Because of the many criticisms of the notion of a theory independent observation language, it is useful to look at a few actual examples of scientific data to see what theories might be implicit and in what way. There are several possibilities. The theories could be previously accepted, under active investigation, or of a kind that has never been systematically examined. The present study explores these possibilities in the data of several scientific journal articles. The conclusion is that theories and (...) regularities of all three kinds can be found as antecedent conditions of data statements. Any descriptive account of scientific inference must find room for such presuppositions. (shrink)