LEVELS OF CAUSE AND EFFECT AS ORGANIZING PRINCIPLES FOR RESEARCH IN ANIMAL BEHAVIOR

Animal behaviour can be explained from several complementary perspectives, and these explanations can be formally categorized into "levels of analysis." Recognizing levels of analysis may prevent unnecessary conflict between complementary explanations and promote the understanding of behaviour from all possible perspectives. However, there is considerable room for debate as to the best framework of levels. The current framework is based on the concept of "proximate and ultimate causes," which merges the functional consequences of a behaviour with its evolutionary causes. Consequently, the use of this framework leads to confusion between causes and effects of behaviours, and leaves no room for studies of the effects of a behaviour other than those with adaptive significance for the animal performing it. A more balanced approach can be encouraged by considering questions about causes and effects of behaviours to be in separate categories, with their own analytical levels. Causes of behaviours can be explained by processes on at least four levels: evolution, ontogeny, fluctuations in an animal's state, and occurrence of behaviour-eliciting stimuli. Effects of behaviour may also occur on several levels, but the levels will depend on what is being affected (i.e., the animal performing the behaviour, other organisms, or the physical environment). The causes of behaviours can be influenced by their effects through feedback mechanisms such as reinforcement and natural selection. However, we should not assume such feedback mechanisms to be important in the causation of all behaviours. If we use the effects of behaviours to explain why they occur, it is important to explicitly recognize the feedback mechanism(s) invoked.