There has been a major shift in the psychology of language from a behavioristic to a cognitive view of learning. One important reason for this shift in thinking was that behavioristic theory simply could not account for language behavior as portrayed in Chomsky's theory of language. Within educational psychology, instructional research has assumed a more cognitive posture as well, particularly research involving the stimulation of prose processing through interspersing questions in text. Originally based on the associative or behavioristic model, early (1965)(1966)(1967)(1968)(1969)(1970) adjunct question research can be Characterized as having a "variables orientation". Within this orientation, primary consideration was given to examining the effects of manipulating question position (before or after text segments) and question frequency on text comprehension. In this early research, inserted questions were invariably pitched at a rote or verbatim learning level. Since the onset (c. 1970) of the cognitive revolution, however, adjunct question research has been focussed on examining the effects of different cognitive levels of inserted questions on text comprehension, and on developing methods for assessing processes produced by adjunct questions. However, the degree to which cognitive levels as operationalyzed in the experiments reviewed here captures the richness of cognition as it occurs in the schools is, of course, open to question. Other areas of cognitively-oriented research include the assessment of the relationship between individual differences and adjunct question treatments, and the generation of questions by subjects while reading. This more recent thrust in adjunct question research can be termed a "processes orientation". Research associated with both the "variables orientation" and the "processes orientation" are reviewed. Finally, some remarks are made concerning the paucity of theory in adjunct question research, the possible use of semantic memory theories as a basis for future research within the adjunct question paradigm, and the need for new question paradigms that more closely "match" question answering in general.* Appreciation is extended to Michael Macdonald-Ross for his critical comments and suggestions on an earlier version of this paper.