At least two important questions are posed by the existence of hippocampal place cells. The first of these has to do with how the complex, abstract properties exhibited by these cells can be explained mechanistically. The second has to do with the implications of place cells for our conception of the broader role of the hippocampus in spatial and other behaviors. Here, evidence is reviewed that: (1) Hippocampal cells show different "maps" (place cell representations) for each environment the animal visits and, in fact, can show multiple maps even for any one environment. The choice of the current map for any one environment depends on environmental, contextual, and event-related variables. (2) Cells in the subiculum and entorhinal cortex also show location-specific firing patterns (like hippocampal place cells), but show the same pattern for each environment the animal visits. A model is presented that is a variant of hippocampus-based path integration models developed by McNaughton and colleagues. In this version, the subiculum and entorhinal cortex work together to form a single, universal map that is used for each environment, and that can exhibit path integration abilities. The universal subicular/entorhinal representation is postulated to assist the hippocampal layer to rapidly form new environment and context specific "maps" for each new environment/temporal context ("episode") the animal experiences. In this view, hippocampal layer activity is always obligatorily spatial, due to the input from the entorhinal universal "map." However, the fact that the hippocampus generates a new map in response to global, non-spatial, contextual attributes of each situation, means that the hippocampus is always coding non-spatial aspects of a situation using its obligatorily spatial code. This brings the hippocampal place cell activity in to line with the broader role that has been postulated for the hippocampus in learning and memory functions.