ABSTRACT
Different methods of assessing word perception degrade different parts of the information in the original stimulus; in particular the brief exposures used in a tachistoscope have different effects on receptors sensitive to high spatial frequencies and upon receptors for low spatial frequencies. In a first experiment, common and rare words, each with predictable and unpredictable letter sequences, were exposed masked by visual noise which had been high-pass or low-pass filtered. With rare words, there was an interaction between the nature of the masking noise and the predictability of the letter sequence. This encouraged a further experiment in which neutral and unpleasant words were presented after removal of alternate letters, as if masked by relatively low frequencies. These words have previously shown an effect of emotionality in a tachistoscope, but not when presented defocused. In the version with alternate letters missing, the unpleasant words were less frequently identified by subjects. Further analyses, and two more experiments, showed that the effect of emotionality was not that of a constant bias against unpleasant words: it appeared to be rather an increased variance on such words, causing them to be chosen more when evidence pointed toward them and less when evidence pointed against them. These results have methodological implications, but also confirm that some important information about words comes from their general shape rather than detail; finally, they show that it is not necessary, in order to explain the effect of word emotionality, to suppose any system feeding back from partial identification of words in order to cut down incoming evidence.
