In an analysis of 20 scholarly journal articles in the hard sciences and 20 from the humanities, text was examined for author comments on hypotheses, probabilities, and evaluations. These include such expressions as "It cannot be denied, it seems likely, it was presumed," etc. Such comments were found to be somewhat more common in science-related articles, with frequency ranging from a high of 1.17 per sentence in a philosophy article to a low of 0.14 per sentence in an organic chemistry article. When analyzed for evidence of rhetorical sections, it appeared that generally, the science articles could be divided into three sections (beginning, middle, end), but the humanities articles had no consistent rhetorical "shape." In science articles, comments were most frequent in end sections, less so in beginnings, and least commonly observed in the middle. Functions of the comments were different in each section. Certain comments were found to be used more often than others and were different for the arts and sciences. It is concluded that the function of comments in academic prose is to express three types of certainty: information that is taken for granted; the purely hypothetical; and logical deduction. (MSE)