"The concept of "cognition" is a sore point in the philosophy of emotion, since there is no consensus and almost no clarity about what it might best mean. Prinz offers a view as good as any and better than most. He argues that "cognitions are states containing representations that are under [direct] organismic control" (49). This sounds about the same as the view that cognitions are products of the will; but though organismic control or the will is itself mysterious, it is possible to see how one might go about looking for evidence that a state was answerable to the will. One can test if a subject can change the state, for example. Furthermore, Prinz goes so far as to brave a hypothesis about the brain areas that may be required for direct control: "I propose that we call a state cognitive just in case it includes representations that are under the control of structures in executive systems, which, in mammals, are found in the prefrontal cortex" (47). With this working notion, Prinz concludes that emotions are not cognitive.
Instead, emotions are perceptions of certain kinds of body states. These body states are ones that reliably track certain kinds of conditions in the environment of the agent. For example, one kind of body state reliably is caused by potential dangers in the environment. An instance of fear is a perception of an instance of this kind of body state. But because the body state perceived is reliably linked with dangers, it is appropriate to say that fear represents potential dangers in the environment. To make this distinction, Prinz introduces the terminology of nominal and real contents, so that the nominal content of the emotion as a representation is the body state, but the real content is the environmental condition that reliably causes the state -- in this case, dangers." - NDPR.