that information about an object, such as its name only become available when earlier processing stages are complete (e.g. Potter and Faulconer, 1975; Riddoch and Humphreys,1988). Potter and Faulconer (1975) showed that participants can access information(semantic) relevant to an objects category much faster than its name. This is considered to be evidence to support a multistage model for object recognition. This sequential model has discrete stages. However this model does not explain a finding from neuropsychology patients. Accounts from Humphreys et al. patients found that some had selective impairment of object naming. They had more difficulty with some objects compared to others. Humphreys therefore suggested an alteration to the model in order to account for this. They say that there are broadly two groups to which all real object belong. Those with a visual prototype whose members are structurally similar visually, and those without a visual prototype, whose members are structurally distinct. The implication of these categories suggests that naming object from the SS group will be slower than those of the structurally distinct group. Humphreys model proposed that the stages for recognition and naming interact with each other; that the process involves a continuous

stream of information pertaining to the object; and that the process are not in a strict order. The cascade model as it is known works via excitation and inhibition of activated/suggested information about the object and its context from memory. Many excitations may occur at once but these are assessed and either used or inhibited, until the object name has been found. Comparisons of many objects fitting the descriptions given are trialled until a match is found. This model and the differentiation of objects on structural characteristic offers a possible explanation to the selective impairment of object naming found in some patients. On the basis of these categories selective dissociation patients appear to have most difficulty with structurally similar objects. They have access to the semantics of the object but fail to name the object. Using Humphreys cascade model they either have to many options presented from the cascade and cannot find a fit or they simply are not offered the name.

However many other factors could account for the naming latency or failure, for instance frequency of the object, the number of syllables, and orthographic regularity.

We have examined the differences in object naming latencies in the context of a picture naming task to see whether there is a difference in the time it takes for participants to name objects from structurally similar categories and structurally distinct categories. We expected that it would take participants more time to name objects from SS groups than SD groups. We attempted to examine the possibility that an alternative explanation could account for naming latencies seen in Humphreys