A two-trial memory task, based on place or object exploration in a Y-maze, was developed to study recognition processes in young and aged Sprague-Dawley rats (Dellu et al., 1992). This paradigm avoids the use of electric shocks or deprivations that may have nonspecific effects, and the task does not require learning of a rule, i.e., involves essentially working memory. A number of behavioral parameters in several animals could be recorded automatically, thus allowing a detailed analysis of behavior and a simultaneous testing of several animals. We extended this task in order to evaluate the role of contextual information in recognition memory. When the contextual environment during acquisition and retrieval trials was unchanged, place recognition remained intact despite interpolated explorations. When the contextual environment during acquisition and retrieval trials was changed, object recognition was impaired. It was clearly shown that the correct choice between two discrete cues (e.g., small objects) is conditional on more diffuse environmental cues. Thus, it can be hypothesized that at least two kinds of recognition processes can be assessed in animals: an automatic process, based on a discrimination between familiarity and novelty, and a second process, more cognitively based and more effortful. In conclusion, the two-trial task in the Y-maze is sensitive, specific, and fast. It allows the study of both spatial and nonspatial memory, allows the study of different levels (automatic vs controlled) of cognitive processes and their neuroanatomical substrates, and may be profitably employed in aging studies and neuropharmacological research in general.