RÉSUMÉ
The use of animals in neurosciences has a long history. It is considered indispensable in areas in which "translational" research is deemed invaluable, such as behavioral pharmacology and comparative psychology. Animal models are being used in pharmacology and genetics to screen for treatment targets, and in the field of experimental psychopathology to understand the neurobehavioral underpinnings of a disorder and of its putative treatment. The centrality of behavioral models betrays the complexity of the epistemic and semantic considerations which are needed to understand what a model is. In this review, such considerations are made, and the breadth of model building and evaluation approaches is extended to include theoretical considerations on the etiology of mental disorders. This expansion is expected to help improve the validity of behavioral models and to increase their translational value. Moreover, the role of theory in improving construct validity creates the need for behavioral scientists to fully engage this process.