To compare empathy in human and nonhuman animals (hereafter “animals”) we need an adequate conceptual framework characterizing similarities and differences between animal and human empathy. Such a framework is still lacking. The core intuition of our proposal is that empathy typically involves the registration of the other’s relevant affective state, the specific situation, and possibly even mental states (or processes). This is combined with a somewhat flexible selection of a response behavior that accounts for the mental states and the situation of the agent. More precisely: We think that we can cluster relevant elements into five main dimensions, starting with three forms of sensitivity for the other: (1) sensitivity for bodily and affective states of others, (2) perceiving others as individuals in a situation, (3) cognitive perspective taking in relation to the other; in addition, we include as relevant the kind of typically (4) other-oriented response, and (5) the level of flexibility which is connected with the response behavior. Each of these dimensions can be categorized by several different sub-abilities. Further, all five dimensions can be realized to a certain degree. Thus, for each dimension of empathy we can characterize to which degree it is realized in an animal; this leads to new evaluation of and questions about empathic behavior, e.g. is the recently described helping behavior of rats (Bartal, Decety & Mason, 2011) empathic behavior? In their famous experiment, a rat was freely moving in an arena while another rat from the same cage was trapped in a restrainer. After several sessions, the rat being able to move freely in the arena learned to open the restrainer (with another rat being in it). In a follow-up experiment, the rat could choose between opening the restrainer with her cage mate being trapped or opening a second retainer enabling access to tasty food, chocolate. The rat seemed to register the distress of the trapped animal and first opened her retainer and let her free before opening the second restrainer. Moreover, the helping rat even shared the chocolate with the other rat (Bartal, Decety & Mason, 2011). Should we interpret this as a case of full-blown empathy or do we need to employ a killjoy interpretation according to which the freeing behavior is just triggered by the egoistic desire for social interaction? Our multidimensional approach offers a new way of interpreting the observations.