ABSTRACT
Chemical ‘substances’ and biological ‘species’ are less artificial than many objects of scientific knowledge, for their samples are things which have many properties accessible to fairly direct inspection. The neglect of the category of evidence clearly derives from the traditional concerns of the philosophy of science, and its paradigm examples of scientific achievement naturally obscure this aspect of scientific work. The artificial and fallible nature of the conclusions of solved scientific problems has an importance beyond the improved historical understanding of natural science. The similarity of structure between these other sorts of projects and a scientific problem should not obscure one very deep difference. The work of a technician is also with particular things and events: but some of the properties which he must understand and control require the intellectually constructed classes of ‘scientific’ knowledge for their explanation or even for their description.
