Total Precedence Relations

A description is given of a method of syntact ical analysis based on the use of precedence relations which are defined over the whole vocabulary of the context-free grammar under investigation. These relations are defined in such a way tha t t r ial-and-error is avoided whenever possible. To the same end, the customary restr ict ion tha t analysis will be lefttoright is not imposed. The results are formulated in such a way tha t all relations can be computed by simple operations on Boolean matrices. This s tudy reveals the existence of an interest ing subset of context-free languages, which is then compared with other well-known subsets.