In this article I attempt to reconstruct David Hume's use of the label ?experimental? to characterise his method in the Treatise. Although its meaning may strike the present-day reader as unusual, such a reconstruction is possible from the background of eighteenth-century practices and concepts of natural inquiry. As I argue, Hume's inquiries into human nature are experimental not primarily because of the way the empirical data he uses are produced, but because of the way those data are theoretically processed. He (...) seems to follow a method of analysis and synthesis quite similar to the one advertised in Newton's Opticks, which profoundly influenced eighteenth-century natural and moral philosophy. This method brings him much closer to the methods of qualitative, chemical investigations than to mechanical approaches to both nature and human nature. (shrink)