The goal of this thesis was to develop a system for the computer analysis of written natural language texts that could also serve as a theroy of human comprehension of natural language. Therefore the construction of this system was guided by four basic assumptions about natural language comprehension. First, the primary goal of comprehension is always to find meanings as soon as possible, Other tasks, such as discovering the syntactic relationships, are performed only when essential to decisions about meaning. Second, an attempt is made to understand each word as soon as it is read, to decide what it means and how it relates to the rest of the text. Third, comprehension means not only understanding what has been seen but also predicting what is likely to be seen next. Fourth, the words of a text provide the cues for finding the information necessary for comprehending that text.