Date of Award

1997

Document Type

Thesis

Publisher

Edith Cowan University

Degree Name

Bachelor of Arts Honours

Faculty

Faculty of Health and Human Sciences

First Supervisor

Don Thomson

Abstract

Visual word recognition studies rely on priming tasks to examine underlying processes within the lexical system. A commonly used method is the lexical decision task, where participants are presented with a letter string that is either a familiar word or a meaningless non word such as fost. Response times are measured for the time taken to decide if the letter string is a word or a non word. The word the participant responds to is the target, while the preceding word is referred to as the prime. There are three types of priming conditions reported here. First, semantic priming where a target in the pair chair-table is recognised faster than the target in the pair horse-table. Semantic priming studies are considered to reflect later processes in word recognition, which can occur after primes have been identified. A second paradigm is orthographic priming, where the target in a word pair sharing letters, such as fable-table, is recognised faster than the target item in the control pair shoot-table, in which no letters overlap. Orthographic priming appears to be more robust in a masked condition. That is, the prime stimulus is presented so briefly and in close proximity to other visual features that it cannot be readily recognised. The reason targets in orthographically related pairs are more likely to be facilitated when the prime is masked is unresolved. This work addresses this question by examining what effect the mask can have on the processing of the prime. There are two opposing views. Firstly, it is assumed word recognition occurs over time, and when a mask appears shortly after a word has been presented, further processing of the prime immediately ceases. However, because the prime has already been perceived by the system to some degree, it is said to be partially activated. This partial activation can persist for a brief period of time, but later processes of recognition do not occur, and the word is never identified by the lexical system. This is referred to as an interruption theory of masking. The alternate account suggests the mask does not disrupt the processing of the prime, rather, it affects the ability a person has in consciously reporting the primes presence. That is, the word may have been identified by the lexical system, but it has not been identified by the conscious system. Determining the true effect of a mask has proved difficult. There are many parameters within the existing models of word recognition that are yet to be accurately identified and described. With a large volume of data from a vast array of different priming designs, theory testing is likely to remain a slow process. This paper aims to take a unique approach of examining both orthographic and semantic priming within the same design, which are considered here to be somewhat opposing forces. Unexpectedly, no orthographic priming was found in a design previously showing a robust effect. The results are examined in terms of an interactive activation model, where an interruption account of the prime did not appear to be supported. An expected result was obtained however in that a semantic priming effect was not found in the masked condition. Subsequent tests attempted to obtain a semantic effect while looking at the relationship between semantic priming and conscious awareness of the primes. The study highlights some of the difficulties in making an unbiased assessment of the "participants" ability to detect masked primes.

Share

 
COinS