⁷From an alternative probability theory viewpoint, impermanence could also be seen as incorporating nonstationarity, which is the technical term for the situation where the world is changing, and a statistical model learned on data in the past may not be valid anymore in the present and even less in the future. Such problems were already alluded to in previous chapters when it was pointed out that humans may be evolutionarily adapted to the African savannah instead of the modern city environment (Chapter 5); or that emotional reactions learned as a child may be far from optimal in an adult since the environment is very different (Chapter 8). It could be argued that nonstationarity is as important as scarcity of the data, and an independent root cause of suffering. However, I don’t pursue that line of argumentation here since I tend to think that any problems with nonstationarity could be avoided if the agent had enough data and computation since it would then be able to predict the nonstationarity (like Laplace’s demon in footnote 15 below); but this is admittedly a complex and controversial point that needs further research.