Food aversion learning has attracted widespread interest because it is a highly adaptive, powerful type of learning with both practical and theoretical ramifications. It has features that make it unusual and robust when compared with other learning paradigms. It has relevance to human problems in that it is likely to contribute to food choice and appetite problems in certain clinical situations. And the robustness of this learning makes it a promising model for neurobiologists interested in understanding neural mechanisms of plasticity. This review provides a broad overview of these aspects of taste aversion learning and points to areas where questions remain and additional research is needed.