Benzodiazepines, used in the clinic as anxiolytics, have in animal models been found specifically to attenuate behavioural suppression caused by response contigent aversive stimuli, non-reward or novelty. The effects have been interpreted in more general terms as "behavioural disinhibition" or "response perseveration" or in more specific terms as reduced "reward delay" or as an attenuation of a "behavioural inhibition system". In a recent publication we have described an experimental test in which decision making in the rat can be studied. The model is derived from ethology, in particular from optimal foraging theory. In order to solve the task, the animal must choose correctly between two options. For each option the probability of its resulting in a reward (water) has to be estimated on the basis of available information and to be related to the cost of performing it. We found that diazepam, in a dose that did not significantly affect the ability to perform the options per se, caused a strong impairment when these options, on the basis of available information, had to be combined into functional sequences in a decision making procedure. The results obtained cannot be explained on the basis of disinhibition or response perseveration. The hypothesis is advanced that benzodiazepines alter decision making in a more nonspecific may, by, for example, affecting the evaluation of the learned significance of stimuli in the environment.