A method for the syntactic analysis of Russian sentences, applied to automatic translation, is divided into a morphological word-by-word phase and a syntactical sentence-by-sentence phase. An idealized canonical stem dictionary is presented, and its significant lexicographic properties are pointed out. This idealized dictionary then serves AS A BASIS FOR EVALUATING THE ACTUAL Harvard Automatic Dictionary. Aspects of morphological analysis of the Russian language and a series of programs written to carry it out are described. To explain the practical problems encountered in an experimental syntactic analysis program, of which a detailed description is given, a new model of natural language is introduced.