Recent decades have seen a number of influential attacks on the comparative psychology of learning and intelligence. Two specific charges have been that the use of distantly related species has prevented us from making valid evolutionary inferences and that learning mechanisms are species-specific adaptations to ecological niches and hence not properly comparable between species. It is argued here that work using distantly related species may yield valuable insights into the structure of intelligence and that the question of whether or not learning mechanisms are niche-specific is one which can only be answered by comparative work in “nonnatural” situations. The problems involved in defining and assessing intelligence are discussed. Experimental work has not succeeded in demonstrating differences in intellect among nonhuman vertebrates. Hence the null hypothesis – that there are no differences in intellect among nonhuman vertebrates – should be adopted; the superiority of human intelligence stems from our possessing a species-specific language-acquisition device. One implication of the null hypothesis is that general problem-solving capacity is independent of niche-specific adaptations. A second implication is that problem-solving may involve relatively simple mechanisms; association formation in particular may play a central role in nonhuman intelligence, allowing the successful detection of causal links between events. Causality is a constraint common to all ecological niches.