One of the most consistent findings of research on discourse is that important text information is better learned than less important information because readers devote more attention to the important information. There is now very good reason to believe that questions cause readers to attend selectively to question-relevant information and that a process supported by the extra attention causes more of the question-relevant information to be learned. However, despite superficial appearances, attention does not lie on the causal path between the interest value of a sentence and the learning of this sentence. Children do pay more attention to interesting sentences and they do learn more interesting sentences. However, a deep analysis suggests that the extra attention is a secondary phenomenon. So far, research on whether attention plays a part in the learning of information important in the light of a reader's perspective has been inconclusive. The problem may not be so much with the concept as with the method of assessing the level of cognitive effort using discrete secondary task probes. (HOD)