COGNITIVE PROCESSES IN LANGUAGE COMPREHENSION : MORPHOLOGICAL AND ORTHOGRAPHIC PROCESSING OF ENGLISH COMPOUND WORDS

Abstract

The aim of this study was to investigate the nature of lexical access and representation of words in the mental lexicon. Specifically, the study investigated the morphological and orthographic processing of English compound words in order to elucidate the processes involved in lexical access. Past research has indicated that access to the representations of transparent and opaque compounds in the mental lexicon are not the same. Transparent compounds such as doorbell appear to undergo decomposition during processing but not opaque compounds such as cocktail. It was hypothesised that the lexical access codes for transparent compounds are their constituents, while the access code for opaque compounds is the whole compound itself. Experiment 1 was conducted to obtain constituent-compound relatedness ratings on a list of compound words, of which 48 were chosen as experimental stimuli after controlling for word frequency, word length, and syllable length. Three other experiments were subsequently conducted using the masked priming paradigm and a naming task. The dependent variable for all three experiments was the voice onset reaction time or naming latency. Experiment 2 investigated whether the constituents would prime the compounds. The results showed that both constituents, or fragments of them, primed the naming latency of transparent compounds. Door and doo primed doorbell, and so did bell and bel. However, opaque compounds were only primed by the first constituent. Cock primed cocktail, but coc, tail, and tai did not. In Experiment 3, the opposite effect was investigated - whether the compounds would prime their constituents. Results showed that transparent compounds primed the naming latency of its constituents, while opaque compounds only facilitated the first constituent. Doorbell primed both door and bell, while cocktail only primed cock. When the compounds were reversed to make them nonwords, both types of compounds exhibited the same effect on the constituents. This time, only the second constituents of the actual compounds were facilitated. Belldoor primed bell but not door, and tailcock primed tail but not cock. Experiment 4 tested the naming latencies of the experimental stimuli without any priming. It showed that there was no difference in the baseline naming latencies of the constituent words, but the naming latencies of opaque compounds were inherently faster than those of transparent compounds. This suggested that additional processing was required for transparent compounds. The evidence that was gathered is generally supportive of the hypothesis. The overall implications are that transparent compounds undergo morphological decomposition and are accessed through their constituents, while opaque compounds are not decomposed and are accessed via the whole compound. The Cohort Model proposed by Sandra (1990) to show the workings of the processes that govern the lexical access of compound words required some modifications to account for the present findings.