I hope this is the right (or rightest) place to ask this question, I wasn't sure whether this question is more appropriate for cogsci or biology. After all, it does concern the biology of emotions.
To my knowledge, it is established research that the brain 'knows' what emotions you are having by examining their manifestations in the body. This absurd rube-goldbergian wiring can for instance be used to force oneself to become happier (at least temporarily) by smiling for a few minutes.
This brings up the question, whether emotions have any identity separate from a cluster of bodily sensations (or even more specifically, muscle contractions) at all.
It is certainly true (by introspection for e.g.) that 'negative' emotions are accompanied by muscle contradictions. On the other hand, there are several different 'negative' emotions that can be easily distinguished by the person having the emotion (e.g. shame and fear). Unfortunately I am not able to simply recall the sensations I had when I was ashamed or fearful from memory and compare them, but I would assume you could even tell from the outside, simply by looking at the person afflicted, which of the two emotions they are going through.
So, what does the research say? Are emotions uniquely recognizable by facial expression/muscle contractions? Or do emotions have some kind of 'hidden variables' in the brain?