Decivilising processes are what happens when civilising processes go into reverse; both terms are used here in a specific technical sense derived from the work of Norbert Elias (The Civilising Process, 2 vols., 1978/1982, orig. 1939). The first part of the paper briefly sketches aspects of the theory of civilising processes necessary for identifying symptoms of their reversal. It is inferred that true decivilising processes would be marked by breaking links and shorter chains of social interdependence, associated with higher levels of danger and incalculability in everyday life, the re-emergence of violence into the public sphere and a decline in mutual identification, reduced pressures on individuals to restrain the expression of impulses (including the freer expression of aggressiveness), changes in socialisation and personality formation, and increasing fantasy-content of modes of knowledge.The second part of the paper examines evidence relating to four candidates for the label `decivilising process'. The debate about the `permissive society' suggests that this was `decivilising' in a very limited sense: it involved a `highly controlled decontrolling of emotional controls'. Evidence of an upturn in violence in contemporary United States and British society is examined next; the data of Gurr, Stone and Dunning suggest the longterm trend was until recently downwards. The Holocaust and other instances of mass murder are discussed next. Finally, it is proposed that the clearest cases of decivilising processes would be those where trends could be observed over at least three generations; evidence should be sought in the collapse of complex societies (such as the Roman Empire) and episodes such as the Thirty Years War and its aftermath, and the so-called `Wild West'.