Investigation into the effect of homophone training on reaction times for a forced choice lexical decision task

Authors Avatar
Investigation into the effect of homophone training on reaction times for a forced choice lexical decision task

Abstract

The study investigated the effect of training with homophones on the pseudohomophone effect when participants were required to search their lexicon for a familiar letter string, this also investigated the ideas put forward by the dual route model where orthographic and phonological processes are both used in the analysis of word strings. The design of the experiment was a between subjects forced choice lexical decision task, where participants were shown two word strings simultaneously and asked to respond as to which was a correct word. Participants were students from the University of Nottingham split into two different groups that were subject too different training conditions. Stimuli were four letter single syllabale word strings including homophones, regular words, non word strings, and pseudohomophones as used in previous research by Underwood (1988). The results obtained did not show a significant difference in reaction times between the two conditions although further analysis did show that the pseudohomophone effect was present. The study concludes that although the results were not significant at a high confidence level they are still positive in supporting the ideas of earlier studies including the dual route model and the important roles of both orthographic and phonological processes in word recognition.

Introduction

A pseudohomophone is a letter string that looks and sounds like a word such as "Bild" (a non word that sounds like a real world, build). The pseudohomophone effect says that it will take longer to distinguish between a real word and a pseudohomophone than to distinguish between a real word and a non word, such as "jate", which is not a pseudohomophone. Rubenstein et al (1971) presented participants with a word that was either a non word, a pseudohomophone or a real world for 2 seconds and asked them to respond as to whether they had just seen a real word or a non word. The results showed the pseudohomophone effect and Rubenstein (1971) suggested that this occurrs because we use a phonological code when deciphering words. The code looks at the phenomes (the smallest units of sound a word can be broken into) to see if the word is phonologically viable for pronunciation. If a phonological match has occurred the model goes onto suggest that our internal dictionary (known as the lexicon) is then searched using an orthographic check for a visual match of the word. Rubenstein et al's (1971) model with phonological then orthographic checks occurring in series means that non words are rejected earlier at the first check and do not proceed to the second. Whereas pseudohomophones pass the brains phonological code and are only rejected after being processed orthographically, causing longer response times when deciding if a pseudohomophone is a word or not.

Coltheart et al (1977) proposed a different model called the dual route model suggesting that orthographic and phonological checks are of equal importance. In their model it is proposed that we use visual cues such as graphemes (smallest units of orthographically recognition in a word such as "ed") if these cues are familiar we may except that it is a word and not need to check its sounding, such as words that are used often. However on other occasions when the word is not so familiar phonological checks are carried out earlier in processing. Coltheart et al (1977) also proposed that after the word has been processed both visually and phonologically if no recognition is made within a certain period of time then the brain will decide that it is not a word. This "time limit" could be extended, giving a slower reaction time if the word is found to be similar to an entry already in the lexicon and explaining the pseudohomophone effect.

Martin (1982) proposed that previous tests of the pseudohomophone effect conducted by Rubenstein (1971) were flawed as they do not take into account that there are more than two categories of non word. He said there was another type of non word, orthographically legal non words, which look like real words but don't sound like them. In his experiments Martin (1982) showed that when letter strings of orthographically legal non words were used in experiments demonstrating the pseudohomophone effect the effect disappeared.
Join now!


Underwood (1988) felt that as Rubenstein (1971) used homophones (words that only differed in one way in spelling or sound) in his experiment and Martin (1982) did not this could have effected the results. Underwood looked at the effects of having two training blocks doing the same experiment. One block was trained using homophones, which encouraged them to think about the sounds of words priming them phonologically. The other block was trained using regular non words with no homophones encouraging them to focus more on how the word was presented visually. This primed them to use more orthographic ...

This is a preview of the whole essay