Many geologic generalizations are observational patterns that are used to organize and interpret our data base. These patterns play an important role in the logical structure of geologic reasoning. In some cases we use the patterns as empirical generalizasions, and in others we use the same patterns as rules in an explanatory framework. The interaction of these two uses allows us to develop an understanding of geologic processes. It is important to keep these uses separate because different types of problems are associated with each. I suggest that observational patterns in geology are used in the same way as law statements in other sciences.