GENERATION EFFECT

High And Low Frequency, Generation Effect

High And Low Frequency, Generation Effect

Introduction

It is a commonplace that words often have more than one meaning. The prevalence of ambiguity in language and the problems it poses for comprehension has prompted a great deal of research on this phenomenon in an effort to understand the process of recognizing words. In particular, much research has focused on lexical access, that part of word recognition involving the sensory analysis o f the input and its contact with a representation or representations in memory. Less research has focused on post-lexical processes that use the results of lexical access for selection, elaboration, and integration with the previous discourse. An understanding of lexical access and its products, however, is required before serious study o f post-lexical processes can be made.

It has been widely found that items which have in some way been internally 'generated' by subjects are remembered better than items which are provided by the experimenter. For example, requiring the subject to provide the answer to 'What is the capital of France?' leads to better recollection that 'Paris' was in the list of items than requiring a yes or no answer to 'Is Paris the capital of France?' This generation effect has been extensively studied and a number of explanations explored (Slameka & Graf, 1978). Slameka McElroy (1982) found no effect of the generation of non-words, and therefore proposed the hypothesis of activation of the lexicon. According to this hypothesis, subjects should be presented in the mental lexicon of a generation effect to occur and, with respect to "read" the state of generation increases the activation of lexical representations, the benefits of this activation of memory.

Nairne, Pusen and Widner (1985) used common, i.e. high frequency (HF) words, uncommon, i.e., low frequency (LF) words and non-words, and obtained a generation effect for only the HF words. They argued that representation in the lexicon is not a sufficient condition for the generation effect to appear and so rejected the lexical activation hypothesis. Instead they proposed what may be termed the associative linkage hypothesis. Non-words, then, should not give rise to a generation effect because they have no representation in that network and LF words should show less of an effect than HF words because they have fewer associative links. While Nairne et al.'s hypothesis predicts the interaction between word frequency and presentation conditions which they found it also predicts that there will be a generation effect for LF words. Nairne et al. employed a recognition test in which LF words usually give better performance than HF words; so much so that it is possible that the interaction in Nairne et al.'s results reflects a ceiling effect with LF words. The effect of lexical frequency is certainly phenomenon replicated in most studies psycholinguistic. It reflects the fact that words appearing more frequently in a given language are seen faster and produce less Error recognition that words appearing less frequently. This effect is consistent both through the experimental tasks (lexical decision, ...