Different types of science definitions (conceptual, descriptive and prescriptive) and their relation to the different philosophies of science are discussed. It is argued that science must involve a search for consensus, and that this is a possible line of demarcation between science and non-science. To illustrate this, the plurality of approaches found in the humanities and social sciences is explored. It is suggested that human activities with the ambition to produce and/or give out knowledge about the world can at least be placed on a spectrum from the most scientific to the least. It is further discussed what it means to be ‘scientific’, and marks of good science are outlined: repeatability, transparency, objectivity, parsimony, the willingness to test and abandon hypotheses, and a non-authoritarian approach to knowledge. Scientists can sometimes behave in a less than perfectly scientific way, and non-scientists may behave scientifically to some extent, and non-science can thus have scientific aspects (or not). Science can only work if there is a web of trust, aided by supportive and safeguarding systems, setting the stage for the upcoming chapters on science in practice.