The approach to animal behaviour adopted by behavioural ecology is based on the investigation of the adaptive function of behaviour. A common assumption is that the action of natural selection on behaviour can be predicted without reference to processes inside the organism. I believe that it is time to combine an analysis based on evolution with one based on mechanisms, where a mechanism might be psychological, physiological or a combination of both. Animals have mechanisms that need to perform well in changing and dangerous environments. In order to understand the evolution of mechanisms, we need a fundamental change in the sort of models that are analysed. Instead of building complex models of optimal behaviour in simple environments, we need to evolve simple mechanisms that perform well in complex environments. This approach can provide a novel and unified perspective on a range of issues involving decisions by animals, including humans. The main objective of the project is to provide a comprehensive view of behaviour that can account for both adaptive and non-adaptive actions. This involves developing a novel theoretical framework based on an understanding of the underlying information-processing rules, combined with an evolutionary perspective that explains how any such rule came into existence in the first place. The theme of coping with uncertain and dangerous environments is used to investigate various features of behaviour such as rationality and self-control. These topics lead to the broader issues of the organisation of thought and emotions. The project also explores the consequences of the evolved behaviour and the implications for conservation and animal welfare.