The paper describes various aspects of automatic morphological disambiguation of Czech realized not by statistical methods, but by grammatical disambiguation rules