Dividing attention without alternation or automaticity.

Abstract

SUMMARY Spelke, Hirst, and Neisser trained two subjects to copy unrelated words at dictation as they read and understood stories. The subjects' success was interpreted as evidence against the hypothesis of a fixed attentional capacity or limited cognitive resources; instead, it was hypothesized, attention is a skill that improves with practice. However, other explanations of these results can be proposed. The present research addressed two such counterhypotheses: that capacity may be alternated between reading and writing and that the writing task may become automatic, and require no capacity at all. Experiment 1 was designed to see whether subjects take intermittent advantage of the redundancy of the stories to switch to the writing task. Some subjects were trained to copy words while reading highly redundant material (short stories); others were trained with less redundant encyclopedia articles. On reaching criterion, each subject was switched to the other type of reading material. Three of the four subjects trained with stories transferred their skill immediately to the encyclopedia, suggesting that they had not been using the redundancy of the stories to accomplish their task. Experiment 2 addressed that automaticity hypothesis. Two subjects were trained to copy complete sentences while reading. Several tests then showed that they understood the meaning of the sentences : (a) They made fewer copying errors with real sentences than with random words; (b) they recalled real sentences better than random words; (c) they integrated information from successive sentences, as demonstrated by a test of recognition memory for new statements whose truth was implied by the original ones. In view of this evidence that the sentences were understood, it is hard to maintain that they were being handled in an automatic way. These results strengthen the hypothesis that the ability to divide attention is constrained primarily by the individual's level of skill, not by the size of a fixed pool of resources. Postulated capacity limits may provide plausible accounts of unskilled performance but fail to explain the achievements of practiced individuals.