Two experiments were conducted to investigate active memory processes during reading. Subjects read two-clause sentences one word at a time at a self-paced rate. Following each sentence a probe word was presented; subjects were to decide if this word occurred in the sentence they had just read. The first experiment examined clausal effects during reading. Reaction times to items from the final clause were shorter than those to items from a previous clause even when the same number of words intervened. The second experiment used the clause effect to address the issue of proniminal reference. Results indicated that a pronoun in the final clause activated the meaning of its antecedent, thus demonstrating that the method is sensitive to both surface and meaning codes in active memory.The complex process through which readers turn printed symbols on a page into meanings has been the subject of much experimentation in both psychology and education. The present study investigates the role of active memory as one aspect of this comprehension process. The usage of the term "active memory" should be viewed as similar to that of "working memory" (cf. Baddeley & Hitch, 1974) or "operational memory" (cf. Posner, 1973) in the sense that it represents a control system with limits on both storage and processing capabilities. The aim of the present research is essentially twofold: To investigate the function of active memory during reading, and to develop a method through which the process can reliably be studied. Two experiments are reported here. The first investigated how active memory might be influenced by clausal boundaries in sentences, and it served to test and validate the method used. The second experiment employed this method to examine the issue of pronominal reference.Most of the research looking at active memory processes for linguistic materials has investigated the processing of auditorily presented stimuli. One of the major theories of sentence processing involving a conThis paper is based on a thesis submitted to the University of Oregon in partial fulfillment of the requirements for the Master of Arts degree. Portions of this research were presented at the meeting of the Western Psychological Association in San Diego, California, on April 6, 1979. I would like to express my appreciation to Michael Posner for his support and guidance throughout this project. I am also grateful to Albert Corbett, Wayne Wickelgren, and Douglas Hintzman for their advice and comments on this research. This work was supported by There has been some experimental evidence supporting this contention. Jarvella (1971) presented listeners with short stories with instructions for comprehension. He interrupted the subjects at various points in the stories, and their task was to write down as much as they could remember verbatim. The results were that subjects' rote recall was best for the material in the clause immediately preceding the interruption and fell off markedly for material that occurred prior to that clause. He found a simila...