There are a variety of different kinds of normative questions that may be asked about science. We can ask, for example, whether laws should be conceived as stating no more than constant conjunctions? Or how should experimental controls be structured? Or is evidential support measurable? Or do theories function only as anti-realists claim that they do — viz. as instruments for the prediction of observable events? Or is there only one correct pattern of scientific finding, or, as some readers of Alasdair Crombie's Styles of Scientific Thinking in the European Tradition might suppose, are there several? These and their like are one familiar group of normative questions about one aspect of scientific enquiry. Let us call this the "epistemological" aspect, because it is this kind of question that needs to be answered in order to make progress towards resolving the larger issue that may be conveniently summarised in the question "What is scientific knowledge?"