A research project investigated the process of reading comprehension through which the reader generates a semantic representation of the message conveyed by a text. The first focus of the project was an examination of the functioning of abstract knowledge in text comprehension. Studies were conducted to explore the activation of proposition schemata and the use of such schemata in constraint checking and encoding implicit propositional arguments. Other studies provided evidence that episode schemata remain activated in memory as episodes are encoded. Additional studies examined the consequences of a mismatch between input and activated schemata, investigated the retrieval of abstract semantic knowledge, reviewed retrieval of abstract world knowledge and prior information in a text in coherent anaphoric references, and explored the use of stored orthographic knowledge, phonetic knowledge, and semantic knowledge in word recognition. The second focus of the project concerned the nature of text representation and the retrieval of prior information in a text. Two studies looked at the role of clause structure in integrating new information and in retrieving information from memory. Additional studies examined a process theory of pronoun disambiguation and employed reading time in processing anaphoric reference to examine frequency and importance as factors in retrieving prior text information. Theories were developed concerning the representation of context in text structure and chunking and consolidation in memory. (Author/FL)