ABSTRACT
It seems to me that there are two distinct strategies, among others, that are available in the comprehension of connected discourse. In one strategy, you try to maintain the integrity of a particular passage as much as possible. I think this strategy is very common, for example, in memory experiments, in which the material is typically not of any general usefulness, and the goal is to try to remember as much as possible. When this strategy is operative, it makes a lot of sense that intrapassage structural relations will be a very potent factor in determining the likelihood of recall for the various elements of the discourse. However, there is another strategy which is probably more typical, and, for educators, one which would possibly be considered more ideal – that the function of comprehension can be to update one’s knowledge. There is a great deal of redundancy in what one reads and hears; a lot of information that one encounters is old and uninteresting information. A lot of it is new. I think, in this second strategy, you focus on the latter information, and use it to update old knowledge. When this strategy is employed, I think it is very possible that the intrapassage structural relations will be subordinated to the pattern of relations which exist in the preexisting knowledge structure which is being updated. In that case, there is no reason to suspect consonant patterns of superordinateness.
